RESUMO
BACKGROUND: A twinning partnership is a formal and substantive collaboration between two districts to improve their performance in providing primary healthcare services. The 'win-win' twinning partnership pairs are categorized under relatively high and low-performing districts. The purpose of this formative evaluation is to use the empirically derived systems model as an analytical framework to systematically document the inputs, throughputs and outputs of the twinning partnership strategy. METHODS: This explanatory sequential mixed method study design was conducted from October 2018 to September 2019, in Amhara, Oromia, Southern, Nations, Nationalities and Peoples' (SNNP) and Tigray regions. The quantitative research approach used an observational design which consists of three measurements: at baseline (October 2018), midterm (March 2019) and end-line (September 2019), and the qualitative approach employed a case study. Qualitative data was collected using interviewer-guided semi-structured interview tools. The data were transcribed verbatim, translated from Amharic and Afan Oromo into English and analyzed through a theoretical framework named the Bergen Model of Collaborative Functioning (BMCF). Quantitative data were extracted from routine health management information system. The results are presented as averages, percentages and graphs. To claim statistical significance, non-parametric tests: Friedman test at (p < 0.05) and Wilcoxon signed ranks test (p < 0.017) were analyzed. RESULTS: The District Health System Performance (DHSP) was determined using data collected from eight districts. At baseline, the mean DHSP score was 50.97, at midterm, it was 60.3 and at end-line, it was 72.07. There was a strong degree and statistically significant relationship between baseline, midterm and end-line DHSP scores (r > 0.978**), using the Friedman test χ2(2) = 16.000, p = 0.001. Post hoc analysis using Wilcoxon signed-rank test was conducted with a Bonferroni correction and the results elicit higher DHSP values from baseline to midterm and from midterm to end-line with significance level set at p < 0.017. The qualitative results of the case study revealed that scanning the mission of the twinning partnership and focusing on a shared vision coupled with mobilizing internal and external resources were the fundamental input elements for successful twinning partnerships at the district level. In addition, the context of pursuing Universal Health Coverage (UHC) through achieving transformed districts can be enhanced through deploying skilled and knowledgeable leadership, defining clear roles and responsibilities for all stakeholders, forming agreed detailed action plans, and effective communication that leads to additive results and synergy. The twinning partnership implementing districts benefit from the formal relationship and accelerate their performances towards meeting the criteria of transformed districts in Ethiopia. CONCLUSIONS: Twinning partnerships help to accelerate the health system's performance in achieving the district transformation criteria. Therefore, scaling up the implementation of the twinning partnership strategy is recommended.