Diane J. Madlon-Kay, MD

From 1% to 4% of full-term infants are readmitted to the hospital for jaundice in the first week of life, representing as many as 109,000 admissions¹ Delayed diagnosis of jaundice puts babies at risk for kernicterus, which had virtually disappeared in the United States but is now on the rise. There are anecdotal reports of 22 full-term infants born in the early 1990s who developed kernicterus after discharge from the hospital within 48 hours of birth.¹ The Joint Commission on Accreditation of Healthcare Organizations (JCAHO) recently issued a Sentinel Event Alert recommending that organizations take steps to raise awareness among neonatal caregivers of the potential for kernicterus and its risk factors by reviewing their current patient care processes with regard to the identification and management of hyperbilirubinemia in newborns and by identifying risk reduction strategies that could enhance the effectiveness of these processes.²

The JCAHO alert cites the American Academy of Pediatrics (AAP) Practice Parameter for Management of Hyperbilirubinemia in the Healthy Term Newborn, which is based on available data and expert consensus, as an example of a guideline for identifying at-risk newborns and their diagnosis and treatment. The AAP guideline suggests checking for jaundice by blanching the skin with digital pressure to reveal its underlying color. The guideline states that clinical assessment must be done in a well-lighted room and suggests that as the bilirubin level rises, the extent of caudal progression may be helpful in quantifying the degree of jaundice.³

The AAP jaundice guideline suggests that the use of an icterometer (transcutaneous jaundice meter) may be helpful in the clinical assessment of jaundice.³ A variety of instruments have been tested in different patient populations.^4-8 A potential role for such devices is their use by parents. The Ingram icterometer (Cascade Health Care Products, Salem, Ore.) is particularly promising because of its low cost ($17) and simplicity.⁵ It is a simple handheld device, made of clear plastic, on which are painted 5 transverse stripes of precise and graded hue. The stripes and spaces between them are 3/16 inch wide and are numbered from 1 (lightest in color) to 5 (darkest). When the icterometer is used, the painted side is pressed against the tip of the infant’s nose until the skin becomes blanched. The yellow color of the blanched skin can then be matched with the yellow stripes on the instrument, and a jaundice score assigned.

The purpose of my study was to determine whether mothers can accurately assess the presence and severity of jaundice in their newborns, both visually and with an icterometer, after hospital discharge. Maternal assessments were compared with bilirubin levels and home health nurse assessments to determine their accuracy. Serum bilirubin levels were used as the reference standard. Maternal comfort with the examination techniques was also assessed.

Methods

This study was approved by the Ramsey (now Regions) Hospital institutional review board. Mothers who gave birth at Regions Hospital in St. Paul, Minn., participated in the study. Mothers on the postpartum ward were invited to participate, but were excluded if they were not proficient in reading English, did not have a telephone, or lived more than 10 miles from the hospital. Infants were excluded if they were in the intensive care nursery, were not discharged on the same day as the mother, or if they received phototherapy. Mothers were advised to follow their health care providers’ instructions about timing for the first follow-up visit, and any provider instructions regarding jaundice.

After obtaining consent, the author or a study assistant showed the mothers how to examine their infants for jaundice by 2 methods. Each mother was instructed to examine her baby in a well-lighted room. First, the mother was shown how to look for jaundice by digitally blanching the skin on the cheek. The mother then documented whether she saw any underlying yellow color on her baby. Next, the mother was shown how to determine the caudal progression of the jaundice and to draw a horizontal line on an illustration of a baby corresponding to where the jaundice ended. The distance from the top of the infant’s head to the line drawn by the mother was used to determine the caudal progression. The mother was then shown how to use the Ingram icterometer and obtain a reading from the baby’s nose. Each mother was given an icterometer and a study booklet to document her examination for a total of 7 days, beginning the day after discharge from the hospital. The study booklet also contained some demographic questions, and questions about the mother’s comfort level with both methods of jaundice assessment. The mother was instructed to return the booklet and icterometer by mail when completed. The mother was sent a $25 gift certificate when the study materials were returned.

Within 7 days of discharge, a home health nurse visited each mother and infant in the home. The nurses were trained in the same methods of clinically assessing jaundice, and they assessed each infant by visually determining the caudal progression and by use of the icterometer. The nurse did not share the results of her examination with the mother. The nurse obtained bilirubin levels from all infants and notified the infants’ health care providers of any bilirubin levels higher than 14 mg/dL.

Standard descriptive statistics were calculated for all variables. Categorical relationships were assessed using kappa and chi-square statistics, as appropriate. All analyses were performed using Statistical Package for Social Sciences for Windows, version 10.0.5.

Results

A total 113 of 177 mothers returned their study packets. Home health nurses visited 96 of the 113 mothers; the other 17 mothers were not visited because they declined the visits or could not be located. Although all babies were to have serum bilirubin levels determined whether or not they appeared jaundiced, only 90 of the 96 infants had the blood test. For the other 6 infants, either insufficient blood was drawn or the mother refused the test. On the day of the nurse’s visits, mothers documented in their study booklets the caudal progression of jaundice (for 56 infants) and icterometer readings (for 55 infants).

The educational levels of the mothers were as follows: 15% completed grade school or less; 40% completed high school; and 45% completed college. The mothers reported being from the following racial and ethnic groups: white, 59%; Hispanic, 16%; black, 14%; Asian, 8%; and other, 3%. A total of 53% of the women were primiparous, 84% completed examination forms for their babies for all 7 days, and 53% assessed their infants as being jaundiced during the study.

On the day of the nurse’s visit, there was moderate agreement between the nurses and the mothers about the presence of jaundice in the infants (= 0.50; P < .001). For those infants with jaundice, there was little agreement on the extent of caudal progression between the nurses and the mothers (correlation = 0.36; P > 0.1), but there was moderate agreement between their icterometer readings (correlation = 0.58; P < .05).

The total serum bilirubin results ranged from 0.8 mg/dL to 18.8 mg/dL, with a mean of 7.4 mg/dL. The mean bilirubin level of infants thought to be jaundiced by their mothers was 11.3 mg/dL, while the mean bilirubin of infants not thought to be jaundiced was 4.8 mg/dL (P < .001).

The mothers’ icterometer readings and determinations of jaundice to the nipple line or below it are compared with bilirubin levels in (Table 1). (Table 2) summarizes the diagnostic accuracy of jaundice extending to the nipple line or below it, and for icterometer readings of 2.5, in identifying bilirubin levels of 12 mg/dL and 17 mg/dL. A bilirubin level of 12 mg/dL is the level at which the AAP guideline suggests considering phototherapy for infants aged 24 to 47 hours, and 17 mg/dL is the level at which phototherapy should be considered for infants older than 72 hours.³

The mothers of the 3 infants with bilirubin levels 17 mg/dL recognized that their infants were jaundiced and determined that the jaundice extended below the nipple line. The icterometer readings obtained by the mothers were 2.5, 3, and 3.5. The corresponding icterometer readings by the nurses were 4.5, 3.5 and 3.

The study booklet contained 6 questions about the mothers’ reactions to the study. Almost all of the mothers (98%) responded that the method for checking for caudal progression of jaundice was explained clearly, and even more (99%) felt the use of the icterometer was explained clearly. A total of 69% of the mothers felt it was “very easy” or “easy” to check for caudal progression, and 80% felt it was “very easy” or “easy” to use the icterometer. Forty-six percent of the mothers reported that checking their babies for jaundice made them “very worried” or “somewhat worried” about their babies’ health. Mothers with less education were significantly more likely to report being worried than mothers with higher education levels (P < .05). However, 93% of the mothers reported that checking their babies for jaundice made them “very reassured” or “somewhat reassured” about their babies’ health.

