Article Type
Changed
Thu, 03/28/2019 - 14:39
Display Headline
Does measuring episiotomy rates really benefit the quality of care our patients receive?
Are these the right metrics to evaluate our obstetric practice?

Like most California institutions performing deliveries, St. Joseph Hospital in Orange, California, started releasing 2016 maternal quality metrics internally at first. Data for the first 9 months of 2016 were distributed in December 2016. These metrics depend on a denominator based on the number of deliveries attributed to each obstetrician.

  • I was very pleased to see that I ranked first in the vaginal birth after cesarean (VBAC) rate at 36.8%.
  • I also was pleased that I ranked the fourth lowest, at 15.9%, for my cesarean delivery rate in the low-risk, nulliparous term singleton vertex (NTSV) population.
  • I was neither pleased nor displeased that I ranked number 29 of 31 physicians at 59.1% (39/66) for the episiotomy rate. The denominator range was 1 to 287. I knew I would hear about this! Sure enough, a medical director asked me how 2 of my metrics could be so good, yet the third be so abysmal.

After the release of the data and the somewhat humorous chastisement by the medical director, I decided to try complying with the new American College of Obstetricians and Gynecologists (ACOG) guidelines1 again beginning in January 2017.

A little personal history

Allow me to date myself. I completed my residency in 1981 and was Board Certified in 1984. My wife refers to me as a “Dinosaur!” As an ObGyn in solo practice, I take my own call.

During my training, episiotomies were commonly performed but were not always necessary. We were taught to not perform an episiotomy if the patient could safely and easily deliver without one. However, if clinically indicated, an episiotomy should be performed. If a 3rd- or 4th-degree laceration occurred, we were taught to anatomically repair it. Nowadays in my practice, these lacerations are rare in conjunction with an episiotomy, and with a controlled delivery of the fetal head.

Our nursing staff will tell you that I resist change. However, I usually attempt and often do adapt the latest national guidelines into my practice. Although I did not agree when restrictive episiotomies became a national goal a few years ago, I tried to adhere to the new national episiotomy recommendations.1 I am meticulous: a standard episiotomy repair that does not involve excessive bleeding usually takes 20 minutes to restore normal anatomy with a simple, straightforward, layered closure.

My episiotomy performance record

In 2015, I restricted my use of episiotomies. When I did not perform one, the patient usually experienced lacerations. These were labial or periurethral as well as complex 3-dimensional “Z” or “Y” shaped vaginal/perineal lacerations, not just the 1st- and 2nd-degree perineal lacerations to which the literature refers.

The problems associated with complex, geometric vaginal lacerations are multifactorial:

  • Lacerations occur at multiple locations.
  • Significant bleeding often occurs. Because the lacerations are in multiple locations, the bleeding cannot be addressed easily, quickly, or at once.
  • Visualization is difficult because of the bleeding, thus further prolonging the repair.
  • These lacerations are often deeper than an episiotomy would have been and are very friable as all the layers have been stretched to their breaking point before tearing.
  • Sometimes the lacerations include avulsion of the hymen with extensive bleeding.
  • Difficult-to-repair lacerations can tear upon suturing, requiring layer upon layer of sutures at the same site. Future scarring and vaginal stricture leading to sexual dysfunction are concerns.
  • At times, the friability and bleeding is so brisk that once the bleeding is controlled and the episiotomy is partially repaired, I can see that it is has not been an anatomic repair. I then have to take it down and re-do the repair with an obscured field from bleeding again!
  • Some repairs are so fragile that when I express retained blood from the uterus and upper vagina after completing the repair, the tissue tears, bleeds, and requires additional restoration.
  • These tears usually require an hour to repair and achieve hemostasis. At times, an assistant and a retractor are necessary.

After 2015, when I spent the year complying with the new guidelines, I returned to my original protocol: I performed an episiotomy only when I thought the patient was going to experience a significant laceration. I did not perform an episiotomy if I thought the mother could deliver easily without one. That is how I attained the 59.1% episiotomy rate in 2016.

