Modeling health care for cost containment: A decision support system comparing multivariate techniques and artificial neural systems.
详细信息
- 作者：Morrison ; Joyce Ann.
- 学历：Doctor
- 年：1995
- 导师：Johnson,John D.,eadvisor
- 毕业院校：The University of Mississippi
- CBH：9615665
- Country：USA
- 语种：English
- FileSize：4709485
- Pages：182
文摘
This country is currently facing a crisis in health care due to escalating costs. Curtailing these spiraling costs has become a focal point for many health care managers. Previously,there have been few studies on predicting treatment costs. This is due in part to the inadequacy of computer technology necessary to isolate,aggregate,and model large volumes of patient information. As the price of health care continues to rise,modern technology must provide answers to questions that were unasked earlier. Fortunately,the 1990s have brought us the technology to process and model patient data into useful information that can become a valuable resource in the business of health care. This research focuses on postmenopausal women as a study group. This group was chosen for two reasons: First,due to increased life expectancy,women now live over one third of their life in postmenopausal years,and with these years comes a myriad of health problems. Secondly,one proposed solution to this problem is the prescription of conjugated estrogens,now one of the most frequently dispensed pharmaceuticals in the United States. The impact of estrogen replacement therapy on the reduction of total treatment costs will be measured. In order to validate the construction of cost models,this research will examine two very large medical databases,Mississippi Medicaid and The Prudential Health. By using the claims data of postmenopausal women,this study will evaluate data modeling techniques and propose a decision support tool for health care managers. Two highly complex and powerful statistical procedures are proposed as methodologies for this prediction model: stepwise multivariate regression and Artificial Neural Systems ANS). The predictive accuracy of these two diverse techniques is presented within a comprehensive,mathematically sound framework. The results of this research offer a new decision support instrument for health care managers. Heretofore,medical science alone has charted pathways to health. Now information science enables us to use the retrospective clinical data to both predict and possibly improve the future through estimating and reducing the cost of health care.