The effect of landscape fires on child hospital visits and admissions: a time-series study in southern Mozambique

BACKGROUND AND AIM Epidemiological evidence linking exposure to landscape fires with child health remains limited. We assessed the association between day-to-day variation in landscape fires and child hospital visits and admissions in Manhiça district, Mozambique, an area affected by forest and cropland fires. METHODS We conducted a time-series analysis (2012-2020) using daily satellite-derived fire data (VIIRS 375m) and number of hospital visits and admissions from an ongoing pediatric morbidity surveillance system (children aged ≤ 15 years). Fires were extracted from the study area and a 100km surrounding buffer. We applied quasi-Poisson regression models controlling for temperature, day of the week, season and long-term trend, and offsetting by annual population-time at risk to examine associations between fires (lags 0-4) and hospital visits and admissions. RESULTS At least 114,443 fires (mean: 44/day; median: 13/day) and a total of 507,034 child hospital visits (mean: 154/day; 39.5% respiratory-linked) and 8,981 admissions (mean: 2.7/day; 26.7% respiratory-linked) occurred in 2012-2020. Almost half of the fires were forest fires. Fires explained 39% of temporal variability in PM2.5 from a local monitoring campaign. Comparing two days with identical values in the adjustment variables but differing by 40 fires (interquartile range) resulted in an increase of all-cause (0.95%; 95%CI: 0.46, 1.44) and respiratory-linked hospital visits (1.14%; 0.57, 1.69) the following day (lag 1). Significant associations were observed for other lags in fires including: lag 0-1 with all-cause (0.47%; 0.17, 0.77) and respiratory-linked visits (0.64%; 0.29, 0.99) and lag 2 with respiratory-linked visits (0.47%; 0.17, 0.77). We did not observe associations for hospital admissions. CONCLUSIONS Landscape fires were associated with all-cause and respiratory-linked hospital visits in children. Improved exposure assessment is needed to better quantify the contribution of air pollution due to landscape fires on child health in regions with limited air pollution monitoring. KEYWORDS Landscape fires, child health, Mozambique