MORTALITY IN PSYCHIATRIC-HOSPITAL PATIENTS - A COHORT ANALYSIS OF PROGNOSTIC FACTORS

Background. This work followed a group of patients living in a psychia tric hospital in Central Italy in 1978 at the time of enforcement of t he Italian reform law (No.180) for closing down mental hospitals. The study had the following aims: a) to compare in terms of mortality pati ents discharged into the community with patients who did not experienc e deinstitutionalization; b) to determine the survival of the cohort o f patients and to analyse prognostic risk factors for death; c) to ana lyse differences in mortality rates between psychiatric patients and t he general population. Methods. The study was designed as an historica l follow-up investigation. Univariate (product limit) and multivariate (proportional hazards model) methods were used to estimate prognostic variables and related death risks. Mortality was assessed using stand ardized mortality ratios (SMR) on the entire cohort as well as after s tratification according to age, sex, cause of death and discharge stat us, assuming the Abruzzo Region's population as standard. Results. Len gth of hospitalization and discharge from hospital are prognostic vari ables for death risk, with relative risks respectively of 4.22 (95% co nfidence interval [CI] :2.41-7.40) for a length of hospitalization of 10-25 years, and 8.13 (95% CI : 4.73-13.88) for non-discharge. The glo bal SMR of the cohort was 2.68 (95% CI : 2.42-3.07). Non-discharged pa tients showed higher SMR than discharged. Excess mortality was found b oth in males and females for circulatory, respiratory and undefined di seases. A significantly lower mortality for cancer was observed in mal e patients. A strong excess mortality was observed in younger patients (20-29 years: SMRmales = 43.57; SMRfemales = 97.52). Conclusions. Lon ger periods of hospitalization and non-discharge from hospital are the main risk factors for death in psychiatric patients, who globally exp erience higher death rates than the general population for a wide spec trum of causes of death, whatever their diagnosis or gender. These fin dings strongly suggest positive actions in order to overcome the effec ts of institutionalization.