Babies of mothers who experience significant psychological distress during pregnancy showed evidence of altered brain development in utero and reduced cognitive outcomes at 18 months, based on data from a pair of studies including approximately 300 women.

In a longitudinal study published in JAMA Network Open, Yao Wu, PhD, of Children’s National Hospital, Washington, and colleagues recruited 97 healthy mother-infant dyads between January 2016 and October 2020 at a single center. Of these, 87 underwent two fetal brain imaging studies each, and 10 completed the first MRI visit, for a total of 184 fetal MRIs.

Neurodevelopment and social-emotional development for infants at 18 months of age was measured using the Bayley Scales of Infant and Toddler Development and Infant-Toddler Social and Emotional Assessment. The mean age of the mothers was 35 years; maternal distress was assessed between 24 and 40 weeks’ gestation using validated self-report questionnaires. Parenting stress was assessed at the 18-month infant testing using the Parenting Stress Index-Short Form.

Overall, prenatal maternal stress was negatively associated with infant cognitive performance (P = .01) at 18 months, mediated by fetal left hippocampal volume.

In addition, increased fetal cortical local gyrification index and sulcal depth measured during reported times of prenatal maternal distress were associated with significantly poorer social-emotional scores and competence scores at age 18 months. The beta coefficients for local gyrification index and sulcal depth were –54.62 and –14.22, respectively, for social-emotional and competence scores, –24.01 and –7.53, respectively; P values were P < .001, P < .002, P = .003, P < .001, respectively.

“Increased cortical gyrification has been suggested in children with dyslexia and autism, and sulcal depth has been associated with the severity of impaired performance on working memory and executive function in adults with schizophrenia,” the researchers wrote in their discussion of the findings.

The current study “extends our previous findings and suggests a critical role for disturbances in emerging fetal cerebral cortical folding development in mediating the association between prenatal maternal distress and neurodevelopmental problems that later manifest in infancy,” they explained.

The researchers also found that prenatal maternal anxiety, stress, and depression were positively associated with all measures of parenting stress at the 18-month testing visit.

The study findings were limited by several factors including the use of self-reports for both maternal distress and infant social-emotional assessment, despite the use of validated questionnaires, and the fact that assessment of maternal distress at specific times may not reflect the entire pregnancy, the researchers noted. Other potential limitations included the inability to use some MRI data because of fetal movement and the homogenous population of relatively highly educated women with access to health care that may not reflect other areas, they said.

“Identifying early brain developmental biomarkers may help improve the identification of infants at risk for later neurodevelopmental impairment who might benefit from early targeted interventions,” the researchers concluded.
 

Technology enhances health and disease models

The effect of the prenatal period on future well-being is recognized, but the current study makes “substantial contributions to prenatal programming science, with implications for ways to transform the prenatal care ecosystem for two-generation impact,” Catherine Monk, PhD, and Cristina R. Fernández, MD, both of Columbia University, New York, wrote in an accompanying editorial.

John Abbott/Columbia University

The developmental origins of health and disease (DOHaD) conceptual model introduced by Dr. David Barker in 1995 were later applied to show that maternal stress, depression, and anxiety affected child prenatal and future development, they said. However, the current study uses cutting-edge neuroscience to directly assess developing fetal brains. The finding of reduced cognitive functioning at 18 months associated with maternal stress is consistent with other findings, they noted.

“Finding an association between maternal prenatal stress and infant cognitive outcomes in the setting of what may be modest stress relative to that of a low-resourced or historically marginalized sample underscores the importance of this research; presumably, with higher stress, and greater social determinants of health burden, the effect sizes would be even greater and of greater concern,” they said.

However, studies such as the current one “have the potential to transform the prenatal and postpartum care ecosystems,” by encouraging a whole-person approach to the care of pregnant women, including attention to mental well-being and quality of life, they emphasized.
 

COVID-19 stress considerations

In a separate study published in Communications Medicine, Yuan-Chiao Lu, MD, also of Children’s National Hospital in Washington, and colleagues found a similar effect of maternal stress on fetal brain development.

The researchers imaged the brains of fetuses before and during the COVID-19 pandemic and interviewed mothers about any distress they experienced during pregnancy.

The study population included 65 women with known COVID-19 exposures who underwent 92 fetal MRIs and 137 prepandemic controls who underwent 182 fetal MRIs. Maternal distress was measured via the Spielberger State Anxiety Inventory, Spielberger Trait Anxiety Inventory, Perceived Stress Scale, and Edinburgh Postnatal Depression Scale.

Overall, scores on measures of stress and depression were significantly higher for women in the pandemic group compared with controls. Of the 173 women for whom maternal distress measures were available, 28% of the prepandemic group and 52% of the pandemic group met criteria for elevated maternal psychological distress, defined as above the threshold for distress on any one of the four measures.

After the researchers controlled for maternal distress, MRI data showed decreases in fetal white matter and in hippocampal and cerebellar volumes in fetuses in the pandemic group compared with controls.

Other signs of impaired brain development were similar to those seen in the JAMA Network Open study, including decreased cortical surface area and local gyrification index, as well as reduced sulcal depth in multiple brain lobes, indicating delayed cerebral cortical gyrification.

The second study was limited by a lack of data on other lifestyle changes during the pandemic that might influence maternal health and fetal development, the researchers noted. Other limitations were the possible lack of generalizability to a range of racial and ethnic populations and geographic areas outside of Washington, and the inability to control for unknown COVID-19 exposures or subclinical infections in controls, they said.

However, the results support findings from previous studies, and provide a unique opportunity to study the effect of prenatal stress on early development, as well as a chance to implement “novel and timely interventions,” the researchers wrote.

“Monitoring the COVID generation of infants for long-term cognitive and health outcomes after birth is warranted and currently underway,” and continued research may inform preventive strategies for pregnant women experiencing multiple stressors beyond the pandemic, they concluded.


 

Interpret pandemic effect with caution

“Research studies, as well as our own daily experiences, have made it abundantly clear that stress is on the rise as a consequence of the COVID-19 pandemic,” said editorial author Dr. Monk, who commented on the second study in an interview. “This is an important public health question: Early identification of pandemic effects on child development can help garner the necessary resources to intervene early, dramatically increasing the likelihood of improving that child’s developmental trajectory,” she said.

“The pandemic is an unprecedented experience that has widespread impact on people’s lives, how could it not also alter gestational biology and the developing brain? That being said, we need to be cautious in that we do not yet know the functional implications of these brain changes for longer-term development,” Dr. Monk said. “Also, we do not know what aspects of women’s pandemic-affected lives had an influence on fetal brain development. The authors found higher stress in pandemic versus nonpandemic women, but not evidence that distress was the mediating variable relating pregnancy during the pandemic to altered brain development,” she explained.

The take-home message for clinicians is to “provide your patients with realistic avenues for neurodevelopmental assessments of their children if they, or you, have concerns,” Dr. Monk said. “However, do not prejudge ‘pandemic babies,’ as not all children will be affected by these potential pandemic effects,” she emphasized. “It is possible to misjudge normal variation in children’s development and unnecessarily raise parents’ anxiety levels. Importantly, this period of brain plasticity means any needed intervention likely can have a big, ameliorating impact,” she added.

“We need follow-up studies looking at pandemic effects on prenatal and postnatal development and what factors protect the fetus and birthing person from the negative influences,” she said.

The JAMA study was supported by the National Institutes of Health and the A. James & Alice B. Clark Foundation. The study in Communications Medicine was supported by the National Institutes of Health, the Intellectual and Developmental Disabilities Research Center, and the A. James & Alice B. Clark Foundation. None of the researchers in either study disclosed conflicts of interest. Dr. Monk disclosed grants from the National Institutes of Health, the Bezos Family Foundation, and the Robin Hood Foundation outside the submitted work.

Babies of mothers who experience significant psychological distress during pregnancy showed evidence of altered brain development in utero and reduced cognitive outcomes at 18 months, based on data from a pair of studies including approximately 300 women.

In a longitudinal study published in JAMA Network Open, Yao Wu, PhD, of Children’s National Hospital, Washington, and colleagues recruited 97 healthy mother-infant dyads between January 2016 and October 2020 at a single center. Of these, 87 underwent two fetal brain imaging studies each, and 10 completed the first MRI visit, for a total of 184 fetal MRIs.

Neurodevelopment and social-emotional development for infants at 18 months of age was measured using the Bayley Scales of Infant and Toddler Development and Infant-Toddler Social and Emotional Assessment. The mean age of the mothers was 35 years; maternal distress was assessed between 24 and 40 weeks’ gestation using validated self-report questionnaires. Parenting stress was assessed at the 18-month infant testing using the Parenting Stress Index-Short Form.

Overall, prenatal maternal stress was negatively associated with infant cognitive performance (P = .01) at 18 months, mediated by fetal left hippocampal volume.

In addition, increased fetal cortical local gyrification index and sulcal depth measured during reported times of prenatal maternal distress were associated with significantly poorer social-emotional scores and competence scores at age 18 months. The beta coefficients for local gyrification index and sulcal depth were –54.62 and –14.22, respectively, for social-emotional and competence scores, –24.01 and –7.53, respectively; P values were P < .001, P < .002, P = .003, P < .001, respectively.

“Increased cortical gyrification has been suggested in children with dyslexia and autism, and sulcal depth has been associated with the severity of impaired performance on working memory and executive function in adults with schizophrenia,” the researchers wrote in their discussion of the findings.

The current study “extends our previous findings and suggests a critical role for disturbances in emerging fetal cerebral cortical folding development in mediating the association between prenatal maternal distress and neurodevelopmental problems that later manifest in infancy,” they explained.

The researchers also found that prenatal maternal anxiety, stress, and depression were positively associated with all measures of parenting stress at the 18-month testing visit.

The study findings were limited by several factors including the use of self-reports for both maternal distress and infant social-emotional assessment, despite the use of validated questionnaires, and the fact that assessment of maternal distress at specific times may not reflect the entire pregnancy, the researchers noted. Other potential limitations included the inability to use some MRI data because of fetal movement and the homogenous population of relatively highly educated women with access to health care that may not reflect other areas, they said.

“Identifying early brain developmental biomarkers may help improve the identification of infants at risk for later neurodevelopmental impairment who might benefit from early targeted interventions,” the researchers concluded.
 

Technology enhances health and disease models

The effect of the prenatal period on future well-being is recognized, but the current study makes “substantial contributions to prenatal programming science, with implications for ways to transform the prenatal care ecosystem for two-generation impact,” Catherine Monk, PhD, and Cristina R. Fernández, MD, both of Columbia University, New York, wrote in an accompanying editorial.

John Abbott/Columbia University

The developmental origins of health and disease (DOHaD) conceptual model introduced by Dr. David Barker in 1995 were later applied to show that maternal stress, depression, and anxiety affected child prenatal and future development, they said. However, the current study uses cutting-edge neuroscience to directly assess developing fetal brains. The finding of reduced cognitive functioning at 18 months associated with maternal stress is consistent with other findings, they noted.

“Finding an association between maternal prenatal stress and infant cognitive outcomes in the setting of what may be modest stress relative to that of a low-resourced or historically marginalized sample underscores the importance of this research; presumably, with higher stress, and greater social determinants of health burden, the effect sizes would be even greater and of greater concern,” they said.

However, studies such as the current one “have the potential to transform the prenatal and postpartum care ecosystems,” by encouraging a whole-person approach to the care of pregnant women, including attention to mental well-being and quality of life, they emphasized.
 

COVID-19 stress considerations

In a separate study published in Communications Medicine, Yuan-Chiao Lu, MD, also of Children’s National Hospital in Washington, and colleagues found a similar effect of maternal stress on fetal brain development.

The researchers imaged the brains of fetuses before and during the COVID-19 pandemic and interviewed mothers about any distress they experienced during pregnancy.

The study population included 65 women with known COVID-19 exposures who underwent 92 fetal MRIs and 137 prepandemic controls who underwent 182 fetal MRIs. Maternal distress was measured via the Spielberger State Anxiety Inventory, Spielberger Trait Anxiety Inventory, Perceived Stress Scale, and Edinburgh Postnatal Depression Scale.

Overall, scores on measures of stress and depression were significantly higher for women in the pandemic group compared with controls. Of the 173 women for whom maternal distress measures were available, 28% of the prepandemic group and 52% of the pandemic group met criteria for elevated maternal psychological distress, defined as above the threshold for distress on any one of the four measures.

After the researchers controlled for maternal distress, MRI data showed decreases in fetal white matter and in hippocampal and cerebellar volumes in fetuses in the pandemic group compared with controls.

Other signs of impaired brain development were similar to those seen in the JAMA Network Open study, including decreased cortical surface area and local gyrification index, as well as reduced sulcal depth in multiple brain lobes, indicating delayed cerebral cortical gyrification.

The second study was limited by a lack of data on other lifestyle changes during the pandemic that might influence maternal health and fetal development, the researchers noted. Other limitations were the possible lack of generalizability to a range of racial and ethnic populations and geographic areas outside of Washington, and the inability to control for unknown COVID-19 exposures or subclinical infections in controls, they said.

However, the results support findings from previous studies, and provide a unique opportunity to study the effect of prenatal stress on early development, as well as a chance to implement “novel and timely interventions,” the researchers wrote.

“Monitoring the COVID generation of infants for long-term cognitive and health outcomes after birth is warranted and currently underway,” and continued research may inform preventive strategies for pregnant women experiencing multiple stressors beyond the pandemic, they concluded.


 

Interpret pandemic effect with caution

“Research studies, as well as our own daily experiences, have made it abundantly clear that stress is on the rise as a consequence of the COVID-19 pandemic,” said editorial author Dr. Monk, who commented on the second study in an interview. “This is an important public health question: Early identification of pandemic effects on child development can help garner the necessary resources to intervene early, dramatically increasing the likelihood of improving that child’s developmental trajectory,” she said.

“The pandemic is an unprecedented experience that has widespread impact on people’s lives, how could it not also alter gestational biology and the developing brain? That being said, we need to be cautious in that we do not yet know the functional implications of these brain changes for longer-term development,” Dr. Monk said. “Also, we do not know what aspects of women’s pandemic-affected lives had an influence on fetal brain development. The authors found higher stress in pandemic versus nonpandemic women, but not evidence that distress was the mediating variable relating pregnancy during the pandemic to altered brain development,” she explained.

The take-home message for clinicians is to “provide your patients with realistic avenues for neurodevelopmental assessments of their children if they, or you, have concerns,” Dr. Monk said. “However, do not prejudge ‘pandemic babies,’ as not all children will be affected by these potential pandemic effects,” she emphasized. “It is possible to misjudge normal variation in children’s development and unnecessarily raise parents’ anxiety levels. Importantly, this period of brain plasticity means any needed intervention likely can have a big, ameliorating impact,” she added.

“We need follow-up studies looking at pandemic effects on prenatal and postnatal development and what factors protect the fetus and birthing person from the negative influences,” she said.

The JAMA study was supported by the National Institutes of Health and the A. James & Alice B. Clark Foundation. The study in Communications Medicine was supported by the National Institutes of Health, the Intellectual and Developmental Disabilities Research Center, and the A. James & Alice B. Clark Foundation. None of the researchers in either study disclosed conflicts of interest. Dr. Monk disclosed grants from the National Institutes of Health, the Bezos Family Foundation, and the Robin Hood Foundation outside the submitted work.

Babies of mothers who experience significant psychological distress during pregnancy showed evidence of altered brain development in utero and reduced cognitive outcomes at 18 months, based on data from a pair of studies including approximately 300 women.

In a longitudinal study published in JAMA Network Open, Yao Wu, PhD, of Children’s National Hospital, Washington, and colleagues recruited 97 healthy mother-infant dyads between January 2016 and October 2020 at a single center. Of these, 87 underwent two fetal brain imaging studies each, and 10 completed the first MRI visit, for a total of 184 fetal MRIs.

Neurodevelopment and social-emotional development for infants at 18 months of age was measured using the Bayley Scales of Infant and Toddler Development and Infant-Toddler Social and Emotional Assessment. The mean age of the mothers was 35 years; maternal distress was assessed between 24 and 40 weeks’ gestation using validated self-report questionnaires. Parenting stress was assessed at the 18-month infant testing using the Parenting Stress Index-Short Form.

Overall, prenatal maternal stress was negatively associated with infant cognitive performance (P = .01) at 18 months, mediated by fetal left hippocampal volume.

In addition, increased fetal cortical local gyrification index and sulcal depth measured during reported times of prenatal maternal distress were associated with significantly poorer social-emotional scores and competence scores at age 18 months. The beta coefficients for local gyrification index and sulcal depth were –54.62 and –14.22, respectively, for social-emotional and competence scores, –24.01 and –7.53, respectively; P values were P < .001, P < .002, P = .003, P < .001, respectively.

“Increased cortical gyrification has been suggested in children with dyslexia and autism, and sulcal depth has been associated with the severity of impaired performance on working memory and executive function in adults with schizophrenia,” the researchers wrote in their discussion of the findings.

The current study “extends our previous findings and suggests a critical role for disturbances in emerging fetal cerebral cortical folding development in mediating the association between prenatal maternal distress and neurodevelopmental problems that later manifest in infancy,” they explained.

The researchers also found that prenatal maternal anxiety, stress, and depression were positively associated with all measures of parenting stress at the 18-month testing visit.

The study findings were limited by several factors including the use of self-reports for both maternal distress and infant social-emotional assessment, despite the use of validated questionnaires, and the fact that assessment of maternal distress at specific times may not reflect the entire pregnancy, the researchers noted. Other potential limitations included the inability to use some MRI data because of fetal movement and the homogenous population of relatively highly educated women with access to health care that may not reflect other areas, they said.

“Identifying early brain developmental biomarkers may help improve the identification of infants at risk for later neurodevelopmental impairment who might benefit from early targeted interventions,” the researchers concluded.
 

Technology enhances health and disease models

The effect of the prenatal period on future well-being is recognized, but the current study makes “substantial contributions to prenatal programming science, with implications for ways to transform the prenatal care ecosystem for two-generation impact,” Catherine Monk, PhD, and Cristina R. Fernández, MD, both of Columbia University, New York, wrote in an accompanying editorial.

John Abbott/Columbia University

The developmental origins of health and disease (DOHaD) conceptual model introduced by Dr. David Barker in 1995 were later applied to show that maternal stress, depression, and anxiety affected child prenatal and future development, they said. However, the current study uses cutting-edge neuroscience to directly assess developing fetal brains. The finding of reduced cognitive functioning at 18 months associated with maternal stress is consistent with other findings, they noted.

“Finding an association between maternal prenatal stress and infant cognitive outcomes in the setting of what may be modest stress relative to that of a low-resourced or historically marginalized sample underscores the importance of this research; presumably, with higher stress, and greater social determinants of health burden, the effect sizes would be even greater and of greater concern,” they said.

However, studies such as the current one “have the potential to transform the prenatal and postpartum care ecosystems,” by encouraging a whole-person approach to the care of pregnant women, including attention to mental well-being and quality of life, they emphasized.
 

COVID-19 stress considerations

In a separate study published in Communications Medicine, Yuan-Chiao Lu, MD, also of Children’s National Hospital in Washington, and colleagues found a similar effect of maternal stress on fetal brain development.

The researchers imaged the brains of fetuses before and during the COVID-19 pandemic and interviewed mothers about any distress they experienced during pregnancy.

The study population included 65 women with known COVID-19 exposures who underwent 92 fetal MRIs and 137 prepandemic controls who underwent 182 fetal MRIs. Maternal distress was measured via the Spielberger State Anxiety Inventory, Spielberger Trait Anxiety Inventory, Perceived Stress Scale, and Edinburgh Postnatal Depression Scale.

Overall, scores on measures of stress and depression were significantly higher for women in the pandemic group compared with controls. Of the 173 women for whom maternal distress measures were available, 28% of the prepandemic group and 52% of the pandemic group met criteria for elevated maternal psychological distress, defined as above the threshold for distress on any one of the four measures.

After the researchers controlled for maternal distress, MRI data showed decreases in fetal white matter and in hippocampal and cerebellar volumes in fetuses in the pandemic group compared with controls.

Other signs of impaired brain development were similar to those seen in the JAMA Network Open study, including decreased cortical surface area and local gyrification index, as well as reduced sulcal depth in multiple brain lobes, indicating delayed cerebral cortical gyrification.

The second study was limited by a lack of data on other lifestyle changes during the pandemic that might influence maternal health and fetal development, the researchers noted. Other limitations were the possible lack of generalizability to a range of racial and ethnic populations and geographic areas outside of Washington, and the inability to control for unknown COVID-19 exposures or subclinical infections in controls, they said.

However, the results support findings from previous studies, and provide a unique opportunity to study the effect of prenatal stress on early development, as well as a chance to implement “novel and timely interventions,” the researchers wrote.

“Monitoring the COVID generation of infants for long-term cognitive and health outcomes after birth is warranted and currently underway,” and continued research may inform preventive strategies for pregnant women experiencing multiple stressors beyond the pandemic, they concluded.


 

Interpret pandemic effect with caution

“Research studies, as well as our own daily experiences, have made it abundantly clear that stress is on the rise as a consequence of the COVID-19 pandemic,” said editorial author Dr. Monk, who commented on the second study in an interview. “This is an important public health question: Early identification of pandemic effects on child development can help garner the necessary resources to intervene early, dramatically increasing the likelihood of improving that child’s developmental trajectory,” she said.

“The pandemic is an unprecedented experience that has widespread impact on people’s lives, how could it not also alter gestational biology and the developing brain? That being said, we need to be cautious in that we do not yet know the functional implications of these brain changes for longer-term development,” Dr. Monk said. “Also, we do not know what aspects of women’s pandemic-affected lives had an influence on fetal brain development. The authors found higher stress in pandemic versus nonpandemic women, but not evidence that distress was the mediating variable relating pregnancy during the pandemic to altered brain development,” she explained.

The take-home message for clinicians is to “provide your patients with realistic avenues for neurodevelopmental assessments of their children if they, or you, have concerns,” Dr. Monk said. “However, do not prejudge ‘pandemic babies,’ as not all children will be affected by these potential pandemic effects,” she emphasized. “It is possible to misjudge normal variation in children’s development and unnecessarily raise parents’ anxiety levels. Importantly, this period of brain plasticity means any needed intervention likely can have a big, ameliorating impact,” she added.

“We need follow-up studies looking at pandemic effects on prenatal and postnatal development and what factors protect the fetus and birthing person from the negative influences,” she said.

