Objective — To discuss the contribution of epidemiology and epidemiological methods to psychosomatic medicine. Method — Critical review of the literature, including both philosophical concepts and empirical data. Results — The adjective «psychosomatic» has been used in two different ways: in the so called «psychosomatic» or holistic approach to medicine; and in a narrower approach, referring to particular disorders in which psychological factors were considered to have a fundamental aetiological role. While the ideal of the holistic, «humanistic» or «anthropological» approach should probably be never abandoned, the practical limitations of encompassing models, including Engel's bio-psycho-social model are also obvious: they may be heuristically sterile. On the other side, in relation to the narrow psychosomatic approach, psychogenetic views in the so called «psychosomatic illnesses» have been strongly criticized on empirical grounds. The potential of epidemiological methods to study these illnesses, but also «somatopsychic» disorders is shown in a number of papers in the empirical literature. We have grouped such contributions, including our own experience, in the five categories suggested by M. Shepherd for epidemiological methods in general: 1) the completion of the spectrum of disease; 2) the establishment of outcome; 3) the actuarial assessment of morbid risk; 4) the evaluation of the efficacy of treatment; and 5) the conceptual construction of diagnosis and classification. Conclusions — The contribution of epidemiological data to areas of interest in psychosomatic medicine has been relevant in recent years. The potential of epidemiological methods in this area is very important both to increase knowledge and to improve the quality of clinical practice.

This paper concerns one of the four research projects developed during a training course in clinical epidemiology managed by the Lombardy training centers IREF. Objectives — To compare the recommandations for treatment concerning 9 vignettes derived from the Australian Quality Assurance Project. Setting — Six Mental Health Services of Regione Lombardia. Design and Participants — For each vignette, all psychiatrists working in the 6 Mental Health Services were asked to fill in a questionnaire about treatment location, psychopharmacology, psychotheraphy, priority between psychotherapy and psychopharmacology and degree of difficulty in answering. Results — 44 out of 52 target psychiatrists took part to the study. Remarkable variability for treatment location and psychotherapies; moderate variation for psychodrugs prescriptions and a good agreement for diagnoses were observed. In drugs prescription an eccess of association was observed. The most prevalent model of psychotherapy was the psychodynamic, followed by the cognitivebehavioural and the family-systemic. There was a tendency toward a flexible approach, as suggested by recommendations of different psychotherapeutic models according to the nature of the disorder. No case were judged very difficult; only in 3 cases a judgement of «somewhat difficult» was expressed by more than 20% (but less than 30%) of the psychiatrists. Conclusions — Studies of this type are very easy to carry out and give useful information for continuous training programs and Continuous Quality Improvement projects.

Objective — To asses the capacity of the Parental Bonding Instrument (PBI) to discriminate between normal subjects and clinical samples and between with different psychiatric diagnosis. Design — The present paper analyzes the studies published between 1979 and 1995, which have used the PBI in normal subjects and clinical samples and have reported the respective means and standard deviations obtained on the two PBI dimensions: affection and control. Multiple comparisons were carried out between the mean scores of affection and control of: 1) samples with the same psychiatric diagnosis (intragroup comparison); 2) samples with different psychiatric diagnoses (intergroup comparison); 3) normal subjects and clinical samples. Results — Of the 46 studies with normal and clinical subjects, 23 studies were selected for the analysis, reporting means and standard deviations and specifying the diagnostic criteria. Samples with the same psychiatric diagnosis had similar affection and control scores. With the exception of bipolar affective disorders and avoidant personality disorders, the prevalent parental style was for all diagnostic groups the affectionless control style. Within the affectionless control style, the PBI discriminated between panic attacks, borderline personality and drug addiction but not between schizophrenia, unipolar depression and anxiety disorder. The PBI discriminated also between normal subjects samples and samples with anxiety disorder, schizophrenia, bipolar affective disorder, personality disorder and drug addiction respectively. Conclusion — The results confirm previous suggestions from single studies that the perceived parental style as measured by the PBI can be considered a good predictor for the presence of psychiatric disorders excluding panic attacks, avoidant personality disorders and unipolar affective disorders. Although the different diagnostic groups do not differ in their perceived parental style (affectionless control), significant differences between some diagnostic groups within this category suggest that the PBI might have some specificity as well.

