← Back to Research Papers

Why are the implementation of and access to child developmental assessment within routine healthcare so difficult in Ethiopia? A descriptive qualitative study.

Authors: Gebeyehu S, Mommers M, Hussen S, Spigt M
Journal: PloS one
mental health psychology open access

Abstract

Data Envelopment Analysis (DEA), originally proposed by Charnes et al. [], has become one of the most widely used nonparametric methods for evaluating the relative performance of decision-making units (DMUs) operating with multiple inputs and outputs. Owing to its ability to endogenously determine input and output weights without imposing a priori price information or functional assumptions, DEA has been extensively applied in banking, healthcare, education, transportation, energy, and public-sector performance evaluation. A distinctive feature of DEA is that efficiency scores are generated through endogenous shadow-price systems. Consequently, DEA efficiency is fundamentally different from conventional productivity indicators based on externally specified aggregation rules. The estimated efficiency of a DMU depends not only on its observed input-output bundle but also on the production technology used to construct the production possibility set and the endogenous valuation system implied by the optimization model. Alterations in returns-to-scale assumptions, orientation specifications, weight restrictions, or secondary-goal formulations may substantially change efficiency estimates even when the underlying production data remain unchanged. DEA efficiency should therefore be viewed as both technology-dependent and weight-dependent. Although this dependence has long been recognized in the DEA literature, relatively little is known about how technology and endogenous valuation systems jointly contribute to observed self-efficiency differences and, more importantly, how their effects can be quantified. Existing studies generally regard model dependence and weight flexibility as sources of sensitivity, robustness concerns, or ranking instability. Consequently, DEA self-efficiency is usually interpreted as a single composite measure, whereas the informational mechanisms responsible for its formation remain largely hidden.