Predictors of nurse and midwife performance in a Ugandan district: a mixed-methods study of individual, organizational, and community factors

A well-performing health workforce, defined as one that is available, competent, productive, and responsive to patient needs, is central to achieving Sustainable Development Goals SDG 3.8 on Universal Health Coverage. In Uganda, district-level health performance reports have indicated persistent challenges. Addressing health workforce performance gaps is therefore a critical health systems issue for achieving district-level health targets. This study assessed factors affecting the performance of professional nurses and midwives defined as enrolled, registered, and degree-level practitioners licensed by the Uganda Nurses and Midwives Council, in Lira District, Northern Uganda.

A cross-sectional convergent parallel mixed-methods study was conducted from April 2017 to May 2018. A structured questionnaire was administered to 156 randomly selected nurses (n = 98) and midwives (n = 58) across all government (n = 24) and private-not-for-profit (n = 6) facilities. Performance was measured across four dimensions: competency, productivity, availability, and responsiveness. Principal Component Analysis (PCA) reduced independent variables. Linear regression identified predictors of performance. Qualitative data from 20 key informant interviews (KIIs) and three Focus Group Discussions (n = 30) were analyzed thematically.

The majority of respondents were female (83.3%), certificate holders (72.4%), and had 1–10 years of experience (54.5%). PCA yielded six components (C1-C6) explaining 85.9% of variance. Performance levels varied across dimensions: half of the respondents (50.0%) had competency scores in the 0–50% range, while productivity (60.3%), availability (51.3%), and responsiveness (75.0%) scores were predominantly in the 51–75% range. Linear regression identified distinct predictors for each dimension: Competency was significantly predicted by C1 (Poor Clinical Practice, β=-0.010, p = 0.012), C2 (Adherence to Systems, β = 0.066, p < 0.001), C3 (Organizational Commitment, β = 0.039, p < 0.001), and C4 (Participatory Environment, β = 0.080, p < 0.001). Productivity was predicted by C1 (β=-0.029, p < 0.001), C2 (β = 0.030, p = 0.006), C3 (β = 0.044, p = 0.005), and C4 (β = 0.131, p < 0.001). Availability was predicted by C2 (β = 0.090, p < 0.001), C4 (β = 0.131, p < 0.001), C5 (Unmet Training Needs, β=-0.060, p < 0.001), and C6 (Politicized Recruitment, β=-0.020, p = 0.004). Responsiveness was predicted by C2 (β = 0.070, p < 0.001), C4 (β = 0.139, p < 0.001), and C5 (β=-0.047, p < 0.001). Qualitative findings from 20 key informant interviews and 3 FGDs revealed key contextual factors: poor health-seeking behaviors (mentioned in 85% of FGDs), political interference in recruitment/promotion (reported by 70% of key informants), lack of community ownership of health facilities (discussed in 80% of FGDs), and negative staff attitudes (raised in 85% of FGDs).

Individual factors (clinical practices, motivation), organizational factors (working environment, systems adherence, resource availability), and political economy factors (political interference in recruitment/promotion) significantly predict nurse and midwife performance across all dimensions. Community-level factors (poor health-seeking behaviors, negative mutual perceptions) further undermine performance and service utilization. A multi-level intervention addressing individual capacity, organizational support, depoliticized human resource management, and community-health system linkages is urgently needed to improve workforce performance and service delivery. This requires coordinated action from national policymakers, district managers, facility leaders, and communities.