Another try

After the 2016 hospital data were released, I decided to comply with the new guidelines1 again beginning in January 2017. Here I share the details of 3 deliveries that occurred in 2017:

  1. A 30-year-old woman (G2P1) planned to have a repeat cesarean delivery. At 38 3/7 weeks’ gestation, she was admitted in active labor with the cervix dilated to 5 cm. She requested a VBAC. After successful vaginal delivery without episiotomy of a 7 lb 5 oz infant, there were bilateral periurethral and right labia minora abrasions/lacerations.
  2. A 21-year-old woman (G1P0) at 40 4/7 weeks’ gestation was admitted in early labor. The cervix was 2-cm dilated and 70% effaced after spontaneous rupture of membranes. I exercised my clinical judgment and performed a midline episiotomy. A 9 lb 3 oz infant was delivered by vaginal delivery.
  3. A 16-year-old woman (G1P0) at 41 weeks’ gestation was admitted for induction of labor with an unripe cervix. I was delayed, and the laborist performed a vaginal delivery after 1 attempt at vacuum extraction and no episiotomy. The 7 lb 3 oz baby had Apgar scores of 4 and 9 at 1 and 5 minutes, respectively. There was significant bleeding from bilateral vaginal lacerations with bilateral hymeneal avulsions.

What is the benefit?

Are we really benefitting our patients by restricting the use of episiotomy? Consider these questions:

  • Should we delay the mother’s bonding with her baby for an hour’s complex repair versus 20 minutes for a simple, layered episiotomy repair?
  • In a busy labor and delivery unit, should resources be tied up for this extra time? With all due respect to the national experts advocating this recommendation, are they in the trenches performing deliveries and spending hours repairing complex lacerations?
  • Should we not use our clinical judgment instead of allowing the mother to experience an extensive vaginal/perineal laceration after a vaginal delivery of a 9- or 10-lb baby?
  • Where are the long-term data showing that it is better for a woman to stretch and attenuate her perineal and vaginal muscles to the breaking point, and then tear?
  • Do all the additional sutures lead to vaginal scarring, vaginal stricture, and sexual dysfunction in later years?
  • Which protocol better enables the mother to maintain pelvic organ support and avoid pelvic organ prolapse and stress urinary incontinence?

In Williams Obstetrics, the authors state: “We are of the view that episiotomy should be applied selectively for the appropriate indications. The final rule is that there is no substitute for surgical judgment and common sense.”2

Consider other metrics

Patients might be better served by measuring quality and safety metrics other than episiotomy. These might include, for example, measuring:

  • the use of prophylactic oxytocin after the anterior shoulder is delivered in order to decrease the risk of postpartum hemorrhage, as advocated by the California Maternal Quality Care Collaborative3
  • the number of patients admitted before active labor and those receiving an epidural before active labor (with the aim of decreasing the primary cesarean rate in the NTSV population)
  • the number of patients in an advanced stage of labor whose labor pattern has become dysfunctional, in whom no interventions have been instituted to improve the labor pattern, and who subsequently deliver by primary cesarean.

Recommendation

I recommend that performing an episiotomy should be an individual clinical decision for the individual patient by the individual obstetrician, and not a national mandate. We can provide quality care to our patients by performing selective episiotomies when clinically necessary, and not avoid them to an extreme that harms our patients. In my opinion, using the episiotomy rate as a quality metric should be abandoned.

Share your thoughts! Send your Letter to the Editor to [email protected]. Please include your name and the city and state in which you practice.

References
  1. CambriaAmerican College of Obstetricians and Gynecologists’ Committee on Practice Bulletins—Obstetrics. Practice Bulletin No. 165: Prevention and management of obstetric lacerations at vaginal delivery. Obstet Gynecol. 2016;128(1):e1Times New Roman MT Std–e15.
  2. Vaginal delivery. In: Cunnigham FG, Leveno KJ, Bloom SL, et al, eds. Williams Obstetrics. 24th ed. New York, NY: McGraw-Hill Medical; 2014:550.
  3. CMQCC: California Maternal Quality Care Collaborative website. https://www.cmqcc.org/. Accessed March 12, 2018.
Author and Disclosure Information

Dr. Kanofsky is in private ObGyn practice in Orange, California. He is a member of the Perinatal Safety Work Group and the Perinatal Collaborative Committee at St. Joseph Hospital, Orange, and is Chairman, Division of Gynecology, Department of Surgery, at CHOC Children’s Hospital of Orange County.

The author reports that he is a consultant to the St. Joseph Hospital Clinical Information Systems Project.