The JAMA study was supported by the National Institutes of Health and the A. James & Alice B. Clark Foundation. The study in Communications Medicine was supported by the National Institutes of Health, the Intellectual and Developmental Disabilities Research Center, and the A. James & Alice B. Clark Foundation. None of the researchers in either study disclosed conflicts of interest. Dr. Monk disclosed grants from the National Institutes of Health, the Bezos Family Foundation, and the Robin Hood Foundation outside the submitted work.

Less than 10% of parents followed recommended safe sleep practices for their infants aged 12 months and younger at both sleep onset and after nighttime waking, based on data from a survey of 1,500 parents published in Pediatrics.

Sleep-related death remains a major cause of infant mortality in the United States despite the early success of public health campaigns for safe sleep practices, such as “Back to Sleep,” and many parents persist in unsafe practices such as prone positioning and bed-sharing, Mersine A. Bryan, MD, of the University of Washington, Seattle, and colleagues wrote. “Though nighttime waking is common for infants, less attention has been paid to the safety of second-sleep practices.”

To examine the prevalence and safety of infant second-sleep practices, the researchers used a cross-sectional online survey to collect information on sleep practices from parents of infants aged 12 months and younger; 74% of the respondents were female, 65% were White, 12% were Black, and 17% were Hispanic. The mean age of the infants was 6.6 months, and 24% were aged 3 months and younger.

The survey included parent reports of three safe sleep practices based on the American Academy of Pediatrics 2016 Safe Infant Sleep Guidelines: supine infant sleep position, use of a separate sleep space (vs. bed sharing), and use of an approved surface/safe location (such as a bassinet, crib, cradle, or play yard vs. an adult bed).

Parents were asked to report sleep practices at sleep onset and at nighttime waking, and the researchers used a composite score to determine safe practices were met at each of these two time points.

Of the 1,500 participants, 581 (39%), reported any second-sleep practice. Of the 482 who reported on all three sleep practices at both time points, 29% met all three safe sleep criteria at sleep onset and 9% met all three safe sleep criteria at sleep onset and nighttime waking.

Of the parents who reported second sleep practices, 39% reported changes in practice after nighttime waking from sleep onset. Significantly more parents who switched practices between sleep onset and nighttime waking shifted from a safer to a less safe practice, the researchers noted.

For positioning, 67% of respondents overall reported placing infants on their backs at sleep onset. Among the 564 who reported a second sleep position, 42% placed infants on their backs again; 13% switched from supine to nonsupine positions and 7% changed from nonsupine to supine.

For sleep spaces, 72% of participants overall reported a separate sleep space for infants at sleep onset. Of the 508 who reported on second-sleep spaces, 54% kept infants in a separate space after nighttime waking, 18% shifted to a shared space after nighttime waking. Of those in shared spaces at sleep onset, 8% shifted to separate spaces after nighttime waking.

For sleep location, 71% of respondents overall used an approved sleep surface at sleep onset. Of the 560 who reported sleep location at both time points, 42% remained in a safe location after nighttime waking, while 30% were moved from a safe to an unsafe location, and 10% of those in an unsafe location were moved from an unsafe to a safe location.

In a multivariate analysis, the researchers examined the demographics associated with changes in sleep practice after nighttime waking. Parents younger than 25 years, first-time parents, those who identified as Black non-Hispanic or Hispanic, smokers, and those with preterm infants (less than 37 weeks’ gestation) were more likely to change sleep practices after nighttime waking. However, parents who reported a safe sleep practice at sleep onset were more likely to do so after nighttime waking.

“We hypothesize that expansion of existing strategies to promote infant safe sleep practices to include sleep practices after nighttime waking can have a positive impact on infant safe sleep,” the researchers wrote.

The study findings were limited by several factors including the use of an online survey, which limited the study population to those with internet and computer access, and the reliance on self-reports and only two time points, the researchers noted. Other limitations included the inclusion of only three of the AAP sleep recommendations, and the inclusion of only English speakers.

However, the results were strengthened by the large, diverse, and geographically representative sample of parents.

“When advising families about infant sleep, pediatricians should discuss nighttime wakings with parents because they are common and reinforce the need for safe sleep practices every time,” the researchers noted.
 

Increase opportunities for education

The current study is important because infants continue to die or experience life-long catastrophic health outcomes as a result of not following safe sleep practices, Cathy Haut, DNP, CPNP-AC, CPNP-PC, a pediatric nurse practitioner in Rehoboth Beach, Del., said in an interview.

“I am not surprised by the study findings,” said Dr. Haut, who was not involved in the study. “As a pediatric nurse practitioner for over 35 years, I see infant sleep as a continuing challenge for families. In today’s fast-paced world, multiple priorities leave parents few resources for managing their own well-being, with adequate sleep being one health requirement that is often not met for them.”

To improve safe sleep practices, “it is imperative for health care providers in any setting to address safe sleep practices for infants and children,” said Dr. Haut. “In addition to safety, opportunity for adequate hours of sleep is also important.” She acknowledged that, “in the office setting, time is a huge barrier to completing comprehensive anticipatory guidance. When parents ask questions about sleep, they are often doing everything they can to physically make it through the night with a crying infant. Enforcing safe practices at this point is extremely difficult.”

However, some opportunities for safe sleep education include the prenatal period when parents can take time to listen and plan, not just for feeding preferences but for safe infant sleep practices, Dr. Haut noted.

“When sleep is a problem, families can be invited back to the office for additional counseling and education, which allows more time than within a scheduled health visit,” Dr. Haut emphasized. “Finally, enhanced public awareness is an aspect of learning. In my career I have seen the devastating results of suffocation while cosleeping as well as injuries from falling from a bed or inappropriate sleeping space, and other poor outcomes from inadequate support for safe sleep habits.”

As for additional research, studies are needed to include larger populations and “to further quantify positive outcomes of following safe sleeping practices,” said Dr. Haut. The results of these studies should be made available to the general public, not only to health care professionals. 

The study was supported by Seattle Children’s Research Institute. The researchers had no financial conflicts to disclose. Dr. Haut had no financial conflicts to disclose and serves on the editorial advisory board of Pediatric News.

Less than 10% of parents followed recommended safe sleep practices for their infants aged 12 months and younger at both sleep onset and after nighttime waking, based on data from a survey of 1,500 parents published in Pediatrics.

Sleep-related death remains a major cause of infant mortality in the United States despite the early success of public health campaigns for safe sleep practices, such as “Back to Sleep,” and many parents persist in unsafe practices such as prone positioning and bed-sharing, Mersine A. Bryan, MD, of the University of Washington, Seattle, and colleagues wrote. “Though nighttime waking is common for infants, less attention has been paid to the safety of second-sleep practices.”

To examine the prevalence and safety of infant second-sleep practices, the researchers used a cross-sectional online survey to collect information on sleep practices from parents of infants aged 12 months and younger; 74% of the respondents were female, 65% were White, 12% were Black, and 17% were Hispanic. The mean age of the infants was 6.6 months, and 24% were aged 3 months and younger.

The survey included parent reports of three safe sleep practices based on the American Academy of Pediatrics 2016 Safe Infant Sleep Guidelines: supine infant sleep position, use of a separate sleep space (vs. bed sharing), and use of an approved surface/safe location (such as a bassinet, crib, cradle, or play yard vs. an adult bed).

Parents were asked to report sleep practices at sleep onset and at nighttime waking, and the researchers used a composite score to determine safe practices were met at each of these two time points.

Of the 1,500 participants, 581 (39%), reported any second-sleep practice. Of the 482 who reported on all three sleep practices at both time points, 29% met all three safe sleep criteria at sleep onset and 9% met all three safe sleep criteria at sleep onset and nighttime waking.

Of the parents who reported second sleep practices, 39% reported changes in practice after nighttime waking from sleep onset. Significantly more parents who switched practices between sleep onset and nighttime waking shifted from a safer to a less safe practice, the researchers noted.

For positioning, 67% of respondents overall reported placing infants on their backs at sleep onset. Among the 564 who reported a second sleep position, 42% placed infants on their backs again; 13% switched from supine to nonsupine positions and 7% changed from nonsupine to supine.

For sleep spaces, 72% of participants overall reported a separate sleep space for infants at sleep onset. Of the 508 who reported on second-sleep spaces, 54% kept infants in a separate space after nighttime waking, 18% shifted to a shared space after nighttime waking. Of those in shared spaces at sleep onset, 8% shifted to separate spaces after nighttime waking.

For sleep location, 71% of respondents overall used an approved sleep surface at sleep onset. Of the 560 who reported sleep location at both time points, 42% remained in a safe location after nighttime waking, while 30% were moved from a safe to an unsafe location, and 10% of those in an unsafe location were moved from an unsafe to a safe location.

In a multivariate analysis, the researchers examined the demographics associated with changes in sleep practice after nighttime waking. Parents younger than 25 years, first-time parents, those who identified as Black non-Hispanic or Hispanic, smokers, and those with preterm infants (less than 37 weeks’ gestation) were more likely to change sleep practices after nighttime waking. However, parents who reported a safe sleep practice at sleep onset were more likely to do so after nighttime waking.

“We hypothesize that expansion of existing strategies to promote infant safe sleep practices to include sleep practices after nighttime waking can have a positive impact on infant safe sleep,” the researchers wrote.

The study findings were limited by several factors including the use of an online survey, which limited the study population to those with internet and computer access, and the reliance on self-reports and only two time points, the researchers noted. Other limitations included the inclusion of only three of the AAP sleep recommendations, and the inclusion of only English speakers.

However, the results were strengthened by the large, diverse, and geographically representative sample of parents.

“When advising families about infant sleep, pediatricians should discuss nighttime wakings with parents because they are common and reinforce the need for safe sleep practices every time,” the researchers noted.
 

Increase opportunities for education

The current study is important because infants continue to die or experience life-long catastrophic health outcomes as a result of not following safe sleep practices, Cathy Haut, DNP, CPNP-AC, CPNP-PC, a pediatric nurse practitioner in Rehoboth Beach, Del., said in an interview.

“I am not surprised by the study findings,” said Dr. Haut, who was not involved in the study. “As a pediatric nurse practitioner for over 35 years, I see infant sleep as a continuing challenge for families. In today’s fast-paced world, multiple priorities leave parents few resources for managing their own well-being, with adequate sleep being one health requirement that is often not met for them.”

To improve safe sleep practices, “it is imperative for health care providers in any setting to address safe sleep practices for infants and children,” said Dr. Haut. “In addition to safety, opportunity for adequate hours of sleep is also important.” She acknowledged that, “in the office setting, time is a huge barrier to completing comprehensive anticipatory guidance. When parents ask questions about sleep, they are often doing everything they can to physically make it through the night with a crying infant. Enforcing safe practices at this point is extremely difficult.”

However, some opportunities for safe sleep education include the prenatal period when parents can take time to listen and plan, not just for feeding preferences but for safe infant sleep practices, Dr. Haut noted.

“When sleep is a problem, families can be invited back to the office for additional counseling and education, which allows more time than within a scheduled health visit,” Dr. Haut emphasized. “Finally, enhanced public awareness is an aspect of learning. In my career I have seen the devastating results of suffocation while cosleeping as well as injuries from falling from a bed or inappropriate sleeping space, and other poor outcomes from inadequate support for safe sleep habits.”

As for additional research, studies are needed to include larger populations and “to further quantify positive outcomes of following safe sleeping practices,” said Dr. Haut. The results of these studies should be made available to the general public, not only to health care professionals. 

The study was supported by Seattle Children’s Research Institute. The researchers had no financial conflicts to disclose. Dr. Haut had no financial conflicts to disclose and serves on the editorial advisory board of Pediatric News.

Less than 10% of parents followed recommended safe sleep practices for their infants aged 12 months and younger at both sleep onset and after nighttime waking, based on data from a survey of 1,500 parents published in Pediatrics.

Sleep-related death remains a major cause of infant mortality in the United States despite the early success of public health campaigns for safe sleep practices, such as “Back to Sleep,” and many parents persist in unsafe practices such as prone positioning and bed-sharing, Mersine A. Bryan, MD, of the University of Washington, Seattle, and colleagues wrote. “Though nighttime waking is common for infants, less attention has been paid to the safety of second-sleep practices.”

To examine the prevalence and safety of infant second-sleep practices, the researchers used a cross-sectional online survey to collect information on sleep practices from parents of infants aged 12 months and younger; 74% of the respondents were female, 65% were White, 12% were Black, and 17% were Hispanic. The mean age of the infants was 6.6 months, and 24% were aged 3 months and younger.

The survey included parent reports of three safe sleep practices based on the American Academy of Pediatrics 2016 Safe Infant Sleep Guidelines: supine infant sleep position, use of a separate sleep space (vs. bed sharing), and use of an approved surface/safe location (such as a bassinet, crib, cradle, or play yard vs. an adult bed).

Parents were asked to report sleep practices at sleep onset and at nighttime waking, and the researchers used a composite score to determine safe practices were met at each of these two time points.

Of the 1,500 participants, 581 (39%), reported any second-sleep practice. Of the 482 who reported on all three sleep practices at both time points, 29% met all three safe sleep criteria at sleep onset and 9% met all three safe sleep criteria at sleep onset and nighttime waking.

Of the parents who reported second sleep practices, 39% reported changes in practice after nighttime waking from sleep onset. Significantly more parents who switched practices between sleep onset and nighttime waking shifted from a safer to a less safe practice, the researchers noted.

For positioning, 67% of respondents overall reported placing infants on their backs at sleep onset. Among the 564 who reported a second sleep position, 42% placed infants on their backs again; 13% switched from supine to nonsupine positions and 7% changed from nonsupine to supine.

For sleep spaces, 72% of participants overall reported a separate sleep space for infants at sleep onset. Of the 508 who reported on second-sleep spaces, 54% kept infants in a separate space after nighttime waking, 18% shifted to a shared space after nighttime waking. Of those in shared spaces at sleep onset, 8% shifted to separate spaces after nighttime waking.

For sleep location, 71% of respondents overall used an approved sleep surface at sleep onset. Of the 560 who reported sleep location at both time points, 42% remained in a safe location after nighttime waking, while 30% were moved from a safe to an unsafe location, and 10% of those in an unsafe location were moved from an unsafe to a safe location.

In a multivariate analysis, the researchers examined the demographics associated with changes in sleep practice after nighttime waking. Parents younger than 25 years, first-time parents, those who identified as Black non-Hispanic or Hispanic, smokers, and those with preterm infants (less than 37 weeks’ gestation) were more likely to change sleep practices after nighttime waking. However, parents who reported a safe sleep practice at sleep onset were more likely to do so after nighttime waking.

“We hypothesize that expansion of existing strategies to promote infant safe sleep practices to include sleep practices after nighttime waking can have a positive impact on infant safe sleep,” the researchers wrote.

The study findings were limited by several factors including the use of an online survey, which limited the study population to those with internet and computer access, and the reliance on self-reports and only two time points, the researchers noted. Other limitations included the inclusion of only three of the AAP sleep recommendations, and the inclusion of only English speakers.

However, the results were strengthened by the large, diverse, and geographically representative sample of parents.

“When advising families about infant sleep, pediatricians should discuss nighttime wakings with parents because they are common and reinforce the need for safe sleep practices every time,” the researchers noted.
 

Increase opportunities for education

The current study is important because infants continue to die or experience life-long catastrophic health outcomes as a result of not following safe sleep practices, Cathy Haut, DNP, CPNP-AC, CPNP-PC, a pediatric nurse practitioner in Rehoboth Beach, Del., said in an interview.

“I am not surprised by the study findings,” said Dr. Haut, who was not involved in the study. “As a pediatric nurse practitioner for over 35 years, I see infant sleep as a continuing challenge for families. In today’s fast-paced world, multiple priorities leave parents few resources for managing their own well-being, with adequate sleep being one health requirement that is often not met for them.”

To improve safe sleep practices, “it is imperative for health care providers in any setting to address safe sleep practices for infants and children,” said Dr. Haut. “In addition to safety, opportunity for adequate hours of sleep is also important.” She acknowledged that, “in the office setting, time is a huge barrier to completing comprehensive anticipatory guidance. When parents ask questions about sleep, they are often doing everything they can to physically make it through the night with a crying infant. Enforcing safe practices at this point is extremely difficult.”

However, some opportunities for safe sleep education include the prenatal period when parents can take time to listen and plan, not just for feeding preferences but for safe infant sleep practices, Dr. Haut noted.

“When sleep is a problem, families can be invited back to the office for additional counseling and education, which allows more time than within a scheduled health visit,” Dr. Haut emphasized. “Finally, enhanced public awareness is an aspect of learning. In my career I have seen the devastating results of suffocation while cosleeping as well as injuries from falling from a bed or inappropriate sleeping space, and other poor outcomes from inadequate support for safe sleep habits.”

As for additional research, studies are needed to include larger populations and “to further quantify positive outcomes of following safe sleeping practices,” said Dr. Haut. The results of these studies should be made available to the general public, not only to health care professionals. 

The study was supported by Seattle Children’s Research Institute. The researchers had no financial conflicts to disclose. Dr. Haut had no financial conflicts to disclose and serves on the editorial advisory board of Pediatric News.

Here we go again, and again, and again.

There just aren’t enough tears, and before the bodies of 19 small children are identified, the political noise starts up. Mass shootings are a part of the American landscape, but when they happen at schools, we all feel a distinct sense of violation and gaping grief. Those children are so innocent, so deserving of a right to live their lives, hold their place with their families, create their own legacies, and die of natural causes at a ripe old age. And those children could have been our children. There was nothing special about them; they were just sent to school that day like every child who is sent to school every day.

Here is how the politics goes: The Republicans will blame the Democrats and the Democrats will blame the Republicans. Is Rachel Maddow at fault, or is it Tucker Carlson? Social media accounts blamed both of them for the racially motivated mass murder in a Buffalo grocery store on May 14.

Mass murders were previously defined as a shooting where four or more victims are killed, excluding the shooter, in a public place that is not related to the commission of another crime. In 2012, the definition was changed to include events with three victims. This definition excludes gang violence and the murder of family members.

When it comes to explaining mass murder, the camps divide: They are the result of some combination of mental illness, easy access to firearms, and terrorism and hate. For psychiatry, there is a unique place in the argument – half of all mass shooters have exhibited signs or symptoms of psychiatric illness, and for those who want to deflect the issue away from issues related to the regulation of firearms, it becomes easy to blame “mental illness,” as though that explains it all. Either the gunman “snapped” in such a way that no one could have predicted, or the mental health system is at fault for not preventing it.

There are many ways to be emotionally disturbed; mental illness is only one of them, and there is no psychiatric diagnosis that includes the symptom of shooting strangers, or shooting children. The vast majority of people, including nearly all psychiatrists, will never know someone who perpetrates a mass shooting. 

Can mental illness give us any insight into these senseless killings? Sometimes, but rarely. Take John Hinckley Jr., who shot President Ronald Reagan as a means to impress actress Jodie Foster. Sometimes these killings are motivated by delusional beliefs. But the planning and preparation that goes into most mass shootings involves a degree of organization and forethought that we don’t typically see in those with severe psychotic disorders.

The other psychological explanation that satisfies some of a nonmedical population is that these killers “just snap.” This, too, is a term that is not included in our diagnostic vocabulary, but it remains a way for some to explain that which can’t be explained. If mental illness, however, is the cause of mass murders, then more stringent gun control is unnecessary. Every state already has a mechanism to prevent those with criminal and specified psychiatric histories from buying legal firearms, and it may be inevitable that these screens are not perfect. 

The next line of political thinking moves to the psychiatric “if only.” If only there were more state hospital beds and if only it were easier to compel people with psychiatric disorders to get treatment against their will, then we could eliminate these crimes. The Virginia tech shooter was mandated to get outpatient psychiatric treatment after a brief hospitalization, yet he never went and there was no mechanism in place to track him.

In cases where a person with a psychotic illness has a history of repeated violent episodes after stopping medications, it does make sense to mandate treatment, not because they are likely to shoot strangers, but because some people do become violent when they are ill and mental illness is believed to play a role in 10% of murders.

Mass murders remain rare, and while advocates for legislation that would make it easier to mandate involuntary care have cited violence prevention as a reason, it is hard to imagine that we would force people to get care because they “might” commit such a crime – unless there was convincing evidence that someone was at risk of committing such a heinous act.

For those who oppose stronger gun control laws, the “what if” may circulate around the need for even more firearms. What if teachers carried guns? What if schools were more heavily policed? What if the criminals were made to be afraid?

We are left with the fact that other countries do not see these numbers of mass shooting events, yet mental illness is ubiquitous. While the presence of psychiatric disorders does little to explain school shootings, we still have no understanding of what motivated the Sandy Hook killer, and it remains to be seen what we will come to understand about the gunman in Uvalde, Texas.

Mental illness is not unique to the United States; however, the number of available firearms is. In a country of 323 million people (including children and people who live in institutions where they have no access to firearms), there are estimated to be over 400 million guns in the United States, 98% of which are owned by civilians. 

Hate crimes and terrorism are another explanation for mass murders. In these instances, the gunman makes his motive obvious: There are social media announcements, or the site of the shooting is a synagogue, a mosque, or a location where the victims are of a specific race or religion. But hate may come out of a psychotic illness, and easy access to firearms allows for these crimes to continue. 

Firearms are now the No. 1 cause of mortality in children. Very few of these deaths are the result of mass murders. Many more are from accidental deaths, targeted crime, or suicide. Still, school shootings rip at our hearts. Neither the victims nor their grieving families have any role in the act, and suffering leaves its mark on families, communities, and all of us.
 

Are there answers?

In many states, physicians can now request emergency removal of firearms from the home of someone who is both mentally ill and threatening either suicide or homicide. During the era when high-capacity firearms were banned, from 1994 to 2004, mass murders decreased in our country. While most gunmen use legal firearms they have purchased, I would contend that “smart guns” – firearms that allow only the legal owner to operate them based on biometrics – would prevent some mass shootings and many accidents, crimes, and suicides. Universal background checks and tracking gun purchases in the way we monitor controlled medications, or even Sudafed, might allow authorities to predict who might be at risk of committing these heinous acts.

In his newly released book, Trigger Points: Inside the Mission to Stop Mass Murders in America, journalist Mark Follman argues for a proactive community approach using threat assessment methods and providing wraparound services to those who are deemed to be at risk for violent acts. Mr. Follman’s voice is one of the few out there saying that these events are not random and are, in fact, preventable.

