Comparison of pregnancy and neonatal outcomes in a retrospective full pregnancy history survey versus population-based prospective records: a validation study in rural Sarlahi District, Nepal

Abstract

Introduction: Countries without complete civil registration and vital statistics systems rely on retrospective full pregnancy history surveys (FPH) to estimate incidence of pregnancy and mortality outcomes, including stillbirth and neonatal death. Yet surveys are subject to biases that impact demographic estimates, and few studies have quantified these effects. We compare data from an FPH vs. prospective records from a population-based cohort to estimate validity for maternal recall of live births, stillbirths, and neonatal deaths in a rural population in Sarlahi District, Nepal.

Methods: We used prospective data, collected through frequent visits of women from early pregnancy through the neonatal period, from a population-based randomized trial spanning 2010-2017. We randomly selected 76 trial participants from three pregnancy outcome groups: live birth (n = 26), stillbirth (n = 25), or neonatal death (n = 25). Data collectors administered the Nepal 2016 Demographic and Health Surveys (DHS)-VII pregnancy history survey between October 22, 2021, and November 18, 2021. We compared total pregnancy outcomes and numbers of pregnancy and neonatal outcomes between the two data sources. We matched pregnancy outcomes dates in the two sources within ± 30 days and calculated measures of validity for adverse outcomes.

Results: Among 76 participants, we recorded 122 pregnancy outcomes in the prospective data and 104 outcomes in the FPH within ± 30 days of each woman's total observation period in the trial. Among 226 outcomes, we observed 65 live births that survived to 28 days, 25 stillbirths, and 32 live births followed by neonatal death in the prospective data and participants reported 63 live births that survived to 28 days, 15 stillbirths, and 26 live births followed by neonatal death in the pregnancy history survey. Sixty-two FPH outcomes were matched by date within ± 30 days to an outcome in prospective data. Stillbirth, neonatal death, higher parity, and delivery at a health facility were associated with likelihood of a non-matched pregnancy outcome.

Conclusions: Stillbirth and neonatal deaths were underestimated overall by the FPH, potentially underestimating the burden of mortality in this population. There is a need to develop tools to reduce or adjust for biases and errors in retrospective surveys to improve reporting of pregnancy and mortality outcomes.

Fig. 1

Participant flowchart and outcomes by data source

Fig. 2

Pregnancy outcomes from prospective data and pregnancy history survey. The graphs below display pregnancy outcomes from the prospective data and pregnancy history survey that occurred within the prospective follow-up period for each participant plus 30 days before and after their observation. Specifically, this period was defined as the time from 30 days before the participant’s enrollment date in the prospective study and 30 days after either the last birth visit or last pregnancy surveillance visit (whichever was later) in the prospective study. This yielded 226 pregnancy outcomes, including 122 from the prospective data and 104 from the pregnancy history survey. We display pregnancy outcomes that did not match by date with ± 30 days, including those inside and outside the prospective follow-up range defined above, in Additional file 1: Fig. S1. The 76 participants enrolled in the pregnancy history survey included women with at least one live birth that survived to 28 days (n = 26), stillbirth (n = 25), or live birth followed by neonatal death (n = 25), according to the prospective data, between January 4, 2015, and May 25, 2017

Fig. 2

Pregnancy outcomes from prospective data and pregnancy history survey. The graphs below display pregnancy outcomes from the prospective data and pregnancy history survey that occurred within the prospective follow-up period for each participant plus 30 days before and after their observation. Specifically, this period was defined as the time from 30 days before the participant’s enrollment date in the prospective study and 30 days after either the last birth visit or last pregnancy surveillance visit (whichever was later) in the prospective study. This yielded 226 pregnancy outcomes, including 122 from the prospective data and 104 from the pregnancy history survey. We display pregnancy outcomes that did not match by date with ± 30 days, including those inside and outside the prospective follow-up range defined above, in Additional file 1: Fig. S1. The 76 participants enrolled in the pregnancy history survey included women with at least one live birth that survived to 28 days (n = 26), stillbirth (n = 25), or live birth followed by neonatal death (n = 25), according to the prospective data, between January 4, 2015, and May 25, 2017