Background: Recent systematic reviews have encouraged the psychiatric research community to re-evaluate the contours of schizophrenia epidemiology. In particular, understanding the incidence, prevalence and mortality in schizophrenia are vital indicators to inform decision-making for policy formulation, health planning, and research related to risk factor epidemiology.
Aims: The aims of the study were to conduct systematic reviews of prevalence and mortality in schizophrenia, evaluate methods of summarising epidemiological data, and examine the interrelationship between the incidence and prevalence of schizophrenia.
Methods: The dissertation is comprised of four parts. The first two parts involve major systematic reviews that build on the expertise of the candidate and his supervisors. Part One presents a systematic review of the prevalence of schizophrenia. The distribution of various types of prevalence estimates obtained from population based studies were summarized using measures of central tendency with quantiles and moments. The data were analysed using several pre¬specified hypotheses. Part Two presents a systematic review of mortality associated with schizophrenia. Population based data on Standardized Mortality Ratios were collated and key features of these distributions were summarized. In keeping with previous systematic reviews, these data were interrogated using several pre¬specified hypotheses. Part Three of the dissertation compares two methods of data synthesis; (a) the distribution median, and 10 to 90% quantiles derived from a cumulative plot of estimates versus (b) pooled estimates based on conventional meta-analysis. Pooled estimates of the incidence, prevalence and mortality were obtained, the general concordance between the two methods explored from a qualitative perspective, and the strengths and weaknesses of the alternate approaches examined. Finally, in Part Four of the dissertation, the relationship between appropriately matched incidence and prevalence estimates was examined based on both empirical and modelled data using exploratory modelling approaches. Analyses were performed using SAS 9.2 (SAS Institute Inc., Cary, NC, USA), and DISMOD II (WHO 2004) software.
Results: The main findings of Part One (Saha et al., 2005) were: (i) the median prevalence (per 1,000 persons; 10¬90% quantile) for point, period, lifetime, and lifetime morbid risk were 4.6 (1.9–10.0), 3.3 (1.3– 8.2), 4.0 (1.6–12.1), and 7.2 (3.1–27.1), respectively, (ii) there was substantial variation of the prevalence of schizophrenia between sites, (iii) the prevalence of schizophrenia in migrants was higher compared to native¬born individuals (the migrant¬to¬native-born ratio median (10–90% quantile) being
1.8 (0.9–6.4)), and (iv) when sites were grouped by economic status, prevalence estimates from ‘‘least developed’’ countries were significantly lower than those from both ‘‘emerging’’ and ‘‘developed’’ countries (p = 0.04).
Results of Part Two (Saha et al., 2007): The main results were that (i) the median standardised mortality ratio (SMR) for persons for ‘all¬cause’ mortality was 2.58 (1.18-5.76) with a corresponding random¬effects pooled SMR (95% CI) of 2.50 (2.18, 2.43), (ii) there was no sex difference, (iii) suicide was associated with the highest SMR (12.86), while most of the major causes of death categories were also found to be elevated in schizophrenia, and (iv) the SMRs for ‘all¬cause’ mortality have increased over recent decades (p=.03).
Results of Part Three (Saha et al., 2008b): It was demonstrated that meta¬analysis when used with appropriate caution, can complement the synthesis of frequency data in epidemiology, however researchers interested in variation of the frequency measures should not rely on meta¬analysis alone.
Results of Part Four (Saha et al., 2008a): In respect to modelling the incidence and prevalence of schizophrenia using published data revealed that the consistency between published versus modelled estimates was poor.
Conclusions: There is a wealth of published data on the epidemiology of schizophrenia. Systematic reviews of prevalence and mortality and modelling based on these data can provide insights into the ‘dynamics’ of schizophrenia epidemiology.