Do psychiatric registries include all persons with schizophrenia in the general population? A population-based longitudinal study

Abstract

Background

Psychiatric hospitalization registries are utilized to investigate the incidence and prevalence of schizophrenia for both research and administrative purposes. The assumption behind this is that most individuals with schizophrenia will be hospitalized at least once in their life-time.

Method

In an epidemiological survey conducted in the 1980s, a population-based sample (n = 4914) of Israel-born individuals then aged 25–34 were screened in the community, and 29 (0.6%) were subsequently diagnosed by psychiatrists using SADS/RDC criteria. Twenty four years later we linked data from the epidemiological survey with the Israeli National Psychiatric Hospitalization Registry.

Results

Twenty seven of the 29 individuals (93%) diagnosed with schizophrenia in the survey were identified in the hospitalization registry with the same diagnosis. Fifty-two (1.0%) participants not diagnosed during the survey with schizophrenia were identified in the psychiatric hospitalization registry 24 years later with schizophrenia. The majority of them were diagnosed with other psychiatric disorders in the survey. If all diagnoses of schizophrenia are accepted at face value, the lifetime prevalence rate would be 1.8% for this cohort.

Conclusion

The overwhelming majority of individuals diagnosed with schizophrenia at ages 25–34 in an epidemiological survey were present in the Psychiatric Hospitalization Registry. However, the assessment of life-time rates of schizophrenia at these ages is problematic because some future cases are asymptomatic, others have premorbid non-psychotic disorders, while in others it is difficult to differentiate between affective disorders and schizophrenia. Availability of psychiatric services and hospitalization policy must be considered when generalizing these findings to other countries.

Introduction

Genetic factors(Cannon, 2005, Lichtenstein et al., 2009, Williams et al., 2009), subtle developmental delays (Maki et al., 2005, Murray et al., 2006, Jaaskelainen et al., 2008), exposure to environmental factors (Kirkbride et al., 2008, Oh and Petronis, 2008), and/or gene–environment interactions (Weiser et al., 2001, van Os and Murray, 2008, van Os et al., 2008) have been identified as putative risk factors of schizophrenia (Welham et al., 2009). Due to the low prevalence of schizophrenia, the large number of potential risk factors and the long follow-up period required, reliance on prospective cohorts is often not practical. Hence, alternative research strategies have been used, including the use of psychiatric hospitalization registries created for statistical and administrative purposes that contain demographic and diagnostic data. These registries have been utilized for research, employing both cross-sectional and historical-prospective designs (Jones et al., 1994, Mortensen et al., 1999, Dinesh et al., 2004). However, this approach would be of limited value if a large proportion of individuals with schizophrenia were not included in these registries (Kendell et al., 1993, Kohn et al., 2004).

Previous studies sought to identify people with schizophrenia [18] or psychotic disorders [19] who may not be included in hospitalization-based case registries, by identifying patients receiving ambulatory treatment in a given catchment area and following them up to ascertain those who were eventually hospitalized. One of those studies (Geddes and Kendell, 1995) reported that 6.7% of the patients with schizophrenia had never been hospitalized while a more recent study (Jorgensen et al., 2010) reported that 25% of patients received outpatient care only. However, persons with schizophrenia who do not attend ambulatory mental health services (Shad et al., 2007) would not be included in such studies.

A literature search did not yield any study which identified a sample based on a birth cohort of individuals in the community, whose diagnoses were determined independently from the utilization of psychiatric services, and were subsequently followed over time to ascertain later hospitalization. To meet such requirements we utilized data from a birth cohort of 4914 participants from Israel who were screened for psychopathology and diagnosed in the mid 1980s at ages 25–34 (Dohrenwend et al., 1992). Psychiatric hospitalizations were ascertained using a nation-wide psychiatric hospitalization case registry 24 years later, when the participants were 50–60 years old and had passed through the age of risk for schizophrenia.