In psychiatry, we struggle with school shootings such as the one we just saw in Uvalde. Our own hearts ache as we hold our children close and empathize with the loss of strangers who have been through the unthinkable. We help our patients as they process their emotions. And we wonder whether any of our patients might ever do anything so horrific. The feelings get complicated, the sadness and anger intermingle while the frustration builds, and we are left with our fears and the hope that if that very rare person were to walk through our office door, we would know what to do.

Dr. Miller is a coauthor of Committed: The Battle Over Involuntary Psychiatric Care (Johns Hopkins University Press, 2016). She has a private practice and is assistant professor of psychiatry and behavioral sciences at Johns Hopkins in Baltimore. A version of this article first appeared on Medscape.com.

Here we go again, and again, and again.

There just aren’t enough tears, and before the bodies of 19 small children are identified, the political noise starts up. Mass shootings are a part of the American landscape, but when they happen at schools, we all feel a distinct sense of violation and gaping grief. Those children are so innocent, so deserving of a right to live their lives, hold their place with their families, create their own legacies, and die of natural causes at a ripe old age. And those children could have been our children. There was nothing special about them; they were just sent to school that day like every child who is sent to school every day.

Here is how the politics goes: The Republicans will blame the Democrats and the Democrats will blame the Republicans. Is Rachel Maddow at fault, or is it Tucker Carlson? Social media accounts blamed both of them for the racially motivated mass murder in a Buffalo grocery store on May 14.

Mass murders were previously defined as a shooting where four or more victims are killed, excluding the shooter, in a public place that is not related to the commission of another crime. In 2012, the definition was changed to include events with three victims. This definition excludes gang violence and the murder of family members.

When it comes to explaining mass murder, the camps divide: They are the result of some combination of mental illness, easy access to firearms, and terrorism and hate. For psychiatry, there is a unique place in the argument – half of all mass shooters have exhibited signs or symptoms of psychiatric illness, and for those who want to deflect the issue away from issues related to the regulation of firearms, it becomes easy to blame “mental illness,” as though that explains it all. Either the gunman “snapped” in such a way that no one could have predicted, or the mental health system is at fault for not preventing it.

There are many ways to be emotionally disturbed; mental illness is only one of them, and there is no psychiatric diagnosis that includes the symptom of shooting strangers, or shooting children. The vast majority of people, including nearly all psychiatrists, will never know someone who perpetrates a mass shooting. 

Can mental illness give us any insight into these senseless killings? Sometimes, but rarely. Take John Hinckley Jr., who shot President Ronald Reagan as a means to impress actress Jodie Foster. Sometimes these killings are motivated by delusional beliefs. But the planning and preparation that goes into most mass shootings involves a degree of organization and forethought that we don’t typically see in those with severe psychotic disorders.

The other psychological explanation that satisfies some of a nonmedical population is that these killers “just snap.” This, too, is a term that is not included in our diagnostic vocabulary, but it remains a way for some to explain that which can’t be explained. If mental illness, however, is the cause of mass murders, then more stringent gun control is unnecessary. Every state already has a mechanism to prevent those with criminal and specified psychiatric histories from buying legal firearms, and it may be inevitable that these screens are not perfect. 

The next line of political thinking moves to the psychiatric “if only.” If only there were more state hospital beds and if only it were easier to compel people with psychiatric disorders to get treatment against their will, then we could eliminate these crimes. The Virginia tech shooter was mandated to get outpatient psychiatric treatment after a brief hospitalization, yet he never went and there was no mechanism in place to track him.

In cases where a person with a psychotic illness has a history of repeated violent episodes after stopping medications, it does make sense to mandate treatment, not because they are likely to shoot strangers, but because some people do become violent when they are ill and mental illness is believed to play a role in 10% of murders.

Mass murders remain rare, and while advocates for legislation that would make it easier to mandate involuntary care have cited violence prevention as a reason, it is hard to imagine that we would force people to get care because they “might” commit such a crime – unless there was convincing evidence that someone was at risk of committing such a heinous act.

For those who oppose stronger gun control laws, the “what if” may circulate around the need for even more firearms. What if teachers carried guns? What if schools were more heavily policed? What if the criminals were made to be afraid?

We are left with the fact that other countries do not see these numbers of mass shooting events, yet mental illness is ubiquitous. While the presence of psychiatric disorders does little to explain school shootings, we still have no understanding of what motivated the Sandy Hook killer, and it remains to be seen what we will come to understand about the gunman in Uvalde, Texas.

Mental illness is not unique to the United States; however, the number of available firearms is. In a country of 323 million people (including children and people who live in institutions where they have no access to firearms), there are estimated to be over 400 million guns in the United States, 98% of which are owned by civilians. 

Hate crimes and terrorism are another explanation for mass murders. In these instances, the gunman makes his motive obvious: There are social media announcements, or the site of the shooting is a synagogue, a mosque, or a location where the victims are of a specific race or religion. But hate may come out of a psychotic illness, and easy access to firearms allows for these crimes to continue. 

Firearms are now the No. 1 cause of mortality in children. Very few of these deaths are the result of mass murders. Many more are from accidental deaths, targeted crime, or suicide. Still, school shootings rip at our hearts. Neither the victims nor their grieving families have any role in the act, and suffering leaves its mark on families, communities, and all of us.
 

Are there answers?

In many states, physicians can now request emergency removal of firearms from the home of someone who is both mentally ill and threatening either suicide or homicide. During the era when high-capacity firearms were banned, from 1994 to 2004, mass murders decreased in our country. While most gunmen use legal firearms they have purchased, I would contend that “smart guns” – firearms that allow only the legal owner to operate them based on biometrics – would prevent some mass shootings and many accidents, crimes, and suicides. Universal background checks and tracking gun purchases in the way we monitor controlled medications, or even Sudafed, might allow authorities to predict who might be at risk of committing these heinous acts.

In his newly released book, Trigger Points: Inside the Mission to Stop Mass Murders in America, journalist Mark Follman argues for a proactive community approach using threat assessment methods and providing wraparound services to those who are deemed to be at risk for violent acts. Mr. Follman’s voice is one of the few out there saying that these events are not random and are, in fact, preventable.

In psychiatry, we struggle with school shootings such as the one we just saw in Uvalde. Our own hearts ache as we hold our children close and empathize with the loss of strangers who have been through the unthinkable. We help our patients as they process their emotions. And we wonder whether any of our patients might ever do anything so horrific. The feelings get complicated, the sadness and anger intermingle while the frustration builds, and we are left with our fears and the hope that if that very rare person were to walk through our office door, we would know what to do.

Dr. Miller is a coauthor of Committed: The Battle Over Involuntary Psychiatric Care (Johns Hopkins University Press, 2016). She has a private practice and is assistant professor of psychiatry and behavioral sciences at Johns Hopkins in Baltimore. A version of this article first appeared on Medscape.com.

Here we go again, and again, and again.

There just aren’t enough tears, and before the bodies of 19 small children are identified, the political noise starts up. Mass shootings are a part of the American landscape, but when they happen at schools, we all feel a distinct sense of violation and gaping grief. Those children are so innocent, so deserving of a right to live their lives, hold their place with their families, create their own legacies, and die of natural causes at a ripe old age. And those children could have been our children. There was nothing special about them; they were just sent to school that day like every child who is sent to school every day.

Here is how the politics goes: The Republicans will blame the Democrats and the Democrats will blame the Republicans. Is Rachel Maddow at fault, or is it Tucker Carlson? Social media accounts blamed both of them for the racially motivated mass murder in a Buffalo grocery store on May 14.

Mass murders were previously defined as a shooting where four or more victims are killed, excluding the shooter, in a public place that is not related to the commission of another crime. In 2012, the definition was changed to include events with three victims. This definition excludes gang violence and the murder of family members.

When it comes to explaining mass murder, the camps divide: They are the result of some combination of mental illness, easy access to firearms, and terrorism and hate. For psychiatry, there is a unique place in the argument – half of all mass shooters have exhibited signs or symptoms of psychiatric illness, and for those who want to deflect the issue away from issues related to the regulation of firearms, it becomes easy to blame “mental illness,” as though that explains it all. Either the gunman “snapped” in such a way that no one could have predicted, or the mental health system is at fault for not preventing it.

There are many ways to be emotionally disturbed; mental illness is only one of them, and there is no psychiatric diagnosis that includes the symptom of shooting strangers, or shooting children. The vast majority of people, including nearly all psychiatrists, will never know someone who perpetrates a mass shooting. 

Can mental illness give us any insight into these senseless killings? Sometimes, but rarely. Take John Hinckley Jr., who shot President Ronald Reagan as a means to impress actress Jodie Foster. Sometimes these killings are motivated by delusional beliefs. But the planning and preparation that goes into most mass shootings involves a degree of organization and forethought that we don’t typically see in those with severe psychotic disorders.

The other psychological explanation that satisfies some of a nonmedical population is that these killers “just snap.” This, too, is a term that is not included in our diagnostic vocabulary, but it remains a way for some to explain that which can’t be explained. If mental illness, however, is the cause of mass murders, then more stringent gun control is unnecessary. Every state already has a mechanism to prevent those with criminal and specified psychiatric histories from buying legal firearms, and it may be inevitable that these screens are not perfect. 

The next line of political thinking moves to the psychiatric “if only.” If only there were more state hospital beds and if only it were easier to compel people with psychiatric disorders to get treatment against their will, then we could eliminate these crimes. The Virginia tech shooter was mandated to get outpatient psychiatric treatment after a brief hospitalization, yet he never went and there was no mechanism in place to track him.

In cases where a person with a psychotic illness has a history of repeated violent episodes after stopping medications, it does make sense to mandate treatment, not because they are likely to shoot strangers, but because some people do become violent when they are ill and mental illness is believed to play a role in 10% of murders.

Mass murders remain rare, and while advocates for legislation that would make it easier to mandate involuntary care have cited violence prevention as a reason, it is hard to imagine that we would force people to get care because they “might” commit such a crime – unless there was convincing evidence that someone was at risk of committing such a heinous act.

For those who oppose stronger gun control laws, the “what if” may circulate around the need for even more firearms. What if teachers carried guns? What if schools were more heavily policed? What if the criminals were made to be afraid?

We are left with the fact that other countries do not see these numbers of mass shooting events, yet mental illness is ubiquitous. While the presence of psychiatric disorders does little to explain school shootings, we still have no understanding of what motivated the Sandy Hook killer, and it remains to be seen what we will come to understand about the gunman in Uvalde, Texas.

Mental illness is not unique to the United States; however, the number of available firearms is. In a country of 323 million people (including children and people who live in institutions where they have no access to firearms), there are estimated to be over 400 million guns in the United States, 98% of which are owned by civilians. 

Hate crimes and terrorism are another explanation for mass murders. In these instances, the gunman makes his motive obvious: There are social media announcements, or the site of the shooting is a synagogue, a mosque, or a location where the victims are of a specific race or religion. But hate may come out of a psychotic illness, and easy access to firearms allows for these crimes to continue. 

Firearms are now the No. 1 cause of mortality in children. Very few of these deaths are the result of mass murders. Many more are from accidental deaths, targeted crime, or suicide. Still, school shootings rip at our hearts. Neither the victims nor their grieving families have any role in the act, and suffering leaves its mark on families, communities, and all of us.
 

Are there answers?

In many states, physicians can now request emergency removal of firearms from the home of someone who is both mentally ill and threatening either suicide or homicide. During the era when high-capacity firearms were banned, from 1994 to 2004, mass murders decreased in our country. While most gunmen use legal firearms they have purchased, I would contend that “smart guns” – firearms that allow only the legal owner to operate them based on biometrics – would prevent some mass shootings and many accidents, crimes, and suicides. Universal background checks and tracking gun purchases in the way we monitor controlled medications, or even Sudafed, might allow authorities to predict who might be at risk of committing these heinous acts.

In his newly released book, Trigger Points: Inside the Mission to Stop Mass Murders in America, journalist Mark Follman argues for a proactive community approach using threat assessment methods and providing wraparound services to those who are deemed to be at risk for violent acts. Mr. Follman’s voice is one of the few out there saying that these events are not random and are, in fact, preventable.

In psychiatry, we struggle with school shootings such as the one we just saw in Uvalde. Our own hearts ache as we hold our children close and empathize with the loss of strangers who have been through the unthinkable. We help our patients as they process their emotions. And we wonder whether any of our patients might ever do anything so horrific. The feelings get complicated, the sadness and anger intermingle while the frustration builds, and we are left with our fears and the hope that if that very rare person were to walk through our office door, we would know what to do.

Dr. Miller is a coauthor of Committed: The Battle Over Involuntary Psychiatric Care (Johns Hopkins University Press, 2016). She has a private practice and is assistant professor of psychiatry and behavioral sciences at Johns Hopkins in Baltimore. A version of this article first appeared on Medscape.com.

A cohort study has found increases in mortality rates among women with non-endometrioid uterine carcinoma, despite incident rates that have stabilized. After correction with hysterectomy, mortality risk was about doubled for Black women, compared with White women, and these results could not be explained by differences in cancer subtype or cancer stage at diagnosis. Non-endometroid uterine carcinoma represents 15%-20% of uterine cancers diagnosed and carries a worse prognosis.

“We do not know why non-endometrioid subtypes are disproportionately increasing among all women, nor do we understand why they are so much more common among non-Hispanic Black women. We need more research to identify risk factors and exposures more specifically associated with non-endometrioid cancers to better understand the strong increases in this subtype among all women and the particularly high rates and recent increases in non-Hispanic black women,” said lead author Megan Clarke, PhD, MHS, the study’s lead author and a cancer epidemiologist with the National Cancer Institute.

The study was published online in JAMA Oncology.

“Physicians should be aware that both incidence and mortality rates of non-endometrioid cancers are on the rise. Because these subtypes are rarer than endometrioid uterine cancers, physicians may be less familiar with diagnosing and treating these aggressive types of cancers. Increasing awareness among clinicians and patients regarding the signs and symptoms of uterine cancer (such as postmenopausal bleeding) and the differences in histologic subtypes among racial and ethnic groups may promote earlier diagnosis and timely referral to appropriate treatment,” Dr. Clarke said.

Previous studies based on death certificates found increased mortality, especially in Black women, but they were limited by an inability to link mortality to tumor characteristics. To address this, the researchers linked mortality data to records of 208,587 women diagnosed with uterine cancer between 2000 and 2017, drawn from the U.S. Surveillance, Epidemiology, and End Results (SEER) Program.

Black women represented 9.7% of cases, but they suffered 17.7% of uterine cancer deaths. Overall, mortality from uterine corpus cancer increased by 1.8% per year (95% confidence interval, 1.5%-2.9%). Non-endometroid cancers increased at 2.7% per year (95% CI, 1.8%-3.6%), and this was higher in Asian (3.4%; 95% CI, 0.3%-6.6%), Black (3.5%; 95% CI, 2.2%-4.9%), Hispanic (6.7%; 95% CI, 1.9%-11.8%), and White women (1.5%; 95% CI, 0.6%-2.4%).

Mortality increased 1.8% per year overall for uterine cancer and 2.7% per year for non-endometrioid uterine cancer. There was no increase in mortality seen in endometrioid cancers.

“The concerning rise in deaths from non-endometrioid cancers warrants clinical attention. Our findings suggest that there may be several factors contributing to racial disparities in uterine cancer mortality. Higher mortality rates among non-Hispanic Black women are partly attributable to higher incidence of tumors with aggressive subtypes and advanced stages. However, non-Hispanic Black women in our study who were diagnosed with less aggressive subtypes and early-stage disease also had the highest mortality rates,” said Dr. Clarke.

That suggests that inequities of treatment and high-quality care may be at least partly to blame, since those factors are known to contribute to differences in uterine cancer outcomes. “Other factors including comorbidities, health care facility characteristics, treatment preferences and adherence, patient and provider communication, provider bias, discrimination and structural racism, and potential biologic differences in response to treatment need to be better understood in terms of how they influence racial disparities,” Dr. Clarke said.

Dr. Clarke reported no relevant disclosures.

A cohort study has found increases in mortality rates among women with non-endometrioid uterine carcinoma, despite incident rates that have stabilized. After correction with hysterectomy, mortality risk was about doubled for Black women, compared with White women, and these results could not be explained by differences in cancer subtype or cancer stage at diagnosis. Non-endometroid uterine carcinoma represents 15%-20% of uterine cancers diagnosed and carries a worse prognosis.

“We do not know why non-endometrioid subtypes are disproportionately increasing among all women, nor do we understand why they are so much more common among non-Hispanic Black women. We need more research to identify risk factors and exposures more specifically associated with non-endometrioid cancers to better understand the strong increases in this subtype among all women and the particularly high rates and recent increases in non-Hispanic black women,” said lead author Megan Clarke, PhD, MHS, the study’s lead author and a cancer epidemiologist with the National Cancer Institute.

The study was published online in JAMA Oncology.

“Physicians should be aware that both incidence and mortality rates of non-endometrioid cancers are on the rise. Because these subtypes are rarer than endometrioid uterine cancers, physicians may be less familiar with diagnosing and treating these aggressive types of cancers. Increasing awareness among clinicians and patients regarding the signs and symptoms of uterine cancer (such as postmenopausal bleeding) and the differences in histologic subtypes among racial and ethnic groups may promote earlier diagnosis and timely referral to appropriate treatment,” Dr. Clarke said.

Previous studies based on death certificates found increased mortality, especially in Black women, but they were limited by an inability to link mortality to tumor characteristics. To address this, the researchers linked mortality data to records of 208,587 women diagnosed with uterine cancer between 2000 and 2017, drawn from the U.S. Surveillance, Epidemiology, and End Results (SEER) Program.

Black women represented 9.7% of cases, but they suffered 17.7% of uterine cancer deaths. Overall, mortality from uterine corpus cancer increased by 1.8% per year (95% confidence interval, 1.5%-2.9%). Non-endometroid cancers increased at 2.7% per year (95% CI, 1.8%-3.6%), and this was higher in Asian (3.4%; 95% CI, 0.3%-6.6%), Black (3.5%; 95% CI, 2.2%-4.9%), Hispanic (6.7%; 95% CI, 1.9%-11.8%), and White women (1.5%; 95% CI, 0.6%-2.4%).

Mortality increased 1.8% per year overall for uterine cancer and 2.7% per year for non-endometrioid uterine cancer. There was no increase in mortality seen in endometrioid cancers.

“The concerning rise in deaths from non-endometrioid cancers warrants clinical attention. Our findings suggest that there may be several factors contributing to racial disparities in uterine cancer mortality. Higher mortality rates among non-Hispanic Black women are partly attributable to higher incidence of tumors with aggressive subtypes and advanced stages. However, non-Hispanic Black women in our study who were diagnosed with less aggressive subtypes and early-stage disease also had the highest mortality rates,” said Dr. Clarke.

That suggests that inequities of treatment and high-quality care may be at least partly to blame, since those factors are known to contribute to differences in uterine cancer outcomes. “Other factors including comorbidities, health care facility characteristics, treatment preferences and adherence, patient and provider communication, provider bias, discrimination and structural racism, and potential biologic differences in response to treatment need to be better understood in terms of how they influence racial disparities,” Dr. Clarke said.

Dr. Clarke reported no relevant disclosures.

A cohort study has found increases in mortality rates among women with non-endometrioid uterine carcinoma, despite incident rates that have stabilized. After correction with hysterectomy, mortality risk was about doubled for Black women, compared with White women, and these results could not be explained by differences in cancer subtype or cancer stage at diagnosis. Non-endometroid uterine carcinoma represents 15%-20% of uterine cancers diagnosed and carries a worse prognosis.

“We do not know why non-endometrioid subtypes are disproportionately increasing among all women, nor do we understand why they are so much more common among non-Hispanic Black women. We need more research to identify risk factors and exposures more specifically associated with non-endometrioid cancers to better understand the strong increases in this subtype among all women and the particularly high rates and recent increases in non-Hispanic black women,” said lead author Megan Clarke, PhD, MHS, the study’s lead author and a cancer epidemiologist with the National Cancer Institute.

The study was published online in JAMA Oncology.

“Physicians should be aware that both incidence and mortality rates of non-endometrioid cancers are on the rise. Because these subtypes are rarer than endometrioid uterine cancers, physicians may be less familiar with diagnosing and treating these aggressive types of cancers. Increasing awareness among clinicians and patients regarding the signs and symptoms of uterine cancer (such as postmenopausal bleeding) and the differences in histologic subtypes among racial and ethnic groups may promote earlier diagnosis and timely referral to appropriate treatment,” Dr. Clarke said.

Previous studies based on death certificates found increased mortality, especially in Black women, but they were limited by an inability to link mortality to tumor characteristics. To address this, the researchers linked mortality data to records of 208,587 women diagnosed with uterine cancer between 2000 and 2017, drawn from the U.S. Surveillance, Epidemiology, and End Results (SEER) Program.

Black women represented 9.7% of cases, but they suffered 17.7% of uterine cancer deaths. Overall, mortality from uterine corpus cancer increased by 1.8% per year (95% confidence interval, 1.5%-2.9%). Non-endometroid cancers increased at 2.7% per year (95% CI, 1.8%-3.6%), and this was higher in Asian (3.4%; 95% CI, 0.3%-6.6%), Black (3.5%; 95% CI, 2.2%-4.9%), Hispanic (6.7%; 95% CI, 1.9%-11.8%), and White women (1.5%; 95% CI, 0.6%-2.4%).

Mortality increased 1.8% per year overall for uterine cancer and 2.7% per year for non-endometrioid uterine cancer. There was no increase in mortality seen in endometrioid cancers.

“The concerning rise in deaths from non-endometrioid cancers warrants clinical attention. Our findings suggest that there may be several factors contributing to racial disparities in uterine cancer mortality. Higher mortality rates among non-Hispanic Black women are partly attributable to higher incidence of tumors with aggressive subtypes and advanced stages. However, non-Hispanic Black women in our study who were diagnosed with less aggressive subtypes and early-stage disease also had the highest mortality rates,” said Dr. Clarke.

That suggests that inequities of treatment and high-quality care may be at least partly to blame, since those factors are known to contribute to differences in uterine cancer outcomes. “Other factors including comorbidities, health care facility characteristics, treatment preferences and adherence, patient and provider communication, provider bias, discrimination and structural racism, and potential biologic differences in response to treatment need to be better understood in terms of how they influence racial disparities,” Dr. Clarke said.

