Center for Healthcare Policy and Research

The California Hospital Outcomes Program is an initiative mandated by the State of California, and conducted by the Office of Statewide Health Planning and Development (OSHPD), to develop public reports comparing hospital outcomes for selected conditions and treatments in hospitals throughout the state. Over the last decade, OSHPD has reported hospital mortality rates for heart attack and community-acquired pneumonia. A separate OSHPD program produces reports on hospital and surgeon outcomes for heart bypass surgery (www.oshpd.ca.gov).

This is the first public report that OSHPD has published on maternal hospital care in California. The report is based on analysis of Patient Discharge Data (PDD) records submitted to OSHPD by licensed acute care hospitals, as well as Vital Statistics (VS) birth certificate records submitted to the California Department of Public Health. The delivery patients were admitted to the hospital between October 1999 and November 2001.

The quality of hospital performance for maternity care was assessed by estimating each hospital’s rate of two undesirable outcomes: severe perineal lacerations (tears) and postpartum maternal readmissions. Severe perineal lacerations, also described as 3rd or 4th degree, are common but often painful complications of vaginal births. Postpartum maternal readmissions reflect rare but serious complications that occur within 6 weeks after delivery, and require that a woman be readmitted to receive intravenous fluids, powerful antibiotics, surgery, or close monitoring. Both of these quality indicators were risk-adjusted to account for differences in patients’ underlying risk of these undesirable outcomes. Each hospital’s risk-adjusted rate was then compared with the statewide average, which serves as a benchmark. Hospitals are defined as “better” if their risk-adjusted laceration or readmission rates were statistically significantly lower than the state rate and “worse” if their rates were higher.

To provide more information for women and their families, this report also shows each hospital’s vaginal or cesarean delivery rate for two important groups of women: low-risk women who are admitted for their first deliveries, and high-risk women who have had at least one prior cesarean delivery. We focus on these two groups of women because their risk of cesarean delivery is particularly high, and because that risk varies widely across hospitals. By contrast, women who have had prior vaginal deliveries, and no prior cesarean deliveries, tend to have a very low risk of cesarean delivery with subsequent pregnancies, no matter where they go for hospital care. Women with other high risk factors, such as having babies that present feet-first (footling breech) or buttocks-first (breech) instead of head-first, tend to have a very high risk of cesarean delivery no matter where they go.

Key findings from this report include:

• During the study period, 860,588 eligible women were admitted to acute care, nonfederal hospitals in California for delivery of a live baby. Of these women, 4,029 (0.47%) were readmitted to any hospital within 6 weeks after delivery because of a postpartum complication.

• Across the 301 eligible hospitals, the number of eligible deliveries during the study period ranged from 1 to 12,811, and the number of readmissions ranged from 0 to 89. The risk-adjusted readmission rate was 0% for 17 hospitals, 0.01% to 0.49% for 159 hospitals, 0.5% to 0.99% for 110 hospitals, 1.0% to 1.49% for 10 hospitals, and 1.5% or greater for 5 hospitals.

• Sixteen hospitals had significantly fewer readmissions than were expected, based on the characteristics of their patients, whereas fourteen hospitals had significantly more readmissions than were expected.

• During the study period, 651,640 eligible women were admitted to acute care nonfederal hospitals in California and underwent an eligible vaginal delivery. Of these women, 31,331 (4.81%) experienced a third or fourth degree tear.

• Across the 301 eligible hospitals, the number of eligible vaginal deliveries during the study period ranged from 1 to 9,815, and the number of third or fourth degree tears ranged from 0 to 597. The risk-adjusted laceration rate was less than 5% for 180 hospitals, 5% to 10% for 110 hospitals, 10% to 15% for 8 hospitals, and 15% or greater for 3 hospitals.

• Seventy-three hospitals had significantly fewer lacerations than were expected, based on the characteristics of their patients, whereas fifty-six hospitals had significantly more lacerations than were expected.

• There was a weak but consistent association between the risk-adjusted readmission rate (among all deliveries) and the risk-adjusted laceration rate (among vaginal deliveries) at the hospital level. For example, 8 of the 14 hospitals rated as “worse than expected” for postpartum readmissions were also rated as “worse than expected” for tears. Similarly, 8 of the 16 hospitals rated as “better than expected” for postpartum readmissions were also rated as “better than expected” for tears. The concordance between these indicators at the hospital level was surprisingly strong.

• It is critical that all hospitals providing maternal care implement the “best practice” guidelines supported by the medical community. OSHPD’s Postpartum Maternal Outcomes Validation Study suggested that many postpartum readmissions could be prevented through careful evaluation of every patient before discharge and prompt attention to early signs of infection. Other clinical and epidemiological studies (summarized below) have suggested that many perineal tears could be prevented by minimizing use of forceps and episiotomy and avoiding certain positions during labor.

• Coding problems do not appear to cause substantial bias in these analyses, but still need to be addressed by California hospitals. For example, about 0.3% of vaginal delivery records and 0.5% of cesarean delivery records had prohibited combinations of 5th digit ICD-9-CM codes, leading to confusion about whether the affected records were antepartum, childbirth, or postpartum records.