 |
 |
 |
|
|
Fuller Torrey concluded his review of over 70 prevalence studies of schizophrenia with a compelling statement: “As tragic a disease as schizophrenia is, it is also one of the greatest intellectual challenges of contemporary medicine” (1, p 606). In the same review, Torrey identified a series of beguiling questions that remained unanswered: How substantial are geographical differences in the prevalence of schizophrenic disorders? Do pockets of high or low prevalence exist? Have there been robust changes in the incidence of schizophrenic disorders over time? In the current systematic review of studies published between the years 1980 and 2000, we revisit such questions and relate them to mental health policy and practice. Prevalence vs Incidence“Prevalence” quantifies the proportion of individuals in a population who have a disease during a specific time period. “Incidence” refers to the number of new cases of disease that develop in a population of individuals at risk during a specific time period. While both prevalence and incidence rates have similar denominators (that is, the population at risk) and include new cases in their numerator, prevalence rates also contain existing cases in the numerator. Thus, although incidence rates can approach prevalence rates in diseases with short duration, this is not the case with schizophrenia. For example, the US Epidemiological Catchment Area (ECA) study found the incidence rate for schizophrenia to be only 27% of the prevalence rate (2). Incidence and prevalence data also have different uses. Because prevalence data provide a snapshot of the burden of disease on society at a specific time, they can be used to inform planning efforts and to estimate ideal resource allocations. Incidence is more useful in examining changes in the disease risk in different populations over time. Moreover, because incidence studies do not mix old and new cases, they are better able to examine causal relations. For example, they can help determine whether potential risk factors, such as birth complications, lead to the subsequent development of schizophrenia. An understanding of causal relations can lead to the development of effective prevention and early intervention efforts. Readers interested in further examination of the methods and applications of incidence and prevalence research are referred to Hennekens and others (3). Why A Systematic Review?Previous reviews looked at older studies with widely varying methodologies and found 6- to 14-fold variation in incidence and prevalence rates of schizophrenia (1,4). However, individual studies that use identical methodologies across study sites, such as the International Study of Schizophrenia (ISoS) (5), have found only 2- to 3-fold variation worldwide. To help decrease the likelihood that variation of rates is an artifact of methodology, we used an a priori protocol that minimizes missing relevant studies, maximizes comparability across studies, and restricts inclusion of studies to those using high-quality methods. If this systematic process reduces the variation in studies observed in previous reviews and leads to findings closer to the range of the ISoS, we can be more confident about the conclusions drawn from our analyses. MethodsSearch Strategy to Identify Relevant Articles Study Selection 1. Study design. Studies were to be either community surveys of the general population using probability sampling techniques or those that surveyed the entire population of a defined area. We also included studies using key informant methodology, which involves establishing a list of services and agencies in a defined area that are likely sites of contact for potential cases, if it was evident that the case-finding covered an extensive network of mental and nonmental health services. Because the key-informant method aims to ascertain all cases within a specified area, studies that identified cases from treatment settings only were excluded. After we completed the screening stage of the review, we modified these criteria for incidence studies, because only a single study could be included. Incidence studies were eligible for inclusion if they used case-register methods that surveyed, at minimum, primary care general medical services. 2. Study population. We included prevalence studies that covered the entire age range of the general population, as well as studies that focused only on adults (for example, aged 18 to 65 years). However, we included only those incidence studies that examined subjects aged 15 years and over. We did not include studies meeting our eligibility criteria but falling outside these age ranges, because it did not seem feasible for our present review, especially given the relative lack of studies examining other age groups (such as children and the elderly). 3. Sample size. We included all studies having denominator sample sizes of 450 or more. 4. Diagnostic criteria. We selected only studies using operationalized diagnostic criteria and case identification based on either standardized instruments or clinician diagnosis. Where studies used ICD or DSM classification systems, we included only those using ICD-9 or DSM-III and later criteria. We excluded studies published prior to 1980, because we judged that they would be unlikely to meet the above criteria. We also excluded studies if the case definition was not explicit. We scrutinized studies incorporating medical records for inclusion of corroborating diagnostic methods, and we excluded studies relying solely on medical records. We initially applied the above eligibility criteria to the citations and abstracts generated by the search. Based on this information, we excluded publications definitely not meeting the inclusion criteria. When an article met the inclusion criteria, or when there was not enough information to definitely exclude it, we retrieved the full text. We then reviewed these potentially relevant articles to determine whether the inclusion criteria were in fact met. We excluded studies not meeting the eligibility criteria at this stage and documented the reasons for exclusion. Studies selected for inclusion in the review were formally abstracted, as described below.
|
||||||