Dr. Clarke reported no relevant disclosures.

Moderate consumption of coffee, with or without sugar, is associated with a reduced risk of death, according to prospective cohort study.

Among more than 170,000 people in the United Kingdom, those who drank about two to four cups of coffee a day, with or without sugar, had a lower rate of death than those who didn’t drink coffee, reported lead author Dan Liu, MD, of the department of epidemiology at Southern Medical University, Guangzhou, China.

“Previous observational studies have suggested an association between coffee intake and reduced risk for death, but they did not distinguish between coffee consumed with sugar or artificial sweeteners and coffee consumed without,” Dr. Liu, who is also of the department of public health and preventive medicine, Jinan University, Guangzhou, China, and colleagues wrote in Annals of Internal Medicine.

To learn more, the investigators turned to the UK Biobank, which recruited approximately half a million participants in the United Kingdom between 2006 and 2010 to undergo a variety of questionnaires, interviews, physical measurements, and medical tests. Out of this group, 171,616 participants completed at least one dietary questionnaire and met the criteria for the present study, including lack of cancer or cardiovascular disease upon enrollment.

Results from these questionnaires showed that 55.4% of participants drank coffee without any sweetener, 14.3% drank coffee with sugar, 6.1% drank coffee with artificial sweetener, and 24.2% did not drink coffee at all. Coffee drinkers were further sorted into groups based on how many cups of coffee they drank per day.
 

Coffee drinkers were significantly less likely to die from any cause

Over the course of about 7 years, 3,177 of the participants died, including 1,725 who died from cancer and 628 who died from cardiovascular disease.

After accounting for other factors that might impact risk of death, like lifestyle choices, the investigators found that coffee drinkers were significantly less likely to die from any cause, cardiovascular disease, or cancer, than those who didn’t drink coffee at all. This benefit was observed across types of coffee, including ground, instant, and decaffeinated varieties. The protective effects of coffee were most apparent in people who drank about two to four cups a day, among whom death was about 30% less likely, regardless of whether they added sugar to their coffee or not. Individuals who drank coffee with artificial sweetener did not live significantly longer than those who drank no coffee at all; however, the investigators suggested that this result may have been skewed by higher rates of negative health factors, such as obesity and hypertension, in the artificial sweetener group.

Dr. Liu and colleagues noted that their findings align with previous studies linking coffee consumption with survival. Like those other studies, the present data revealed a “U-shaped” benefit curve, in which moderate coffee consumption was associated with longer life, whereas low or no consumption and high consumption were not.
 

Experts caution against drinking sweetened beverages despite new findings

Although the present findings suggested that adding sugar did not eliminate the health benefits of coffee, Dr. Liu and colleagues still cautioned against sweetened beverages, citing widely known associations between sugar consumption and poor health.

In an accompanying editorial, Christina C. Wee, MD, MPH, deputy editor of Annals of Internal Medicine, pointed out a key detail from the data: the amount of sugar added to coffee in the U.K. study may be dwarfed by the amount consumed by some coffee drinkers across the pond.

“The average dose of added sugar per cup of sweetened coffee [in the study] was only a little over a teaspoon, or about 4 grams,” Dr. Wee wrote. “This is a far cry from the 15 grams of sugar in an 8-ounce cup of caramel macchiato at a popular U.S. coffee chain.”

Still, Dr. Wee, an associate professor of medicine at Harvard Medical School, Boston, and director of the obesity research program in the division of general medicine at Beth Israel Deaconess Medical Center, Boston, suggested that your typical coffee drinker can feel safe in their daily habit.

“The evidence does not suggest a need for most coffee drinkers – particularly those who drink it with no or modest amounts of sugar – to eliminate coffee,” she wrote. “So drink up – but it would be prudent to avoid too many caramel macchiatos while more evidence brews.”

Estefanía Toledo, MD, MPH, PhD, of the department of preventive medicine and public health at the University of Navarra, Pamplona, Spain, offered a similar takeaway.

Courtesy University of Navarra

“For those who enjoy drinking coffee, are not pregnant or lactating, and do not have special health conditions, coffee consumption could be considered part of a healthy lifestyle,” Dr. Toledo said in a written comment. “I would recommend adding as little sugar as possible to coffee until more evidence has been accrued.”

Dr. Toledo, who previously published a study showing a link between coffee and extended survival, noted that moderate coffee consumption has “repeatedly” been associated with lower rates of “several chronic diseases” and death, but there still isn’t enough evidence to recommend coffee for those who don’t already drink it.

More long-term research is needed, Dr. Toledo said, ideally with studies comparing changes in coffee consumption and health outcomes over time. These may not be forthcoming, however, as such trials are “not easy and feasible to conduct.”

David Kao, MD, assistant professor of medicine-cardiology and medical director of the school of medicine at the University of Colorado at Denver, Aurora, said that the study conducted by Dr. Liu and colleagues is a “very well-executed analysis” that strengthens our confidence in the safety of long-term coffee consumption, even for patients with heart disease.

Dr. Kao, who recently published an analysis showing that higher coffee intake is associated with a lower risk of heart failure, refrained from advising anyone to up their coffee quota.

“I remain cautious about stating too strongly that people should increase coffee intake purely to improve survival,” Dr. Kao said in a written comment. “That said, it does not appear harmful to increase it some, until you drink consistently more than six to seven cups per day.”

The study was supported by the National Natural Science Foundation of China, the Young Elite Scientist Sponsorship Program by CAST, the Guangdong Basic and Applied Basic Research Foundation, and others. Dr. Toledo and Dr. Kao disclosed no relevant conflicts of interest.

Moderate consumption of coffee, with or without sugar, is associated with a reduced risk of death, according to prospective cohort study.

Among more than 170,000 people in the United Kingdom, those who drank about two to four cups of coffee a day, with or without sugar, had a lower rate of death than those who didn’t drink coffee, reported lead author Dan Liu, MD, of the department of epidemiology at Southern Medical University, Guangzhou, China.

“Previous observational studies have suggested an association between coffee intake and reduced risk for death, but they did not distinguish between coffee consumed with sugar or artificial sweeteners and coffee consumed without,” Dr. Liu, who is also of the department of public health and preventive medicine, Jinan University, Guangzhou, China, and colleagues wrote in Annals of Internal Medicine.

To learn more, the investigators turned to the UK Biobank, which recruited approximately half a million participants in the United Kingdom between 2006 and 2010 to undergo a variety of questionnaires, interviews, physical measurements, and medical tests. Out of this group, 171,616 participants completed at least one dietary questionnaire and met the criteria for the present study, including lack of cancer or cardiovascular disease upon enrollment.

Results from these questionnaires showed that 55.4% of participants drank coffee without any sweetener, 14.3% drank coffee with sugar, 6.1% drank coffee with artificial sweetener, and 24.2% did not drink coffee at all. Coffee drinkers were further sorted into groups based on how many cups of coffee they drank per day.
 

Coffee drinkers were significantly less likely to die from any cause

Over the course of about 7 years, 3,177 of the participants died, including 1,725 who died from cancer and 628 who died from cardiovascular disease.

After accounting for other factors that might impact risk of death, like lifestyle choices, the investigators found that coffee drinkers were significantly less likely to die from any cause, cardiovascular disease, or cancer, than those who didn’t drink coffee at all. This benefit was observed across types of coffee, including ground, instant, and decaffeinated varieties. The protective effects of coffee were most apparent in people who drank about two to four cups a day, among whom death was about 30% less likely, regardless of whether they added sugar to their coffee or not. Individuals who drank coffee with artificial sweetener did not live significantly longer than those who drank no coffee at all; however, the investigators suggested that this result may have been skewed by higher rates of negative health factors, such as obesity and hypertension, in the artificial sweetener group.

Dr. Liu and colleagues noted that their findings align with previous studies linking coffee consumption with survival. Like those other studies, the present data revealed a “U-shaped” benefit curve, in which moderate coffee consumption was associated with longer life, whereas low or no consumption and high consumption were not.
 

Experts caution against drinking sweetened beverages despite new findings

Although the present findings suggested that adding sugar did not eliminate the health benefits of coffee, Dr. Liu and colleagues still cautioned against sweetened beverages, citing widely known associations between sugar consumption and poor health.

In an accompanying editorial, Christina C. Wee, MD, MPH, deputy editor of Annals of Internal Medicine, pointed out a key detail from the data: the amount of sugar added to coffee in the U.K. study may be dwarfed by the amount consumed by some coffee drinkers across the pond.

“The average dose of added sugar per cup of sweetened coffee [in the study] was only a little over a teaspoon, or about 4 grams,” Dr. Wee wrote. “This is a far cry from the 15 grams of sugar in an 8-ounce cup of caramel macchiato at a popular U.S. coffee chain.”

Still, Dr. Wee, an associate professor of medicine at Harvard Medical School, Boston, and director of the obesity research program in the division of general medicine at Beth Israel Deaconess Medical Center, Boston, suggested that your typical coffee drinker can feel safe in their daily habit.

“The evidence does not suggest a need for most coffee drinkers – particularly those who drink it with no or modest amounts of sugar – to eliminate coffee,” she wrote. “So drink up – but it would be prudent to avoid too many caramel macchiatos while more evidence brews.”

Estefanía Toledo, MD, MPH, PhD, of the department of preventive medicine and public health at the University of Navarra, Pamplona, Spain, offered a similar takeaway.

Courtesy University of Navarra

“For those who enjoy drinking coffee, are not pregnant or lactating, and do not have special health conditions, coffee consumption could be considered part of a healthy lifestyle,” Dr. Toledo said in a written comment. “I would recommend adding as little sugar as possible to coffee until more evidence has been accrued.”

Dr. Toledo, who previously published a study showing a link between coffee and extended survival, noted that moderate coffee consumption has “repeatedly” been associated with lower rates of “several chronic diseases” and death, but there still isn’t enough evidence to recommend coffee for those who don’t already drink it.

More long-term research is needed, Dr. Toledo said, ideally with studies comparing changes in coffee consumption and health outcomes over time. These may not be forthcoming, however, as such trials are “not easy and feasible to conduct.”

David Kao, MD, assistant professor of medicine-cardiology and medical director of the school of medicine at the University of Colorado at Denver, Aurora, said that the study conducted by Dr. Liu and colleagues is a “very well-executed analysis” that strengthens our confidence in the safety of long-term coffee consumption, even for patients with heart disease.

Dr. Kao, who recently published an analysis showing that higher coffee intake is associated with a lower risk of heart failure, refrained from advising anyone to up their coffee quota.

“I remain cautious about stating too strongly that people should increase coffee intake purely to improve survival,” Dr. Kao said in a written comment. “That said, it does not appear harmful to increase it some, until you drink consistently more than six to seven cups per day.”

The study was supported by the National Natural Science Foundation of China, the Young Elite Scientist Sponsorship Program by CAST, the Guangdong Basic and Applied Basic Research Foundation, and others. Dr. Toledo and Dr. Kao disclosed no relevant conflicts of interest.

Moderate consumption of coffee, with or without sugar, is associated with a reduced risk of death, according to prospective cohort study.

Among more than 170,000 people in the United Kingdom, those who drank about two to four cups of coffee a day, with or without sugar, had a lower rate of death than those who didn’t drink coffee, reported lead author Dan Liu, MD, of the department of epidemiology at Southern Medical University, Guangzhou, China.

“Previous observational studies have suggested an association between coffee intake and reduced risk for death, but they did not distinguish between coffee consumed with sugar or artificial sweeteners and coffee consumed without,” Dr. Liu, who is also of the department of public health and preventive medicine, Jinan University, Guangzhou, China, and colleagues wrote in Annals of Internal Medicine.

To learn more, the investigators turned to the UK Biobank, which recruited approximately half a million participants in the United Kingdom between 2006 and 2010 to undergo a variety of questionnaires, interviews, physical measurements, and medical tests. Out of this group, 171,616 participants completed at least one dietary questionnaire and met the criteria for the present study, including lack of cancer or cardiovascular disease upon enrollment.

Results from these questionnaires showed that 55.4% of participants drank coffee without any sweetener, 14.3% drank coffee with sugar, 6.1% drank coffee with artificial sweetener, and 24.2% did not drink coffee at all. Coffee drinkers were further sorted into groups based on how many cups of coffee they drank per day.
 

Coffee drinkers were significantly less likely to die from any cause

Over the course of about 7 years, 3,177 of the participants died, including 1,725 who died from cancer and 628 who died from cardiovascular disease.

After accounting for other factors that might impact risk of death, like lifestyle choices, the investigators found that coffee drinkers were significantly less likely to die from any cause, cardiovascular disease, or cancer, than those who didn’t drink coffee at all. This benefit was observed across types of coffee, including ground, instant, and decaffeinated varieties. The protective effects of coffee were most apparent in people who drank about two to four cups a day, among whom death was about 30% less likely, regardless of whether they added sugar to their coffee or not. Individuals who drank coffee with artificial sweetener did not live significantly longer than those who drank no coffee at all; however, the investigators suggested that this result may have been skewed by higher rates of negative health factors, such as obesity and hypertension, in the artificial sweetener group.

Dr. Liu and colleagues noted that their findings align with previous studies linking coffee consumption with survival. Like those other studies, the present data revealed a “U-shaped” benefit curve, in which moderate coffee consumption was associated with longer life, whereas low or no consumption and high consumption were not.
 

Experts caution against drinking sweetened beverages despite new findings

Although the present findings suggested that adding sugar did not eliminate the health benefits of coffee, Dr. Liu and colleagues still cautioned against sweetened beverages, citing widely known associations between sugar consumption and poor health.

In an accompanying editorial, Christina C. Wee, MD, MPH, deputy editor of Annals of Internal Medicine, pointed out a key detail from the data: the amount of sugar added to coffee in the U.K. study may be dwarfed by the amount consumed by some coffee drinkers across the pond.

“The average dose of added sugar per cup of sweetened coffee [in the study] was only a little over a teaspoon, or about 4 grams,” Dr. Wee wrote. “This is a far cry from the 15 grams of sugar in an 8-ounce cup of caramel macchiato at a popular U.S. coffee chain.”

Still, Dr. Wee, an associate professor of medicine at Harvard Medical School, Boston, and director of the obesity research program in the division of general medicine at Beth Israel Deaconess Medical Center, Boston, suggested that your typical coffee drinker can feel safe in their daily habit.

“The evidence does not suggest a need for most coffee drinkers – particularly those who drink it with no or modest amounts of sugar – to eliminate coffee,” she wrote. “So drink up – but it would be prudent to avoid too many caramel macchiatos while more evidence brews.”

Estefanía Toledo, MD, MPH, PhD, of the department of preventive medicine and public health at the University of Navarra, Pamplona, Spain, offered a similar takeaway.

Courtesy University of Navarra

“For those who enjoy drinking coffee, are not pregnant or lactating, and do not have special health conditions, coffee consumption could be considered part of a healthy lifestyle,” Dr. Toledo said in a written comment. “I would recommend adding as little sugar as possible to coffee until more evidence has been accrued.”

Dr. Toledo, who previously published a study showing a link between coffee and extended survival, noted that moderate coffee consumption has “repeatedly” been associated with lower rates of “several chronic diseases” and death, but there still isn’t enough evidence to recommend coffee for those who don’t already drink it.

More long-term research is needed, Dr. Toledo said, ideally with studies comparing changes in coffee consumption and health outcomes over time. These may not be forthcoming, however, as such trials are “not easy and feasible to conduct.”

David Kao, MD, assistant professor of medicine-cardiology and medical director of the school of medicine at the University of Colorado at Denver, Aurora, said that the study conducted by Dr. Liu and colleagues is a “very well-executed analysis” that strengthens our confidence in the safety of long-term coffee consumption, even for patients with heart disease.

Dr. Kao, who recently published an analysis showing that higher coffee intake is associated with a lower risk of heart failure, refrained from advising anyone to up their coffee quota.

“I remain cautious about stating too strongly that people should increase coffee intake purely to improve survival,” Dr. Kao said in a written comment. “That said, it does not appear harmful to increase it some, until you drink consistently more than six to seven cups per day.”

The study was supported by the National Natural Science Foundation of China, the Young Elite Scientist Sponsorship Program by CAST, the Guangdong Basic and Applied Basic Research Foundation, and others. Dr. Toledo and Dr. Kao disclosed no relevant conflicts of interest.

“It must be mental illness. My mind cannot possibly conceive of an alternative. A rational healthy mind cannot be capable of this, Doc.”

These were the opening words of one of many discussions that I had with patients in the wake of yet another gut-wrenching tragedy where we saw innocent children and their teachers murdered in school.

Courtesy of Mena Mirhom

Do any mass shooters have untreated mental illness?

There are data to suggest that of 14,785 murders studied in a database to track death by guns, 1,315 mass shootings were identified. Of these mass shootings, 11% of the shooters had serious mental illness. Could we have diagnosed those cases earlier? Intervened sooner? Offered more effective treatment? Certainly. Would that have explain away the rest of the cases? Unfortunately, no.

What is it, then?

The scary answer is that the people who are capable of doing this are not so far away. They are not the folks that we would image locking up in a “psych ward” and throwing away the key. They are, rather, people who are lonely, neglected, rejected, bullied, and broken down by life. Anger, hatred, racism, and evil may be ailments of the soul, but they are not mental illnesses. The carnage they produce is just as tangible. As a psychiatrist, I must admit to you that I do not have a good medication to treat these manifestations of the human condition.

What do we do as a society?

Gun reform is the first obvious and essential answer, without which little else is truly as impactful. We must advocate for it and fight tirelessly.

But at the time you will read this article, your disgruntled coworker will be able to walk into a local store in a moment of despair, anguish, and hopelessness and purchase a semiautomatic weapon of war.

What if we were to start seeing, as a society, that our lives are interwoven? What if we saw that our health is truly interdependent? The COVID-19 pandemic shattered many things in our lives, but one element in particular is our radical individualism. We saw that the choices you make certainly affect me and vice versa. We saw that public health is just that – a public matter, not a private one. We saw that there are some areas of our lives that force us to come together for our own survival.

Perhaps politicians will not save us here. Perhaps kindness will. Empathy can be as potent as legislation, and compassion as impactful as a Twitter hashtag. We each know a lonely coworker, an isolated neighbor, a bullied student, or someone beaten down by life.

What if some of the prevention is in fact in our hands? Together.

“Darkness cannot drive out darkness. Only light can do that. Hate cannot drive out hate; only love can do that.” – Reverend Dr. Martin Luther King, Jr.

Mena Mirhom, MD, is an assistant professor of psychiatry at Columbia University and teaches writing to public psychiatry fellows. He is a board-certified psychiatrist and a consultant for the National Basketball Players Association, treating NBA players and staff.

A version of this article first appeared on Medscape.com.

“It must be mental illness. My mind cannot possibly conceive of an alternative. A rational healthy mind cannot be capable of this, Doc.”

These were the opening words of one of many discussions that I had with patients in the wake of yet another gut-wrenching tragedy where we saw innocent children and their teachers murdered in school.

Courtesy of Mena Mirhom

Do any mass shooters have untreated mental illness?

There are data to suggest that of 14,785 murders studied in a database to track death by guns, 1,315 mass shootings were identified. Of these mass shootings, 11% of the shooters had serious mental illness. Could we have diagnosed those cases earlier? Intervened sooner? Offered more effective treatment? Certainly. Would that have explain away the rest of the cases? Unfortunately, no.

What is it, then?

The scary answer is that the people who are capable of doing this are not so far away. They are not the folks that we would image locking up in a “psych ward” and throwing away the key. They are, rather, people who are lonely, neglected, rejected, bullied, and broken down by life. Anger, hatred, racism, and evil may be ailments of the soul, but they are not mental illnesses. The carnage they produce is just as tangible. As a psychiatrist, I must admit to you that I do not have a good medication to treat these manifestations of the human condition.

What do we do as a society?

Gun reform is the first obvious and essential answer, without which little else is truly as impactful. We must advocate for it and fight tirelessly.

But at the time you will read this article, your disgruntled coworker will be able to walk into a local store in a moment of despair, anguish, and hopelessness and purchase a semiautomatic weapon of war.

What if we were to start seeing, as a society, that our lives are interwoven? What if we saw that our health is truly interdependent? The COVID-19 pandemic shattered many things in our lives, but one element in particular is our radical individualism. We saw that the choices you make certainly affect me and vice versa. We saw that public health is just that – a public matter, not a private one. We saw that there are some areas of our lives that force us to come together for our own survival.

Perhaps politicians will not save us here. Perhaps kindness will. Empathy can be as potent as legislation, and compassion as impactful as a Twitter hashtag. We each know a lonely coworker, an isolated neighbor, a bullied student, or someone beaten down by life.

What if some of the prevention is in fact in our hands? Together.

“Darkness cannot drive out darkness. Only light can do that. Hate cannot drive out hate; only love can do that.” – Reverend Dr. Martin Luther King, Jr.

Mena Mirhom, MD, is an assistant professor of psychiatry at Columbia University and teaches writing to public psychiatry fellows. He is a board-certified psychiatrist and a consultant for the National Basketball Players Association, treating NBA players and staff.

A version of this article first appeared on Medscape.com.

“It must be mental illness. My mind cannot possibly conceive of an alternative. A rational healthy mind cannot be capable of this, Doc.”

These were the opening words of one of many discussions that I had with patients in the wake of yet another gut-wrenching tragedy where we saw innocent children and their teachers murdered in school.

Courtesy of Mena Mirhom

Do any mass shooters have untreated mental illness?

There are data to suggest that of 14,785 murders studied in a database to track death by guns, 1,315 mass shootings were identified. Of these mass shootings, 11% of the shooters had serious mental illness. Could we have diagnosed those cases earlier? Intervened sooner? Offered more effective treatment? Certainly. Would that have explain away the rest of the cases? Unfortunately, no.

What is it, then?

The scary answer is that the people who are capable of doing this are not so far away. They are not the folks that we would image locking up in a “psych ward” and throwing away the key. They are, rather, people who are lonely, neglected, rejected, bullied, and broken down by life. Anger, hatred, racism, and evil may be ailments of the soul, but they are not mental illnesses. The carnage they produce is just as tangible. As a psychiatrist, I must admit to you that I do not have a good medication to treat these manifestations of the human condition.

What do we do as a society?

Gun reform is the first obvious and essential answer, without which little else is truly as impactful. We must advocate for it and fight tirelessly.

