Objective: Implicit in comparisons of unadjusted cesarean rates for hospitals and providers is the assumption that differences result from management practices rather than differences in case mix. This study proposes a method for comparison of cesarean rates that takes the effect of case mix into account.
Methods: All women delivered of infants at our institution from December 1, 1994, through July 31, 1995, were classified according to whether they received care from community-based practitioners (N=3913) or from the hospital-based practice that serves a higher-risk population (N=1556). Women were categorized according to both obstetric history (nulliparas, multiparas without a previous cesarean, multiparas with a previous cesarean) and the presence of obstetric conditions influencing the risk of cesarean delivery (multiple birth, breech presentation or transverse lie, preterm, no trial of labor for a medical indication). We determined the percent of women in each parity-obstetric condition subgroup and calculated a standardized cesarean rate for the hospital-based practice using the case mix of the community-based practitioners as the standard.
Results: The crude cesarean rate was higher for the hospital-based practice (24.4%) than for the community-based practitioners (21.5%), a rate difference of 2.9% (95% confidence interval=0.4%, 5.4%; P=.02). However, the proportion of women falling into categories conferring a high risk of cesarean delivery (multiple pregnancy, breech presentation or transverse lie, preterm, no trial of labor permitted) was twice as high for the hospital-based practice (24.4% hospital, 12.1% community). The standardization indicates that if the hospital-based practitioners had the same case mix as community-based practitioners, their overall cesarean rate would be 20.1%, similar to the 21.5% rate of community providers (rate difference=-1.4%, 95% confidence interval =-3.1%, 0.3%; P=.11).
Conclusion: Standardization for case mix provides a mechanism for distinguishing differences in cesarean rates resulting from case mix from those relating to differences in practice. The methodology is not complex and could be applied to facilitate fairer comparisons of rates among providers and across institutions.