Using sequence analysis to visualize exposure to pregnancy in the postpartum period

TL;DR

This study uses sequence analysis to identify patterns of postpartum exposure to pregnancy among Ethiopian women, revealing distinct behavioral and biological trajectories.

Contribution

The novel use of sequence and cluster analyses provides new insights into postpartum reproductive behavior patterns in a low-resource setting.

Findings

Five distinct postpartum exposure trajectories were identified, including clusters based on sexual activity, menses return, and contraceptive use.

The largest cluster (50%) involved resumption of sex without menses or contraception around three months postpartum.

Family-planning adopters (39%) showed varied patterns depending on whether menses had returned.

Abstract

Exposure to pregnancy during the postpartum period is shaped by biological and behavioral determinants, such as resumption of sexual activity, return of menses, and contraceptive use dynamics.

We implement sequence and cluster analyses to generate new insights about exposure to pregnancy during the postpartum period using unique longitudinal data in a low-resource setting.

We used population-based data from a sample of 1,935 Ethiopian women who provided reports on factors influencing exposure to pregnancy in the year following childbirth. We used sequence and cluster analyses to characterize patterns of women’s reproductive behaviors during the postpartum period.

We identified five postpartum trajectories of exposure to pregnancy: (1) no sex; (2) family-planning adopters, no menses; (3) family-planning adopters, return of menses; (4) sex, no menses, no family planning; and (5) sex, menses, no family planning. The ‘sex, no menses, no family planning’ cluster (50% of the sample) was characterized by resumption of sexual activity around three months postpartum, amenorrhea, and no contraceptive adoption. Women in the two ‘family-planning adopters’ clusters (39%) resumed sexual activity and adopted contraception around three months postpartum but differ by return of menses. The ‘no sex’ cluster (5%) was characterized by no sexual activity, contraceptive use, or menses.

Sequence analysis offers new insights into a critical reproductive window by emphasizing the dynamic biological and behavioral states that influence distinct patterns of postpartum exposure to pregnancy. Establishing longitudinal trajectories of exposure to pregnancy has research and programmatic implications that include a more holistic understanding of postpartum fecundity and measuring unmet need for postpartum family planning.

Figures