TABLE 1
Maternal assessment of jaundice, by caudal progression and icterometer readings, compared with serum bilirubin levels

TABLE 2
Diagnostic accuracy of maternal visual assessment of jaundice and of the Ingram icterometer

Discussion

The ability of mothers to detect and respond to jaundice in their newborns after discharge from the hospital has not been previously studied. Opinions about the value of parental education regarding jaundice vary markedly. The AAP recommends that all mothers be able to recognize signs of jaundice before discharge.⁹ Others are skeptical that such education will be helpful: “Experience suggests that asking mothers to observe infants for the development of jaundice is not satisfactory. Despite such instructions, it is difficult for many parents to recognize significant jaundice.”¹⁰

Several studies have documented that jaundice is first seen in the face and progresses caudally to the trunk and extremities.^11-13 These studies also found good correlation between serum bilirubin levels and the advancement of dermal icterus. In a previous study, parents were able to accurately assess the caudal progression of jaundice while their babies were in the hospital.¹⁴ However, the bilirubin levels in that study were relatively low, reflecting the brief hospital stay of most of the infants. In contrast, a recent study concluded that the clinical examination for jaundice by nurses and physicians had poor reliability and only moderate correlation with bilirubin levels.¹⁵ The authors did conclude, however, that finding no jaundice below the nipple line reliably predicted that an infant would have a bilirubin concentration of less than 12.0 mg/dL. In this study, finding no jaundice below the nipple line reliably predicted that an infant would have a bilirubin concentration of less than 17.0 mg/dL.

Because of the relatively small number of infants having bilirubin levels high enough to require potential intervention, the measures of diagnostic accuracy in the tables should be interpreted with caution. However, the results of my study confirm several prior reports that restricting bilirubin testing to infants with icterometer readings 2.5 would have safely eliminated many unnecessary tests.^6,14,16 Although most of the infants in my study were white, the efficacy of the icterometer has also been documented in Asian and black newborns.¹⁷

Previous studies have shown that neonatal jaundice and its treatments are associated with an increased risk of maternal behaviors consistent with the vulnerable child syndrome.^18,19 This syndrome was originally described in 1964 in children whose parents believed that their child had suffered a “close call,” and thereafter perceived the child as vulnerable to serious injury or accident.¹⁸ Frequent blood tests to monitor bilirubin levels, supplementation or replacement of breast milk with formula, the physical separation of the mother and infant because of phototherapy, and prolonged hospitalization may create the impression that the infant is seriously ill, despite reassurances from medical personnel. Therefore, the mothers were asked whether the study itself served as a source of anxiety. Almost half of the mothers in this study reported that checking their babies for jaundice made them very or somewhat worried about their babies’ health. Some of the women must have felt ambivalent, however, because almost all of them (93%) also reported that checking their babies for jaundice made them very or somewhat reassured about their babies’ health. Most of the 48 comments written by the mothers in the study booklets were very positive.

Conclusions

One of the strategies recommended by the JCAHO to reduce the risk of kernicterus is to provide parents with adequate educational materials about newborn infants that include information about jaundice.² The message given to parents should be consistent, and should reassure mothers that most jaundiced infants are basically healthy. My study results suggest that it may also be useful for parents to be shown how to visually assess jaundice or to be given an Ingram icterometer to monitor their infants for jaundice after hospital discharge. Further study is needed to determine the optimal method of parental education about newborn jaundice.

Acknowledgments

This study was funded by a grant from the Ramsey Foundation. The author thanks Laura Lantz, Pamela Ristau, Kim Stone, Annette Swain, Mary Jo Feely, and the nurses at Integrated Home Care for their assistance with this project.

From 1% to 4% of full-term infants are readmitted to the hospital for jaundice in the first week of life, representing as many as 109,000 admissions¹ Delayed diagnosis of jaundice puts babies at risk for kernicterus, which had virtually disappeared in the United States but is now on the rise. There are anecdotal reports of 22 full-term infants born in the early 1990s who developed kernicterus after discharge from the hospital within 48 hours of birth.¹ The Joint Commission on Accreditation of Healthcare Organizations (JCAHO) recently issued a Sentinel Event Alert recommending that organizations take steps to raise awareness among neonatal caregivers of the potential for kernicterus and its risk factors by reviewing their current patient care processes with regard to the identification and management of hyperbilirubinemia in newborns and by identifying risk reduction strategies that could enhance the effectiveness of these processes.²

The JCAHO alert cites the American Academy of Pediatrics (AAP) Practice Parameter for Management of Hyperbilirubinemia in the Healthy Term Newborn, which is based on available data and expert consensus, as an example of a guideline for identifying at-risk newborns and their diagnosis and treatment. The AAP guideline suggests checking for jaundice by blanching the skin with digital pressure to reveal its underlying color. The guideline states that clinical assessment must be done in a well-lighted room and suggests that as the bilirubin level rises, the extent of caudal progression may be helpful in quantifying the degree of jaundice.³

The AAP jaundice guideline suggests that the use of an icterometer (transcutaneous jaundice meter) may be helpful in the clinical assessment of jaundice.³ A variety of instruments have been tested in different patient populations.^4-8 A potential role for such devices is their use by parents. The Ingram icterometer (Cascade Health Care Products, Salem, Ore.) is particularly promising because of its low cost ($17) and simplicity.⁵ It is a simple handheld device, made of clear plastic, on which are painted 5 transverse stripes of precise and graded hue. The stripes and spaces between them are 3/16 inch wide and are numbered from 1 (lightest in color) to 5 (darkest). When the icterometer is used, the painted side is pressed against the tip of the infant’s nose until the skin becomes blanched. The yellow color of the blanched skin can then be matched with the yellow stripes on the instrument, and a jaundice score assigned.

The purpose of my study was to determine whether mothers can accurately assess the presence and severity of jaundice in their newborns, both visually and with an icterometer, after hospital discharge. Maternal assessments were compared with bilirubin levels and home health nurse assessments to determine their accuracy. Serum bilirubin levels were used as the reference standard. Maternal comfort with the examination techniques was also assessed.

Methods

This study was approved by the Ramsey (now Regions) Hospital institutional review board. Mothers who gave birth at Regions Hospital in St. Paul, Minn., participated in the study. Mothers on the postpartum ward were invited to participate, but were excluded if they were not proficient in reading English, did not have a telephone, or lived more than 10 miles from the hospital. Infants were excluded if they were in the intensive care nursery, were not discharged on the same day as the mother, or if they received phototherapy. Mothers were advised to follow their health care providers’ instructions about timing for the first follow-up visit, and any provider instructions regarding jaundice.

After obtaining consent, the author or a study assistant showed the mothers how to examine their infants for jaundice by 2 methods. Each mother was instructed to examine her baby in a well-lighted room. First, the mother was shown how to look for jaundice by digitally blanching the skin on the cheek. The mother then documented whether she saw any underlying yellow color on her baby. Next, the mother was shown how to determine the caudal progression of the jaundice and to draw a horizontal line on an illustration of a baby corresponding to where the jaundice ended. The distance from the top of the infant’s head to the line drawn by the mother was used to determine the caudal progression. The mother was then shown how to use the Ingram icterometer and obtain a reading from the baby’s nose. Each mother was given an icterometer and a study booklet to document her examination for a total of 7 days, beginning the day after discharge from the hospital. The study booklet also contained some demographic questions, and questions about the mother’s comfort level with both methods of jaundice assessment. The mother was instructed to return the booklet and icterometer by mail when completed. The mother was sent a $25 gift certificate when the study materials were returned.

