Changes in the Number and Types of Interventions Used in Childbirth Following the 2003 and 2011 Resident Duty-Hour Reforms

In July 2003, the Accreditation Council for Graduate Medical Education (ACGME) issued new nationwide limitations on the number of hours residents could work in a week as well as the maximum duration of consecutive shifts. ACGME enacted a revised policy in 2011 that further restricted the work hours of first-year residents and also intensified the Council’s oversight (Rosenbaum & Lamas, 2012). The effectiveness of these reforms in reducing medical errors and improving patient outcomes has been mixed. One potential unintended consequence of this reform is an increase in uncertainty in decision-making due to an increased number of handoffs between providers. Another result of the policy change may be that residents face end-of-shift pressures to finish patient care tasks that may result in additional interventions. This study examines the differences in number of procedures used in childbirth to test these hypotheses.

We analyzed mothers who had not had a previous c-section using the 2000-2014 Nationwide Inpatient Samples (NIS) from the Healthcare Cost and Utilization Project (HCUP) of the Agency for Healthcare Research and Quality (AHRQ). Births were classified as high, medium, and low-risk by a comprehensive risk-assessment using Clinical Classifications Software codes for comorbid diagnoses. Predelivery lengths of stay were calculated using procedure dates as well as estimates based on expected trends. The response variables of interest are the total number of birth-related interventions administered during each hospital birth, whether the mother delivered via C-section, and whether any procedures designed to hasten delivery were used during birth. Changes in outcome variables were measured between teaching and non-teaching hospitals before and after the duty-hour reforms using differences-in-differences analysis.

Between 2000 and 2012, there were 7,059,123 births with no indication of prior C-section recorded in the HCUP NIS that contained information on all variables of interest. In the entire cohort, 17% of mothers delivered via C-section, the average number of birth-related procedures (including forceps and vacuum delivery, membrane rupture, episiotomy, and other birth-related procedures) was 1.59 (SD 1.05), and 75% underwent at least one procedure to hasten delivery. Using the differences-in-differences model to assess the impact of the 2003 reform, we see a significant decrease in the total number of procedures and the likelihood of using a procedure to hasten delivery among the cohort of low risk women. We do not find persistent trends among the medium and high-risk cohorts as a result of the 2003 reform. When examining the impact of the 2011 reform using differences-in-differences, we see a significant 1% reduction in the number of C-sections [95% CI -0.013% - -0.007%], a simultaneous 1.7% increase in other procedures to hasten delivery [95% CI 1.1% - 2.3%], and an overall suggestive increase in the number of procedures. These trends are strongest in the low-risk cohort with short pre-delivery lengths of stay, and provide preliminary evidence for the hypothesis that residents may increase the number of delivery-hastening procedures as they face end-of-shift pressures to avoid a costly patient handoff.