Publications
Topics
Sections
Author and Disclosure Information

Dr. Kanofsky is in private ObGyn practice in Orange, California. He is a member of the Perinatal Safety Work Group and the Perinatal Collaborative Committee at St. Joseph Hospital, Orange, and is Chairman, Division of Gynecology, Department of Surgery, at CHOC Children’s Hospital of Orange County.

The author reports that he is a consultant to the St. Joseph Hospital Clinical Information Systems Project.

Author and Disclosure Information

Dr. Kanofsky is in private ObGyn practice in Orange, California. He is a member of the Perinatal Safety Work Group and the Perinatal Collaborative Committee at St. Joseph Hospital, Orange, and is Chairman, Division of Gynecology, Department of Surgery, at CHOC Children’s Hospital of Orange County.

The author reports that he is a consultant to the St. Joseph Hospital Clinical Information Systems Project.

Are these the right metrics to evaluate our obstetric practice?
Are these the right metrics to evaluate our obstetric practice?

Like most California institutions performing deliveries, St. Joseph Hospital in Orange, California, started releasing 2016 maternal quality metrics internally at first. Data for the first 9 months of 2016 were distributed in December 2016. These metrics depend on a denominator based on the number of deliveries attributed to each obstetrician.

  • I was very pleased to see that I ranked first in the vaginal birth after cesarean (VBAC) rate at 36.8%.
  • I also was pleased that I ranked the fourth lowest, at 15.9%, for my cesarean delivery rate in the low-risk, nulliparous term singleton vertex (NTSV) population.
  • I was neither pleased nor displeased that I ranked number 29 of 31 physicians at 59.1% (39/66) for the episiotomy rate. The denominator range was 1 to 287. I knew I would hear about this! Sure enough, a medical director asked me how 2 of my metrics could be so good, yet the third be so abysmal.

After the release of the data and the somewhat humorous chastisement by the medical director, I decided to try complying with the new American College of Obstetricians and Gynecologists (ACOG) guidelines1 again beginning in January 2017.

A little personal history

Allow me to date myself. I completed my residency in 1981 and was Board Certified in 1984. My wife refers to me as a “Dinosaur!” As an ObGyn in solo practice, I take my own call.

During my training, episiotomies were commonly performed but were not always necessary. We were taught to not perform an episiotomy if the patient could safely and easily deliver without one. However, if clinically indicated, an episiotomy should be performed. If a 3rd- or 4th-degree laceration occurred, we were taught to anatomically repair it. Nowadays in my practice, these lacerations are rare in conjunction with an episiotomy, and with a controlled delivery of the fetal head.

Our nursing staff will tell you that I resist change. However, I usually attempt and often do adapt the latest national guidelines into my practice. Although I did not agree when restrictive episiotomies became a national goal a few years ago, I tried to adhere to the new national episiotomy recommendations.1 I am meticulous: a standard episiotomy repair that does not involve excessive bleeding usually takes 20 minutes to restore normal anatomy with a simple, straightforward, layered closure.

My episiotomy performance record

In 2015, I restricted my use of episiotomies. When I did not perform one, the patient usually experienced lacerations. These were labial or periurethral as well as complex 3-dimensional “Z” or “Y” shaped vaginal/perineal lacerations, not just the 1st- and 2nd-degree perineal lacerations to which the literature refers.

The problems associated with complex, geometric vaginal lacerations are multifactorial:

  • Lacerations occur at multiple locations.
  • Significant bleeding often occurs. Because the lacerations are in multiple locations, the bleeding cannot be addressed easily, quickly, or at once.
  • Visualization is difficult because of the bleeding, thus further prolonging the repair.
  • These lacerations are often deeper than an episiotomy would have been and are very friable as all the layers have been stretched to their breaking point before tearing.
  • Sometimes the lacerations include avulsion of the hymen with extensive bleeding.
  • Difficult-to-repair lacerations can tear upon suturing, requiring layer upon layer of sutures at the same site. Future scarring and vaginal stricture leading to sexual dysfunction are concerns.
  • At times, the friability and bleeding is so brisk that once the bleeding is controlled and the episiotomy is partially repaired, I can see that it is has not been an anatomic repair. I then have to take it down and re-do the repair with an obscured field from bleeding again!
  • Some repairs are so fragile that when I express retained blood from the uterus and upper vagina after completing the repair, the tissue tears, bleeds, and requires additional restoration.
  • These tears usually require an hour to repair and achieve hemostasis. At times, an assistant and a retractor are necessary.

