Comparison of Brief Pregnancy Preference Measures with the Desire to Avoid Pregnancy Scale among Kenyan Women Living with HIV

Abstract

Accurate assessment of pregnancy preferences can inform patient-centered reproductive health (RH) care, yet preferences are complex, dynamic, and difficult to measure, particularly for women living with HIV (WLWH). The Desire to Avoid Pregnancy (DAP) scale is a measure of pregnancy preferences that is validated, but its length and scoring complexity may limit clinical implementation. To identify more feasible alternatives, we compared two single-item measures (pregnancy planning and importance of pregnancy prevention) to the multidimensional DAP scale using baseline data from 2,048 Kenyan WLWH enrolled in the Mobile WACh Empower cluster randomized clinical trial. Preferences were categorized as preference to avoid pregnancy, ambivalence/indifference, and preference toward pregnancy. Categorical agreement between both single-item measures and DAP (as the comparator) was assessed using Kendall’s coefficient of concordance (W) and percent agreement; sensitivity and specificity for each single-item measure’s three pregnancy preference categories compared to DAP were estimated using generalized estimating equations with Poisson distributions and cluster-robust standard errors. Median age was 33 years and 72% reported being married. Using Kendall’s W concordance thresholds of >0.7 (strong), 0.3-0.7 (moderate), <0.3 (weak), concordance with DAP was moderate for the planning and importance measures (W = 0.64 and 0.56, respectively). Category-specific agreement (planning vs importance) was 80% vs 74% for preference to avoid pregnancy, 77% vs 88% for preference toward pregnancy, and 28% vs 16% for ambivalence/indifference, respectively (for all). Sensitivity for preference to avoid pregnancy was high (planning 0.83; importance 0.96) but low for ambivalence/indifference (0.35; 0.05) and preference toward pregnancy (0.14; 0.11). Specificity was modest for preference to avoid pregnancy (0.47; 0.15) but high for both ambivalence/indifference (0.79; 0.94) and preference toward pregnancy (1.0 for both). Sensitivity varied more by covariates than specificity for both measures. Notably, three or more children and affirmative preference for another child lowered sensitivity, while undecided preference for another child increased sensitivity. Single-item measures demonstrated high sensitivity for identifying women preferring to avoid pregnancy and may be useful in routine care to identify contraceptive needs. However, neither single-item measure adequately captured ambivalence/indifference relative to DAP, highlighting the continued need for concise, valid measures to support individualized RH care for WLWH.