A Framework for Scheduling and Optimization of Healthcare Resources

Abstract

Healthcare is known as the coordinated delivery of medical services to individuals and populations to ascertain their well-being. Healthcare resources are considered essential assets for superior functional outcomes of a healthcare system. These resources help healthcare organizations in decision-making and to improve the healthcare system to achieve efficiency, equity, quality, and patients' trust. Planning of healthcare resources for attaining sustainable and improved patient’s care is a major challenge for healthcare administrators. Proper planning for optimum utilization of resources ultimately promotes quality of services and reflects the healthcare provider's performance and commitment towards patient care. Growing population is also contributing in the increased demand of timely medical services. Since healthcare resources are inherently limited, varying health needs generate disparities between increasing demand and the healthcare delivery process, thereby affecting the utilization and efficiency of health systems' resources and sustainability. Ultimately both health outcomes quality and healthcare performance suffer a lot. Hence, there is a requirement of methods and systems those can help in managing and utilizing healthcare resources to enhance patient care and overall performance. Scheduling and optimization of healthcare resources is an active research area now a days. Researchers working in the field of Scheduling and optimization of healthcare resources study the needs of patients and community, thereby explore how decisions should be made so that the demands of medical services can be aligned. Analysis of healthcare data is carried out to gain insights and making decisions for the overall management of resources. Researchers have applied various soft computing and conventional computation methods for xiv resource-related decision-making. In such data-driven analysis, evolutionary computing methods add a firm hand. In this thesis, we have focused on scheduling and optimization of healthcare resources such as waiting list management, classification of patients and selection of healthcare personnel using soft computing approaches. The overall objective of this research is to develop a framework containing modules for surgical waiting list prioritization, surgical patients clustering, and surgical team selection. In this research, a decision-making model PSWL-CCI is proposed to prioritize patients on the surgical waiting list. The proposed model addresses two critical issues: First, to prioritize patients from the surgical waiting list. Second, to refine and optimize the cosine consistency index (CCI) of inconsistent pair-wise comparison matrix (PCM) and obtain consistent priorities. The cosine maximization method (CM) with the analytic hierarchy process (AHP) is used to compute the priority of patients from the surgical waiting list, and a hybrid algorithm, HMWCA (Hybrid modified water cycle algorithm), is proposed to improve and optimize the cosine consistency index (CCI) of inconsistent pair-wise comparison matrix (PCM). The proposed hybrid algorithm exploits the features of three traditional algorithms, namely the evaporation-based water cycle algorithm (ER-WCA), genetic algorithm, and 2-opt heuristic algorithm. In the proposed algorithm (HMWCA), the concept of salt concentration and absorption is applied to the evaporation rate of ER-WCA that improves the modified water cycle algorithm (MWCA). The performance of the proposed algorithm is tested on different inconsistent PCMs and compared with existing algorithms. The optimized CCI values obtained by the proposed algorithm are validated through paired sample t-test also. Finally, the proposed model is validated through a case study of a real patient dataset xv from an orthopedic surgery department of a multispecialty hospital in India. The proposed model is compared with existing prioritization methods. The experimental results reveal that the proposed model and associated algorithm significantly improve the CCI values and generate optimum priorities for the patients of the surgical waiting list. Next, focusing on hospital’s existing surgical record management procedure (SRMP), we have proposed an efficient clustering algorithm to arrange patients in the optimal number of distinct groups on the basis of similarity of characteristics. The proposed clustering algorithm is based on population-based metaheuristic artificial electric field algorithm (AEFA) and is designed to deal with the mixed dataset. The proposed algorithm utilizes real-encoded variable-length cluster representation scheme to illustrate the candidate solution, which enables the algorithm to find optimal number of clusters automatically. The concepts of threshold setting and cut-off ratio are used in the optimization process to further refine the clusters. In the proposed algorithm, similarity among data points and different cluster centers is measured using Euclidean distance (for numeric attributes) and probability of co-occurrence (for categorical attributes). The performance of the proposed algorithm is tested using real-life datasets and compared with existing mixed data clustering algorithms on the basis of two parameters: average accuracy and standard deviation. The proposed algorithm is validated using an unpaired t-test also. Finally, the proposed algorithm is validated using a case study of real postoperative surgical mixed data set obtained from the surgical department of a multispecialty hospital in India. It is observed from the results that the proposed clustering algorithm arrange the patients into optimal subgroups efficiently Finally, a decision-making model is proposed to select an optimal list of surgical teams. The proposed model addresses two critical issues: First, to improve the existing surgical history management system of a multi-specialty hospital (MSH), and second, to select an optimal list of surgical teams for a new surgical patient. Therefore, two modules: the surgical history management (SHM) module and the surgical team selection (STS) module, are introduced in the proposed model. The SHM module aims to arrange surgical patients into optimal sub-groups on the basis of similar characteristics. It helps the STS module in the selection of the optimal list of the surgical team for a newly referred surgical patient. In the STS module, a population-based meta-heuristic algorithm (AEFA) for multi-objective optimization is proposed to select the optimal surgical team. The performance of the proposed clustering algorithm is tested on reallife datasets and compared with the existing clustering algorithm. Further, the performance of proposed multi-optimization algorithm tested on benchmark functions and compared with existing multi-objective optimization algorithms. Finally, the proposed model is applied to a real postoperative surgical dataset from the orthopedic surgery department of MSH in India. The experimental results show that the proposed model proves its efficiency in selecting the optimum surgical teams for a patient