Content area
Abstract
Purpose is to improve methodology of measuring contraceptive failure rates using retrospective fertility surveys. It derives and applies an upward correction factor to use-failure rates to account for unreported contraceptive failures which resulted in abortions. A simple model of pregnancy and abortion reporting process leads to a correction factor which requires only (1) the probability of abortion as a pregnancy outcome and (2) abortion reporting on the fertility survey. The correction procedure adapts readily to multivariate life table methodology.
The 1982 National Survey of Family Growth provides retrospective contraceptive use-intervals for a sample of women 15-44 for the period 1979-1981. A combination of state and national induced abortion data for 1980 furnishes abortion probabilities by age, race, marital status, and contraceptive used (if any) at conception. About 34% of abortions are the result of contraceptive failure. Users who fail with the Pill and diaphragm are most likely, and rhythm users least likely, to have an induced abortion.
Abortion reporting on the 1982 NSFG varies by age, race, and marital status. Overall reporting of induced abortion is 50%. Whites report more of their abortions than nonwhites, married women more than unmarried, older more than younger. Of the bivariate combinations, white 30-34 year olds have the highest abortion reporting, 76%.
An earlier study of contraceptive failure rates from the 1982 NSFG (uncorrected for induced abortion reporting) finds that rates vary by age, marital status, parity, income, and other variables, but not by race. In addition, it finds surprisingly that failure rates for the rhythm method are comparable to condom and diaphragm rates.
This re-analysis correcting for abortion under-reporting paints a different picture of the covariates of contraceptive failure. After correction, failure rates vary by age and race, but not by marital status. The standardized failure rates in the first 12 months of use are: Pill 3.7%, IUD 5.8%, Condom 10.3%, Diaphragm 19.8%, Spermicides 24.9%, and Rhythm 12.2%. Rhythm failure rates appear comparable to those for condom. The adjustment procedure strengthens this finding because women who fail with rhythm are least likely to have an induced abortion.
