Developing a Lean Service Quality Improvement Model to Enhance Medical Tourism in Healthcare Sector: A Case of Ethiopia
No Thumbnail Available
Date
2025-06
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Ababa University
Abstract
Healthcare is a sector with unique features where defects and errors cannot be tolerated. Its service delivery is one of the principals and most complex systems on the globe due to rapidly growing pressure, waiting time, an aging population, increasing patients flow, limited resources and competing social needs to sustain life. Thus, the sector has turned its attention to a lean approach due to a growing influence in reducing waiting time to enhance service quality and increasing flexibility. The researcher begins the investigation by discussing the current healthcare challenges long waiting time (95 minutes in averages), long length of stay (15 days in average), lack of access to some services in the country, and driving factors that have contributed to the improvement of service quality, particularly at St. Paul’s Hospital Millennium Medical College. Hence, the study's primary goal is to develop lean quality service improvement model aimed at enhancing the patient satisfaction in the Ethiopian healthcare sector, specifically within the case healthcare, thereby increasing medical tourism. The study utilized various databases to conduct a comprehensive assessment of the literature and enhance service quality.
A thorough review of the literature was conducted to investigate the concepts, practices, and challenges related to healthcare service quality improvement. The review process identified literature gaps, including the dimensions of healthcare service quality improvement, service quality and lean thinking integration, the absence of service design thinking in service quality improvement, and their impact on patient satisfaction and medical tourism, which prior work has not sufficiently addressed. The study adopted a mixed-methods approach, incorporating both primary and secondary data gathering methods. The Define, Measure, Analyse, Improve, and Control (DMAIC) technique was also used. To quantify service quality gaps, patient expectations and perceptions were gathered during the Define Phase. During the Measure Phase, the high weighted scores obtained from patient input were evaluated using Quality Function Deployment (QFD). In the Analyse Phase, a cause-and-effect diagram was used to determine the underlying reasons behind these high results.
In the Improve Phase, inputs, resources, and methods were considered major causes, along with possible solutions. In the Control Phase, machine learning tools such as random forest, neural networks (NN), and support vector machines (SVM) were employed to predict healthcare patient waiting times, ensuring sustained service quality and patient satisfaction. Performance comparison metrics, including Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and coefficient of determination (R²), were assessed to identify the predictive model's accuracy.
The study employs a dual approach of Structural Equation Modelling (SEM) and Artificial Neural Networks (ANN) to develop a model that identifies dimensions significantly impacting patient satisfaction and medical tourism. Additionally, sensitivity analysis was used to rank these dimensions, providing better insights and alternatives. A total of 225 patient data points were collected from respondents through a questionnaire to develop a lean service quality framework using SPSS, AMOS, and Artificial Neural Network (ANN). Among the three models for predicting waiting time, the support vector machine model demonstrated better prediction accuracy compared to the neural network and random forest models when assessing actual data. The support vector machine effectively mimics waiting time, significantly improving service quality. Based on the developed model, five components with significant factor loadings exceeding 0.50 have been identified. These components are service quality, lean thinking, lean service quality, patient satisfaction, and medical tourism considerations. The results were further analyzed to assess the model's fit. A decent model fit is implied by the RMSEA value of 0.05, which is below the permissible limit of 0.08. At 0.82, the Adjusted Goodness of Fit Index (AGFI) score is nearly at its suggested level. Tucker-Lewis Index, or TLI, is 0.96; the Normed-Fit Index (NFI) is 0.92; and the Comparative Fit Index (CFI) is 0.97. Since the NFI, CFI, TLI, and IFI values are higher than the suggested value of 0.9, they indicate strong model fits.
The structural equation modelling is used as the input unit of the artificial neural network model to detect both nonlinear and linear relationships without robust speculative or theoretical bases. Artificial neural networks identify all linear, nonlinear, and non-compensatory relationships by avoiding assumptions in distribution and model development. The RMSE value of the artificial neural network model (0.88) indicates good predictive accuracy for the lean service quality improvement model. Regarding the dimension of healthcare lean service quality improvement, there is no weak effect in the verified model of the artificial neural network. The normalized importance levels for all dimensions are greater than 80%, which is very high. It showed a significant relationship with both structural equation modelling and artificial neural network model analysis. Hence, the dimensions in model 1, 2, and 3 indicated a significant relationship with each other.The study proposed the integration of lean thinking and service quality due to lack of available literature and the need for additional research on lean service quality adoption gaps, waiting time prediction, and improvement opportunities in the healthcare sector. The study is also original in its focus on developing a lean service quality model using a dual approach within the context of the healthcare sector. It contributes to healthcare lean service quality improvement based on quality function deployment, Six Sigma, and an artificial intelligence approach. It also ranked the importance levels of the lean service quality improvement dimensions through artificial neural network sensitivity analysis.
Description
Keywords
Healthcare service quality, lean thinking, medical tourism, service quality, six-sigma, DMAIC, QFD, SEM, Machine learning