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.