Predicting In-hospital Mortality of Pediatric Surgical Admissions in Central Malawi

Kamuzu Central Hospital (KCH), a tertiary care facility, serves as 1 of 2 pediatric surgical centers in Malawi, a resource limited country in southeastern Africa. We sought to understand the impact of clinical and access-related factors on in-hospital mortality at KCH.

Pediatric surgical admissions (<15 years) to KCH from April 1-December 1, 2023, were retrospectively analyzed. Patients were stratified as neonatal (admitted in the first 30 days of life) and non-neonatal. ArcGIS was used to calculate approximate distances traveled to care and to create district-level choropleth maps. Multivariate logistic regression analyses were conducted to identify relevant patient and pre-admission characteristics predictive of in-hospital mortality.

Of 1044 pediatric surgical admissions, 211 (20.2 %) were neonatal, and 833 (79.8 %) were non-neonatal. In-hospital mortality was 23.8 % (n = 50) for neonatal admissions and 2 % (n = 17) for non-neonatal admissions. Increasing distance from home to KCH (OR = 0.22, p = 0.01) and increasing age (OR = 0.93, p < 0.01) conferred a protective effect on in-hospital mortality for neonatal admissions. For non-neonatal patients, readmission (OR = 4.69, p = 0.02) and high level of care on admission (OR = 14.79, p < 0.001) increased the odds of mortality. A distance effect was not seen in the non-neonatal population.

There is a 12-fold increase of in-hospital mortality among admitted neonates with surgical conditions compared to older children. A protective effect of distance in this group may represent the natural selection of healthier neonates being able to travel from farther distances successfully. Older children who were previously admitted or required higher levels of care were at increased risk of death.