But at the time you will read this article, your disgruntled coworker will be able to walk into a local store in a moment of despair, anguish, and hopelessness and purchase a semiautomatic weapon of war.

What if we were to start seeing, as a society, that our lives are interwoven? What if we saw that our health is truly interdependent? The COVID-19 pandemic shattered many things in our lives, but one element in particular is our radical individualism. We saw that the choices you make certainly affect me and vice versa. We saw that public health is just that – a public matter, not a private one. We saw that there are some areas of our lives that force us to come together for our own survival.

Perhaps politicians will not save us here. Perhaps kindness will. Empathy can be as potent as legislation, and compassion as impactful as a Twitter hashtag. We each know a lonely coworker, an isolated neighbor, a bullied student, or someone beaten down by life.

What if some of the prevention is in fact in our hands? Together.

“Darkness cannot drive out darkness. Only light can do that. Hate cannot drive out hate; only love can do that.” – Reverend Dr. Martin Luther King, Jr.

Mena Mirhom, MD, is an assistant professor of psychiatry at Columbia University and teaches writing to public psychiatry fellows. He is a board-certified psychiatrist and a consultant for the National Basketball Players Association, treating NBA players and staff.

A version of this article first appeared on Medscape.com.

From Medically Home Group, Boston, MA.

Brick-and-mortar hospitals in the United States have historically been considered the dominant setting for providing care to patients. The coordination and delivery of care has previously been bound to physical hospitals largely because multidisciplinary services were only accessible in an individual location. While the fundamental make-up of these services remains unchanged, these services are now available in alternate settings. Some of these services include access to a patient care team, supplies, diagnostics, pharmacy, and advanced therapeutic interventions. Presently, the physical environment is becoming increasingly irrelevant as the core of what makes the traditional hospital—the professional staff, collaborative work processes, and the dynamics of the space—have all been translated into a modern digitally integrated environment. The elements necessary to providing safe, effective care in a physical hospital setting are now available in a patient’s home.

Impetus for the Model

As hospitals reconsider how and where they deliver patient care because of limited resources, the hospital-at-home model has gained significant momentum and interest. This model transforms a home into a hospital. The inpatient acute care episode is entirely substituted with an intensive at-home hospital admission enabled by technology, multidisciplinary teams, and ancillary services. Furthermore, patients requiring post-acute support can be transitioned to their next phase of care seamlessly. Given the nationwide nursing shortage, aging population, challenges uncovered by the COVID-19 pandemic, rising hospital costs, nurse/provider burnout related to challenging work environments, and capacity constraints, a shift toward the combination of virtual and in-home care is imperative. The hospital-at-home model has been associated with superior patient outcomes, including reduced risks of delirium, improved functional status, improved patient and family member satisfaction, reduced mortality, reduced readmissions, and significantly lower costs.¹ COVID-19 alone has unmasked major facility-based deficiencies and limitations of our health care system. While the pandemic is not the impetus for the hospital-at-home model, the extended stress of this event has created a unique opportunity to reimagine and transform our health care delivery system so that it is less fragmented and more flexible.

Nursing in the Model

Nursing is central to the hospital-at-home model. Virtual nurses provide meticulous care plan oversight, assessment, and documentation across in-home service providers, to ensure holistic, safe, transparent, and continuous progression toward care plan milestones. The virtual nurse monitors patients using in-home technology that is set up at the time of admission. Connecting with patients to verify social and medical needs, the virtual nurse advocates for their patients and uses these technologies to care and deploy on-demand hands-on services to the patient. Service providers such as paramedics, infusion nurses, or home health nurses may be deployed to provide services in the patient’s home. By bringing in supplies, therapeutics, and interdisciplinary team members, the capabilities of a brick-and-mortar hospital are replicated in the home. All actions that occur wherever the patient is receiving care are overseen by professional nursing staff; in short, virtual nurses are the equivalent of bedside nurses in the brick-and-mortar health care facilities.

Potential Benefits

There are many benefits to the hospital-at-home model (Table). This health care model can be particularly helpful for patients who require frequent admission to acute care facilities, and is well suited for patients with a range of conditions, including those with COVID-19, pneumonia, cellulitis, or congestive heart failure. This care model helps eliminate some of the stressors for patients who have chronic illnesses or other conditions that require frequent hospital admissions. Patients can independently recover at home and can also be surrounded by their loved ones and pets while recovering. This care approach additionally eliminates the risk of hospital-acquired infections and injuries. The hospital-at-home model allows for increased mobility,² as patients are familiar with their surroundings, resulting in reduced onset of delirium. Additionally, patients with improved mobility performance are less likely to experience negative health outcomes.³ There is less chance of sleep disruption as the patient is sleeping in their own bed—no unfamiliar roommate, no call bells or health care personnel frequently coming into the room. The in-home technology set up for remote patient monitoring is designed with the user in mind. Ease of use empowers the patient to collaborate with their care team on their own terms and center the priorities of themselves and their families.

Positive Outcomes

The hospital-at-home model is associated with positive outcomes. The authors of a systematic review identified 10 randomized controlled trials of hospital-at-home programs (with a total of 1372 patients), but were able to obtain data for only 5 of these trials (with a total of 844 patients).⁴ They found a 38% reduction in 6-month mortality for patients who received hospital care at home, as well as significantly higher patient satisfaction across a range of medical conditions, including patients with cellulitis and community-acquired pneumonia, as well as elderly patients with multiple medical conditions. The authors concluded that hospital care at home was less expensive than admission to an acute care hospital.⁴ Similarly, a meta-analysis done by Caplan et al⁵ that included 61 randomized controlled trials concluded that hospital at home is associated with reductions in mortality, readmission rates, and cost, and increases in patient and caregiver satisfaction. Levine et al² found reduced costs and utilization with home hospitalization compared to in-hospital care, as well as improved patient mobility status.

The home is the ideal place to empower patients and caregivers to engage in self-management.² Receiving hospital care at home eliminates the need for dealing with transportation arrangements, traffic, road tolls, and time/scheduling constraints, or finding care for a dependent family member, some of the many stressors that may be experienced by patients who require frequent trips to the hospital. For patients who may not be clinically suitable candidates for hospital at home, such as those requiring critical care intervention and support, the brick-and-mortar hospital is still the appropriate site of care. The hospital-at-home model helps prevent bed shortages in brick-and-mortar hospital settings by allowing hospital care at home for patients who meet preset criteria. These patients can be hospitalized in alternative locations such as their own homes or the residence of a friend. This helps increase health system capacity as well as resiliency.

In addition to expanding safe and appropriate treatment spaces, the hospital-at-home model helps increase access to care for patients during nonstandard hours, including weekends, holidays, or when the waiting time in the emergency room is painfully long. Furthermore, providing care in the home gives the clinical team valuable insight into the patient’s daily life and routine. Performing medication reconciliation with the medicine cabinet in sight and dietary education in a patient’s kitchen are powerful touch points.² For example, a patient with congestive heart failure who must undergo diuresis is much more likely to meet their care goals when their home diet is aligned with the treatment goal. By being able to see exactly what is in a patient’s pantry and fridge, the care team can create a much more tailored approach to sodium intake and fluid management. Providers can create and execute true patient-centric care as they gain direct insight into the patient’s lifestyle, which is clearly valuable when creating care plans for complex chronic health issues.

Although there are clear benefits to hospital at home, how to best implement and scale this model presents a challenge. In addition to educating patients and families about this model of care, health care systems must expand their hospital-at-home programs and provide education about this model to clinical staff and trainees, and insurers must create reimbursement paradigms. Patients meeting eligibility criteria to enroll in hospital at home is the easiest hurdle, as hospital-at-home programs function best when they enroll and service as many patients as possible, including underserved populations.

Upfront Costs and Cost Savings

While there are upfront costs to set up technology and coordinate services, hospital at home also provides significant total cost savings when compared to coordination associated with brick-and-mortar admission. Hospital care accounts for about one-third of total medical expenditures and is a leading cause of debt.² Eliminating fixed hospital costs such as facility, overhead, and equipment costs through adoption of the hospital-at-home model can lead to a reduction in expenditures. It has been found that fewer laboratory and diagnostic tests are ordered for hospital-at-home patients when compared to similar patients in brick-and-mortar hospital settings, with comparable or better clinical patient outcomes.⁶ Furthermore, it is estimated that there are cost savings of 19% to 30% when compared to traditional inpatient care.⁶ Without legislative action, upon the end of the current COVID-19 public health emergency, the Centers for Medicare & Medicaid Service’s Acute Hospital Care at Home waiver will terminate. This could slow down scaling of the model.However, over the past 2 years there has been enough buy-in from major health systems and patients to continue the momentum of the model’s growth. When setting up a hospital-at-home program, it would be wise to consider a few factors: where in the hospital or health system entity structure the hospital-at-home program will reside, which existing resources can be leveraged within the hospital or health system, and what are the state or federal regulatory requirements for such a program. This type of program continues to fill gaps within the US health care system, meeting the needs of widely overlooked populations and increasing access to essential ancillary services.

Conclusion

It is time to consider our bias toward hospital-first options when managing the care needs of our patients. Health care providers have the option to advocate for holistic care, better experience, and better outcomes. Home-based options are safe, equitable, and patient-centric. Increased costs, consumerism, and technology have pushed us to think about alternative approaches to patient care delivery, and the pandemic created a unique opportunity to see just how far the health care system could stretch itself with capacity constraints, insufficient resources, and staff shortages. In light of new possibilities, it is time to reimagine and transform our health care delivery system so that it is unified, seamless, cohesive, and flexible.

Corresponding author: Payal Sharma, DNP, MSN, RN, FNP-BC, CBN; [email protected].

Disclosures: None reported.

From Medically Home Group, Boston, MA.

Brick-and-mortar hospitals in the United States have historically been considered the dominant setting for providing care to patients. The coordination and delivery of care has previously been bound to physical hospitals largely because multidisciplinary services were only accessible in an individual location. While the fundamental make-up of these services remains unchanged, these services are now available in alternate settings. Some of these services include access to a patient care team, supplies, diagnostics, pharmacy, and advanced therapeutic interventions. Presently, the physical environment is becoming increasingly irrelevant as the core of what makes the traditional hospital—the professional staff, collaborative work processes, and the dynamics of the space—have all been translated into a modern digitally integrated environment. The elements necessary to providing safe, effective care in a physical hospital setting are now available in a patient’s home.

Impetus for the Model

As hospitals reconsider how and where they deliver patient care because of limited resources, the hospital-at-home model has gained significant momentum and interest. This model transforms a home into a hospital. The inpatient acute care episode is entirely substituted with an intensive at-home hospital admission enabled by technology, multidisciplinary teams, and ancillary services. Furthermore, patients requiring post-acute support can be transitioned to their next phase of care seamlessly. Given the nationwide nursing shortage, aging population, challenges uncovered by the COVID-19 pandemic, rising hospital costs, nurse/provider burnout related to challenging work environments, and capacity constraints, a shift toward the combination of virtual and in-home care is imperative. The hospital-at-home model has been associated with superior patient outcomes, including reduced risks of delirium, improved functional status, improved patient and family member satisfaction, reduced mortality, reduced readmissions, and significantly lower costs.¹ COVID-19 alone has unmasked major facility-based deficiencies and limitations of our health care system. While the pandemic is not the impetus for the hospital-at-home model, the extended stress of this event has created a unique opportunity to reimagine and transform our health care delivery system so that it is less fragmented and more flexible.

Nursing in the Model

Nursing is central to the hospital-at-home model. Virtual nurses provide meticulous care plan oversight, assessment, and documentation across in-home service providers, to ensure holistic, safe, transparent, and continuous progression toward care plan milestones. The virtual nurse monitors patients using in-home technology that is set up at the time of admission. Connecting with patients to verify social and medical needs, the virtual nurse advocates for their patients and uses these technologies to care and deploy on-demand hands-on services to the patient. Service providers such as paramedics, infusion nurses, or home health nurses may be deployed to provide services in the patient’s home. By bringing in supplies, therapeutics, and interdisciplinary team members, the capabilities of a brick-and-mortar hospital are replicated in the home. All actions that occur wherever the patient is receiving care are overseen by professional nursing staff; in short, virtual nurses are the equivalent of bedside nurses in the brick-and-mortar health care facilities.

Potential Benefits

There are many benefits to the hospital-at-home model (Table). This health care model can be particularly helpful for patients who require frequent admission to acute care facilities, and is well suited for patients with a range of conditions, including those with COVID-19, pneumonia, cellulitis, or congestive heart failure. This care model helps eliminate some of the stressors for patients who have chronic illnesses or other conditions that require frequent hospital admissions. Patients can independently recover at home and can also be surrounded by their loved ones and pets while recovering. This care approach additionally eliminates the risk of hospital-acquired infections and injuries. The hospital-at-home model allows for increased mobility,² as patients are familiar with their surroundings, resulting in reduced onset of delirium. Additionally, patients with improved mobility performance are less likely to experience negative health outcomes.³ There is less chance of sleep disruption as the patient is sleeping in their own bed—no unfamiliar roommate, no call bells or health care personnel frequently coming into the room. The in-home technology set up for remote patient monitoring is designed with the user in mind. Ease of use empowers the patient to collaborate with their care team on their own terms and center the priorities of themselves and their families.

Positive Outcomes

The hospital-at-home model is associated with positive outcomes. The authors of a systematic review identified 10 randomized controlled trials of hospital-at-home programs (with a total of 1372 patients), but were able to obtain data for only 5 of these trials (with a total of 844 patients).⁴ They found a 38% reduction in 6-month mortality for patients who received hospital care at home, as well as significantly higher patient satisfaction across a range of medical conditions, including patients with cellulitis and community-acquired pneumonia, as well as elderly patients with multiple medical conditions. The authors concluded that hospital care at home was less expensive than admission to an acute care hospital.⁴ Similarly, a meta-analysis done by Caplan et al⁵ that included 61 randomized controlled trials concluded that hospital at home is associated with reductions in mortality, readmission rates, and cost, and increases in patient and caregiver satisfaction. Levine et al² found reduced costs and utilization with home hospitalization compared to in-hospital care, as well as improved patient mobility status.

The home is the ideal place to empower patients and caregivers to engage in self-management.² Receiving hospital care at home eliminates the need for dealing with transportation arrangements, traffic, road tolls, and time/scheduling constraints, or finding care for a dependent family member, some of the many stressors that may be experienced by patients who require frequent trips to the hospital. For patients who may not be clinically suitable candidates for hospital at home, such as those requiring critical care intervention and support, the brick-and-mortar hospital is still the appropriate site of care. The hospital-at-home model helps prevent bed shortages in brick-and-mortar hospital settings by allowing hospital care at home for patients who meet preset criteria. These patients can be hospitalized in alternative locations such as their own homes or the residence of a friend. This helps increase health system capacity as well as resiliency.

In addition to expanding safe and appropriate treatment spaces, the hospital-at-home model helps increase access to care for patients during nonstandard hours, including weekends, holidays, or when the waiting time in the emergency room is painfully long. Furthermore, providing care in the home gives the clinical team valuable insight into the patient’s daily life and routine. Performing medication reconciliation with the medicine cabinet in sight and dietary education in a patient’s kitchen are powerful touch points.² For example, a patient with congestive heart failure who must undergo diuresis is much more likely to meet their care goals when their home diet is aligned with the treatment goal. By being able to see exactly what is in a patient’s pantry and fridge, the care team can create a much more tailored approach to sodium intake and fluid management. Providers can create and execute true patient-centric care as they gain direct insight into the patient’s lifestyle, which is clearly valuable when creating care plans for complex chronic health issues.

Although there are clear benefits to hospital at home, how to best implement and scale this model presents a challenge. In addition to educating patients and families about this model of care, health care systems must expand their hospital-at-home programs and provide education about this model to clinical staff and trainees, and insurers must create reimbursement paradigms. Patients meeting eligibility criteria to enroll in hospital at home is the easiest hurdle, as hospital-at-home programs function best when they enroll and service as many patients as possible, including underserved populations.

Upfront Costs and Cost Savings

While there are upfront costs to set up technology and coordinate services, hospital at home also provides significant total cost savings when compared to coordination associated with brick-and-mortar admission. Hospital care accounts for about one-third of total medical expenditures and is a leading cause of debt.² Eliminating fixed hospital costs such as facility, overhead, and equipment costs through adoption of the hospital-at-home model can lead to a reduction in expenditures. It has been found that fewer laboratory and diagnostic tests are ordered for hospital-at-home patients when compared to similar patients in brick-and-mortar hospital settings, with comparable or better clinical patient outcomes.⁶ Furthermore, it is estimated that there are cost savings of 19% to 30% when compared to traditional inpatient care.⁶ Without legislative action, upon the end of the current COVID-19 public health emergency, the Centers for Medicare & Medicaid Service’s Acute Hospital Care at Home waiver will terminate. This could slow down scaling of the model.However, over the past 2 years there has been enough buy-in from major health systems and patients to continue the momentum of the model’s growth. When setting up a hospital-at-home program, it would be wise to consider a few factors: where in the hospital or health system entity structure the hospital-at-home program will reside, which existing resources can be leveraged within the hospital or health system, and what are the state or federal regulatory requirements for such a program. This type of program continues to fill gaps within the US health care system, meeting the needs of widely overlooked populations and increasing access to essential ancillary services.

Conclusion

It is time to consider our bias toward hospital-first options when managing the care needs of our patients. Health care providers have the option to advocate for holistic care, better experience, and better outcomes. Home-based options are safe, equitable, and patient-centric. Increased costs, consumerism, and technology have pushed us to think about alternative approaches to patient care delivery, and the pandemic created a unique opportunity to see just how far the health care system could stretch itself with capacity constraints, insufficient resources, and staff shortages. In light of new possibilities, it is time to reimagine and transform our health care delivery system so that it is unified, seamless, cohesive, and flexible.

Corresponding author: Payal Sharma, DNP, MSN, RN, FNP-BC, CBN; [email protected].

Disclosures: None reported.

From Medically Home Group, Boston, MA.

Brick-and-mortar hospitals in the United States have historically been considered the dominant setting for providing care to patients. The coordination and delivery of care has previously been bound to physical hospitals largely because multidisciplinary services were only accessible in an individual location. While the fundamental make-up of these services remains unchanged, these services are now available in alternate settings. Some of these services include access to a patient care team, supplies, diagnostics, pharmacy, and advanced therapeutic interventions. Presently, the physical environment is becoming increasingly irrelevant as the core of what makes the traditional hospital—the professional staff, collaborative work processes, and the dynamics of the space—have all been translated into a modern digitally integrated environment. The elements necessary to providing safe, effective care in a physical hospital setting are now available in a patient’s home.

Impetus for the Model

As hospitals reconsider how and where they deliver patient care because of limited resources, the hospital-at-home model has gained significant momentum and interest. This model transforms a home into a hospital. The inpatient acute care episode is entirely substituted with an intensive at-home hospital admission enabled by technology, multidisciplinary teams, and ancillary services. Furthermore, patients requiring post-acute support can be transitioned to their next phase of care seamlessly. Given the nationwide nursing shortage, aging population, challenges uncovered by the COVID-19 pandemic, rising hospital costs, nurse/provider burnout related to challenging work environments, and capacity constraints, a shift toward the combination of virtual and in-home care is imperative. The hospital-at-home model has been associated with superior patient outcomes, including reduced risks of delirium, improved functional status, improved patient and family member satisfaction, reduced mortality, reduced readmissions, and significantly lower costs.¹ COVID-19 alone has unmasked major facility-based deficiencies and limitations of our health care system. While the pandemic is not the impetus for the hospital-at-home model, the extended stress of this event has created a unique opportunity to reimagine and transform our health care delivery system so that it is less fragmented and more flexible.

Nursing in the Model

Nursing is central to the hospital-at-home model. Virtual nurses provide meticulous care plan oversight, assessment, and documentation across in-home service providers, to ensure holistic, safe, transparent, and continuous progression toward care plan milestones. The virtual nurse monitors patients using in-home technology that is set up at the time of admission. Connecting with patients to verify social and medical needs, the virtual nurse advocates for their patients and uses these technologies to care and deploy on-demand hands-on services to the patient. Service providers such as paramedics, infusion nurses, or home health nurses may be deployed to provide services in the patient’s home. By bringing in supplies, therapeutics, and interdisciplinary team members, the capabilities of a brick-and-mortar hospital are replicated in the home. All actions that occur wherever the patient is receiving care are overseen by professional nursing staff; in short, virtual nurses are the equivalent of bedside nurses in the brick-and-mortar health care facilities.

Potential Benefits

There are many benefits to the hospital-at-home model (Table). This health care model can be particularly helpful for patients who require frequent admission to acute care facilities, and is well suited for patients with a range of conditions, including those with COVID-19, pneumonia, cellulitis, or congestive heart failure. This care model helps eliminate some of the stressors for patients who have chronic illnesses or other conditions that require frequent hospital admissions. Patients can independently recover at home and can also be surrounded by their loved ones and pets while recovering. This care approach additionally eliminates the risk of hospital-acquired infections and injuries. The hospital-at-home model allows for increased mobility,² as patients are familiar with their surroundings, resulting in reduced onset of delirium. Additionally, patients with improved mobility performance are less likely to experience negative health outcomes.³ There is less chance of sleep disruption as the patient is sleeping in their own bed—no unfamiliar roommate, no call bells or health care personnel frequently coming into the room. The in-home technology set up for remote patient monitoring is designed with the user in mind. Ease of use empowers the patient to collaborate with their care team on their own terms and center the priorities of themselves and their families.

Positive Outcomes

The hospital-at-home model is associated with positive outcomes. The authors of a systematic review identified 10 randomized controlled trials of hospital-at-home programs (with a total of 1372 patients), but were able to obtain data for only 5 of these trials (with a total of 844 patients).⁴ They found a 38% reduction in 6-month mortality for patients who received hospital care at home, as well as significantly higher patient satisfaction across a range of medical conditions, including patients with cellulitis and community-acquired pneumonia, as well as elderly patients with multiple medical conditions. The authors concluded that hospital care at home was less expensive than admission to an acute care hospital.⁴ Similarly, a meta-analysis done by Caplan et al⁵ that included 61 randomized controlled trials concluded that hospital at home is associated with reductions in mortality, readmission rates, and cost, and increases in patient and caregiver satisfaction. Levine et al² found reduced costs and utilization with home hospitalization compared to in-hospital care, as well as improved patient mobility status.