After 2015, when I spent the year complying with the new guidelines, I returned to my original protocol: I performed an episiotomy only when I thought the patient was going to experience a significant laceration. I did not perform an episiotomy if I thought the mother could deliver easily without one. That is how I attained the 59.1% episiotomy rate in 2016.

Another try

After the 2016 hospital data were released, I decided to comply with the new guidelines1 again beginning in January 2017. Here I share the details of 3 deliveries that occurred in 2017:

  1. A 30-year-old woman (G2P1) planned to have a repeat cesarean delivery. At 38 3/7 weeks’ gestation, she was admitted in active labor with the cervix dilated to 5 cm. She requested a VBAC. After successful vaginal delivery without episiotomy of a 7 lb 5 oz infant, there were bilateral periurethral and right labia minora abrasions/lacerations.
  2. A 21-year-old woman (G1P0) at 40 4/7 weeks’ gestation was admitted in early labor. The cervix was 2-cm dilated and 70% effaced after spontaneous rupture of membranes. I exercised my clinical judgment and performed a midline episiotomy. A 9 lb 3 oz infant was delivered by vaginal delivery.
  3. A 16-year-old woman (G1P0) at 41 weeks’ gestation was admitted for induction of labor with an unripe cervix. I was delayed, and the laborist performed a vaginal delivery after 1 attempt at vacuum extraction and no episiotomy. The 7 lb 3 oz baby had Apgar scores of 4 and 9 at 1 and 5 minutes, respectively. There was significant bleeding from bilateral vaginal lacerations with bilateral hymeneal avulsions.

What is the benefit?

Are we really benefitting our patients by restricting the use of episiotomy? Consider these questions:

  • Should we delay the mother’s bonding with her baby for an hour’s complex repair versus 20 minutes for a simple, layered episiotomy repair?
  • In a busy labor and delivery unit, should resources be tied up for this extra time? With all due respect to the national experts advocating this recommendation, are they in the trenches performing deliveries and spending hours repairing complex lacerations?
  • Should we not use our clinical judgment instead of allowing the mother to experience an extensive vaginal/perineal laceration after a vaginal delivery of a 9- or 10-lb baby?
  • Where are the long-term data showing that it is better for a woman to stretch and attenuate her perineal and vaginal muscles to the breaking point, and then tear?
  • Do all the additional sutures lead to vaginal scarring, vaginal stricture, and sexual dysfunction in later years?
  • Which protocol better enables the mother to maintain pelvic organ support and avoid pelvic organ prolapse and stress urinary incontinence?

In Williams Obstetrics, the authors state: “We are of the view that episiotomy should be applied selectively for the appropriate indications. The final rule is that there is no substitute for surgical judgment and common sense.”2

Consider other metrics

Patients might be better served by measuring quality and safety metrics other than episiotomy. These might include, for example, measuring:

  • the use of prophylactic oxytocin after the anterior shoulder is delivered in order to decrease the risk of postpartum hemorrhage, as advocated by the California Maternal Quality Care Collaborative3
  • the number of patients admitted before active labor and those receiving an epidural before active labor (with the aim of decreasing the primary cesarean rate in the NTSV population)
  • the number of patients in an advanced stage of labor whose labor pattern has become dysfunctional, in whom no interventions have been instituted to improve the labor pattern, and who subsequently deliver by primary cesarean.

Recommendation

I recommend that performing an episiotomy should be an individual clinical decision for the individual patient by the individual obstetrician, and not a national mandate. We can provide quality care to our patients by performing selective episiotomies when clinically necessary, and not avoid them to an extreme that harms our patients. In my opinion, using the episiotomy rate as a quality metric should be abandoned.

Share your thoughts! Send your Letter to the Editor to [email protected]. Please include your name and the city and state in which you practice.

Like most California institutions performing deliveries, St. Joseph Hospital in Orange, California, started releasing 2016 maternal quality metrics internally at first. Data for the first 9 months of 2016 were distributed in December 2016. These metrics depend on a denominator based on the number of deliveries attributed to each obstetrician.

