# Mechanisms for stimulating facility quality improvement: A positive deviance study of Tanzania’s Star Rating Assessment

## TL;DR

This study explores how a national quality improvement program in Tanzania helped healthcare facilities improve their performance through a star rating assessment.

## Contribution

The study identifies mechanisms of improvement through a positive deviance approach in Tanzania's healthcare quality program.

## Key findings

- The Star Rating Assessment (SRA) helped providers find meaning in their clinical work, which was a core mechanism for improvement.
- Stronger relationships and connections between health system actors supported the improvement process.
- Government ownership and implementation were critical to the program's success in Tanzania.

## Abstract

Tanzania launched the nationwide Star Rating Assessment in 2015 and has implemented it as a cornerstone of its national quality improvement program. We use variation in assessment results to explore mechanisms through which facilities improve quality. A mixed-methods positive deviance approach was applied by ranking all primary care facilities included in the Star Rating Assessment (n = 5,595) by their change in score between 2015/16 and 2017/18. The least- and most-improved facilities (n = 27) were selected for qualitative interviews with the highest-ranking provider at the facility on the day of data collection. The dataset was thematically analyzed first as a full-set and then divided by most and least improved to develop a local theory of improvement. Interviews were conducted with 27 facility leaders in 27 primary care facilities. Analysis showed that the SRA helped providers and staff find meaning in their clinical work which served as a core mechanism for improvement. This meaning-making was supported by stronger relationships and connections between health system actors. Connection to leaders came through the creation of a shared national mission to improve quality and an accountability system that allowed respondents to advocate for change. Connections to colleagues came through greater collaboration within facilities and competition between facilities. Connection to patients came through improved knowledge about clinical and interpersonal quality and self-checking against this new knowledge. Inputs to care were important, but not sufficient to drive improvement. Distilling SRA into a replicable improvement model carries with it the risk of losing contextual and programmatic factors which appear to be key to the success of the program in Tanzania. Implementation and ownership by the government of Tanzania was one of these critical elements.