Within 7 days of discharge, a home health nurse visited each mother and infant in the home. The nurses were trained in the same methods of clinically assessing jaundice, and they assessed each infant by visually determining the caudal progression and by use of the icterometer. The nurse did not share the results of her examination with the mother. The nurse obtained bilirubin levels from all infants and notified the infants’ health care providers of any bilirubin levels higher than 14 mg/dL.

Standard descriptive statistics were calculated for all variables. Categorical relationships were assessed using kappa and chi-square statistics, as appropriate. All analyses were performed using Statistical Package for Social Sciences for Windows, version 10.0.5.

Results

A total 113 of 177 mothers returned their study packets. Home health nurses visited 96 of the 113 mothers; the other 17 mothers were not visited because they declined the visits or could not be located. Although all babies were to have serum bilirubin levels determined whether or not they appeared jaundiced, only 90 of the 96 infants had the blood test. For the other 6 infants, either insufficient blood was drawn or the mother refused the test. On the day of the nurse’s visits, mothers documented in their study booklets the caudal progression of jaundice (for 56 infants) and icterometer readings (for 55 infants).

The educational levels of the mothers were as follows: 15% completed grade school or less; 40% completed high school; and 45% completed college. The mothers reported being from the following racial and ethnic groups: white, 59%; Hispanic, 16%; black, 14%; Asian, 8%; and other, 3%. A total of 53% of the women were primiparous, 84% completed examination forms for their babies for all 7 days, and 53% assessed their infants as being jaundiced during the study.

On the day of the nurse’s visit, there was moderate agreement between the nurses and the mothers about the presence of jaundice in the infants (= 0.50; P < .001). For those infants with jaundice, there was little agreement on the extent of caudal progression between the nurses and the mothers (correlation = 0.36; P > 0.1), but there was moderate agreement between their icterometer readings (correlation = 0.58; P < .05).

The total serum bilirubin results ranged from 0.8 mg/dL to 18.8 mg/dL, with a mean of 7.4 mg/dL. The mean bilirubin level of infants thought to be jaundiced by their mothers was 11.3 mg/dL, while the mean bilirubin of infants not thought to be jaundiced was 4.8 mg/dL (P < .001).

The mothers’ icterometer readings and determinations of jaundice to the nipple line or below it are compared with bilirubin levels in (Table 1). (Table 2) summarizes the diagnostic accuracy of jaundice extending to the nipple line or below it, and for icterometer readings of 2.5, in identifying bilirubin levels of 12 mg/dL and 17 mg/dL. A bilirubin level of 12 mg/dL is the level at which the AAP guideline suggests considering phototherapy for infants aged 24 to 47 hours, and 17 mg/dL is the level at which phototherapy should be considered for infants older than 72 hours.³

The mothers of the 3 infants with bilirubin levels 17 mg/dL recognized that their infants were jaundiced and determined that the jaundice extended below the nipple line. The icterometer readings obtained by the mothers were 2.5, 3, and 3.5. The corresponding icterometer readings by the nurses were 4.5, 3.5 and 3.

The study booklet contained 6 questions about the mothers’ reactions to the study. Almost all of the mothers (98%) responded that the method for checking for caudal progression of jaundice was explained clearly, and even more (99%) felt the use of the icterometer was explained clearly. A total of 69% of the mothers felt it was “very easy” or “easy” to check for caudal progression, and 80% felt it was “very easy” or “easy” to use the icterometer. Forty-six percent of the mothers reported that checking their babies for jaundice made them “very worried” or “somewhat worried” about their babies’ health. Mothers with less education were significantly more likely to report being worried than mothers with higher education levels (P < .05). However, 93% of the mothers reported that checking their babies for jaundice made them “very reassured” or “somewhat reassured” about their babies’ health.

TABLE 1
Maternal assessment of jaundice, by caudal progression and icterometer readings, compared with serum bilirubin levels

TABLE 2
Diagnostic accuracy of maternal visual assessment of jaundice and of the Ingram icterometer

Discussion

The ability of mothers to detect and respond to jaundice in their newborns after discharge from the hospital has not been previously studied. Opinions about the value of parental education regarding jaundice vary markedly. The AAP recommends that all mothers be able to recognize signs of jaundice before discharge.⁹ Others are skeptical that such education will be helpful: “Experience suggests that asking mothers to observe infants for the development of jaundice is not satisfactory. Despite such instructions, it is difficult for many parents to recognize significant jaundice.”¹⁰

Several studies have documented that jaundice is first seen in the face and progresses caudally to the trunk and extremities.^11-13 These studies also found good correlation between serum bilirubin levels and the advancement of dermal icterus. In a previous study, parents were able to accurately assess the caudal progression of jaundice while their babies were in the hospital.¹⁴ However, the bilirubin levels in that study were relatively low, reflecting the brief hospital stay of most of the infants. In contrast, a recent study concluded that the clinical examination for jaundice by nurses and physicians had poor reliability and only moderate correlation with bilirubin levels.¹⁵ The authors did conclude, however, that finding no jaundice below the nipple line reliably predicted that an infant would have a bilirubin concentration of less than 12.0 mg/dL. In this study, finding no jaundice below the nipple line reliably predicted that an infant would have a bilirubin concentration of less than 17.0 mg/dL.

Because of the relatively small number of infants having bilirubin levels high enough to require potential intervention, the measures of diagnostic accuracy in the tables should be interpreted with caution. However, the results of my study confirm several prior reports that restricting bilirubin testing to infants with icterometer readings 2.5 would have safely eliminated many unnecessary tests.^6,14,16 Although most of the infants in my study were white, the efficacy of the icterometer has also been documented in Asian and black newborns.¹⁷

Previous studies have shown that neonatal jaundice and its treatments are associated with an increased risk of maternal behaviors consistent with the vulnerable child syndrome.^18,19 This syndrome was originally described in 1964 in children whose parents believed that their child had suffered a “close call,” and thereafter perceived the child as vulnerable to serious injury or accident.¹⁸ Frequent blood tests to monitor bilirubin levels, supplementation or replacement of breast milk with formula, the physical separation of the mother and infant because of phototherapy, and prolonged hospitalization may create the impression that the infant is seriously ill, despite reassurances from medical personnel. Therefore, the mothers were asked whether the study itself served as a source of anxiety. Almost half of the mothers in this study reported that checking their babies for jaundice made them very or somewhat worried about their babies’ health. Some of the women must have felt ambivalent, however, because almost all of them (93%) also reported that checking their babies for jaundice made them very or somewhat reassured about their babies’ health. Most of the 48 comments written by the mothers in the study booklets were very positive.

Conclusions

One of the strategies recommended by the JCAHO to reduce the risk of kernicterus is to provide parents with adequate educational materials about newborn infants that include information about jaundice.² The message given to parents should be consistent, and should reassure mothers that most jaundiced infants are basically healthy. My study results suggest that it may also be useful for parents to be shown how to visually assess jaundice or to be given an Ingram icterometer to monitor their infants for jaundice after hospital discharge. Further study is needed to determine the optimal method of parental education about newborn jaundice.

Acknowledgments

This study was funded by a grant from the Ramsey Foundation. The author thanks Laura Lantz, Pamela Ristau, Kim Stone, Annette Swain, Mary Jo Feely, and the nurses at Integrated Home Care for their assistance with this project.

