This project comprises the behavioral research component of the U54 Contraceptive Research Center (PI: Georg). In the United States, the proportion of pregnancies that is unintended?either mistimed or unwanted? is high (approximately 45%), but only a small fraction (5%) of unintended pregnancies are due to contraceptive failure. Instead, the vast majority of unintended pregnancies occur to couples who were either inconsistently using contraception (41%) or not using contraception at all (54%). In clinical trials, consistent use of effective contraception reduces a couple's risk of an unintended pregnancy to nearly zero ? that is, the clinical trial/perfect use failure rate for most methods is extremely low (e.g., 2% for condoms, 4% for withdrawal, and .3% for oral contraceptive pills). However, typical use failure rates for couples are much higher (e.g., 18% for condoms, 22% for withdrawal, and 9% for oral contraceptive pills). We know very little about these differences between perfect and typical use failure rates, largely due to data limitations. We propose to use newly available, innovative data that overcomes the prior barriers to understanding these typical use patterns. The Relationship Dynamics and Social Life (RDSL) project collected data on a population-representative random sample of 953 women ages 18 and 19 in a Michigan county. Following a face-to-face 50-minute interview conducted by a professional interviewer at the University of Michigan's Survey Research Center, the young women completed online or phone follow-up interviews weekly for 2.5 years (130 weeks). Weekly measures of contraceptive use and consistency provide an unprecedented opportunity to understand the dynamics of typical contraceptive use practices among the 2,500 couples included in the study. We propose an innovative data reduction technique ? sequence analysis ? to describe, summarize, and classify typical use patterns. Because the data are based on a population that is diverse in terms of race/ethnicity, socioeconomic characteristics, and adolescent experiences related to pregnancy, we will also analyze these patterns for population sub-groups. Further, because the data also includes measures of relationship characteristics with weekly precision, we propose previously impossible analyses of how the characteristics of couples that change weekly (e.g., conflict, sexual exclusivity, etc.) influence their weekly contraceptive use and consistency, and contribute to the overall patterns of effective (or ineffective) contraceptive use throughout their relationship. Thus, the proposed project investigates a persistent puzzle in understanding couples' contraceptive use behaviors (why typical use failure rates are so high), by combining an innovative dataset (RDSL) with state-of-the-art statistical analysis techniques (sequence analysis, couple-level fixed effects models). The results of this project will contribute to the development of new contraceptive methods that are more likely to be used consistently by couples, thereby increasing effective contraceptive use and reducing unintended pregnancy.
The fraction of pregnancies in the United States that is unintended remains high (~50%). Reducing this rate has been a goal in the Department of Health and Human Services? Healthy People initiative every decade since its inception in 1980. Births from these pregnancies experience lower quality social, economic, and health circumstances, as do their mothers and siblings. Further, because many unintended births are paid for by public insurance programs, they are costly ? at least $11.5 billion in 2006.
