Modeling the incidence rate of hospital re-admissions within 30 days and associated factors among patients with advanced HIV disease in Uganda: comparison of the performance of binary logistic and count regression models

Advanced HIV Disease (AHD) remains a major driver of HIV-related morbidity, readmissions, and mortality, particularly in Sub-Saharan Africa. AHD affects People Living with HIV (PLHIV) with CD4 < 200 cells/mm³ and/or WHO stage 3 or 4, or all children under five years living with HIV. Globally, approximately 18% of AHD patients are readmitted within 30 days post-discharge; rates in Sub-Saharan Africa have been found to reach up to 52.9%. In Uganda, despite Uganda’s 2020 rollout of the AHD Package, data on 30-day readmissions among AHD patients remains limited. Understanding the incidence rate of 30-day readmissions is critical to improving quality of care and reducing system burden. The study used retrospective cohort study design using secondary data of AHD patients aged ≥ 12 years admitted at Kiruddu National Referral Hospital (KNRH) between March 2020 to July 2022. Patients discharged alive were included, excluding in-hospital deaths. Model performance for the binary logistic regression and count regression (Poisson and Negative binomial) models was compared using Bayesian Information Criteria (BIC). The final binary logistic regression model was built using backward stepwise selection with statistical significance determined at a 95% confidence level. Among 855 AHD patients stratified by ART categories (active, defaulted, naïve), the overall 30-day readmission incidence rate was 6.8 (95% CI: 6.7–6.9) per 100 months of observation. AHD patients aged 25–39, 40–49, and 50 + years, active on ART had lower odds of 30-day readmission compared to those aged 12–24 years, (aOR: 0.12, 95% CI: 0.03–0.43; aOR: 0.08, 95% CI: 0.02–0.37; aOR: 0.14, 95% CI: 0.03–0.63). Patients with candidiasis and/or ≥ 3 comorbidities were more likely to be readmitted (aOR: 3.65, 95% CI: 1.02–13.04; aOR: 7.02, 95% CI: 1.32–37.48), after adjusting for other factors. These associations remained consistent across the entire sample of AHD patients. Younger age (12–24 years), candidiasis, and ≥ 3 comorbidities were associated with higher risk of 30-day readmission, among AHD patients active on ART and the entire sample of AHD patients. Binary logistic regression outperformed count models.

Authors

Institutions

Publication Details

Journal
BMC Public Health
Published
2025-10-17
DOI
https://doi.org/10.1186/s12889-025-24791-w
Primary Topic
HIV/AIDS Impact and Responses
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Modeling the incidence rate of hospital re-admissions within 30 days and associated factors among patients with advanced HIV disease in Uganda: comparison of the performance of binary logistic and count regression models

Barbara Eva Kirunda, Mary Mbuliro, David Guwatudde, Nazarius Mbona Tumwesigye et al.
BMC Public Health
HIV/AIDS Impact and Responses
article

Modeling the incidence rate of hospital re-admissions within 30 days and associated factors among patients with advanced HIV disease in Uganda: comparison of the performance of binary logistic and count regression models

Barbara Eva Kirunda, Mary Mbuliro, David Guwatudde, Nazarius Mbona Tumwesigye, Douglas Joseph Musimbaggo
article en

Abstract

Advanced HIV Disease (AHD) remains a major driver of HIV-related morbidity, readmissions, and mortality, particularly in Sub-Saharan Africa. AHD affects People Living with HIV (PLHIV) with CD4 < 200 cells/mm³ and/or WHO stage 3 or 4, or all children under five years living with HIV. Globally, approximately 18% of AHD patients are readmitted within 30 days post-discharge; rates in Sub-Saharan Africa have been found to reach up to 52.9%. In Uganda, despite Uganda’s 2020 rollout of the AHD Package, data on 30-day readmissions among AHD patients remains limited. Understanding the incidence rate of 30-day readmissions is critical to improving quality of care and reducing system burden. The study used retrospective cohort study design using secondary data of AHD patients aged ≥ 12 years admitted at Kiruddu National Referral Hospital (KNRH) between March 2020 to July 2022. Patients discharged alive were included, excluding in-hospital deaths. Model performance for the binary logistic regression and count regression (Poisson and Negative binomial) models was compared using Bayesian Information Criteria (BIC). The final binary logistic regression model was built using backward stepwise selection with statistical significance determined at a 95% confidence level. Among 855 AHD patients stratified by ART categories (active, defaulted, naïve), the overall 30-day readmission incidence rate was 6.8 (95% CI: 6.7–6.9) per 100 months of observation. AHD patients aged 25–39, 40–49, and 50 + years, active on ART had lower odds of 30-day readmission compared to those aged 12–24 years, (aOR: 0.12, 95% CI: 0.03–0.43; aOR: 0.08, 95% CI: 0.02–0.37; aOR: 0.14, 95% CI: 0.03–0.63). Patients with candidiasis and/or ≥ 3 comorbidities were more likely to be readmitted (aOR: 3.65, 95% CI: 1.02–13.04; aOR: 7.02, 95% CI: 1.32–37.48), after adjusting for other factors. These associations remained consistent across the entire sample of AHD patients. Younger age (12–24 years), candidiasis, and ≥ 3 comorbidities were associated with higher risk of 30-day readmission, among AHD patients active on ART and the entire sample of AHD patients. Binary logistic regression outperformed count models.

BMC Public HealthVol. 25(1)
Makerere University (UG)
Openalex Percentile: Top 61%
HIV/AIDS Impact and Responses
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.