From 1% to 4% of full-term infants are readmitted to the hospital for jaundice in the first week of life, representing as many as 109,000 admissions¹ Delayed diagnosis of jaundice puts babies at risk for kernicterus, which had virtually disappeared in the United States but is now on the rise. There are anecdotal reports of 22 full-term infants born in the early 1990s who developed kernicterus after discharge from the hospital within 48 hours of birth.¹ The Joint Commission on Accreditation of Healthcare Organizations (JCAHO) recently issued a Sentinel Event Alert recommending that organizations take steps to raise awareness among neonatal caregivers of the potential for kernicterus and its risk factors by reviewing their current patient care processes with regard to the identification and management of hyperbilirubinemia in newborns and by identifying risk reduction strategies that could enhance the effectiveness of these processes.²

The JCAHO alert cites the American Academy of Pediatrics (AAP) Practice Parameter for Management of Hyperbilirubinemia in the Healthy Term Newborn, which is based on available data and expert consensus, as an example of a guideline for identifying at-risk newborns and their diagnosis and treatment. The AAP guideline suggests checking for jaundice by blanching the skin with digital pressure to reveal its underlying color. The guideline states that clinical assessment must be done in a well-lighted room and suggests that as the bilirubin level rises, the extent of caudal progression may be helpful in quantifying the degree of jaundice.³

The AAP jaundice guideline suggests that the use of an icterometer (transcutaneous jaundice meter) may be helpful in the clinical assessment of jaundice.³ A variety of instruments have been tested in different patient populations.^4-8 A potential role for such devices is their use by parents. The Ingram icterometer (Cascade Health Care Products, Salem, Ore.) is particularly promising because of its low cost ($17) and simplicity.⁵ It is a simple handheld device, made of clear plastic, on which are painted 5 transverse stripes of precise and graded hue. The stripes and spaces between them are 3/16 inch wide and are numbered from 1 (lightest in color) to 5 (darkest). When the icterometer is used, the painted side is pressed against the tip of the infant’s nose until the skin becomes blanched. The yellow color of the blanched skin can then be matched with the yellow stripes on the instrument, and a jaundice score assigned.

The purpose of my study was to determine whether mothers can accurately assess the presence and severity of jaundice in their newborns, both visually and with an icterometer, after hospital discharge. Maternal assessments were compared with bilirubin levels and home health nurse assessments to determine their accuracy. Serum bilirubin levels were used as the reference standard. Maternal comfort with the examination techniques was also assessed.

Methods

This study was approved by the Ramsey (now Regions) Hospital institutional review board. Mothers who gave birth at Regions Hospital in St. Paul, Minn., participated in the study. Mothers on the postpartum ward were invited to participate, but were excluded if they were not proficient in reading English, did not have a telephone, or lived more than 10 miles from the hospital. Infants were excluded if they were in the intensive care nursery, were not discharged on the same day as the mother, or if they received phototherapy. Mothers were advised to follow their health care providers’ instructions about timing for the first follow-up visit, and any provider instructions regarding jaundice.

After obtaining consent, the author or a study assistant showed the mothers how to examine their infants for jaundice by 2 methods. Each mother was instructed to examine her baby in a well-lighted room. First, the mother was shown how to look for jaundice by digitally blanching the skin on the cheek. The mother then documented whether she saw any underlying yellow color on her baby. Next, the mother was shown how to determine the caudal progression of the jaundice and to draw a horizontal line on an illustration of a baby corresponding to where the jaundice ended. The distance from the top of the infant’s head to the line drawn by the mother was used to determine the caudal progression. The mother was then shown how to use the Ingram icterometer and obtain a reading from the baby’s nose. Each mother was given an icterometer and a study booklet to document her examination for a total of 7 days, beginning the day after discharge from the hospital. The study booklet also contained some demographic questions, and questions about the mother’s comfort level with both methods of jaundice assessment. The mother was instructed to return the booklet and icterometer by mail when completed. The mother was sent a $25 gift certificate when the study materials were returned.

Within 7 days of discharge, a home health nurse visited each mother and infant in the home. The nurses were trained in the same methods of clinically assessing jaundice, and they assessed each infant by visually determining the caudal progression and by use of the icterometer. The nurse did not share the results of her examination with the mother. The nurse obtained bilirubin levels from all infants and notified the infants’ health care providers of any bilirubin levels higher than 14 mg/dL.

Standard descriptive statistics were calculated for all variables. Categorical relationships were assessed using kappa and chi-square statistics, as appropriate. All analyses were performed using Statistical Package for Social Sciences for Windows, version 10.0.5.

Results

A total 113 of 177 mothers returned their study packets. Home health nurses visited 96 of the 113 mothers; the other 17 mothers were not visited because they declined the visits or could not be located. Although all babies were to have serum bilirubin levels determined whether or not they appeared jaundiced, only 90 of the 96 infants had the blood test. For the other 6 infants, either insufficient blood was drawn or the mother refused the test. On the day of the nurse’s visits, mothers documented in their study booklets the caudal progression of jaundice (for 56 infants) and icterometer readings (for 55 infants).

The educational levels of the mothers were as follows: 15% completed grade school or less; 40% completed high school; and 45% completed college. The mothers reported being from the following racial and ethnic groups: white, 59%; Hispanic, 16%; black, 14%; Asian, 8%; and other, 3%. A total of 53% of the women were primiparous, 84% completed examination forms for their babies for all 7 days, and 53% assessed their infants as being jaundiced during the study.

On the day of the nurse’s visit, there was moderate agreement between the nurses and the mothers about the presence of jaundice in the infants (= 0.50; P < .001). For those infants with jaundice, there was little agreement on the extent of caudal progression between the nurses and the mothers (correlation = 0.36; P > 0.1), but there was moderate agreement between their icterometer readings (correlation = 0.58; P < .05).

The total serum bilirubin results ranged from 0.8 mg/dL to 18.8 mg/dL, with a mean of 7.4 mg/dL. The mean bilirubin level of infants thought to be jaundiced by their mothers was 11.3 mg/dL, while the mean bilirubin of infants not thought to be jaundiced was 4.8 mg/dL (P < .001).

The mothers’ icterometer readings and determinations of jaundice to the nipple line or below it are compared with bilirubin levels in (Table 1). (Table 2) summarizes the diagnostic accuracy of jaundice extending to the nipple line or below it, and for icterometer readings of 2.5, in identifying bilirubin levels of 12 mg/dL and 17 mg/dL. A bilirubin level of 12 mg/dL is the level at which the AAP guideline suggests considering phototherapy for infants aged 24 to 47 hours, and 17 mg/dL is the level at which phototherapy should be considered for infants older than 72 hours.³

The mothers of the 3 infants with bilirubin levels 17 mg/dL recognized that their infants were jaundiced and determined that the jaundice extended below the nipple line. The icterometer readings obtained by the mothers were 2.5, 3, and 3.5. The corresponding icterometer readings by the nurses were 4.5, 3.5 and 3.

The study booklet contained 6 questions about the mothers’ reactions to the study. Almost all of the mothers (98%) responded that the method for checking for caudal progression of jaundice was explained clearly, and even more (99%) felt the use of the icterometer was explained clearly. A total of 69% of the mothers felt it was “very easy” or “easy” to check for caudal progression, and 80% felt it was “very easy” or “easy” to use the icterometer. Forty-six percent of the mothers reported that checking their babies for jaundice made them “very worried” or “somewhat worried” about their babies’ health. Mothers with less education were significantly more likely to report being worried than mothers with higher education levels (P < .05). However, 93% of the mothers reported that checking their babies for jaundice made them “very reassured” or “somewhat reassured” about their babies’ health.

