It has been proposed that the effectiveness of medical treatment should be measured by mortality from specified diseases for which death is apparently avoidable given appropriate medical intervention. Unfortunately, that proposal was not supported by direct analysis of the relation between health care resources and avoidable mortality--an analysis which is essential for establishing the validity of avoidable mortality as an indicator of the outcome of health care. Thus the proposal does not yet have any basis, theoretical or practical. A more realistic but still oversimple model is now proposed. With presently available data, however, the validity of avoidable mortality is unlikely to be established, even by this model.