Efficiency Determinants in Healthcare: A Systematic Review With an Integrated Canonical Correlation Analysis–Data Envelopment Analysis/Assurance Region Model

Abstract

Goal: Instead of considering many variables for the accurate measurement of healthcare efficiency, working with the select few variables that really affect efficiency will provide more accurate efficiency scores. In addition, calculating the efficiency by weighting the inputs and outputs according to their effect and severity levels will give more realistic results. In this article, a three-step hybrid system with a two-stage CCA (canonical correlation analysis)–DEA/AR (data envelopment analysis/assurance region) model is proposed to obtain results of health efficiency. Methods: Healthcare efficiency studies conducted between 2000 and 2020 were reviewed. In this examination of the input and output variables used in the DEA of 63 previous studies, the 6 inputs and 5 outputs preferred by previous researchers were determined. Afterward, the health efficiency scores of countries represented in the research were calculated with weight-restricted DEA, and CCA was used for a priori statistical analysis in determining the weights. Thus, in this analysis of the preferred outputs and inputs with the help of CCA to estimate the relationship between multiple input and output sets, the variables that had no effect were eliminated and the ones that had an effect were included in DEA/AR with their degree of effectiveness. Principal Findings: For the model proposed here, three inputs and three outputs were identified by following a five-item variable reduction procedure. The numbers of doctors and nurses were identified as the most effective inputs, and infant mortality rates were found to be the most effective outputs. Therefore, health efficiency scores obtained with the proposed CCA–DEA/AR model and the basic DEA are presented together. A review of the results found fewer health-efficient countries with the weight-restricted DEA. This is proof that weighting the variables into the DEA increases the discriminating power of the method. Practical Applications: By applying the proposed model, healthcare administrators can analyze healthcare efficiency accurately and thus improve efficiency by transferring limited resources to the right places according to deficiencies or surpluses identified by the model's inputs. Resources can be allocated at both private and public hospitals in a way that increases healthcare efficiency outputs.