TABLE 1
Maternal assessment of jaundice, by caudal progression and icterometer readings, compared with serum bilirubin levels

TABLE 2
Diagnostic accuracy of maternal visual assessment of jaundice and of the Ingram icterometer

Discussion

The ability of mothers to detect and respond to jaundice in their newborns after discharge from the hospital has not been previously studied. Opinions about the value of parental education regarding jaundice vary markedly. The AAP recommends that all mothers be able to recognize signs of jaundice before discharge.⁹ Others are skeptical that such education will be helpful: “Experience suggests that asking mothers to observe infants for the development of jaundice is not satisfactory. Despite such instructions, it is difficult for many parents to recognize significant jaundice.”¹⁰

Several studies have documented that jaundice is first seen in the face and progresses caudally to the trunk and extremities.^11-13 These studies also found good correlation between serum bilirubin levels and the advancement of dermal icterus. In a previous study, parents were able to accurately assess the caudal progression of jaundice while their babies were in the hospital.¹⁴ However, the bilirubin levels in that study were relatively low, reflecting the brief hospital stay of most of the infants. In contrast, a recent study concluded that the clinical examination for jaundice by nurses and physicians had poor reliability and only moderate correlation with bilirubin levels.¹⁵ The authors did conclude, however, that finding no jaundice below the nipple line reliably predicted that an infant would have a bilirubin concentration of less than 12.0 mg/dL. In this study, finding no jaundice below the nipple line reliably predicted that an infant would have a bilirubin concentration of less than 17.0 mg/dL.

Because of the relatively small number of infants having bilirubin levels high enough to require potential intervention, the measures of diagnostic accuracy in the tables should be interpreted with caution. However, the results of my study confirm several prior reports that restricting bilirubin testing to infants with icterometer readings 2.5 would have safely eliminated many unnecessary tests.^6,14,16 Although most of the infants in my study were white, the efficacy of the icterometer has also been documented in Asian and black newborns.¹⁷

Previous studies have shown that neonatal jaundice and its treatments are associated with an increased risk of maternal behaviors consistent with the vulnerable child syndrome.^18,19 This syndrome was originally described in 1964 in children whose parents believed that their child had suffered a “close call,” and thereafter perceived the child as vulnerable to serious injury or accident.¹⁸ Frequent blood tests to monitor bilirubin levels, supplementation or replacement of breast milk with formula, the physical separation of the mother and infant because of phototherapy, and prolonged hospitalization may create the impression that the infant is seriously ill, despite reassurances from medical personnel. Therefore, the mothers were asked whether the study itself served as a source of anxiety. Almost half of the mothers in this study reported that checking their babies for jaundice made them very or somewhat worried about their babies’ health. Some of the women must have felt ambivalent, however, because almost all of them (93%) also reported that checking their babies for jaundice made them very or somewhat reassured about their babies’ health. Most of the 48 comments written by the mothers in the study booklets were very positive.

Conclusions

One of the strategies recommended by the JCAHO to reduce the risk of kernicterus is to provide parents with adequate educational materials about newborn infants that include information about jaundice.² The message given to parents should be consistent, and should reassure mothers that most jaundiced infants are basically healthy. My study results suggest that it may also be useful for parents to be shown how to visually assess jaundice or to be given an Ingram icterometer to monitor their infants for jaundice after hospital discharge. Further study is needed to determine the optimal method of parental education about newborn jaundice.

Acknowledgments

This study was funded by a grant from the Ramsey Foundation. The author thanks Laura Lantz, Pamela Ristau, Kim Stone, Annette Swain, Mary Jo Feely, and the nurses at Integrated Home Care for their assistance with this project.

Nearly 25 years ago, the American Academy of Pediatrics (AAP) Committee on Drugs described the inaccuracies of administering liquid medication by household spoons.¹ The Committee recommended that physicians advise their community pharmacies to stock appropriate liquid administration devices and insist on the use of such devices when prescribing liquid medications. The committee recommended the use of the oral dosing syringe, which was described as novel and innovative. Since then, a variety of liquid medication dosing devices have been developed and have become widely available, each of which ha sits advantages and disadvantages.²

Limited information is available about the current use of liquid medicine dosing devices in this country. In a 1975 study³ when the oral dosing syringe was still new, 75% of patients used a household teaspoon or kitchen measuring spoon when dosing liquid medication. In a 1989 study from Israel,⁴ 80% of the children were given medications by a household teaspoon. The purpose of our study was to examine the following issues concerning the use of liquid medications: (1) which of the many liquid medication dosing devices are commonly owned and used by families; (2) the ability of potential patients to accurately measure liquids using 3 different dosing devices; (3) their ability to correctly interpret a variety of dosing instructions; and (4) their ability to correctly interpret a pediatric dosing chart.

Methods

Our study was approved by the institutional review board of Ramsey (now Regions) Hospital in St. Paul, Minnesota. In the summer of 1996, one of the investigators (F.S.M.) interviewed a convenience sample of people in the waiting areas of 3 clinics in the St. Paul, Minnesota, area: Ramsey Clinic Maplewood, a small private multispecialty clinic serving a predominately white middle-class suburban population; Ramsey Family Physicians, a residency clinic serving primarily a white lower-socioeconomic population; and West Side Clinic, a community clinic serving mostly Hmong and Hispanic patients. These clinics were chosen to obtain subjects with a variety of socioeconomic and ethnic backgrounds. The interview consisted of several parts. Participants were shown the following liquid dosing devices: cylindrical spoon, medicine cup, oral dosing syringe, oral dropper, andby dispenser. In addition, they were shown a household teaspoon and a measuring spoon ([Figure]). They were asked which of the dosing devices they had in their homes and which they had ever used for dispensing liquid medications. Demographic information was also obtained.

The participants were also tested and scored on their ability to measure liquid medicines and interpret dosing instructions accurately. A total performance score was determined by adding the scores from the following activities, with a maximum obtainable score of 11. The investigator observed the subjects measuring 3 doses of medicine using a cylindrical spoon, medicine cup, and an oral dosing syringe and noted the accuracy of the measurement. The subjects received a score of 0 or 1 if the measurement was done incorrectly or correctly, respectively, for each of the 3 devices. Then the subjects were asked to indicate on a chart what times they would take medicine if it was prescribed every 6 hours, 4 times daily, and 3 times daily. The subjects received a score of 0, 1, or 2 for each of these 3 charts. The score of 0 was given if both the number of doses and the timing were incorrect, a 1 if either the number of doses or the timing was correct, and a 2 if both were correct. Finally, subjects were shown a pediatric dosing chart that had dosing listed by both age and weight and contained a note that dosing by weight is more accurate. Subjects were asked to indicate the correct dose for 2 children. In one example the child’s age and weight matched on the chart, and in the other the age and weight were discordant. A score of 1 was given for each correct reading of the dosing chart.

We calculated descriptive statistics and frequency distributions for all variables. Chi-square analyses were used for categorical comparisons. Grouped t tests and analysis of variance techniques were used to assess the effect of categorical predictors and demographics on total score. Pearson correlations were calculated to assess relationships among continuous variables.

Results

Of the 130 participants, 105 were women. The participants had a mean age of 40 years, a mean education level of 12.5 years, and a mean of 1.1 children in the household. Sixty-eight percent of the subjects were white, 19% Hispanic, 11% African American, and 2% Asian. English was the second language for 13% of the participants.

The liquid dosing devices available in the participants’ homes and the devices they used are shown in the [Table]. The one most frequently used (73%) for measuring liquid medication was a household teaspoon.

