STATISTICAL MODELS FOR MEDICAL MALPRACTICE LIABILITY LOSSES (INSURANCE)

Health-care providers--typically physicians, surgeons, and hospitals--may be liable for damages arising out of the rendering of, or failure to render, professional services when these services depart from an accepted standard of care and result directly in injury to the patient. Medical professional liability insurance provides protection against claims of "medical malpractice" arising from such injuries.;The purpose of this dissertation research is to try to identify factors which influence the frequency and severity of medical malpractice claims. These factors include characteristics of the hospital and its environment, characteristics of the patient, and characteristics of the alleged injury. Two statistical models are developed, one for paid claim frequency and one for paid claim severity (settlement value).;The severity model attempts to explain the variation in claim settlement by variation in corresponding variables, such as seriousness of the injury, age and sex of the patient, and the dates of the injury, report, and closure. For this model, each observation in the data set corresponds to an injury, and the observed settlement value is the total amount paid by all defendants for this injury.;The frequency model attempts to relate the variation in hospital paid claim frequencies to characteristics of those hospitals and their environment; variables considered include number of occupied beds, number of surgeries, number of emergency room visits, city population, and so forth. This model is complicated by the fact that the dependent variable is assumed to have a Poisson distribution, and by several other factors, including the so-called long tail on malpractice claims.;In the early 1970s, the frequency and severity of medical malpractice losses began to rise at alarming rates. The so-called "medical malpractice crisis" gained national prominence by 1974-75, generating widespread discussion in state legislatures, the Congress, and the press. Resulting changes in the medical malpractice insurance market have made it increasingly important to be able to meaningfully evaluate the risk of a small group of health-care providers or even the risk of a single provider.;Although neither model can be said to provide an excellent fit of the data, both produce statistically significant and intuitively satisfying coefficient estimates. (Abstract shortened with permission of author.).