Stochastic scheduling problems in healthcare
Abstract
Inefficient scheduling is the primary source of operational disruption and patient dissatisfaction in any outpatient clinical facility. The situation is further agravated by factors such as patient no-show, cancellation, walk-ins, tardiness etc. This not only affects revenue but diminishes the quality of care. Hence designing an efficient scheduling system is a crucial component to improve operations of a clinic. An effective algorithm is one which imitates the actual reservation mechanism taking into consideration all the above factors. But, unfortunately none of the existing work do the same. The current research develops a sequential clinical scheduling algorithm under general service time for patient no-show, when the total service period is divided into a fixed number of “slots”. It is further extended to a similar problem where the service period is continuous (i.e. not divided into slots). The later can be used to provide a guideline for designing a systematic appointment slot structure. The second part of this dissertation strives to compute the optimal time of access intervention for hemodialysis patients. These patients undergo a surgery to establish a special type of connection, access, between an artery and a vein to facilitate dialysis. But the life of these accesses get shortened due to a phenomenon called stenosis. Statistics show that incidences of such access complications are very high among hemodialysis patients. It is also one of the primary causes of morbidity. But unfortunately no concrete guideline exists delineating a proper treatment plan. The proposed research is the first endeavor, which determines the optimal time of treatment by maximizing the utility of the decision maker. For this purpose an optimal stopping model is developed, the solution for which is obtained by solving the corresponding free boundary problem. Numerical examples show that depending on the risk characteristic of the decision maker, the optimal stopping point obtained from this model vary significantly from KDOQI [43] guidelines.
Degree
Ph.D.
Advisors
Lawley, Purdue University.
Subject Area
Health care management|Operations research
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.