The effect of induction of labor on the odds of primary cesarean delivery
The induction rate in the United States has more than doubled since 1989 to a high of 20.6% (Martinet al., 2003). The purpose of this study was to examine the effect of labor induction on the odds of primary cesarean delivery controlling for factors that confound or modify the effect with a database that provides size and scope. The study design was a retrospective, explanatory study design, using secondary data analysis with a database comprised of hospital discharge data, birth certificate data, AHA Annual Hospital Survey data, and neonatal intensive care survey data (N= 160,533 births). Data from cases with total agreement between the hospital discharge data and the birth certificate data on the two key variables of interest, labor induction and primary cesarean delivery, were used for the analysis. Secondary data analysis of administrative data provided an advantage of having a large volume of demographically diverse cases, thus addressing issues of generalizability and sample size limitations. Multivariable logistic regression methods were deployed as described by Hosmer and Lemeshow (2000) to control for confounding and effect modifying variables that influence the relationship of labor induction on the odds of primary cesarean delivery. Study findings indicated that the effect of labor induction was confounded by maternal age and teaching status. Additional findings included significant (p s 0.01) effect modification with six covariates: baby weight, parity, dystocia, medical indication for induction, gestation, and race and ethnicity. The findings indicate that effect modifying covariates interact with the effect of labor induction on the odds of primary cesarean delivery. When labor induction was examinedcontrolling for effect modifiers and confounders, the odds ratio was 1.165 (95% CI: 0.948; 1.430;p >.05), indicating that labor induction no longer significantly influenced the effect of labor induction on the odds of primary cesarean delivery, independent of other client and system characteristics. An important conclusion was that labor induction increases the odds of primary cesarean delivery under certain conditions dependent on the level of the effect modifying client characteristics.