Ninety-two percent of the participants measured the correct dose when using the oral dosing syringe. Only 85% of the participants measured the correct dose of 1 tablespoon when using a medicine cup. The most frequent error (70%) occurred when the participants mistakenly measured 1 teaspoon instead of 1 tablespoon. Although 92% of the subjects ultimately measured the correct dose using the cylindrical spoon, many subjects spilled the liquid and required several attempts before measuring the correct dose.

Eighty-nine percent of participants noted the correct number of doses and time between doses when asked to indicate the times that they would take a medicine if they were instructed to take it 4 times daily and 3 times daily. Only 38% of the participants correctly indicated this information when they were instructed to take a medication every 6 hours. Participants commonly misinterpreted this instruction as meaning every 6 hours while awake, and indicated 3 rather than 4 doses.

The pediatric dosing chart was correctly interpreted by 87% of the participants for both case scenarios. Twelve percent of those surveyed gave the incorrect dose when the age and weight of the child were discordant, choosing the dose based on the child’s age rather than weight.

The participants’ mean total performance score was 9.5. Women scored significantly better than men. (9.7 vs 8.7, P <.05). Total performance score was significantly correlated with the participants’ education level (Pearson correlation=0.177, P <.05), but not with age or the number of children in the household. Total performance score did not differ significantly by the participants’ native language or ethnic group.

Discussion

In 1975 the AAP Committee on Drugs denounced the use of household teaspoons for administering liquid medications.¹ The volume of household teaspoons can range from 2 to 10 mL. Also, the same spoon when used by different persons may deliver from 3 to 7 mL. Therefore, even household measuring spoons are problematic. Participants in this study used a household teaspoon for measuring liquid medications more often than any other dosing device.

Oral dosing devices such as oral dosing syringes, oral droppers, cylindrical spoons, and medication cups are preferred over the traditional household teaspoon or measuring spoon, because they are more accurate. The advantages and disadvantages of the different devices have been described elsewhere.² The cylindrical spoon has been described as having an increased potential for easy spillage before and during administration of medication.² Participants in our study had problems with spillage with this device.

Study participants’ measurements were less accurate when using a measuring cup than when using a cylindrical spoon or oral dosing syringe, primarily because of confusion between the cup’s markings for tablespoons and teaspoons. A survey of poison control centers found 3 major causes of dosing errors using dispensing cups: (1) confusion of tablespoons for teaspoons; (2) the assumption that the entire cup was the unit of measure; and (3) the misinterpretation that 1 cupful was the recommended dose.⁵ After receiving reports of inappropriately marked plastic dosing cups, the Food and Drug Administration began a public education campaign in 1994 to increase health professional and consumer awareness of misdosing hazards with liquid medicines.⁶

The oral dosing syringe is felt to be the best device for delivery of liquid medication.² Its advantages include accuracy, convenience, availability in various sizes, and relatively low expense. The syringe permits the user to direct the delivery of the medication to the back and side of the mouth of an infant or small child, thus minimizing spillage. It also reduces the risk of possible gagging and aspiration of medication. Only a third of this study’s participants had an oral dosing syringe in their home.

It is alarming that the majority of participants misinterpreted instructions to take a medicine every 6 hours, so that they would take only 3 rather than 4 doses of medicine in a day. This problem of misinterpreting dosing frequency appears to be relatively unrecognized, although it could be an important cause of apparent treatment failures. Studies of medication errors typically focus on mistakes that cause clinical symptoms. In a study of 1108 medication errors in pediatrics reported to poison control centers in France, none involved underdosage errors.⁷

Most participants were able to correctly interpret a pediatric dosing chart, although some errors were noted when the age and weight were discordant. In a previous study, only 40% of caretakers were able to state a correct dose of acetaminophen for their child when given the child’s weight and all package labeling.⁸ Parents often fail to revise medication doses as a child grows older and gains weight and therefore tend to underdose.^4,8,9 Although less serious than overdosing in terms of morbidity, underdosing of acetaminophen may lead to ineffective treatment of fever and unnecessary visits to the clinic or emergency department.

It is not surprising that women and participants with higher education levels had higher total performance scores. In previous studies of liquid medication, the majority of the caretakers giving medication to children were mothers.^3,8-10 Women are therefore more likely to have experience administering liquid medication than men. Fortunately, parental education has been shown to be very effective in eliminating medication dosing errors.¹⁰ In a recent study by McMahon and colleagues of 90 English-speaking and Spanish-speaking families, 100% of them dosed medication correctly when given instructions and a syringe with a line marked at the prescribed dose.

Conclusions

The recommendation for the use of the oral dosing syringe made by the AAP almost 25 years ago is just as relevant today. On the basis of our study results, we make the following additional recommendations to clinicians: (1) when possible, indicate the dosing interval by the number of doses in a day, rather than by the number of hours between doses; and (2) always consider the possibility of a medication dosing error when faced with an apparent treatment failure.

Acknowledgments

Our study was funded by a grant from the American Academy of Family Physicians Foundation and the Minnesota Academy of Family Physicians Foundation.

Nearly 25 years ago, the American Academy of Pediatrics (AAP) Committee on Drugs described the inaccuracies of administering liquid medication by household spoons.¹ The Committee recommended that physicians advise their community pharmacies to stock appropriate liquid administration devices and insist on the use of such devices when prescribing liquid medications. The committee recommended the use of the oral dosing syringe, which was described as novel and innovative. Since then, a variety of liquid medication dosing devices have been developed and have become widely available, each of which ha sits advantages and disadvantages.²

Limited information is available about the current use of liquid medicine dosing devices in this country. In a 1975 study³ when the oral dosing syringe was still new, 75% of patients used a household teaspoon or kitchen measuring spoon when dosing liquid medication. In a 1989 study from Israel,⁴ 80% of the children were given medications by a household teaspoon. The purpose of our study was to examine the following issues concerning the use of liquid medications: (1) which of the many liquid medication dosing devices are commonly owned and used by families; (2) the ability of potential patients to accurately measure liquids using 3 different dosing devices; (3) their ability to correctly interpret a variety of dosing instructions; and (4) their ability to correctly interpret a pediatric dosing chart.

Methods

Our study was approved by the institutional review board of Ramsey (now Regions) Hospital in St. Paul, Minnesota. In the summer of 1996, one of the investigators (F.S.M.) interviewed a convenience sample of people in the waiting areas of 3 clinics in the St. Paul, Minnesota, area: Ramsey Clinic Maplewood, a small private multispecialty clinic serving a predominately white middle-class suburban population; Ramsey Family Physicians, a residency clinic serving primarily a white lower-socioeconomic population; and West Side Clinic, a community clinic serving mostly Hmong and Hispanic patients. These clinics were chosen to obtain subjects with a variety of socioeconomic and ethnic backgrounds. The interview consisted of several parts. Participants were shown the following liquid dosing devices: cylindrical spoon, medicine cup, oral dosing syringe, oral dropper, andby dispenser. In addition, they were shown a household teaspoon and a measuring spoon ([Figure]). They were asked which of the dosing devices they had in their homes and which they had ever used for dispensing liquid medications. Demographic information was also obtained.

The participants were also tested and scored on their ability to measure liquid medicines and interpret dosing instructions accurately. A total performance score was determined by adding the scores from the following activities, with a maximum obtainable score of 11. The investigator observed the subjects measuring 3 doses of medicine using a cylindrical spoon, medicine cup, and an oral dosing syringe and noted the accuracy of the measurement. The subjects received a score of 0 or 1 if the measurement was done incorrectly or correctly, respectively, for each of the 3 devices. Then the subjects were asked to indicate on a chart what times they would take medicine if it was prescribed every 6 hours, 4 times daily, and 3 times daily. The subjects received a score of 0, 1, or 2 for each of these 3 charts. The score of 0 was given if both the number of doses and the timing were incorrect, a 1 if either the number of doses or the timing was correct, and a 2 if both were correct. Finally, subjects were shown a pediatric dosing chart that had dosing listed by both age and weight and contained a note that dosing by weight is more accurate. Subjects were asked to indicate the correct dose for 2 children. In one example the child’s age and weight matched on the chart, and in the other the age and weight were discordant. A score of 1 was given for each correct reading of the dosing chart.