The home is the ideal place to empower patients and caregivers to engage in self-management.² Receiving hospital care at home eliminates the need for dealing with transportation arrangements, traffic, road tolls, and time/scheduling constraints, or finding care for a dependent family member, some of the many stressors that may be experienced by patients who require frequent trips to the hospital. For patients who may not be clinically suitable candidates for hospital at home, such as those requiring critical care intervention and support, the brick-and-mortar hospital is still the appropriate site of care. The hospital-at-home model helps prevent bed shortages in brick-and-mortar hospital settings by allowing hospital care at home for patients who meet preset criteria. These patients can be hospitalized in alternative locations such as their own homes or the residence of a friend. This helps increase health system capacity as well as resiliency.

In addition to expanding safe and appropriate treatment spaces, the hospital-at-home model helps increase access to care for patients during nonstandard hours, including weekends, holidays, or when the waiting time in the emergency room is painfully long. Furthermore, providing care in the home gives the clinical team valuable insight into the patient’s daily life and routine. Performing medication reconciliation with the medicine cabinet in sight and dietary education in a patient’s kitchen are powerful touch points.² For example, a patient with congestive heart failure who must undergo diuresis is much more likely to meet their care goals when their home diet is aligned with the treatment goal. By being able to see exactly what is in a patient’s pantry and fridge, the care team can create a much more tailored approach to sodium intake and fluid management. Providers can create and execute true patient-centric care as they gain direct insight into the patient’s lifestyle, which is clearly valuable when creating care plans for complex chronic health issues.

Although there are clear benefits to hospital at home, how to best implement and scale this model presents a challenge. In addition to educating patients and families about this model of care, health care systems must expand their hospital-at-home programs and provide education about this model to clinical staff and trainees, and insurers must create reimbursement paradigms. Patients meeting eligibility criteria to enroll in hospital at home is the easiest hurdle, as hospital-at-home programs function best when they enroll and service as many patients as possible, including underserved populations.

Upfront Costs and Cost Savings

While there are upfront costs to set up technology and coordinate services, hospital at home also provides significant total cost savings when compared to coordination associated with brick-and-mortar admission. Hospital care accounts for about one-third of total medical expenditures and is a leading cause of debt.² Eliminating fixed hospital costs such as facility, overhead, and equipment costs through adoption of the hospital-at-home model can lead to a reduction in expenditures. It has been found that fewer laboratory and diagnostic tests are ordered for hospital-at-home patients when compared to similar patients in brick-and-mortar hospital settings, with comparable or better clinical patient outcomes.⁶ Furthermore, it is estimated that there are cost savings of 19% to 30% when compared to traditional inpatient care.⁶ Without legislative action, upon the end of the current COVID-19 public health emergency, the Centers for Medicare & Medicaid Service’s Acute Hospital Care at Home waiver will terminate. This could slow down scaling of the model.However, over the past 2 years there has been enough buy-in from major health systems and patients to continue the momentum of the model’s growth. When setting up a hospital-at-home program, it would be wise to consider a few factors: where in the hospital or health system entity structure the hospital-at-home program will reside, which existing resources can be leveraged within the hospital or health system, and what are the state or federal regulatory requirements for such a program. This type of program continues to fill gaps within the US health care system, meeting the needs of widely overlooked populations and increasing access to essential ancillary services.

Conclusion

It is time to consider our bias toward hospital-first options when managing the care needs of our patients. Health care providers have the option to advocate for holistic care, better experience, and better outcomes. Home-based options are safe, equitable, and patient-centric. Increased costs, consumerism, and technology have pushed us to think about alternative approaches to patient care delivery, and the pandemic created a unique opportunity to see just how far the health care system could stretch itself with capacity constraints, insufficient resources, and staff shortages. In light of new possibilities, it is time to reimagine and transform our health care delivery system so that it is unified, seamless, cohesive, and flexible.

Corresponding author: Payal Sharma, DNP, MSN, RN, FNP-BC, CBN; [email protected].

Disclosures: None reported.

The Institute of Medicine brought much-needed attention to the need for process improvement in medicine with its seminal report To Err Is Human: Building a Safer Health System, which was issued in 1999, leading to the quality movement’s call to close health care performance gaps in Crossing the Quality Chasm: A New Health System for the 21st Century.^1,2 Quality improvement science in medicine has evolved over the past 2 decades to include a broad spectrum of approaches, from agile improvement to continuous learning and improvement. Current efforts focus on Lean-based process improvement along with a reduction in variation in clinical practice to align practice with the principles of evidence-based medicine in a patient-centered approach.³ Further, the definition of quality improvement under the Affordable Care Act was framed as an equitable, timely, value-based, patient-centered approach to achieving population-level health goals.⁴ Thus, the science of quality improvement drives the core principles of care delivery improvement, and the rigorous evidence needed to expand innovation is embedded within the same framework.^5,6 In clinical practice, quality improvement projects aim to define gaps and then specific steps are undertaken to improve the evidence-based practice of a specific process. The overarching goal is to enhance the efficacy of the practice by reducing waste within a particular domain. Thus, quality improvement and implementation research eventually unify how clinical practice is advanced concurrently to bridge identified gaps.⁷

System redesign through a patient-centered framework forms the core of an overarching strategy to support system-level processes. Both require a deep understanding of the fields of quality improvement science and implementation science.⁸ Furthermore, aligning clinical research needs, system aims, patients’ values, and clinical care give the new design a clear path forward. Patient-centered improvement includes the essential elements of system redesign around human factors, including communication, physical resources, and updated information during episodes of care. The patient-centered improvement design is juxtaposed with care planning and establishing continuum of care processes.⁹ It is essential to note that safety is rooted within the quality domain as a top priority in medicine.¹⁰ The best implementation methods and approaches are discussed and debated, and the improvement progress continues on multiple fronts.¹¹ Patient safety systems are implemented simultaneously during the redesign phase. Moreover, identifying and testing the health care delivery methods in the era of competing strategic priorities to achieve the desirable clinical outcomes highlights the importance of implementation, while contemplating the methods of dissemination, scalability, and sustainability of the best evidence-based clinical practice.

The cycle of quality improvement research completes the system implementation efforts. The conceptual framework of quality improvement includes multiple areas of care and transition, along with applying the best clinical practices in a culture that emphasizes continuous improvement and learning. At the same time, the operating principles should include continuous improvement in a simple and continuous system of learning as a core concept. Our proposed implementation approach involves taking simple and practical steps while separating the process from the outcomes measures, extracting effectiveness throughout the process. It is essential to keep in mind that building a proactive and systematic improvement environment requires a framework for safety, reliability, and effective care, as well as the alignment of the physical system, communication, and professional environment and culture (Figure).

In summary, system design for quality improvement research should incorporate the principles and conceptual framework that embody effective implementation strategies, with a focus on operational and practical steps. Continuous improvement will be reached through the multidimensional development of current health care system metrics and the incorporation of implementation science methods.

Corresponding author: Ebrahim Barkoudah, MD, MPH, Department of Medicine, Brigham and Women’s Hospital, Boston, MA; [email protected]

Disclosures: None reported.

The Institute of Medicine brought much-needed attention to the need for process improvement in medicine with its seminal report To Err Is Human: Building a Safer Health System, which was issued in 1999, leading to the quality movement’s call to close health care performance gaps in Crossing the Quality Chasm: A New Health System for the 21st Century.^1,2 Quality improvement science in medicine has evolved over the past 2 decades to include a broad spectrum of approaches, from agile improvement to continuous learning and improvement. Current efforts focus on Lean-based process improvement along with a reduction in variation in clinical practice to align practice with the principles of evidence-based medicine in a patient-centered approach.³ Further, the definition of quality improvement under the Affordable Care Act was framed as an equitable, timely, value-based, patient-centered approach to achieving population-level health goals.⁴ Thus, the science of quality improvement drives the core principles of care delivery improvement, and the rigorous evidence needed to expand innovation is embedded within the same framework.^5,6 In clinical practice, quality improvement projects aim to define gaps and then specific steps are undertaken to improve the evidence-based practice of a specific process. The overarching goal is to enhance the efficacy of the practice by reducing waste within a particular domain. Thus, quality improvement and implementation research eventually unify how clinical practice is advanced concurrently to bridge identified gaps.⁷

System redesign through a patient-centered framework forms the core of an overarching strategy to support system-level processes. Both require a deep understanding of the fields of quality improvement science and implementation science.⁸ Furthermore, aligning clinical research needs, system aims, patients’ values, and clinical care give the new design a clear path forward. Patient-centered improvement includes the essential elements of system redesign around human factors, including communication, physical resources, and updated information during episodes of care. The patient-centered improvement design is juxtaposed with care planning and establishing continuum of care processes.⁹ It is essential to note that safety is rooted within the quality domain as a top priority in medicine.¹⁰ The best implementation methods and approaches are discussed and debated, and the improvement progress continues on multiple fronts.¹¹ Patient safety systems are implemented simultaneously during the redesign phase. Moreover, identifying and testing the health care delivery methods in the era of competing strategic priorities to achieve the desirable clinical outcomes highlights the importance of implementation, while contemplating the methods of dissemination, scalability, and sustainability of the best evidence-based clinical practice.

The cycle of quality improvement research completes the system implementation efforts. The conceptual framework of quality improvement includes multiple areas of care and transition, along with applying the best clinical practices in a culture that emphasizes continuous improvement and learning. At the same time, the operating principles should include continuous improvement in a simple and continuous system of learning as a core concept. Our proposed implementation approach involves taking simple and practical steps while separating the process from the outcomes measures, extracting effectiveness throughout the process. It is essential to keep in mind that building a proactive and systematic improvement environment requires a framework for safety, reliability, and effective care, as well as the alignment of the physical system, communication, and professional environment and culture (Figure).

In summary, system design for quality improvement research should incorporate the principles and conceptual framework that embody effective implementation strategies, with a focus on operational and practical steps. Continuous improvement will be reached through the multidimensional development of current health care system metrics and the incorporation of implementation science methods.

Corresponding author: Ebrahim Barkoudah, MD, MPH, Department of Medicine, Brigham and Women’s Hospital, Boston, MA; [email protected]

Disclosures: None reported.

The Institute of Medicine brought much-needed attention to the need for process improvement in medicine with its seminal report To Err Is Human: Building a Safer Health System, which was issued in 1999, leading to the quality movement’s call to close health care performance gaps in Crossing the Quality Chasm: A New Health System for the 21st Century.^1,2 Quality improvement science in medicine has evolved over the past 2 decades to include a broad spectrum of approaches, from agile improvement to continuous learning and improvement. Current efforts focus on Lean-based process improvement along with a reduction in variation in clinical practice to align practice with the principles of evidence-based medicine in a patient-centered approach.³ Further, the definition of quality improvement under the Affordable Care Act was framed as an equitable, timely, value-based, patient-centered approach to achieving population-level health goals.⁴ Thus, the science of quality improvement drives the core principles of care delivery improvement, and the rigorous evidence needed to expand innovation is embedded within the same framework.^5,6 In clinical practice, quality improvement projects aim to define gaps and then specific steps are undertaken to improve the evidence-based practice of a specific process. The overarching goal is to enhance the efficacy of the practice by reducing waste within a particular domain. Thus, quality improvement and implementation research eventually unify how clinical practice is advanced concurrently to bridge identified gaps.⁷

System redesign through a patient-centered framework forms the core of an overarching strategy to support system-level processes. Both require a deep understanding of the fields of quality improvement science and implementation science.⁸ Furthermore, aligning clinical research needs, system aims, patients’ values, and clinical care give the new design a clear path forward. Patient-centered improvement includes the essential elements of system redesign around human factors, including communication, physical resources, and updated information during episodes of care. The patient-centered improvement design is juxtaposed with care planning and establishing continuum of care processes.⁹ It is essential to note that safety is rooted within the quality domain as a top priority in medicine.¹⁰ The best implementation methods and approaches are discussed and debated, and the improvement progress continues on multiple fronts.¹¹ Patient safety systems are implemented simultaneously during the redesign phase. Moreover, identifying and testing the health care delivery methods in the era of competing strategic priorities to achieve the desirable clinical outcomes highlights the importance of implementation, while contemplating the methods of dissemination, scalability, and sustainability of the best evidence-based clinical practice.

The cycle of quality improvement research completes the system implementation efforts. The conceptual framework of quality improvement includes multiple areas of care and transition, along with applying the best clinical practices in a culture that emphasizes continuous improvement and learning. At the same time, the operating principles should include continuous improvement in a simple and continuous system of learning as a core concept. Our proposed implementation approach involves taking simple and practical steps while separating the process from the outcomes measures, extracting effectiveness throughout the process. It is essential to keep in mind that building a proactive and systematic improvement environment requires a framework for safety, reliability, and effective care, as well as the alignment of the physical system, communication, and professional environment and culture (Figure).

In summary, system design for quality improvement research should incorporate the principles and conceptual framework that embody effective implementation strategies, with a focus on operational and practical steps. Continuous improvement will be reached through the multidimensional development of current health care system metrics and the incorporation of implementation science methods.

Corresponding author: Ebrahim Barkoudah, MD, MPH, Department of Medicine, Brigham and Women’s Hospital, Boston, MA; [email protected]

Disclosures: None reported.

From the Department of Cardiothoracic Surgery, Stanford University, Stanford, CA.

Abstract

Complex cardiac patients are often referred for surgery or palliative care based on the risk of perioperative mortality. This decision ignores factors such as quality of life or duration of life in either surgery or the palliative path. Here, we propose a model to numerically assess and compare the value of surgery vs palliation. This model includes quality and duration of life, as well as risk of perioperative mortality, and involves a patient’s preferences in the decision-making process.

For each pathway, surgery or palliative care, a value is calculated and compared to a normal life value (no disease symptoms and normal life expectancy). The formula is adjusted for the risk of operative mortality. The model produces a ratio of the value of surgery to the value of palliative care that signifies the superiority of one or another. This model calculation presents an objective estimated numerical value to compare the value of surgery and palliative care. It can be applied to every decision-making process before surgery. In general, if a procedure has the potential to significantly extend life in a patient who otherwise has a very short life expectancy with palliation only, performing high-risk surgery would be a reasonable option. A model that provides a numerical value for surgery vs palliative care and includes quality and duration of life in each pathway could be a useful tool for cardiac surgeons in decision making regarding high-risk surgery.

Keywords: high-risk surgery, palliative care, quality of life, life expectancy.

Patients with complex cardiovascular disease are occasionally considered inoperable due to the high risk of surgical mortality. When the risk of perioperative mortality (POM) is predicted to be too high, surgical intervention is denied, and patients are often referred to palliative care. The risk of POM in cardiac surgery is often calculated using large-scale databases, such as the Society of Thoracic Surgeons (STS) records. The STS risk models, which are regularly updated, are based on large data sets and incorporate precise statistical methods for risk adjustment.¹ In general, these calculators provide a percentage value that defines the magnitude of the risk of death, and then an arbitrary range is selected to categorize the procedure as low, medium, or high risk or inoperable status. The STS database does not set a cutoff point or range to define “operability.” Assigning inoperable status to a certain risk rate is problematic, with many ethical, legal, and moral implications, and for this reason, it has mostly remained undefined. In contrast, the low- and medium-risk ranges are easier to define. Another limitation encountered in the STS database is the lack of risk data for less common but very high-risk procedures, such as a triple valve replacement.

A common example where risk classification has been defined is in patients who are candidates for surgical vs transcatheter aortic valve replacement. Some groups have described a risk of <4% as low risk, 4% to 8% as intermediate risk, >8% as high risk, and >15% as inoperable²; for some other groups, a risk of POM >50% is considered extreme risk or inoperable.^3,4 This procedure-specific classification is a useful decision-making tool and helps the surgeon perform an initial risk assessment to allocate a specific patient to a group—operable or nonoperable—only by calculating the risk of surgical death. However, this allocation method does not provide any information on how and when death occurs in either group. These 2 parameters of how and when death occurs define the quality of life (QOL) and the duration of life (DOL), respectively, and together could be considered as the value of life in each pathway. A survivor of a high-risk surgery may benefit from good quality and extended life (a high value), or, on the other end of the spectrum, a high-risk patient who does not undergo surgery is spared the mortality risk of the surgery but dies sooner (low value) with symptoms due to the natural course of the untreated disease.

The central question is, if a surgery is high risk but has the potential of providing a good value (for those who survive it), what QOL and DOL values are acceptable to risk or to justify accepting and proceeding with a risky surgery? Or how high a POM risk is justified to proceed with surgery rather than the alternative palliative care with a certain quality and duration? It is obvious that a decision-making process that is based on POM cannot compare the value of surgery (Vs) and the value of palliation (Vp). Furthermore, it ignores patient preferences and their input, as these are excluded from this decision-making process.

To be able to include QOL and DOL in any decision making, one must precisely describe these parameters. Both QOL and DOL are used for estimation of disease burden by health care administrators, public health experts, insurance agencies, and others. Multiple models have been proposed and used to estimate the overall burden of the disease. Most of the models for this purpose are created for large-scale economic purposes and not for decision making in individual cases.

An important measure is the quality-adjusted life year (QALY). This is an important parameter since it includes both measures of quality and quantity of life.^5,6 QALY is a simplified measure to assess the value of health outcomes, and it has been used in economic calculations to assess mainly the cost-effectiveness of various interventions. We sought to evaluate the utility of a similar method in adding further insight into the surgical decision-making process. In this article, we propose a simple model to compare the value of surgery vs palliative care, similar to QALY. This model includes and adjusts for the quality and the quantity of life, in addition to the risk of POM, in the decision-making process for high-risk patients.

The Model

The 2 decision pathways, surgery and palliative care, are compared for their value. We define the value as the product of QOL and DOL in each pathway and use the severity of the symptoms as a surrogate for QOL. If duration and quality were depicted on the x and y axes of a graph (Figure 1), then the area under the curve would represent the collective value in each situation. Figure 2 shows the timeline and the different pathways with each decision. The value in each situation is calculated in relation to the full value, which is represented as the value of normal life (Vn), that is, life without disease and with normal life expectancy. The values of each decision pathway, the value of surgery (Vs) and the value of palliation (Vp), are then compared to define the benefit for each decision as follows:

If Vs/Vp > 1, the benefit is toward surgery;

If Vs/Vp < 1, the benefit is for palliative care.

Definitions

Both quality and duration of life are presented on a 1-10 scale, 1 being the lowest and 10 the highest value, to yield a product with a value of 100 in normal, disease-free life. Any lower value is presented as a percentage to represent the comparison to the full value. QOL is determined by degradation of full quality with the average level of symptoms. DOL is calculated as a lost time (period of time from death after a specific intervention [surgery or palliation] until death at normal life expectancy) in fraction of full life (death at life expectancy). The Vs is adjusted to exclude the nonsurvivors using the chance of survival (100 – POM risk).

For the DOL under any condition, a 10-year survival rate could be used as a surrogate in this formula. Compared to life expectancy value, using the 10-year survival rate simplifies the calculation since cardiac diseases are more prevalent in older age, close to or beyond the average life expectancy value.

Using the time intervals from the timeline in Figure 2:

dh = time interval from diagnosis to death at life expectancy

dg = time interval from diagnosis to death after successful surgery

df = time interval from diagnosis to death after palliative care

Duration for palliative care:

Duration for surgery:

Adjustment: This value is calculated for those who survive the surgery. To adjust for the POM, it is multiplied by the 100 − POM risk.

Since value is the base for comparison in this model, and it is the product of 2 equally important factors in the formula (severity and duration of symptoms), a factor of 10 was chosen to yield a value of 100, which represents 100% health or absence of symptoms for any duration.

After elimination of normal life expectancy, form the numerator and denominator:

To adjust for surgical outcomes in special circumstances where less than optimal or standard surgical results are expected (eg, in very rare surgeries, limited resource institutions, or suboptimal postoperative surgical care), an optional coefficient R can be added to the numerator (surgical value). This optional coefficient, with values such as 0.8, 0.9 (to degrade the value of surgery) or 1 (standard surgical outcome), adjusts for variability in interinstitutional surgical results or surgeon variability. No coefficient is added to the denominator since palliative care provides minimal differences between clinicians and hospitals. Thus, the final adjusted formula would be as follows:

Example

A 60-year-old patient with a 10% POM risk needs to be allocated to surgical or palliative care. With palliative care, if this patient lived 6 years with average symptoms grade 4, the Vp would be 20; that is, 20% of the normal life value (if he lived 18 years instead without the disease).

Using the formula for calculation of value in each pathway:

With adjustment of 90% survival chance in surgery, 53 × 90% = 48%. In this example, Vs/Vp = 48/20 = 2.4, showing a significant benefit for surgical care. Notably, the unknown value of normal life expectancy is not needed for the calculation of Vs/Vp, since it is the same in both pathways and it is eliminated by calculation in fraction.

Based on this formula, since the duration of surgical symptoms is short, no matter how severe these are, if the potential duration of life after surgery is high (represented by smaller area under the curve in Figure 1), the numerator becomes larger and the value of the surgery grows. For example, if a patient with a 15% risk of POM, which is generally considered inoperable, lives 5 years, as opposed to 2 years with palliative care with mild symptoms (eg 3/10), Vs/Vp would be 2.7, still showing a significant benefit for surgical care.

Discussion

Any surgical intervention is offered with 2 goals in mind, improving QOL and extending DOL. In a high-risk patient, surgery might be declined due to a high risk of POM, and the patient is offered palliative care, which other than providing symptom relief does not change the course of disease and eventually the patient will die due to the untreated disease. In this decision-making method, mostly completed by a care team only, a potential risk of death due to surgery which possibly could cure the patient is traded for immediate survival; however, the symptomatic course ensues until death. This mostly unilateral decision-making process by a care team, which incorporates minimal input from the patient or ignores patient preferences altogether, is based only on POM risk, and roughly includes a single parameter: years of potential life lost (YPLL). YPLL is a measure of premature mortality, and in the setting of surgical intervention, YPLL is the number of years a patient would lose unless a successful surgery were undertaken. Obviously, patients would live longer if a surgery that was intended to save them failed.

In this article, we proposed a simple method to quantify each decision to decide whether to operate or choose surgical care vs palliative care. Since quality and duration of life are both end factors clinicians and patients aspire to in each decision, they can be considered together as the value of each decision. We believe a numerical framework would provide an objective way to assist both the patient at high risk and the care team in the decision-making process.

