Problem Kilbourne, Chesson, and Hilton (1988) found that Physical Evaluation (PE) Board determinations for back-problem cases were related to severity of the problem and length of service. An important question is whether Navy PE Boards uniformly invoke these same two criteria regardless of the type of medical problem. If so, those back-problem results should replicate in the analysis of a different type of medical problem. Objective The main purpose of this study was to examine the extent to which severity of mental health diagnosis (i.e., psychotic [functional and organic] versus nonpsychotic) and length of service predict medical discharge disposition (severance pay versus temporary disability retirement). There was also interest in examining the extent to which mental health diagnoses were consistent from time of hospitalization to time of last PE Board determination. Lastly, this study provided an opportunity to elaborate on the contribution of entitlement costs to the total costs of medical care. Approach The approach of this study was to analyze data from the Navy Enlisted Career/Medical History File regarding the type of medical disability discharge awarded by PE Boards to active duty, enlisted Navy personnel with a mental health problem, in relation to severity of diagnosis, length of service, paygrade, presence of a secondary diagnosis, and number of hospital admissions. The sample (N=1,485) consisted of all incidences of PE Board dispositions of either severance pay or temporary disability retirement between 1981 and 1984, inclusive, for active duty, enlisted Navy personnel who had been hospitalized with a mental health problem as the primary diagnosis. 0* Findings Findings indicated that severity of mental health diagnosis and length of For service significantly predicted medical discharge disability. Hence, results }I for mental health problems replicated results for back problems, wherein the 0 same two primary criteria (i.e., the severity rule and the 20-year service ion rule) considered by PE Boards significantl predicted medical discharge __i .;t, i t On/