PROBLEMConsiderable research effort has already been expended on the search for predictors of length of psychiatric hospital stay(4-5 , l o ) . Research of this kind has usually been undertaken because of the potential clinical usefulness of the early identification of developing chronicity and because predictive information would contribute to a better understanding of the nature and cause of various psychiatric disorders.Although some minimal life history information has frequently been included in prediction studies, the uneven availability of these data has severely limited the extent to which they could be included. The present study was undertaken to study the hypothesis that life history variables, if assessed comprehensively and systematically, would contribute substantially to the prediction of length of hospital stay. The availability of the Minnesota-Briggs History Record (M-B) ( l ) , an objective multiple choice social history questionnaire completed by a relative of the patient, made possible the investigation of the relationship between history variables, as defined by item scores, and length of continuous hospital stay during the first year after admission.
METHODThe general strategy of the study involved the collection of social history inventory responses of the relatives of newly admitted psychiatric patients, the correlation of each item score with the number of consecutive weeks spent by the patients in the hospital during the first year after admission, and the use of multiple regression analysis to discover relatively independent predictors and to determine what proportion of outcome variation could be accounted for by a linear combination of these items.The sample of 192 patients included all admissions to the V. A. Hospital, St. Cloud, Minnesota, between March, 1961, and February, 1962 with the following exceptions: Patients who were over 60, who carried a diagnosis of Central Nervous System pathology, who had been hospitalized during the six months prior to admission, or for whom a usable M-B History Record was not obtained. The mean sample age was 38.8 (SD 8.0), 58% were never married, and 70% were diagnosed as schizophrenics.Within a few days after admission to the hospital, the patients' next of kin were mailed the M-B History Record. The return rate of usable History Records was 88%. Mothers of patients completed 68% of the forms.The M-B History Record is a 175 item multiple-choice questionnaire covering the usual areas considered important in the conventional psychiatric history. Since the last 36 items apply only to married patients, and three of the remaining 139 items are inappropriate for correlational analysis, 136 M-B items were correlated with the length of stay criterion. Marital status, rural-urban residence, and diagnosis were also included. Promising items were selected for regression analyses described below.