The 2 parameters we consider are DOL and QOL. DOL, or survival, can be extracted from large-scale data using statistical methods that have been developed to predict survival under various conditions, such as Kaplan-Meier curves. These methods present the chance of survival in percentages in a defined time frame, such as a 5- or 10-year period.

While the DOL is a numerical parameter and quantifiable, the QOL is a more complex entity. This subjective parameter bears multiple definitions, aspects, and categories, and therefore multiple scales for quantification of QOL have been proposed. These scales have been used extensively for the purpose of health determination in health care policy and economic planning. Most scales acknowledge that QOL is multifactorial and includes interrelated aspects such as mental and socioeconomic factors. We have also noticed that QOL is better determined by the palliative care team than surgeons, so including these care providers in the decision-making process might reduce surgeon bias.

Since our purpose here is only to assist with the decision on medical intervention, we focus on physical QOL. Multiple scales are used to assess health-related QOL, such as the Assessment of Quality of Life (AQoL)-8D,⁷ EuroQol-5 Dimension (EQ-5D),⁸ 15D,⁹ and the 36-Item Short Form Survey (SF-36).¹⁰ These complex scales are built for systematic reviews, and they are not practical for a clinical user. To simplify and keep this practical, we define QOL by using the severity or grade of symptoms related to the disease the patient has on a scale of 0 to 10. The severity of symptoms can be easily determined using available scales. An applicable scale for this purpose is the Edmonton Symptom Assessment Scale (ESAS), which has been in use for years and has evolved as a useful tool in the medical field.¹¹

Once DOL and QOL are determined on a 1-10 scale, the multiplied value then provides a product that we consider a value. The highest value hoped for in each decision is the achievement of the best QOL and DOL, a value of 100. In Figure 1, a graphic presentation of value in each decision is best seen as the area under the curve. As shown, a successful surgery, even when accompanied by significant symptoms during initial recovery, has a chance (100 – risk of POM%) to gain a larger area under curve (value) by achieving a longer life with no or fewer symptoms. However, in palliative care, progressing disease and even palliated symptoms with a shorter life expectancy impose a large burden on the patient and a much lower value. Note that in this calculation, life expectancy, which is an important but unpredictable factor, is initially included; however, by ratio comparison, it is eliminated, simplifying the calculation further.

Using this formula in different settings reveals that high-risk surgery has a greater potential to reduce YPLL in the general population. Based on this formula, compared to a surgery with potential to significantly extend DOL, a definite shorter and symptomatic life course with palliative care makes it a significantly less favorable option. In fact, in the cardiovascular field, palliative care has minimal or no effect on natural history, as the mechanism of illness is mechanical, such as occlusion of coronary arteries or valve dysfunction, leading eventually to heart failure and death. In a study by Xu et al, although palliative care reduced readmission rates and improved symptoms on a variety of scales, there was no effect on mortality and QOL in patients with heart failure.¹²

No model in this field has proven to be ideal, and this model bears multiple limitations as well. We have used severity of symptoms as a surrogate for QOL based on the fact that cardiac patients with different pathologies who are untreated will have a common final pathway with development of heart failure symptoms that dictate their QOL. Also, grading QOL is a difficult task at times. Even a model such as QALY, which is one of the most used, is not a perfect model and is not free of problems.⁶ The difference in surgical results and life expectancy between sexes and ethnic groups might be a source of bias in this formula. Also, multiple factors directly and indirectly affect QOL and DOL and create inaccuracies; therefore, making an exact science from an inexact one naturally relies on multiple assumptions. Although it has previously been shown that most POM occurs in a short period of time after cardiac surgery,¹³ long-term complications that potentially degrade QOL are not included in this model. By applying this model, one must assume indefinite economic resources. Moreover, applying a single mathematical model in a biologic system and in the general population has intrinsic shortcomings, and it must overlook many other factors (eg, ethical, legal). For example, it will be hard to justify a failed surgery with 15% risk of POM undertaken to eliminate the severe long-lasting symptoms of a disease, while the outcome of a successful surgery with a 20% risk of POM that adds life and quality would be ignored in the current health care system. Thus, regardless of the significant potential, most surgeons would waive a surgery based solely on the percentage rate of POM, perhaps using other terms such as ”peri-nonoperative mortality.”

Conclusion

We have proposed a simple and practical formula for decision making regarding surgical vs palliative care in high-risk patients. By assigning a value that is composed of QOL and DOL in each pathway and including the risk of POM, a ratio of values provides a numerical estimation that can be used to show preference over a specific decision. An advantage of this formula, in addition to presenting an arithmetic value that is easier to understand, is that it can be used in shared decision making with patients. We emphasize that this model is only a preliminary concept at this time and has not been tested or validated for clinical use. Validation of such a model will require extensive work and testing within a large-scale population. We hope that this article will serve as a starting point for the development of other models, and that this formula will become more sophisticated with fewer limitations through larger multidisciplinary efforts in the future.

Corresponding author: Rabin Gerrah, MD, Good Samaritan Regional Medical Center, 3640 NW Samaritan Drive, Suite 100B, Corvallis, OR 97330; [email protected].

Disclosures: None reported.

From the Department of Cardiothoracic Surgery, Stanford University, Stanford, CA.

Abstract

Complex cardiac patients are often referred for surgery or palliative care based on the risk of perioperative mortality. This decision ignores factors such as quality of life or duration of life in either surgery or the palliative path. Here, we propose a model to numerically assess and compare the value of surgery vs palliation. This model includes quality and duration of life, as well as risk of perioperative mortality, and involves a patient’s preferences in the decision-making process.

For each pathway, surgery or palliative care, a value is calculated and compared to a normal life value (no disease symptoms and normal life expectancy). The formula is adjusted for the risk of operative mortality. The model produces a ratio of the value of surgery to the value of palliative care that signifies the superiority of one or another. This model calculation presents an objective estimated numerical value to compare the value of surgery and palliative care. It can be applied to every decision-making process before surgery. In general, if a procedure has the potential to significantly extend life in a patient who otherwise has a very short life expectancy with palliation only, performing high-risk surgery would be a reasonable option. A model that provides a numerical value for surgery vs palliative care and includes quality and duration of life in each pathway could be a useful tool for cardiac surgeons in decision making regarding high-risk surgery.

Keywords: high-risk surgery, palliative care, quality of life, life expectancy.

Patients with complex cardiovascular disease are occasionally considered inoperable due to the high risk of surgical mortality. When the risk of perioperative mortality (POM) is predicted to be too high, surgical intervention is denied, and patients are often referred to palliative care. The risk of POM in cardiac surgery is often calculated using large-scale databases, such as the Society of Thoracic Surgeons (STS) records. The STS risk models, which are regularly updated, are based on large data sets and incorporate precise statistical methods for risk adjustment.¹ In general, these calculators provide a percentage value that defines the magnitude of the risk of death, and then an arbitrary range is selected to categorize the procedure as low, medium, or high risk or inoperable status. The STS database does not set a cutoff point or range to define “operability.” Assigning inoperable status to a certain risk rate is problematic, with many ethical, legal, and moral implications, and for this reason, it has mostly remained undefined. In contrast, the low- and medium-risk ranges are easier to define. Another limitation encountered in the STS database is the lack of risk data for less common but very high-risk procedures, such as a triple valve replacement.

A common example where risk classification has been defined is in patients who are candidates for surgical vs transcatheter aortic valve replacement. Some groups have described a risk of <4% as low risk, 4% to 8% as intermediate risk, >8% as high risk, and >15% as inoperable²; for some other groups, a risk of POM >50% is considered extreme risk or inoperable.^3,4 This procedure-specific classification is a useful decision-making tool and helps the surgeon perform an initial risk assessment to allocate a specific patient to a group—operable or nonoperable—only by calculating the risk of surgical death. However, this allocation method does not provide any information on how and when death occurs in either group. These 2 parameters of how and when death occurs define the quality of life (QOL) and the duration of life (DOL), respectively, and together could be considered as the value of life in each pathway. A survivor of a high-risk surgery may benefit from good quality and extended life (a high value), or, on the other end of the spectrum, a high-risk patient who does not undergo surgery is spared the mortality risk of the surgery but dies sooner (low value) with symptoms due to the natural course of the untreated disease.

The central question is, if a surgery is high risk but has the potential of providing a good value (for those who survive it), what QOL and DOL values are acceptable to risk or to justify accepting and proceeding with a risky surgery? Or how high a POM risk is justified to proceed with surgery rather than the alternative palliative care with a certain quality and duration? It is obvious that a decision-making process that is based on POM cannot compare the value of surgery (Vs) and the value of palliation (Vp). Furthermore, it ignores patient preferences and their input, as these are excluded from this decision-making process.

To be able to include QOL and DOL in any decision making, one must precisely describe these parameters. Both QOL and DOL are used for estimation of disease burden by health care administrators, public health experts, insurance agencies, and others. Multiple models have been proposed and used to estimate the overall burden of the disease. Most of the models for this purpose are created for large-scale economic purposes and not for decision making in individual cases.

An important measure is the quality-adjusted life year (QALY). This is an important parameter since it includes both measures of quality and quantity of life.^5,6 QALY is a simplified measure to assess the value of health outcomes, and it has been used in economic calculations to assess mainly the cost-effectiveness of various interventions. We sought to evaluate the utility of a similar method in adding further insight into the surgical decision-making process. In this article, we propose a simple model to compare the value of surgery vs palliative care, similar to QALY. This model includes and adjusts for the quality and the quantity of life, in addition to the risk of POM, in the decision-making process for high-risk patients.

The Model

The 2 decision pathways, surgery and palliative care, are compared for their value. We define the value as the product of QOL and DOL in each pathway and use the severity of the symptoms as a surrogate for QOL. If duration and quality were depicted on the x and y axes of a graph (Figure 1), then the area under the curve would represent the collective value in each situation. Figure 2 shows the timeline and the different pathways with each decision. The value in each situation is calculated in relation to the full value, which is represented as the value of normal life (Vn), that is, life without disease and with normal life expectancy. The values of each decision pathway, the value of surgery (Vs) and the value of palliation (Vp), are then compared to define the benefit for each decision as follows:

If Vs/Vp > 1, the benefit is toward surgery;

If Vs/Vp < 1, the benefit is for palliative care.

Definitions

Both quality and duration of life are presented on a 1-10 scale, 1 being the lowest and 10 the highest value, to yield a product with a value of 100 in normal, disease-free life. Any lower value is presented as a percentage to represent the comparison to the full value. QOL is determined by degradation of full quality with the average level of symptoms. DOL is calculated as a lost time (period of time from death after a specific intervention [surgery or palliation] until death at normal life expectancy) in fraction of full life (death at life expectancy). The Vs is adjusted to exclude the nonsurvivors using the chance of survival (100 – POM risk).

For the DOL under any condition, a 10-year survival rate could be used as a surrogate in this formula. Compared to life expectancy value, using the 10-year survival rate simplifies the calculation since cardiac diseases are more prevalent in older age, close to or beyond the average life expectancy value.

Using the time intervals from the timeline in Figure 2:

dh = time interval from diagnosis to death at life expectancy

dg = time interval from diagnosis to death after successful surgery

df = time interval from diagnosis to death after palliative care

Duration for palliative care:

Duration for surgery:

Adjustment: This value is calculated for those who survive the surgery. To adjust for the POM, it is multiplied by the 100 − POM risk.

Since value is the base for comparison in this model, and it is the product of 2 equally important factors in the formula (severity and duration of symptoms), a factor of 10 was chosen to yield a value of 100, which represents 100% health or absence of symptoms for any duration.

After elimination of normal life expectancy, form the numerator and denominator:

To adjust for surgical outcomes in special circumstances where less than optimal or standard surgical results are expected (eg, in very rare surgeries, limited resource institutions, or suboptimal postoperative surgical care), an optional coefficient R can be added to the numerator (surgical value). This optional coefficient, with values such as 0.8, 0.9 (to degrade the value of surgery) or 1 (standard surgical outcome), adjusts for variability in interinstitutional surgical results or surgeon variability. No coefficient is added to the denominator since palliative care provides minimal differences between clinicians and hospitals. Thus, the final adjusted formula would be as follows:

Example

A 60-year-old patient with a 10% POM risk needs to be allocated to surgical or palliative care. With palliative care, if this patient lived 6 years with average symptoms grade 4, the Vp would be 20; that is, 20% of the normal life value (if he lived 18 years instead without the disease).

Using the formula for calculation of value in each pathway:

With adjustment of 90% survival chance in surgery, 53 × 90% = 48%. In this example, Vs/Vp = 48/20 = 2.4, showing a significant benefit for surgical care. Notably, the unknown value of normal life expectancy is not needed for the calculation of Vs/Vp, since it is the same in both pathways and it is eliminated by calculation in fraction.

Based on this formula, since the duration of surgical symptoms is short, no matter how severe these are, if the potential duration of life after surgery is high (represented by smaller area under the curve in Figure 1), the numerator becomes larger and the value of the surgery grows. For example, if a patient with a 15% risk of POM, which is generally considered inoperable, lives 5 years, as opposed to 2 years with palliative care with mild symptoms (eg 3/10), Vs/Vp would be 2.7, still showing a significant benefit for surgical care.

Discussion

Any surgical intervention is offered with 2 goals in mind, improving QOL and extending DOL. In a high-risk patient, surgery might be declined due to a high risk of POM, and the patient is offered palliative care, which other than providing symptom relief does not change the course of disease and eventually the patient will die due to the untreated disease. In this decision-making method, mostly completed by a care team only, a potential risk of death due to surgery which possibly could cure the patient is traded for immediate survival; however, the symptomatic course ensues until death. This mostly unilateral decision-making process by a care team, which incorporates minimal input from the patient or ignores patient preferences altogether, is based only on POM risk, and roughly includes a single parameter: years of potential life lost (YPLL). YPLL is a measure of premature mortality, and in the setting of surgical intervention, YPLL is the number of years a patient would lose unless a successful surgery were undertaken. Obviously, patients would live longer if a surgery that was intended to save them failed.

In this article, we proposed a simple method to quantify each decision to decide whether to operate or choose surgical care vs palliative care. Since quality and duration of life are both end factors clinicians and patients aspire to in each decision, they can be considered together as the value of each decision. We believe a numerical framework would provide an objective way to assist both the patient at high risk and the care team in the decision-making process.

The 2 parameters we consider are DOL and QOL. DOL, or survival, can be extracted from large-scale data using statistical methods that have been developed to predict survival under various conditions, such as Kaplan-Meier curves. These methods present the chance of survival in percentages in a defined time frame, such as a 5- or 10-year period.

While the DOL is a numerical parameter and quantifiable, the QOL is a more complex entity. This subjective parameter bears multiple definitions, aspects, and categories, and therefore multiple scales for quantification of QOL have been proposed. These scales have been used extensively for the purpose of health determination in health care policy and economic planning. Most scales acknowledge that QOL is multifactorial and includes interrelated aspects such as mental and socioeconomic factors. We have also noticed that QOL is better determined by the palliative care team than surgeons, so including these care providers in the decision-making process might reduce surgeon bias.

Since our purpose here is only to assist with the decision on medical intervention, we focus on physical QOL. Multiple scales are used to assess health-related QOL, such as the Assessment of Quality of Life (AQoL)-8D,⁷ EuroQol-5 Dimension (EQ-5D),⁸ 15D,⁹ and the 36-Item Short Form Survey (SF-36).¹⁰ These complex scales are built for systematic reviews, and they are not practical for a clinical user. To simplify and keep this practical, we define QOL by using the severity or grade of symptoms related to the disease the patient has on a scale of 0 to 10. The severity of symptoms can be easily determined using available scales. An applicable scale for this purpose is the Edmonton Symptom Assessment Scale (ESAS), which has been in use for years and has evolved as a useful tool in the medical field.¹¹

Once DOL and QOL are determined on a 1-10 scale, the multiplied value then provides a product that we consider a value. The highest value hoped for in each decision is the achievement of the best QOL and DOL, a value of 100. In Figure 1, a graphic presentation of value in each decision is best seen as the area under the curve. As shown, a successful surgery, even when accompanied by significant symptoms during initial recovery, has a chance (100 – risk of POM%) to gain a larger area under curve (value) by achieving a longer life with no or fewer symptoms. However, in palliative care, progressing disease and even palliated symptoms with a shorter life expectancy impose a large burden on the patient and a much lower value. Note that in this calculation, life expectancy, which is an important but unpredictable factor, is initially included; however, by ratio comparison, it is eliminated, simplifying the calculation further.

Using this formula in different settings reveals that high-risk surgery has a greater potential to reduce YPLL in the general population. Based on this formula, compared to a surgery with potential to significantly extend DOL, a definite shorter and symptomatic life course with palliative care makes it a significantly less favorable option. In fact, in the cardiovascular field, palliative care has minimal or no effect on natural history, as the mechanism of illness is mechanical, such as occlusion of coronary arteries or valve dysfunction, leading eventually to heart failure and death. In a study by Xu et al, although palliative care reduced readmission rates and improved symptoms on a variety of scales, there was no effect on mortality and QOL in patients with heart failure.¹²

No model in this field has proven to be ideal, and this model bears multiple limitations as well. We have used severity of symptoms as a surrogate for QOL based on the fact that cardiac patients with different pathologies who are untreated will have a common final pathway with development of heart failure symptoms that dictate their QOL. Also, grading QOL is a difficult task at times. Even a model such as QALY, which is one of the most used, is not a perfect model and is not free of problems.⁶ The difference in surgical results and life expectancy between sexes and ethnic groups might be a source of bias in this formula. Also, multiple factors directly and indirectly affect QOL and DOL and create inaccuracies; therefore, making an exact science from an inexact one naturally relies on multiple assumptions. Although it has previously been shown that most POM occurs in a short period of time after cardiac surgery,¹³ long-term complications that potentially degrade QOL are not included in this model. By applying this model, one must assume indefinite economic resources. Moreover, applying a single mathematical model in a biologic system and in the general population has intrinsic shortcomings, and it must overlook many other factors (eg, ethical, legal). For example, it will be hard to justify a failed surgery with 15% risk of POM undertaken to eliminate the severe long-lasting symptoms of a disease, while the outcome of a successful surgery with a 20% risk of POM that adds life and quality would be ignored in the current health care system. Thus, regardless of the significant potential, most surgeons would waive a surgery based solely on the percentage rate of POM, perhaps using other terms such as ”peri-nonoperative mortality.”

Conclusion

We have proposed a simple and practical formula for decision making regarding surgical vs palliative care in high-risk patients. By assigning a value that is composed of QOL and DOL in each pathway and including the risk of POM, a ratio of values provides a numerical estimation that can be used to show preference over a specific decision. An advantage of this formula, in addition to presenting an arithmetic value that is easier to understand, is that it can be used in shared decision making with patients. We emphasize that this model is only a preliminary concept at this time and has not been tested or validated for clinical use. Validation of such a model will require extensive work and testing within a large-scale population. We hope that this article will serve as a starting point for the development of other models, and that this formula will become more sophisticated with fewer limitations through larger multidisciplinary efforts in the future.

Corresponding author: Rabin Gerrah, MD, Good Samaritan Regional Medical Center, 3640 NW Samaritan Drive, Suite 100B, Corvallis, OR 97330; [email protected].

Disclosures: None reported.

From the Department of Cardiothoracic Surgery, Stanford University, Stanford, CA.

Abstract

Complex cardiac patients are often referred for surgery or palliative care based on the risk of perioperative mortality. This decision ignores factors such as quality of life or duration of life in either surgery or the palliative path. Here, we propose a model to numerically assess and compare the value of surgery vs palliation. This model includes quality and duration of life, as well as risk of perioperative mortality, and involves a patient’s preferences in the decision-making process.

For each pathway, surgery or palliative care, a value is calculated and compared to a normal life value (no disease symptoms and normal life expectancy). The formula is adjusted for the risk of operative mortality. The model produces a ratio of the value of surgery to the value of palliative care that signifies the superiority of one or another. This model calculation presents an objective estimated numerical value to compare the value of surgery and palliative care. It can be applied to every decision-making process before surgery. In general, if a procedure has the potential to significantly extend life in a patient who otherwise has a very short life expectancy with palliation only, performing high-risk surgery would be a reasonable option. A model that provides a numerical value for surgery vs palliative care and includes quality and duration of life in each pathway could be a useful tool for cardiac surgeons in decision making regarding high-risk surgery.

Keywords: high-risk surgery, palliative care, quality of life, life expectancy.

Patients with complex cardiovascular disease are occasionally considered inoperable due to the high risk of surgical mortality. When the risk of perioperative mortality (POM) is predicted to be too high, surgical intervention is denied, and patients are often referred to palliative care. The risk of POM in cardiac surgery is often calculated using large-scale databases, such as the Society of Thoracic Surgeons (STS) records. The STS risk models, which are regularly updated, are based on large data sets and incorporate precise statistical methods for risk adjustment.¹ In general, these calculators provide a percentage value that defines the magnitude of the risk of death, and then an arbitrary range is selected to categorize the procedure as low, medium, or high risk or inoperable status. The STS database does not set a cutoff point or range to define “operability.” Assigning inoperable status to a certain risk rate is problematic, with many ethical, legal, and moral implications, and for this reason, it has mostly remained undefined. In contrast, the low- and medium-risk ranges are easier to define. Another limitation encountered in the STS database is the lack of risk data for less common but very high-risk procedures, such as a triple valve replacement.

A common example where risk classification has been defined is in patients who are candidates for surgical vs transcatheter aortic valve replacement. Some groups have described a risk of <4% as low risk, 4% to 8% as intermediate risk, >8% as high risk, and >15% as inoperable²; for some other groups, a risk of POM >50% is considered extreme risk or inoperable.^3,4 This procedure-specific classification is a useful decision-making tool and helps the surgeon perform an initial risk assessment to allocate a specific patient to a group—operable or nonoperable—only by calculating the risk of surgical death. However, this allocation method does not provide any information on how and when death occurs in either group. These 2 parameters of how and when death occurs define the quality of life (QOL) and the duration of life (DOL), respectively, and together could be considered as the value of life in each pathway. A survivor of a high-risk surgery may benefit from good quality and extended life (a high value), or, on the other end of the spectrum, a high-risk patient who does not undergo surgery is spared the mortality risk of the surgery but dies sooner (low value) with symptoms due to the natural course of the untreated disease.

