Assessing the cost-effectiveness of facility-based emergency care in low resource settings

Werner, Kalin

Assessing the cost-effectiveness of facility-based emergency care in low resource settings

Doctoral Thesis

2021

Abstract

Background Emergency conditions comprise nearly half of the total global burden of disease, and disproportionately affect low-resource settings (LRS). This burden of life-threatening yet treatable conditions can be ameliorated by effective, timely emergency care (EC) interventions, so significantly that the Disease Control Priorities project estimates over half of deaths in the lowest-income countries could be addressed though the implementation of effective EC. Interest in developing better facility-based EC is expanding rapidly, yet there is a large gap in the cost-effectiveness literature to support informed resource allocation. Distinguishing the "value for money" of EC is crucial, especially in contexts of extreme resources limitations. Developing robust and setting-specific data on the cost of implementing EC cultivates the ability to understand the impacts of, and plan improvements in, EC in LRS. The aim of this thesis was to investigate the cost-effectiveness of interventions forming a systematic approach to EC in health facilities in LRS. Aims and Objectives The primary aim of this research was to investigate the cost-effectiveness of implementing the WHO emergency care toolkit to reduce mortality related to emergency conditions in health facilities in an LRS. To achieve this aim, the following objectives were established: synthesise evidence relating to the costeffectiveness of EC in LRS, enumerate context specific costs of delivering facility-based EC, and retrospectively study the impact of implementing a low-cost set of EC interventions in low-resource EC naïve health facilities on cost and outcome (mortality), to derive a measure of cost effectiveness. Methods The dissertation is comprised of 3 studies. First, Chapter 3 undertakes a systematic review of literature on EC interventions in LRS, using PRISMA guidelines and the Consolidated Health Economics Evaluation Reporting Standards (CHEERS) checklist. Secondly, to enumerate context specific costs of delivering facility-based EC, data were collected over a 4-week period in Uganda using direct activity-based costing and presented in Chapter 6. Measures of central tendency were derived by condition and by intervention. Variations in cost between conditions were explored using a Kruskal-Wallis H test and a post-hoc Nemenyi test were performed to make pairwise comparisons between conditions. Third, in Chapter 7 a cost-effectiveness analysis model was developed using Microsoft Excel to calculate both the costs and effects of scalable investments strengthening facility-based EC on morbidity and mortality. Costs and consequences associated with piloting the WHO Emergency Care Toolkit package of interventions in Uganda were analysed using the decision tree model. Pre- and post-intervention groups were compared from a societal perspective. Cost and health outcomes were discounted using a microsimulation and parameter uncertainty assessed using Monte-Carlo simulation probabilistic sensitivity analyses. Results 35 studies were included in the final review; most were methodologically weak and focused on singleintervention analyses rather than intervention packages or system changes. This body of literature represented only 24 of 137 low- and middle-income countries (LMICs) globally, and was heterogeneous in methods, settings, and presentation of results of the identified studies. Accordingly, formulating a general conclusion about the wider implication of the findings on the cost–effectiveness of EC is problematic. The overall median (IQR) cost of care across all conditions was $15.53 (14.44 to 19.22). A Krauskal-Wallis test yielded statistically significant difference in cost values between sentinel conditions (H=94.89, p=1.20E-19). At a P value of < .05, the post-hoc Nemenyi test revealed paediatric diarrhoea has a statistically significant lower median cost compared to all other conditions, but did not yield any significant differences in median cost between the remaining four sentinel conditions. In running the decision tree model with a 1753 patient cohort, sampled 10000 times, the intervention averted 509 DALYs over standard care. The model found implementing the WHO Toolkit saved $664,231 ($658,552 to $669,910), and yielded an additional 27 lives saved, or an additional 1,826 life years. Conclusions and relevance This dissertation makes important conceptual, analytical and empirical contributions in exploring the application of local economic evidence-informed priority setting to ensure that decisions made around EC are guided by the populations they serve. In conducting one of the first cost-effectiveness analyses of investments that create a systematic approach to facility-based EC, we found that this is a very low-cost, high-yield intervention. In many cases it may not only be cost-effective, but actually cost saving. This finding is especially relevant in LRS contexts where associated additional costs may be considered affordable given the high burden of emergency conditions.

Keywords

Emergency Medicine

Reference:

Collections

PhD / Doctoral

Full item page