Survival Analysis of Tuberculosis Patients Among National Health Insurance Participants, 2015–2024

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Indonesia ranks second globally in TB incidence, with an estimated 1 million new cases annually. The majority of TB patient care costs in Indonesia are covered through the National Health Insurance program administered by BPJS Kesehatan. This research aims to analyze the time to death pattern of TB patients enrolled in JKN, identify the factors influencing it, and compare the performance of the Cox PH and AFT models. The data used are BPJS Kesehatan sample data from the 2015–2024 period, with an observation window from January 1, 2022 to December 31, 2024. The analysis was conducted on 111,946 patients, consisting of 10,083 patients (9.01%) who died and 101,863 patients (90.99%) who were censored. Analysis was carried out sequentially using the Kaplan-Meier method, the Cox PH Model, and the AFT Model with the Weibull distribution selected based on the best AIC, BIC, and Log-Likelihood values. Results showed a mean survival time of 879 days among TB patients. Age was the most dominant factor; patients above 60 years of age had a hazard ratio 8.09 times higher and a survival time 76% shorter compared to the 0–35 age group. Female patients had a survival time 28% longer than male patients. The PBI JK segment showed the highest mortality risk, while patients in Domicile 4 and 5 had shorter survival times compared to Domicile 1. The Schoenfeld residual test rejected the proportional hazards assumption, rendering the Weibull AFT model the most appropriate and best-fitting for this study.