Introduction — Many studies have been carried out recently to investigate the relationship between health conditions and family and self characteristics. Objectives — To identify family characteristics (such as type of family, socio-economic status, parent immigration, affection for parents, parents' psychological and physical health) and self characteristics (age, drug use, satisfaction for several life circumstances, psychosomatic symptomatology) that are differently distributed by sex and levels of psychological and physical distress among teen-agers. Methods — We carried out a cross-sectional study on a sample of teen-agers attending high schools in Pavia (Italy), using a self-administered questionnaire. The students were divided in four groups having different levels of psychological and physical distress, based on GHQ-30 (psychological distress indicator), on the number of hospital admissions and consultations to a physician in the last year (the last two are physical distress indicator). Data were analysed applying the multivariate analisy of Canonical Variate. Results — 1346 students were sampled, but only 1189 questionnaire were analysed: 36.8% regarding males and 63.2% females. The Canonical Variate analysis indicated that psycosomatic symptomatology, satisfaction for several life circumstances and affection for parents are important for describing the four distress groups. Conclusions — Only affection for parents has an important role on psychological and physical distress of adolescents, while family characteristics traditionally considered associated with psychological and physical distress in teen-agers (such as living with one or without parents, low socio-economic status) are not associated.

Objective — The implementation of a simple methodology to estimate full costs of services provided by a public mental health centre. Setting. CPS (NHS Mental Centre) Ussl 35, Magenta, Lombardy Region. Method — To estimate full costs of 16 types of service we followed a two step procedure. The first step was to estimate all costs attributable to the CPS. In the second one, we allocated this estimate to each type of service provided. We attributed to the CPS the following cost items: personnel, utilities (telephone, electricity, water, heating and cleaning), land & building, transports (for services provided outside the clinic) and a share of general cost of the USSL to which the CPS belongs. Full cost of each service was then calculated on the base of the yearly number of services provided and the time spent by each health professional. Results — In 1995, the CPS provided 14, 562 services. Total costs amounted to L 1,356 million, and more than three quarters of this amount was attributable to the personnel working at the CPS. Unit costs ranged from L 5,300 (drug administration) to L 442,400 (family therapy involving two professionals for 90 minutes) The unit cost of psychiatric visits, psychologist consultations and nurse domiciliary visits were L 105,300, L 106,600 and L 78,000, respectively. Conclusions — This approach requires accessible data and is relatively simple to manage. Some refinements are required, especially to improve the methodology for the determination and the allocation of overheads. However, we are convinced that this cost accounting procedure provides acceptable estimates of the services provided by the CPS. These estimates suggest that charges to be used to fund NHS providers may be too low, especially if fee-for-service will be the main funding source.

Objective — To describe requests of admissions to eight General Hospital Psychiatric Wards (SPDC) in Lombardy, Italy, during November 1995. Design — Descriptive prospective multicenter study. Setting — SPDCs of Busto Arsizio (VA), Calcinate (BG), Desio (MI), Magenta (MI), Merate (LC), Milano San Paolo I, Pavia, Treviglio (BG). The global catchment area sums up to 11% of the whole regional area, and to 18% of the population. Main outcome measures — We used a previously developed flowchart with two major key points: who decided to go to the hospital? Did a doctor confirm this initiative? Main sociodemographic characteristics, ICD10 diagnosis and previous psychiatric admissions were collected for each admission. Results — Admissions were 315, patients 246. In 9.5% of cases patients asked for admission without any medical advice. In one third of cases the ward psychiatrist was the first doctor to visit the patient. Compulsive admissions (TSO) were 45 (14.3%), although patient's initiative lacked in 55.6% of cases. A referral from Outpatient Departments (CPS) was present in 28.2%. First-ever admitted were 63 (25.6%): 20.7% sent by CPS, 16% by GPs, 11% by other non psychiatric wards. Conclusions — Although Goldberg & Huxley's model described General Hospital Psychiatric Wards as the last level of intervention, our data show that Italian SPDCs work as «front line» services: less than one admission out of two were referred by a psychiatrist. Problems raised by «self-referred» patients are conspicuous and an evaluation of the filtering function of CPSs seems necessary. In this regard, a comparison with different modalities of Department organization could be useful.