Investigators at Ayder Comprehensive Specialized Hospital reported a clear association between late-pregnancy food insecurity and reduced neonatal birthweight. The sample included 286 mother–newborn pairs recruited between December 28, 2024 and March 8, 2025. Birthweight was measured immediately after delivery using calibrated balanced scales. The study team found an overall low birth weight (LBW) prevalence of 25.5% in the cohort. Short, stark contrast: 48.6% of newborns from food-insecure mothers were LBW versus 11.3% among infants born to food-secure mothers. Longer analytic estimates quantified that difference: propensity score analyses produced an Average Treatment Effect (ATE) of –204.0 grams (95% CI: –289.3 to –118.7; p < 0.001). The Average Treatment Effect on the Treated (ATET) estimated a –218.3 g difference (95% CI: –411.2 to –25.3; p = 0.027). Those numbers represent average shifts in birthweight associated with household food insecurity measured in late pregnancy in this facility cohort.
The distribution of food insecurity severity was also reported. Sixty-one point nine percent of mothers were classified as food-secure; 38.1% were food-insecure, divided into mild (17.1%), moderate (14.0%), and severe (7.0%) categories using the Household Food Insecurity Access Scale (HFIAS). The HFIAS demonstrated excellent internal consistency in this dataset (Cronbach’s alpha = 0.910). The investigators used crude mean comparisons and more complex adjustments to explore dose-response and robustness; those parallel approaches consistently indicated lower birthweights among infants of food-insecure mothers.
The study was a facility-based cross-sectional analysis conducted at a referral hospital in Mekelle, Tigray Region. Mothers with twin pregnancies, severe fetal congenital malformations, or who were critically ill and uncooperative were excluded. Systematic random sampling recruited every third delivery until 286 participants were enrolled. Maternal sociodemographic details, obstetric history, anthropometrics including mid-upper arm circumference (MUAC), and clinical variables were collected by trained midwives through interviewer-administered questionnaires and record review. The outcome variable was the newborn’s first measured weight in grams.
Household food security in the prior 30 days was measured with the 9-item HFIAS, which classifies households into food secure, mildly, moderately, or severely food insecure based on item-specific scoring rules. The authors explicitly noted that measuring food security in the immediate postpartum period primarily reflects late third-trimester exposure and does not capture food security across the entire pregnancy.
To reduce observed confounding, investigators used propensity score matching (PSM) implemented in STATA 19. The logistic treatment model used a caliper of 0.1 and nearest-neighbor 1:1 matching. Standard errors were estimated with the Abadie-Imbens robust method. Covariates entered into the treatment model were chosen for prior evidence and biological plausibility and included gestational age (preterm vs term), maternal age categories, household income brackets, presence of obstetric complications, chronic medical illness, parity, abortion history, gestational weight gain (<10 kg vs ≥10 kg), antenatal care contacts (<4 vs ≥4), pregnancy intention (planned vs unplanned), and MUAC (<23 cm vs ≥23 cm). The matching procedure produced a balanced weighted sample; balance diagnostics were evaluated using standardized mean differences and variance ratios with pre-specified thresholds.
Additional analyses included subgroup, sensitivity, and robust multiple linear regression models to test effect stability and detect potential dose-response relationships. Investigators reported using Levene’s test for variance equality in mean comparisons, Chi-square testing for association with LBW, and standard diagnostic checks for multicollinearity and model fit. The dataset and STATA command files were made available as supplementary material.
The study team concluded that late-pregnancy household food insecurity was significantly associated with reduced neonatal birthweight and increased LBW prevalence in this hospital-based cohort. They recommended routine assessment of nutritional and food security status during antenatal care and suggested integrating social and behavioural nutrition interventions and nutrition education before and during pregnancy. Policy recommendations included ensuring food security and measures to empower women to improve nutritional conditions.
Limitations that the authors reported include the cross-sectional, facility-based design and the timing of exposure measurement: because household food security was assessed in the immediate postpartum period, the measure primarily reflects late third-trimester status and does not represent food security across the entire gestation. The study did not report longitudinal measures of maternal diet or specific micronutrient intakes, nor did it report community-level sampling outside the hospital setting. Investigators also noted there was no specific external funding for the work. Data files and analytic code were provided as supplements, and ethical clearance was obtained from the Institutional Research Ethics Review Committee of Mekelle University (MU-IRB 2419/2024).
Unanswered questions remain. The source did not report detailed timing or content of nutrition interventions, nor long-term infant growth outcomes following birth. Future work could evaluate whether addressing household food access earlier in pregnancy alters the magnitude of the birthweight effect observed here.