Assessment of the performance of primary healthcare in eight African countries: based on the Composite Fuzzy Index.
African countries face a dual burden of infectious and non-communicable diseases, placing substantial pressure on primary healthcare (PHC) systems. Existing indicator-weighting approaches are limited in addressing redundancy and non-linear relationships. In addition, conventional composite indices often rely on linear aggregation assumptions that may not adequately reflect complex relationships among PHC performance indicators. This study proposes a novel weighting method and composite index to improve facility-level assessment of PHC performance.
We analysed nationally representative healthcare facility data from Service Provision Assessments conducted in eight African countries (Ethiopia, the Democratic Republic of the Congo, Tanzania, Malawi, Kenya, Namibia, Rwanda and Uganda). Based on the Primary Health Care Performance Initiative framework, we constructed an indicator system comprising 34 indicators across five subdomains within the Service Delivery domain. Two composite indices were calculated: the Arithmetic Mean Index (AMI) and the Composite Fuzzy Index (CFI). AMI was derived using a simple arithmetic mean, while CFI employed a distance-based aggregation approach with indicator weights estimated using Multivariate Adaptive Regression Splines. The associations between the composite indices and patient-reported satisfaction were examined, with patient satisfaction calculated as a weighted average across surveyed patients.
A total of 6921 facilities were included. Overall performance across indicators was low, with particularly poor scores observed in community engagement within the population health management subdomain (mean=0.08). AMI ranged from 0.238 to 0.806, and CFI ranged from 0.237 to 0.705. Facility-level CFI scores showed a positive association with patient satisfaction (Spearman's ρ=0.039, p<0.01; Kendall's τ=0.027, p<0.01), whereas AMI showed no significant association.
PHC performance in African countries remains suboptimal across multiple dimensions. The proposed weighting and aggregation approach better captures inter-relationships among indicators and provides a more plausible assessment of PHC performance than traditional composite indices.
We analysed nationally representative healthcare facility data from Service Provision Assessments conducted in eight African countries (Ethiopia, the Democratic Republic of the Congo, Tanzania, Malawi, Kenya, Namibia, Rwanda and Uganda). Based on the Primary Health Care Performance Initiative framework, we constructed an indicator system comprising 34 indicators across five subdomains within the Service Delivery domain. Two composite indices were calculated: the Arithmetic Mean Index (AMI) and the Composite Fuzzy Index (CFI). AMI was derived using a simple arithmetic mean, while CFI employed a distance-based aggregation approach with indicator weights estimated using Multivariate Adaptive Regression Splines. The associations between the composite indices and patient-reported satisfaction were examined, with patient satisfaction calculated as a weighted average across surveyed patients.
A total of 6921 facilities were included. Overall performance across indicators was low, with particularly poor scores observed in community engagement within the population health management subdomain (mean=0.08). AMI ranged from 0.238 to 0.806, and CFI ranged from 0.237 to 0.705. Facility-level CFI scores showed a positive association with patient satisfaction (Spearman's ρ=0.039, p<0.01; Kendall's τ=0.027, p<0.01), whereas AMI showed no significant association.
PHC performance in African countries remains suboptimal across multiple dimensions. The proposed weighting and aggregation approach better captures inter-relationships among indicators and provides a more plausible assessment of PHC performance than traditional composite indices.
Authors
Hu Hu, Zaid Zaid, Zhao Zhao, Tshimanga Tshimanga, Frumence Frumence, Nkhoma Nkhoma, Huang Huang
View on Pubmed