Effectiveness evaluation in mental health facilities La valutazione dell'effectiveness dei servizi per la salute mentale

The aim of this study is to measure, using the HoNOS scale and multilevel analysis, patient problems at admission to mental health services and to verify to what extent these problems have changed at discharge. Materials and methods. The study involved 4712 psychiatric patients (2073 male, 2639 female; mean age 45.2 years). The key indicator was the HoNOS scale score, completed at admission and at discharge: the difference between these scores represented the change in the patient's health status. The variables were collected by the Lombardy Region Informative System, using the software "Psiche", which registers epi- demiological data on patient characteristics and psychiatric performances in all the Departments of Mental Health (DMH) in Lombardy (Italy). In order to blind the data analysis, participating DMH were coded us- ing pairs of alphabet letters. Data were analyzed by the multilevel statistical model, an innovative statistical methodology recently applied to the health care setting. Results. Variables positively correlated to mental health improvement included: non-working class (unem- ployed or retired), lower level of instruction, previous participation in therapy programs located in hospital, residential o semiresidential facilities programs, previous DMH treatment in a Hospital of great or middle dimension, severity of illness at admission. Negative factors were: treatment in DMH with a Residential Center for Psychiatric Therapy (CRT), treatment in a DMH with a case-mix greater than average (mean re- gional value = 1, lower or higher values indicate the percentage of shifting from 1). A DMH effect was ob- served: at univariate analysis, which considered only the difference between admission and discharge score, the most effective DMH was the unit coded as LL; after risk adjustment performed by multilevel analysis, the most effective DMH resulted DD. The least effective DMH was AA. Conclusions. Multilevel analysis is a useful tool for mental health service evaluation. With the right indica- tors it can contribute to quality improvement of the mental health services, allowing a comparison across different kinds of centers and hospitals.