Provider-identified barriers to performance at seven Nigerian accident & emergency units: A cross-sectional study

Nigeria hosts much of Africa's morbidity and mortality from emergency medical conditions. We surveyed providers at seven Nigerian Accident & Emergency (A&E) units about (i) their unit's ability to manage six major types of emergency medical condition (sentinel conditions) and (ii) barriers to performing key functions (signal functions) to manage sentinel conditions. Here, we present our analysis of provider-reported barriers to signal function performance.

503 Health Providers at 7 A&E units, across 7 states, were surveyed using a modified African Federation of Emergency Medicine (AFEM) Emergency Care Assessment Tool (ECAT). Providers indicating suboptimal performance ascribed this performance to any of eight multiple-choice barriers [infrastructural issues, absent and broken equipment, inadequate training, inadequate personnel, requirement of out-of-pocket payment, non-indication of that signal function for the sentinel condition, and hospital-specific policies barring signal function performance] or an open-ended "other" response. The average number of endorsements for each barrier was obtained for each sentinel condition. Differences in barrier endorsement were compared across site, barrier type and sentinel condition using a three-way ANOVA test. Open-ended responses were evaluated using inductive thematic analysis. Sentinel conditions were Shock, Respiratory Failure, Altered Mental Status, Pain, Trauma, and Maternal & Child Health. Study sites were the University of Calabar Teaching Hospital, the Lagos University Teaching Hospital, the Federal Medical Center, Katsina, the National Hospital Abuja, the Federal Teaching Hospital Gombe, the University of Ilorin Teaching Hospital (Kwara), and the Federal Medical Center Owerri (Imo).

Barrier distribution varied widely by study site. Just three study sites shared any one barrier to signal function performance as their most common. The two barriers most commonly endorsed were (i) non-indication of, and (ii) insufficient infrastructure to perform signal functions. A three-way ANOVA test found significant differences in barrier endorsement by barrier type, study site and sentinel condition (p<0.05). Thematic analysis of open-ended responses highlighted (i) considerations disfavoring signal function performance and (ii) lack of experience with signal functions as barriers to signal function performance. Interrater reliability, calculated using Fleiss' Kappa, was found to be 0.5 across 11 initial codes and 0.51 for our two final themes.

Provider perspective varied with regards to barriers to care. Despite these differences, the trends seen for infrastructure reflect the importance of sustained investment in Nigerian health infrastructure. The high level of endorsement seen for the non-indication barrier may signal need for better ECAT adaptation for local practice & education, and for improved Nigerian emergency medical education and training. A low endorsement was seen for patient-facing costs, despite the high burden of Nigerian private expenditure on healthcare, indicating limited representation of patient-facing barriers. Analysis of open-ended responses was limited by the brevity and ambiguity of these responses on the ECAT. Further investigation is needed towards better representation of patient-facing barriers and qualitative approaches to evaluating Nigerian emergency care provision.