Is household WASH access score associated with child developmental z-scores at 24 months among children born to undernourished versus adequately nourished mothers, and what is the magnitude of the indirect association of maternal undernutrition with child developmental z-scores through WASH access, in a prospective cohort of mother-child dyads followed from pregnancy through 24 months postpartum?

Pregnant women aged 18–49 years in their second trimester (gestational age 14–27 weeks confirmed by last menstrual period or ultrasound) attending antenatal care at a selected public health facility, and their subsequent live-born singleton children followed to 24 months. Exclusion criteria: multiple pregnancy, severe obstetric complications requiring specialist referral, known fetal anomaly, or plans to relocate outside the study area within 24 months.

A single public hospital or health centre catchment area in a peri-urban or rural setting with documented variability in household WASH access and moderate prevalence of maternal undernutrition, reflecting conditions described in [[4]] and [[1]].

Systematic sampling of eligible pregnant women at antenatal care: every kth woman presenting at ANC on enrolment days, where k is determined by the expected ANC volume and target enrolment to achieve the required sample over the 9-month enrolment window. This removes the self-selection bias of convenience sampling while remaining operationally feasible within a single facility. Women are enrolled regardless of WASH status; WASH variability is expected from the catchment population and is not used as a selection criterion, avoiding post-treatment selection on the mediator.

Sample size for the b-path of the mediation model (WASH score → child developmental z-score, adjusting for maternal undernutrition and covariates), which is the most demanding path. Using the formula for multiple linear regression n ≥ (z_{1-α/2} + z_{1-β})² / f² + (k+1), where f² = 0.02 (small-to-medium effect per Cohen), α = 0.05 (two-sided), power = 0.80, and k = 10 covariates (maternal undernutrition, WASH score, maternal age, education, parity, household wealth, dietary diversity, ANC visits, child sex, birth weight): f² = 0.02 gives (1.96 + 0.84)² / 0.02 = 392, plus 11 = 403 mother-child dyads needed at the 24-month outcome. Adjusting for 25% attrition over 24 months (consistent with birth-cohort retention challenges noted in [[2]]): enrol 403 / 0.75 ≈ 537 pregnant women. A sensitivity check for the indirect effect (a×b) using the Monte Carlo method with a-path ≈ 0.25, b-path ≈ 0.20, and residual correlation r ≈ 0.30 confirms ≥80% power at n ≈ 250 complete cases, well below the 403 required for the b-path, so the b-path drives the sample.

Primary prospective data collection: (1) structured interviewer-administered questionnaire at antenatal enrolment (maternal anthropometry, WASH assessment, socio-demographics, dietary diversity, PHQ-9, HFIAS); (2) household WASH observation checklist at enrolment and 6 months postpartum; (3) follow-up assessments at child ages 6, 15, and 24 months for developmental testing by trained assessors (per [[2]]); (4) maternal and child anthropometry at each visit; (5) child morbidity recall (2-week recall) at each follow-up; (6) birth records for birth weight and gestational age.

30–36 months (3–4 months setup and ethics, 9 months enrolment, 24 months follow-up, 3–4 months analysis).

Moderate-to-high: ~USD 50,000–80,000 for personnel (trained developmental assessors, fieldworkers, data manager), anthropometric equipment, WASH observation tools, developmental assessment kits, transport for follow-up visits, data tablets, participant retention incentives, and PHQ-9/HFIAS instruments. Requires external grant funding.

Full IRB approval required from the host institution and relevant national ethics committee. Written informed consent from pregnant women at enrolment, with re-consent at major follow-up milestones. Special attention to consent for child developmental testing. Maternal depression screening (PHQ-9) requires a referral pathway for women scoring above clinical thresholds. Data safety monitoring for adverse events during follow-up. Minimal risk but longitudinal contact requires clear withdrawal protocols.