We calculated descriptive statistics and frequency distributions for all variables. Chi-square analyses were used for categorical comparisons. Grouped t tests and analysis of variance techniques were used to assess the effect of categorical predictors and demographics on total score. Pearson correlations were calculated to assess relationships among continuous variables.

Results

Of the 130 participants, 105 were women. The participants had a mean age of 40 years, a mean education level of 12.5 years, and a mean of 1.1 children in the household. Sixty-eight percent of the subjects were white, 19% Hispanic, 11% African American, and 2% Asian. English was the second language for 13% of the participants.

The liquid dosing devices available in the participants’ homes and the devices they used are shown in the [Table]. The one most frequently used (73%) for measuring liquid medication was a household teaspoon.

Ninety-two percent of the participants measured the correct dose when using the oral dosing syringe. Only 85% of the participants measured the correct dose of 1 tablespoon when using a medicine cup. The most frequent error (70%) occurred when the participants mistakenly measured 1 teaspoon instead of 1 tablespoon. Although 92% of the subjects ultimately measured the correct dose using the cylindrical spoon, many subjects spilled the liquid and required several attempts before measuring the correct dose.

Eighty-nine percent of participants noted the correct number of doses and time between doses when asked to indicate the times that they would take a medicine if they were instructed to take it 4 times daily and 3 times daily. Only 38% of the participants correctly indicated this information when they were instructed to take a medication every 6 hours. Participants commonly misinterpreted this instruction as meaning every 6 hours while awake, and indicated 3 rather than 4 doses.

The pediatric dosing chart was correctly interpreted by 87% of the participants for both case scenarios. Twelve percent of those surveyed gave the incorrect dose when the age and weight of the child were discordant, choosing the dose based on the child’s age rather than weight.

The participants’ mean total performance score was 9.5. Women scored significantly better than men. (9.7 vs 8.7, P <.05). Total performance score was significantly correlated with the participants’ education level (Pearson correlation=0.177, P <.05), but not with age or the number of children in the household. Total performance score did not differ significantly by the participants’ native language or ethnic group.

Discussion

In 1975 the AAP Committee on Drugs denounced the use of household teaspoons for administering liquid medications.¹ The volume of household teaspoons can range from 2 to 10 mL. Also, the same spoon when used by different persons may deliver from 3 to 7 mL. Therefore, even household measuring spoons are problematic. Participants in this study used a household teaspoon for measuring liquid medications more often than any other dosing device.

Oral dosing devices such as oral dosing syringes, oral droppers, cylindrical spoons, and medication cups are preferred over the traditional household teaspoon or measuring spoon, because they are more accurate. The advantages and disadvantages of the different devices have been described elsewhere.² The cylindrical spoon has been described as having an increased potential for easy spillage before and during administration of medication.² Participants in our study had problems with spillage with this device.

Study participants’ measurements were less accurate when using a measuring cup than when using a cylindrical spoon or oral dosing syringe, primarily because of confusion between the cup’s markings for tablespoons and teaspoons. A survey of poison control centers found 3 major causes of dosing errors using dispensing cups: (1) confusion of tablespoons for teaspoons; (2) the assumption that the entire cup was the unit of measure; and (3) the misinterpretation that 1 cupful was the recommended dose.⁵ After receiving reports of inappropriately marked plastic dosing cups, the Food and Drug Administration began a public education campaign in 1994 to increase health professional and consumer awareness of misdosing hazards with liquid medicines.⁶

The oral dosing syringe is felt to be the best device for delivery of liquid medication.² Its advantages include accuracy, convenience, availability in various sizes, and relatively low expense. The syringe permits the user to direct the delivery of the medication to the back and side of the mouth of an infant or small child, thus minimizing spillage. It also reduces the risk of possible gagging and aspiration of medication. Only a third of this study’s participants had an oral dosing syringe in their home.

It is alarming that the majority of participants misinterpreted instructions to take a medicine every 6 hours, so that they would take only 3 rather than 4 doses of medicine in a day. This problem of misinterpreting dosing frequency appears to be relatively unrecognized, although it could be an important cause of apparent treatment failures. Studies of medication errors typically focus on mistakes that cause clinical symptoms. In a study of 1108 medication errors in pediatrics reported to poison control centers in France, none involved underdosage errors.⁷

Most participants were able to correctly interpret a pediatric dosing chart, although some errors were noted when the age and weight were discordant. In a previous study, only 40% of caretakers were able to state a correct dose of acetaminophen for their child when given the child’s weight and all package labeling.⁸ Parents often fail to revise medication doses as a child grows older and gains weight and therefore tend to underdose.^4,8,9 Although less serious than overdosing in terms of morbidity, underdosing of acetaminophen may lead to ineffective treatment of fever and unnecessary visits to the clinic or emergency department.

It is not surprising that women and participants with higher education levels had higher total performance scores. In previous studies of liquid medication, the majority of the caretakers giving medication to children were mothers.^3,8-10 Women are therefore more likely to have experience administering liquid medication than men. Fortunately, parental education has been shown to be very effective in eliminating medication dosing errors.¹⁰ In a recent study by McMahon and colleagues of 90 English-speaking and Spanish-speaking families, 100% of them dosed medication correctly when given instructions and a syringe with a line marked at the prescribed dose.

Conclusions

The recommendation for the use of the oral dosing syringe made by the AAP almost 25 years ago is just as relevant today. On the basis of our study results, we make the following additional recommendations to clinicians: (1) when possible, indicate the dosing interval by the number of doses in a day, rather than by the number of hours between doses; and (2) always consider the possibility of a medication dosing error when faced with an apparent treatment failure.

Acknowledgments

Our study was funded by a grant from the American Academy of Family Physicians Foundation and the Minnesota Academy of Family Physicians Foundation.

Nearly 25 years ago, the American Academy of Pediatrics (AAP) Committee on Drugs described the inaccuracies of administering liquid medication by household spoons.¹ The Committee recommended that physicians advise their community pharmacies to stock appropriate liquid administration devices and insist on the use of such devices when prescribing liquid medications. The committee recommended the use of the oral dosing syringe, which was described as novel and innovative. Since then, a variety of liquid medication dosing devices have been developed and have become widely available, each of which ha sits advantages and disadvantages.²

Limited information is available about the current use of liquid medicine dosing devices in this country. In a 1975 study³ when the oral dosing syringe was still new, 75% of patients used a household teaspoon or kitchen measuring spoon when dosing liquid medication. In a 1989 study from Israel,⁴ 80% of the children were given medications by a household teaspoon. The purpose of our study was to examine the following issues concerning the use of liquid medications: (1) which of the many liquid medication dosing devices are commonly owned and used by families; (2) the ability of potential patients to accurately measure liquids using 3 different dosing devices; (3) their ability to correctly interpret a variety of dosing instructions; and (4) their ability to correctly interpret a pediatric dosing chart.

