Condition based scheduling for diabetic patients

Sudaratana Wongweragiat, Purdue University

Abstract

Routine visiting of medical providers is required for diabetic patients for checkups and assessments. Currently, there are no specific guidelines for physicians to schedule clinic appointments for the patients. Using classification tree (CT) and multiple linear regression (MLR) analyses on a set of data collected from a group of diabetic patients, condition based scheduling determines return visits based on patient conditions. The findings derived from CT and MLR suggest that 20-30% of patients classified as improved may be able to reduce their visiting frequency while keeping their improving diabetic condition. This may result in more efficient resource utilization by transferring the unnecessary care from these patients to others who really need it. The study also shows that shortening of return visiting intervals may not significantly lead to a better diabetic condition among 15-20% of patients with no improvement. Further research may try to explore other factors in dealing with these patients. However, 25-40% of patients without improvement have fewer visits than the level predicted by the models to experience improvement. Therefore, physicians may simply try to schedule more appointments for these patients. Although the findings are not served as a standard guideline, they may be beneficial as additional suggestions for physicians in making decisions on return visits, which should contribute to more efficiency and higher quality of the overall healthcare delivery for diabetic patients. Moreover, the method may be applied to other chronic diseases for which routine visiting is required.

Degree

Ph.D.

Advisors

Yih, Purdue University.

Subject Area

Statistics|Biomedical engineering|Industrial engineering

COinS