  • I was very pleased to see that I ranked first in the vaginal birth after cesarean (VBAC) rate at 36.8%.
  • I also was pleased that I ranked the fourth lowest, at 15.9%, for my cesarean delivery rate in the low-risk, nulliparous term singleton vertex (NTSV) population.
  • I was neither pleased nor displeased that I ranked number 29 of 31 physicians at 59.1% (39/66) for the episiotomy rate. The denominator range was 1 to 287. I knew I would hear about this! Sure enough, a medical director asked me how 2 of my metrics could be so good, yet the third be so abysmal.

After the release of the data and the somewhat humorous chastisement by the medical director, I decided to try complying with the new American College of Obstetricians and Gynecologists (ACOG) guidelines1 again beginning in January 2017.

A little personal history

Allow me to date myself. I completed my residency in 1981 and was Board Certified in 1984. My wife refers to me as a “Dinosaur!” As an ObGyn in solo practice, I take my own call.

During my training, episiotomies were commonly performed but were not always necessary. We were taught to not perform an episiotomy if the patient could safely and easily deliver without one. However, if clinically indicated, an episiotomy should be performed. If a 3rd- or 4th-degree laceration occurred, we were taught to anatomically repair it. Nowadays in my practice, these lacerations are rare in conjunction with an episiotomy, and with a controlled delivery of the fetal head.

Our nursing staff will tell you that I resist change. However, I usually attempt and often do adapt the latest national guidelines into my practice. Although I did not agree when restrictive episiotomies became a national goal a few years ago, I tried to adhere to the new national episiotomy recommendations.1 I am meticulous: a standard episiotomy repair that does not involve excessive bleeding usually takes 20 minutes to restore normal anatomy with a simple, straightforward, layered closure.

My episiotomy performance record

In 2015, I restricted my use of episiotomies. When I did not perform one, the patient usually experienced lacerations. These were labial or periurethral as well as complex 3-dimensional “Z” or “Y” shaped vaginal/perineal lacerations, not just the 1st- and 2nd-degree perineal lacerations to which the literature refers.

The problems associated with complex, geometric vaginal lacerations are multifactorial:

  • Lacerations occur at multiple locations.
  • Significant bleeding often occurs. Because the lacerations are in multiple locations, the bleeding cannot be addressed easily, quickly, or at once.
  • Visualization is difficult because of the bleeding, thus further prolonging the repair.
  • These lacerations are often deeper than an episiotomy would have been and are very friable as all the layers have been stretched to their breaking point before tearing.
  • Sometimes the lacerations include avulsion of the hymen with extensive bleeding.
  • Difficult-to-repair lacerations can tear upon suturing, requiring layer upon layer of sutures at the same site. Future scarring and vaginal stricture leading to sexual dysfunction are concerns.
  • At times, the friability and bleeding is so brisk that once the bleeding is controlled and the episiotomy is partially repaired, I can see that it is has not been an anatomic repair. I then have to take it down and re-do the repair with an obscured field from bleeding again!
  • Some repairs are so fragile that when I express retained blood from the uterus and upper vagina after completing the repair, the tissue tears, bleeds, and requires additional restoration.
  • These tears usually require an hour to repair and achieve hemostasis. At times, an assistant and a retractor are necessary.

After 2015, when I spent the year complying with the new guidelines, I returned to my original protocol: I performed an episiotomy only when I thought the patient was going to experience a significant laceration. I did not perform an episiotomy if I thought the mother could deliver easily without one. That is how I attained the 59.1% episiotomy rate in 2016.

Another try

After the 2016 hospital data were released, I decided to comply with the new guidelines1 again beginning in January 2017. Here I share the details of 3 deliveries that occurred in 2017:

  1. A 30-year-old woman (G2P1) planned to have a repeat cesarean delivery. At 38 3/7 weeks’ gestation, she was admitted in active labor with the cervix dilated to 5 cm. She requested a VBAC. After successful vaginal delivery without episiotomy of a 7 lb 5 oz infant, there were bilateral periurethral and right labia minora abrasions/lacerations.
  2. A 21-year-old woman (G1P0) at 40 4/7 weeks’ gestation was admitted in early labor. The cervix was 2-cm dilated and 70% effaced after spontaneous rupture of membranes. I exercised my clinical judgment and performed a midline episiotomy. A 9 lb 3 oz infant was delivered by vaginal delivery.
  3. A 16-year-old woman (G1P0) at 41 weeks’ gestation was admitted for induction of labor with an unripe cervix. I was delayed, and the laborist performed a vaginal delivery after 1 attempt at vacuum extraction and no episiotomy. The 7 lb 3 oz baby had Apgar scores of 4 and 9 at 1 and 5 minutes, respectively. There was significant bleeding from bilateral vaginal lacerations with bilateral hymeneal avulsions.