Methods

Our study was approved by the institutional review board of Ramsey (now Regions) Hospital in St. Paul, Minnesota. In the summer of 1996, one of the investigators (F.S.M.) interviewed a convenience sample of people in the waiting areas of 3 clinics in the St. Paul, Minnesota, area: Ramsey Clinic Maplewood, a small private multispecialty clinic serving a predominately white middle-class suburban population; Ramsey Family Physicians, a residency clinic serving primarily a white lower-socioeconomic population; and West Side Clinic, a community clinic serving mostly Hmong and Hispanic patients. These clinics were chosen to obtain subjects with a variety of socioeconomic and ethnic backgrounds. The interview consisted of several parts. Participants were shown the following liquid dosing devices: cylindrical spoon, medicine cup, oral dosing syringe, oral dropper, andby dispenser. In addition, they were shown a household teaspoon and a measuring spoon ([Figure]). They were asked which of the dosing devices they had in their homes and which they had ever used for dispensing liquid medications. Demographic information was also obtained.

The participants were also tested and scored on their ability to measure liquid medicines and interpret dosing instructions accurately. A total performance score was determined by adding the scores from the following activities, with a maximum obtainable score of 11. The investigator observed the subjects measuring 3 doses of medicine using a cylindrical spoon, medicine cup, and an oral dosing syringe and noted the accuracy of the measurement. The subjects received a score of 0 or 1 if the measurement was done incorrectly or correctly, respectively, for each of the 3 devices. Then the subjects were asked to indicate on a chart what times they would take medicine if it was prescribed every 6 hours, 4 times daily, and 3 times daily. The subjects received a score of 0, 1, or 2 for each of these 3 charts. The score of 0 was given if both the number of doses and the timing were incorrect, a 1 if either the number of doses or the timing was correct, and a 2 if both were correct. Finally, subjects were shown a pediatric dosing chart that had dosing listed by both age and weight and contained a note that dosing by weight is more accurate. Subjects were asked to indicate the correct dose for 2 children. In one example the child’s age and weight matched on the chart, and in the other the age and weight were discordant. A score of 1 was given for each correct reading of the dosing chart.

We calculated descriptive statistics and frequency distributions for all variables. Chi-square analyses were used for categorical comparisons. Grouped t tests and analysis of variance techniques were used to assess the effect of categorical predictors and demographics on total score. Pearson correlations were calculated to assess relationships among continuous variables.

Results

Of the 130 participants, 105 were women. The participants had a mean age of 40 years, a mean education level of 12.5 years, and a mean of 1.1 children in the household. Sixty-eight percent of the subjects were white, 19% Hispanic, 11% African American, and 2% Asian. English was the second language for 13% of the participants.

The liquid dosing devices available in the participants’ homes and the devices they used are shown in the [Table]. The one most frequently used (73%) for measuring liquid medication was a household teaspoon.

Ninety-two percent of the participants measured the correct dose when using the oral dosing syringe. Only 85% of the participants measured the correct dose of 1 tablespoon when using a medicine cup. The most frequent error (70%) occurred when the participants mistakenly measured 1 teaspoon instead of 1 tablespoon. Although 92% of the subjects ultimately measured the correct dose using the cylindrical spoon, many subjects spilled the liquid and required several attempts before measuring the correct dose.

Eighty-nine percent of participants noted the correct number of doses and time between doses when asked to indicate the times that they would take a medicine if they were instructed to take it 4 times daily and 3 times daily. Only 38% of the participants correctly indicated this information when they were instructed to take a medication every 6 hours. Participants commonly misinterpreted this instruction as meaning every 6 hours while awake, and indicated 3 rather than 4 doses.

The pediatric dosing chart was correctly interpreted by 87% of the participants for both case scenarios. Twelve percent of those surveyed gave the incorrect dose when the age and weight of the child were discordant, choosing the dose based on the child’s age rather than weight.

The participants’ mean total performance score was 9.5. Women scored significantly better than men. (9.7 vs 8.7, P <.05). Total performance score was significantly correlated with the participants’ education level (Pearson correlation=0.177, P <.05), but not with age or the number of children in the household. Total performance score did not differ significantly by the participants’ native language or ethnic group.

Discussion

In 1975 the AAP Committee on Drugs denounced the use of household teaspoons for administering liquid medications.¹ The volume of household teaspoons can range from 2 to 10 mL. Also, the same spoon when used by different persons may deliver from 3 to 7 mL. Therefore, even household measuring spoons are problematic. Participants in this study used a household teaspoon for measuring liquid medications more often than any other dosing device.

Oral dosing devices such as oral dosing syringes, oral droppers, cylindrical spoons, and medication cups are preferred over the traditional household teaspoon or measuring spoon, because they are more accurate. The advantages and disadvantages of the different devices have been described elsewhere.² The cylindrical spoon has been described as having an increased potential for easy spillage before and during administration of medication.² Participants in our study had problems with spillage with this device.

Study participants’ measurements were less accurate when using a measuring cup than when using a cylindrical spoon or oral dosing syringe, primarily because of confusion between the cup’s markings for tablespoons and teaspoons. A survey of poison control centers found 3 major causes of dosing errors using dispensing cups: (1) confusion of tablespoons for teaspoons; (2) the assumption that the entire cup was the unit of measure; and (3) the misinterpretation that 1 cupful was the recommended dose.⁵ After receiving reports of inappropriately marked plastic dosing cups, the Food and Drug Administration began a public education campaign in 1994 to increase health professional and consumer awareness of misdosing hazards with liquid medicines.⁶

The oral dosing syringe is felt to be the best device for delivery of liquid medication.² Its advantages include accuracy, convenience, availability in various sizes, and relatively low expense. The syringe permits the user to direct the delivery of the medication to the back and side of the mouth of an infant or small child, thus minimizing spillage. It also reduces the risk of possible gagging and aspiration of medication. Only a third of this study’s participants had an oral dosing syringe in their home.

It is alarming that the majority of participants misinterpreted instructions to take a medicine every 6 hours, so that they would take only 3 rather than 4 doses of medicine in a day. This problem of misinterpreting dosing frequency appears to be relatively unrecognized, although it could be an important cause of apparent treatment failures. Studies of medication errors typically focus on mistakes that cause clinical symptoms. In a study of 1108 medication errors in pediatrics reported to poison control centers in France, none involved underdosage errors.⁷

Most participants were able to correctly interpret a pediatric dosing chart, although some errors were noted when the age and weight were discordant. In a previous study, only 40% of caretakers were able to state a correct dose of acetaminophen for their child when given the child’s weight and all package labeling.⁸ Parents often fail to revise medication doses as a child grows older and gains weight and therefore tend to underdose.^4,8,9 Although less serious than overdosing in terms of morbidity, underdosing of acetaminophen may lead to ineffective treatment of fever and unnecessary visits to the clinic or emergency department.

It is not surprising that women and participants with higher education levels had higher total performance scores. In previous studies of liquid medication, the majority of the caretakers giving medication to children were mothers.^3,8-10 Women are therefore more likely to have experience administering liquid medication than men. Fortunately, parental education has been shown to be very effective in eliminating medication dosing errors.¹⁰ In a recent study by McMahon and colleagues of 90 English-speaking and Spanish-speaking families, 100% of them dosed medication correctly when given instructions and a syringe with a line marked at the prescribed dose.

Conclusions

The recommendation for the use of the oral dosing syringe made by the AAP almost 25 years ago is just as relevant today. On the basis of our study results, we make the following additional recommendations to clinicians: (1) when possible, indicate the dosing interval by the number of doses in a day, rather than by the number of hours between doses; and (2) always consider the possibility of a medication dosing error when faced with an apparent treatment failure.

Acknowledgments

Our study was funded by a grant from the American Academy of Family Physicians Foundation and the Minnesota Academy of Family Physicians Foundation.