On the Stability of Reported Pregnancy Intentions from Pregnancy to 1 Year Postnatally: Impact of Choice of Measure, Timing of Assessment, Women's Characteristics and Outcome of Pregnancy.

OBJECTIVES: Retrospective, cross-sectional estimates of pregnancy intention, as used in the Demographic Health Survey (DHS), are the global norm. The London Measure of Unplanned Pregnancy (LMUP) is a newer, psychometrically validated measure which may be more reliable. This paper assesses the reliability of the LMUP and the DHS question over the first postnatal year and explores the effects of maternal characteristics or pregnancy outcome on reported pregnancy intention. METHODS: We compared the test-retest reliability of the LMUP (using the AC coefficient) and DHS question (using the weighted Kappa) over the first postnatal year using data from Malawian women. We investigated the effect of maternal characteristics and pregnancy outcome using t-tests, Chi squared or Fisher's exact tests, and calculated odds ratios to estimate effect size. RESULTS: The DHS question was associated with a statistically significant decrease in the prevalence of unplanned pregnancies from 1-to-12 months postnatally; the LMUP was not. The LMUP had moderate to substantial reliability (0.51-0.66); the DHS had moderate reliability (0.56-0.58). The LMUP's stability was not related to any of the factors examined; the stability of the DHS varied by marital status (p = 0.033), number of children (p = 0.048) and postnatal depression (p < 0.001). Both underestimated unintended pregnancy postnatally vis-à-vis the LMUP in pregnancy. CONCLUSIONS FOR PRACTICE: The LMUP is a more reliable measure of pregnancy intention than the DHS in the first postnatal year and does not vary by maternal characteristics or pregnancy outcome. The LMUP should become the gold-standard for measuring pregnancy intention and should be collected in pregnancy or at the first postnatal opportunity.