The central question is, if a surgery is high risk but has the potential of providing a good value (for those who survive it), what QOL and DOL values are acceptable to risk or to justify accepting and proceeding with a risky surgery? Or how high a POM risk is justified to proceed with surgery rather than the alternative palliative care with a certain quality and duration? It is obvious that a decision-making process that is based on POM cannot compare the value of surgery (Vs) and the value of palliation (Vp). Furthermore, it ignores patient preferences and their input, as these are excluded from this decision-making process.

To be able to include QOL and DOL in any decision making, one must precisely describe these parameters. Both QOL and DOL are used for estimation of disease burden by health care administrators, public health experts, insurance agencies, and others. Multiple models have been proposed and used to estimate the overall burden of the disease. Most of the models for this purpose are created for large-scale economic purposes and not for decision making in individual cases.

An important measure is the quality-adjusted life year (QALY). This is an important parameter since it includes both measures of quality and quantity of life.^5,6 QALY is a simplified measure to assess the value of health outcomes, and it has been used in economic calculations to assess mainly the cost-effectiveness of various interventions. We sought to evaluate the utility of a similar method in adding further insight into the surgical decision-making process. In this article, we propose a simple model to compare the value of surgery vs palliative care, similar to QALY. This model includes and adjusts for the quality and the quantity of life, in addition to the risk of POM, in the decision-making process for high-risk patients.

The Model

The 2 decision pathways, surgery and palliative care, are compared for their value. We define the value as the product of QOL and DOL in each pathway and use the severity of the symptoms as a surrogate for QOL. If duration and quality were depicted on the x and y axes of a graph (Figure 1), then the area under the curve would represent the collective value in each situation. Figure 2 shows the timeline and the different pathways with each decision. The value in each situation is calculated in relation to the full value, which is represented as the value of normal life (Vn), that is, life without disease and with normal life expectancy. The values of each decision pathway, the value of surgery (Vs) and the value of palliation (Vp), are then compared to define the benefit for each decision as follows:

If Vs/Vp > 1, the benefit is toward surgery;

If Vs/Vp < 1, the benefit is for palliative care.

Definitions

Both quality and duration of life are presented on a 1-10 scale, 1 being the lowest and 10 the highest value, to yield a product with a value of 100 in normal, disease-free life. Any lower value is presented as a percentage to represent the comparison to the full value. QOL is determined by degradation of full quality with the average level of symptoms. DOL is calculated as a lost time (period of time from death after a specific intervention [surgery or palliation] until death at normal life expectancy) in fraction of full life (death at life expectancy). The Vs is adjusted to exclude the nonsurvivors using the chance of survival (100 – POM risk).

For the DOL under any condition, a 10-year survival rate could be used as a surrogate in this formula. Compared to life expectancy value, using the 10-year survival rate simplifies the calculation since cardiac diseases are more prevalent in older age, close to or beyond the average life expectancy value.

Using the time intervals from the timeline in Figure 2:

dh = time interval from diagnosis to death at life expectancy

dg = time interval from diagnosis to death after successful surgery

df = time interval from diagnosis to death after palliative care

Duration for palliative care:

Duration for surgery:

Adjustment: This value is calculated for those who survive the surgery. To adjust for the POM, it is multiplied by the 100 − POM risk.

Since value is the base for comparison in this model, and it is the product of 2 equally important factors in the formula (severity and duration of symptoms), a factor of 10 was chosen to yield a value of 100, which represents 100% health or absence of symptoms for any duration.

After elimination of normal life expectancy, form the numerator and denominator:

To adjust for surgical outcomes in special circumstances where less than optimal or standard surgical results are expected (eg, in very rare surgeries, limited resource institutions, or suboptimal postoperative surgical care), an optional coefficient R can be added to the numerator (surgical value). This optional coefficient, with values such as 0.8, 0.9 (to degrade the value of surgery) or 1 (standard surgical outcome), adjusts for variability in interinstitutional surgical results or surgeon variability. No coefficient is added to the denominator since palliative care provides minimal differences between clinicians and hospitals. Thus, the final adjusted formula would be as follows:

Example

A 60-year-old patient with a 10% POM risk needs to be allocated to surgical or palliative care. With palliative care, if this patient lived 6 years with average symptoms grade 4, the Vp would be 20; that is, 20% of the normal life value (if he lived 18 years instead without the disease).

Using the formula for calculation of value in each pathway:

With adjustment of 90% survival chance in surgery, 53 × 90% = 48%. In this example, Vs/Vp = 48/20 = 2.4, showing a significant benefit for surgical care. Notably, the unknown value of normal life expectancy is not needed for the calculation of Vs/Vp, since it is the same in both pathways and it is eliminated by calculation in fraction.

Based on this formula, since the duration of surgical symptoms is short, no matter how severe these are, if the potential duration of life after surgery is high (represented by smaller area under the curve in Figure 1), the numerator becomes larger and the value of the surgery grows. For example, if a patient with a 15% risk of POM, which is generally considered inoperable, lives 5 years, as opposed to 2 years with palliative care with mild symptoms (eg 3/10), Vs/Vp would be 2.7, still showing a significant benefit for surgical care.

Discussion

Any surgical intervention is offered with 2 goals in mind, improving QOL and extending DOL. In a high-risk patient, surgery might be declined due to a high risk of POM, and the patient is offered palliative care, which other than providing symptom relief does not change the course of disease and eventually the patient will die due to the untreated disease. In this decision-making method, mostly completed by a care team only, a potential risk of death due to surgery which possibly could cure the patient is traded for immediate survival; however, the symptomatic course ensues until death. This mostly unilateral decision-making process by a care team, which incorporates minimal input from the patient or ignores patient preferences altogether, is based only on POM risk, and roughly includes a single parameter: years of potential life lost (YPLL). YPLL is a measure of premature mortality, and in the setting of surgical intervention, YPLL is the number of years a patient would lose unless a successful surgery were undertaken. Obviously, patients would live longer if a surgery that was intended to save them failed.

In this article, we proposed a simple method to quantify each decision to decide whether to operate or choose surgical care vs palliative care. Since quality and duration of life are both end factors clinicians and patients aspire to in each decision, they can be considered together as the value of each decision. We believe a numerical framework would provide an objective way to assist both the patient at high risk and the care team in the decision-making process.

The 2 parameters we consider are DOL and QOL. DOL, or survival, can be extracted from large-scale data using statistical methods that have been developed to predict survival under various conditions, such as Kaplan-Meier curves. These methods present the chance of survival in percentages in a defined time frame, such as a 5- or 10-year period.

While the DOL is a numerical parameter and quantifiable, the QOL is a more complex entity. This subjective parameter bears multiple definitions, aspects, and categories, and therefore multiple scales for quantification of QOL have been proposed. These scales have been used extensively for the purpose of health determination in health care policy and economic planning. Most scales acknowledge that QOL is multifactorial and includes interrelated aspects such as mental and socioeconomic factors. We have also noticed that QOL is better determined by the palliative care team than surgeons, so including these care providers in the decision-making process might reduce surgeon bias.

Since our purpose here is only to assist with the decision on medical intervention, we focus on physical QOL. Multiple scales are used to assess health-related QOL, such as the Assessment of Quality of Life (AQoL)-8D,⁷ EuroQol-5 Dimension (EQ-5D),⁸ 15D,⁹ and the 36-Item Short Form Survey (SF-36).¹⁰ These complex scales are built for systematic reviews, and they are not practical for a clinical user. To simplify and keep this practical, we define QOL by using the severity or grade of symptoms related to the disease the patient has on a scale of 0 to 10. The severity of symptoms can be easily determined using available scales. An applicable scale for this purpose is the Edmonton Symptom Assessment Scale (ESAS), which has been in use for years and has evolved as a useful tool in the medical field.¹¹

Once DOL and QOL are determined on a 1-10 scale, the multiplied value then provides a product that we consider a value. The highest value hoped for in each decision is the achievement of the best QOL and DOL, a value of 100. In Figure 1, a graphic presentation of value in each decision is best seen as the area under the curve. As shown, a successful surgery, even when accompanied by significant symptoms during initial recovery, has a chance (100 – risk of POM%) to gain a larger area under curve (value) by achieving a longer life with no or fewer symptoms. However, in palliative care, progressing disease and even palliated symptoms with a shorter life expectancy impose a large burden on the patient and a much lower value. Note that in this calculation, life expectancy, which is an important but unpredictable factor, is initially included; however, by ratio comparison, it is eliminated, simplifying the calculation further.

Using this formula in different settings reveals that high-risk surgery has a greater potential to reduce YPLL in the general population. Based on this formula, compared to a surgery with potential to significantly extend DOL, a definite shorter and symptomatic life course with palliative care makes it a significantly less favorable option. In fact, in the cardiovascular field, palliative care has minimal or no effect on natural history, as the mechanism of illness is mechanical, such as occlusion of coronary arteries or valve dysfunction, leading eventually to heart failure and death. In a study by Xu et al, although palliative care reduced readmission rates and improved symptoms on a variety of scales, there was no effect on mortality and QOL in patients with heart failure.¹²

No model in this field has proven to be ideal, and this model bears multiple limitations as well. We have used severity of symptoms as a surrogate for QOL based on the fact that cardiac patients with different pathologies who are untreated will have a common final pathway with development of heart failure symptoms that dictate their QOL. Also, grading QOL is a difficult task at times. Even a model such as QALY, which is one of the most used, is not a perfect model and is not free of problems.⁶ The difference in surgical results and life expectancy between sexes and ethnic groups might be a source of bias in this formula. Also, multiple factors directly and indirectly affect QOL and DOL and create inaccuracies; therefore, making an exact science from an inexact one naturally relies on multiple assumptions. Although it has previously been shown that most POM occurs in a short period of time after cardiac surgery,¹³ long-term complications that potentially degrade QOL are not included in this model. By applying this model, one must assume indefinite economic resources. Moreover, applying a single mathematical model in a biologic system and in the general population has intrinsic shortcomings, and it must overlook many other factors (eg, ethical, legal). For example, it will be hard to justify a failed surgery with 15% risk of POM undertaken to eliminate the severe long-lasting symptoms of a disease, while the outcome of a successful surgery with a 20% risk of POM that adds life and quality would be ignored in the current health care system. Thus, regardless of the significant potential, most surgeons would waive a surgery based solely on the percentage rate of POM, perhaps using other terms such as ”peri-nonoperative mortality.”

Conclusion

We have proposed a simple and practical formula for decision making regarding surgical vs palliative care in high-risk patients. By assigning a value that is composed of QOL and DOL in each pathway and including the risk of POM, a ratio of values provides a numerical estimation that can be used to show preference over a specific decision. An advantage of this formula, in addition to presenting an arithmetic value that is easier to understand, is that it can be used in shared decision making with patients. We emphasize that this model is only a preliminary concept at this time and has not been tested or validated for clinical use. Validation of such a model will require extensive work and testing within a large-scale population. We hope that this article will serve as a starting point for the development of other models, and that this formula will become more sophisticated with fewer limitations through larger multidisciplinary efforts in the future.

Corresponding author: Rabin Gerrah, MD, Good Samaritan Regional Medical Center, 3640 NW Samaritan Drive, Suite 100B, Corvallis, OR 97330; [email protected].

Disclosures: None reported.

A 75-year-old male is seen in the emergency department for shortness of breath. D-dimer is very high, and a CT scan chest reveals bilateral pulmonary emboli without right heart strain. He is admitted and started on enoxaparin 1 mg/kg subcutaneously every 12 hours.

By the next morning, he is feeling better and wants to discuss discharge to home and follow-up plans.

Two months ago he presented with abdominal pain and evaluation revealed he had a pancreatic head mass with liver metastases. A liver biopsy was positive for adenocarcinoma consistent with pancreas primary. CA 19-9 level was 1,200 U/mL and he was started on FOLFIRINOX chemotherapy – which he has tolerated well thus far. CA 19-9 and follow-up CT scan show early response to chemotherapy.

Of course, this case raises many questions. Given how successful some directed biomarker-positive therapies are now, you would want to know his microsatellite instability (MSI)/progressive death–ligand 1 (PD-L1) and BRCA mutation status. A high PD-L1 positivity or MSI deficiency would suggest immunoantibody therapy and a BRCA mutation might suggest a poly (ADP-ribose) polymerase inhibitor could play a role.

However, let’s use this case to discuss his venous thromboembolism (VTE) .

Studies show that metastatic cancer patients on chemotherapy might experience a VTE episode of deep vein thrombosis (DVT) or pulmonary embolism (PE) or both as high as 20% of the time during their cancer course and therapy. This patient would be among those who experience the highest incidence of VTE because of the liver metastasis from the pancreatic adenocarcinoma.

So, what to do? Standard treatment of his pulmonary emboli would include either enoxaparin therapeutic dosing 1 mg/kg subcutaneously q12H or 1.5 mg/kg q24H for 3 months. At 3 months, repeat a CT chest scan to show resolution of pulmonary emboli and/or DVT or both, and repeat D-dimer, which should now be well under 1.

But then, there is a second decision to make: Can you stop anticoagulation if his clots have resolved? The answer is yes. If the clots were provoked and the provoking feature is gone you can stop anticoagulation. Patients with pregnancy, on a birth control pill, or on a long trip where immobilization occurred for a extended time (such as driving or flying) can have anticoagulation stopped because the provoking feature is gone, but this is not true in this case. This patient’s pancreas cancer and chemotherapy are ongoing and he will be at increased risk to clot once again if anticoagulation is stopped.

Should this patient have a hypercoagulable workup which might include protein C, protein S, and antithrombin levels? Remember this is quite rare and patients with these deficiencies usually present in their teens or 20s with increased clotting issues. The more common hypercoagulable workup would include checking for factor V Leiden and prothrombin G20210A mutations, as well as acquired antiphospholipid antibodies such as beta₂ glycoprotein I, anticardiolipin, and the lupus inhibitor. However, in this 75-year-old cancer patient, these are not necessary or even relevant since his VTE was clearly provoked by metastatic cancer on chemotherapy.

Unfortunately, with metastatic active cancer, anticoagulation would need to be continued at full or possibly half therapeutic dose. Of course, enoxaparin injections can get tiresome for the patient and data suggest the same result can be achieved either with initial management or by continuing anticoagulation management using either rivaroxaban or apixaban.

Wouldn’t it have been better if this patient had never experienced VTE in the first place? Is that possible?

Yes, data suggest that it is. Higher-risk patients like this one could benefit from prophylactic anticoagulation. The Khorana predictive model gives us a simple clinical means to evaluate this and decide who might be at highest VTE risk and who could benefit from low-dose preventive anticoagulation.

In summary, cancer patients undergoing treatment for metastatic disease are at increased risk for symptomatic VTE. Once diagnosed, therapy is usually very effective, but may need to be prolonged as long as the cancer is still active or else, the VTE could recur. Preventive therapy for high-risk patients would be reasonable.

Dr. Henry is a medical oncologist with the Abramson Cancer Center at the University of Pennsylvania, Philadelphia.

A 75-year-old male is seen in the emergency department for shortness of breath. D-dimer is very high, and a CT scan chest reveals bilateral pulmonary emboli without right heart strain. He is admitted and started on enoxaparin 1 mg/kg subcutaneously every 12 hours.

By the next morning, he is feeling better and wants to discuss discharge to home and follow-up plans.

Two months ago he presented with abdominal pain and evaluation revealed he had a pancreatic head mass with liver metastases. A liver biopsy was positive for adenocarcinoma consistent with pancreas primary. CA 19-9 level was 1,200 U/mL and he was started on FOLFIRINOX chemotherapy – which he has tolerated well thus far. CA 19-9 and follow-up CT scan show early response to chemotherapy.

Of course, this case raises many questions. Given how successful some directed biomarker-positive therapies are now, you would want to know his microsatellite instability (MSI)/progressive death–ligand 1 (PD-L1) and BRCA mutation status. A high PD-L1 positivity or MSI deficiency would suggest immunoantibody therapy and a BRCA mutation might suggest a poly (ADP-ribose) polymerase inhibitor could play a role.

However, let’s use this case to discuss his venous thromboembolism (VTE) .

Studies show that metastatic cancer patients on chemotherapy might experience a VTE episode of deep vein thrombosis (DVT) or pulmonary embolism (PE) or both as high as 20% of the time during their cancer course and therapy. This patient would be among those who experience the highest incidence of VTE because of the liver metastasis from the pancreatic adenocarcinoma.

So, what to do? Standard treatment of his pulmonary emboli would include either enoxaparin therapeutic dosing 1 mg/kg subcutaneously q12H or 1.5 mg/kg q24H for 3 months. At 3 months, repeat a CT chest scan to show resolution of pulmonary emboli and/or DVT or both, and repeat D-dimer, which should now be well under 1.

But then, there is a second decision to make: Can you stop anticoagulation if his clots have resolved? The answer is yes. If the clots were provoked and the provoking feature is gone you can stop anticoagulation. Patients with pregnancy, on a birth control pill, or on a long trip where immobilization occurred for a extended time (such as driving or flying) can have anticoagulation stopped because the provoking feature is gone, but this is not true in this case. This patient’s pancreas cancer and chemotherapy are ongoing and he will be at increased risk to clot once again if anticoagulation is stopped.

Should this patient have a hypercoagulable workup which might include protein C, protein S, and antithrombin levels? Remember this is quite rare and patients with these deficiencies usually present in their teens or 20s with increased clotting issues. The more common hypercoagulable workup would include checking for factor V Leiden and prothrombin G20210A mutations, as well as acquired antiphospholipid antibodies such as beta₂ glycoprotein I, anticardiolipin, and the lupus inhibitor. However, in this 75-year-old cancer patient, these are not necessary or even relevant since his VTE was clearly provoked by metastatic cancer on chemotherapy.

Unfortunately, with metastatic active cancer, anticoagulation would need to be continued at full or possibly half therapeutic dose. Of course, enoxaparin injections can get tiresome for the patient and data suggest the same result can be achieved either with initial management or by continuing anticoagulation management using either rivaroxaban or apixaban.

Wouldn’t it have been better if this patient had never experienced VTE in the first place? Is that possible?

Yes, data suggest that it is. Higher-risk patients like this one could benefit from prophylactic anticoagulation. The Khorana predictive model gives us a simple clinical means to evaluate this and decide who might be at highest VTE risk and who could benefit from low-dose preventive anticoagulation.

In summary, cancer patients undergoing treatment for metastatic disease are at increased risk for symptomatic VTE. Once diagnosed, therapy is usually very effective, but may need to be prolonged as long as the cancer is still active or else, the VTE could recur. Preventive therapy for high-risk patients would be reasonable.

Dr. Henry is a medical oncologist with the Abramson Cancer Center at the University of Pennsylvania, Philadelphia.

A 75-year-old male is seen in the emergency department for shortness of breath. D-dimer is very high, and a CT scan chest reveals bilateral pulmonary emboli without right heart strain. He is admitted and started on enoxaparin 1 mg/kg subcutaneously every 12 hours.

By the next morning, he is feeling better and wants to discuss discharge to home and follow-up plans.

Two months ago he presented with abdominal pain and evaluation revealed he had a pancreatic head mass with liver metastases. A liver biopsy was positive for adenocarcinoma consistent with pancreas primary. CA 19-9 level was 1,200 U/mL and he was started on FOLFIRINOX chemotherapy – which he has tolerated well thus far. CA 19-9 and follow-up CT scan show early response to chemotherapy.

Of course, this case raises many questions. Given how successful some directed biomarker-positive therapies are now, you would want to know his microsatellite instability (MSI)/progressive death–ligand 1 (PD-L1) and BRCA mutation status. A high PD-L1 positivity or MSI deficiency would suggest immunoantibody therapy and a BRCA mutation might suggest a poly (ADP-ribose) polymerase inhibitor could play a role.

However, let’s use this case to discuss his venous thromboembolism (VTE) .

Studies show that metastatic cancer patients on chemotherapy might experience a VTE episode of deep vein thrombosis (DVT) or pulmonary embolism (PE) or both as high as 20% of the time during their cancer course and therapy. This patient would be among those who experience the highest incidence of VTE because of the liver metastasis from the pancreatic adenocarcinoma.

So, what to do? Standard treatment of his pulmonary emboli would include either enoxaparin therapeutic dosing 1 mg/kg subcutaneously q12H or 1.5 mg/kg q24H for 3 months. At 3 months, repeat a CT chest scan to show resolution of pulmonary emboli and/or DVT or both, and repeat D-dimer, which should now be well under 1.

But then, there is a second decision to make: Can you stop anticoagulation if his clots have resolved? The answer is yes. If the clots were provoked and the provoking feature is gone you can stop anticoagulation. Patients with pregnancy, on a birth control pill, or on a long trip where immobilization occurred for a extended time (such as driving or flying) can have anticoagulation stopped because the provoking feature is gone, but this is not true in this case. This patient’s pancreas cancer and chemotherapy are ongoing and he will be at increased risk to clot once again if anticoagulation is stopped.

Should this patient have a hypercoagulable workup which might include protein C, protein S, and antithrombin levels? Remember this is quite rare and patients with these deficiencies usually present in their teens or 20s with increased clotting issues. The more common hypercoagulable workup would include checking for factor V Leiden and prothrombin G20210A mutations, as well as acquired antiphospholipid antibodies such as beta₂ glycoprotein I, anticardiolipin, and the lupus inhibitor. However, in this 75-year-old cancer patient, these are not necessary or even relevant since his VTE was clearly provoked by metastatic cancer on chemotherapy.

Unfortunately, with metastatic active cancer, anticoagulation would need to be continued at full or possibly half therapeutic dose. Of course, enoxaparin injections can get tiresome for the patient and data suggest the same result can be achieved either with initial management or by continuing anticoagulation management using either rivaroxaban or apixaban.

Wouldn’t it have been better if this patient had never experienced VTE in the first place? Is that possible?

Yes, data suggest that it is. Higher-risk patients like this one could benefit from prophylactic anticoagulation. The Khorana predictive model gives us a simple clinical means to evaluate this and decide who might be at highest VTE risk and who could benefit from low-dose preventive anticoagulation.

In summary, cancer patients undergoing treatment for metastatic disease are at increased risk for symptomatic VTE. Once diagnosed, therapy is usually very effective, but may need to be prolonged as long as the cancer is still active or else, the VTE could recur. Preventive therapy for high-risk patients would be reasonable.

Dr. Henry is a medical oncologist with the Abramson Cancer Center at the University of Pennsylvania, Philadelphia.