What question did this study set out to answer?

The aim is to analyze the spatiotemporal distribution of drug-resistant tuberculosis (DR-TB) in Indonesia and identify associated risk factors.

February 14, 2026Open Access

Spatiotemporal epidemiology, geographic hotspots, and risk factor associations of drug-resistant tuberculosis incidence in Indonesia: a Bayesian hierarchical modelling approach

Key Points

The aim is to analyze the spatiotemporal distribution of drug-resistant tuberculosis (DR-TB) in Indonesia and identify associated risk factors.
Nationwide retrospective analysis of DR-TB notification data from 2017 to 2022 across 514 districts.
Utilized multivariable Bayesian spatiotemporal regression models for analysis.
Identified hotspots and assessed associations with socioeconomic and health system factors.
DR-TB incidence is highest in individuals aged 25-54 years, primarily in urban areas.
Hotspots are located in urbanized regions such as Jabodetabek and Greater Surabaya.
First-line treatment success rates are protective against DR-TB incidence (IRR: 0.508).
Higher poverty rates and improved sanitation access correlate positively with DR-TB incidence (IRR: 1.028, 1.006).
Increased municipal human development index is linked to higher DR-TB rates (IRR: 1.068).

Abstract

Abstract Background Indonesia ranks among the countries with the highest burden of drug-resistant tuberculosis (DR-TB), contributing approximately 7.4% of global cases, many of which are likely underdiagnosed. To support targeted public health surveillance and control efforts, this study aimed to characterize the spatiotemporal distribution of DR-TB incidence in Indonesia, identify geographic hotspots, and examine associations with health system and socioeconomic factors. Methods We conducted a nationwide retrospective analysis using annual DR-TB notification data from 2017 to 2022 across all 514 districts, obtained from the national tuberculosis information system. Multivariable Bayesian spatiotemporal regression models were fitted under alternative likelihood assumptions and space-time random effect structures. Model selection criteria were used to identify the best-fitting models for hotspot detection and estimation of risk factor associations. Results DR-TB predominantly affected individuals aged 25–54 years, aligning with the working-age population. Hotspots were concentrated in urbanized regions, including the Jabodetabek megacity, Greater Surabaya, and districts in South Sumatra. The best-fitting model identified a protective association between first-line treatment success rates and DR-TB incidence incidence rate ratio (IRR): 0.508; 95% credible interval (CrI): 0.368–0.702. In contrast, DR-TB incidence was positively associated with the proportion of the population living below the poverty line (IRR: 1.028; 95% CrI: 1.013–1.044), households with improved sanitation access (IRR: 1.006; 95% CrI: 1.002–1.010), and increased municipal human development index (IRR: 1.068; 95% CrI: 1.049–1.094). Conclusions DR-TB hotspots were primarily concentrated in urban areas, highlighting the need for targeted interventions. Improving first-line tuberculosis treatment success rates and addressing socioeconomic drivers, such as poverty, are critical for controlling DR-TB. Public health policies should prioritize workplace-based support for improving treatment adherence, provide safeguards for TB patients affected by poverty, and underscore the importance of a multisectoral TB surveillance and control program. Graphical Abstract

KI fragen

Bookmark

View Full Paper