What is the benefit?

Are we really benefitting our patients by restricting the use of episiotomy? Consider these questions:

  • Should we delay the mother’s bonding with her baby for an hour’s complex repair versus 20 minutes for a simple, layered episiotomy repair?
  • In a busy labor and delivery unit, should resources be tied up for this extra time? With all due respect to the national experts advocating this recommendation, are they in the trenches performing deliveries and spending hours repairing complex lacerations?
  • Should we not use our clinical judgment instead of allowing the mother to experience an extensive vaginal/perineal laceration after a vaginal delivery of a 9- or 10-lb baby?
  • Where are the long-term data showing that it is better for a woman to stretch and attenuate her perineal and vaginal muscles to the breaking point, and then tear?
  • Do all the additional sutures lead to vaginal scarring, vaginal stricture, and sexual dysfunction in later years?
  • Which protocol better enables the mother to maintain pelvic organ support and avoid pelvic organ prolapse and stress urinary incontinence?

In Williams Obstetrics, the authors state: “We are of the view that episiotomy should be applied selectively for the appropriate indications. The final rule is that there is no substitute for surgical judgment and common sense.”2

Consider other metrics

Patients might be better served by measuring quality and safety metrics other than episiotomy. These might include, for example, measuring:

  • the use of prophylactic oxytocin after the anterior shoulder is delivered in order to decrease the risk of postpartum hemorrhage, as advocated by the California Maternal Quality Care Collaborative3
  • the number of patients admitted before active labor and those receiving an epidural before active labor (with the aim of decreasing the primary cesarean rate in the NTSV population)
  • the number of patients in an advanced stage of labor whose labor pattern has become dysfunctional, in whom no interventions have been instituted to improve the labor pattern, and who subsequently deliver by primary cesarean.

Recommendation

I recommend that performing an episiotomy should be an individual clinical decision for the individual patient by the individual obstetrician, and not a national mandate. We can provide quality care to our patients by performing selective episiotomies when clinically necessary, and not avoid them to an extreme that harms our patients. In my opinion, using the episiotomy rate as a quality metric should be abandoned.

Share your thoughts! Send your Letter to the Editor to [email protected]. Please include your name and the city and state in which you practice.

References
  1. CambriaAmerican College of Obstetricians and Gynecologists’ Committee on Practice Bulletins—Obstetrics. Practice Bulletin No. 165: Prevention and management of obstetric lacerations at vaginal delivery. Obstet Gynecol. 2016;128(1):e1Times New Roman MT Std–e15.
  2. Vaginal delivery. In: Cunnigham FG, Leveno KJ, Bloom SL, et al, eds. Williams Obstetrics. 24th ed. New York, NY: McGraw-Hill Medical; 2014:550.
  3. CMQCC: California Maternal Quality Care Collaborative website. https://www.cmqcc.org/. Accessed March 12, 2018.
References
  1. CambriaAmerican College of Obstetricians and Gynecologists’ Committee on Practice Bulletins—Obstetrics. Practice Bulletin No. 165: Prevention and management of obstetric lacerations at vaginal delivery. Obstet Gynecol. 2016;128(1):e1Times New Roman MT Std–e15.
  2. Vaginal delivery. In: Cunnigham FG, Leveno KJ, Bloom SL, et al, eds. Williams Obstetrics. 24th ed. New York, NY: McGraw-Hill Medical; 2014:550.
  3. CMQCC: California Maternal Quality Care Collaborative website. https://www.cmqcc.org/. Accessed March 12, 2018.
Publications
Publications
Topics
Article Type
Display Headline
Does measuring episiotomy rates really benefit the quality of care our patients receive?
Display Headline
Does measuring episiotomy rates really benefit the quality of care our patients receive?
Sections
Disallow All Ads
Content Gating
No Gating (article Unlocked/Free)
Alternative CME
Disqus Comments
Default
Eyebrow Default
Value-Based Medicine
Use ProPublica