Inequalities in tuberculosis control in Ethiopia: A district-level distributional modelling analysis.

Background: Implementation studies indicate that the addition of tuberculosis diagnosis and treatment services into the community health extension workers' tasks-that is 'task-shifting'-improved case detection and treatment outcomes in Ethiopia. Given resource and operational constraints, only a limited number of areas can be targeted by an expanded task-shifting program. Therefore, we mapped the distributional disparities in tuberculosis services across regions and districts and modelled the equity pathways towards optimising national scale-up of this task-shifting intervention in Ethiopia.

Methods: We used data from various sources including District Health Information Software 2; demographic, geospatial and topographic data; and previously published implementation study findings. We developed methods to integrate these datasets and to calculate the proportion of health facilities with tuberculosis services, the district population to health centre ratio, and the proportion of district population living within 2h walking distance from a health centre. Equity and disparities were then measured in terms of: tuberculosis services coverage; health centre adequacy, that is the district population served by health centres; and spatial access adequacy, that is the district population with health centre access within a two-hour walking distance. Subsequently, districts were ranked according to these measures to allow prioritisation of the health extension worker task-shifting intervention.

Results: Tuberculosis services coverage varied from 54% in Afar region to 100% in Harari region, and health centre inadequacy ranged from 10% of districts in Benishangul-Gumuz to 87% in Sidama. After spatial access adjustment, health centre inadequacy ranged from 7% of districts in Sidama to 91% in Somali; and tuberculosis services inadequacy from 7% of districts in Sidama to 97% in Afar. Task-shifting implemented in inadequate districts (55% of all districts) could raise national case detection rate from 66% (currently) to 88% and treatment success rate from 93% to 99%; Benishangul-Gumuz achieving the largest increase of all regions.

Conclusions: Access to effective tuberculosis services presents substantial disparities across districts in Ethiopia, due to both health system and tuberculosis-specific factors. Jointly considering both types of factors would enable prioritisation of districts where health extension workers would be most impactful.