92: The effect of hindsight and outcome bias on the interpretation and management of intrapartum fetal heart rate (FHR) tracings: A randomized trial

Abstract

To determine whether obstetricians who are asked to evaluate management decisions regarding abnormal FHR patterns will be more likely to assess the quality of medical care as negligent if they are told that there was an adverse, rather than normal, neonatal outcome. Obstetricians from four institutions were invited to participate in a research study regarding intrapartum management. Each case included clinical details, segments of the intrapartum FHR tracing, and outcomes including umbilical artery pH and presence or absence of neonatal morbidity (Table 1). Reviewers were asked to categorize the FHR tracing, assess the timing for the decision for cesarean delivery (CD), and finally rate the quality of care as either “below standard of care” or “acceptable practice.” All participants received the same clinical data and outcomes for both Case #1 and Case #2. Case #1 involved a scenario with a normal umbilical pH, whereas case #2 resulted in an abnormal umbilical pH (< 7.00). For Case #3, all participants received the same opening clinical vignette and FHR tracings; but then were randomly allocated to receive one of two alternative outcomes (Scenario A = normal umbilical pH and no neonatal complications or Scenario B = metabolic acidemia and neonatal encephalopathy). We estimated that N= 84 would be needed to show a 30% difference in the frequency of the judgment of obstetrical management as “below standard of care” between those receiving alternative outcomes (90% power and 2-sided alpha = 0.05). A total of 128 obstetricians participated; 70 were randomly allocated to Scenario A and 58 to Scenario B for Case #3. There were no differences in the assessment of the appropriateness and quality of care for the first 2 cases (Table 2). For Case #3, belief that there was an adverse outcome resulted in a worse categorization of the FHR tracing, more frequent perception that CD was performed too late, and more likely judgment that the standard of care had been violated (Table 2). The findings were similar regardless of academic institution, attending or trainee status, or MFM subspecialty. This study confirms the significant impact of hindsight and outcome bias on the interpretation of FHR tracings and the retrospective assessment of the quality of obstetrical care.View Large Image Figure ViewerDownload Hi-res image Download (PPT)