Eleven percent of primary and 65% of repeat cesarean deliveries performed in the United States in 2001 may not have been clinically necessary, according to an analysis of data from a large, nationally representative database.1 White women were significantly less likely than blacks to have a primary cesarean that was potentially unnecessary, but the reverse was true for repeat procedures. The likelihood that a cesarean delivery may have been unnecessary also varied by a number of other maternal characteristics, as well as by some hospital characteristics; for a number of factors, associations with primary procedures were in the opposite direction of those for repeat cesareans.

The analysis was based on information from the 2001 Healthcare Cost and Utilization Project National Inpatient Sample database, which contains records of stays in more than 1,000 hospitals in 33 states. All women who had a singleton live birth were included in the calculations. Those who delivered by cesarean were classified according to whether this was their first such delivery or a repeat procedure; if a woman's discharge record did not document any of 24 standard indications for cesarean, the procedure was categorized as potentially unnecessary. Notably, to ensure that their findings were as comparable as possible to those of earlier work, the analysts did not consider previous surgical delivery an indication for cesarean.

Nationwide, the analysts estimate, 540,174 primary and 371,863 repeat cesareans were performed in 2001. Of these, 11% and 65%, respectively, were potentially unnecessary. The descriptive data suggest some degree of variation in these rates by maternal and hospital characteristics. For example, the proportion of primary cesareans that may not have been necessary was as high as 14% among black women and 16% among Medicare recipients; it was as low as 9% among women living in the West. By contrast, black women and Medicare recipients had below-average rates of potentially unnecessary repeat cesareans (62% and 59%, respectively), and the rate reached 70% in rural hospitals and in government-owned facilities. The analysts used multiple logistic regression to explore the characteristics associated with potentially unnecessary cesareans.

In the multivariate analysis, white women, Hispanic women and women of other or unknown ethnicities were significantly less likely than blacks to have a primary cesarean that was potentially unnecessary (odds ratios, 0.7-0.8). Women aged 35 and older had an elevated likelihood of this outcome, as did women who were admitted to the hospital on a weekend (1.1 for each). The odds that a primary cesarean may have been unnecessary were higher among Medicare recipients than among women with private insurance (1.4), and higher among Southerners than among women from the West (1.2).

With repeat cesareans, however, the odds that a procedure was potentially unnecessary were elevated among white women (odds ratio, 1.1), and were reduced among women aged 35 or older (0.8) and those whose hospitalization began on a weekend (0.7). In addition, repeat cesareans performed in rural hospitals were more likely than those done in urban, teaching hospitals to be potentially unnecessary (1.3). Findings at a marginal level of statistical significance suggest a reduced likelihood that procedures performed for women receiving Medicare and those done in private hospitals may not have been necessary, and an increased likelihood for those performed in urban, nonteaching hospitals.

In commenting on the findings, the analysts emphasize that they cannot draw "definitive conclusions about whether the potentially unnecessary cesareans were clinically unnecessary." Nor can they identify the factors—including physicians' preferences, women's preferences and clinical concerns—underlying the racial and other differences observed in the likelihood of potentially unnecessary cesareans. Nevertheless, they conclude that "a large number of cesareans are not supported by documentation of recognized clinical indicators," and that eliminating such procedures would reduce the costs of delivery and the risk of adverse maternal outcomes.

D. Hollander


1. Kabir AA et al., Racial differences in cesareans: an analysis of U.S. 2001 national inpatient sample data, Obstetrics & Gynecology, 2005, 105(4):710-718.