Equity and accuracy in medical malpractice insurance pricing

Abstract

This study examines alternative classification approaches for setting medical malpractice insurance premiums. Insurers generally form risk classification categories on factors other than the physician's own loss experience. Our analysis of such classification approaches indicates different but no more categories than now used. An actuarially-fair premium-setting scheme based on the frequency and severity of the individual physician's losses would substantially penalize adverse experience. Alternatively, premiums could be set for groups of physicians, such as hospital medical staffs. Our simulations suggest that even staffs at rather small hospitals may be large enough to be experience-rated.

The Case for Experience Rating in Medical Malpractice Insurance: An Empirical Evaluation

“…The latter is obviously a stylized account of the actuarial calculations that insurance companies construct, but it incorporates all available information, short of assigning individual random effects. Given the apparent simplicity of prevailing practices in the pricing of malpractice insurance (Sloan and Hassan, 1990), this assumption gives a conservative benchmark to compare with experience rating.…”

“…Second, a higher volume of patient visits is expected to increase the frequency of claims, but perhaps at a diminishing rate. This aspect is left out of earlier studies, with the exception of Sloan and Hassan (1990). The mean-scaled number of patients seen by the doctor during an average week ("patient volume") captures the level of physician activity and the authors add its square for possible nonlinearities.…”

“…The mean-scaled number of patients seen by the doctor during an average week ("patient volume") captures the level of physician activity and the authors add its square for possible nonlinearities. 14 Finally, other researchers (Cooil, 1991;Sloan and Hassan, 1990) found that female physicians tend to have lower frequencies of suits, so a dummy for gender was also included.…”

“…Analyses of data on the frequency of claims in Pennsylvania, New York, New Jersey, and Florida find heterogeneity of risk within a given physician specialty (Rolph, 1981(Rolph, , 1991Nye and Hofflander, 1988;Venezian, Nye, and Hofflander, 1989;Ellis, Gallup, and McGuire;1990). 2 Sloan and Hassan (1990) show that residual heterogeneity within risk classification groups is reduced but not eliminated when the a priori risk groups are based on more information than just physician specialty. Their study provides the best support to date for the presence of residual heterogeneity in claims.…”

“…From the perspective of experience rating, however, these repeat offenders may pose the greatest concern. Furthermore, Sloan and Hassan (1990) do not assess the performance of their model in out-of-sample prediction, which is crucial to the question of whether measured risk differences are persistent.…”