Schizophrenia is a debilitating mental illness. The prevalence of schizophrenia is of concern and highlights the chronicity of the illness (Kaplan & Sadock, 1998; Warner & de Girolamo, 1995). It is therefore important that greater understanding ofthe variables affecting and potentially contributing to schizophrenia is sought. This study deals with three separate areas: 1. Creating a database and describing the socio-demographic profile of 615 certified patients diagnosed with schizophrenia between the years 1995 and 1996, hospitalised at Fort Napier. 2. Exploring whether a meaningful geographical model of socio-demographic variables of schizophrenia can be built, and whether it can determine areas of high schizophrenia sensitivity. Such a perspective of schizophrenia (exploring the combined effect of all the variables connected to schizophrenia) has not been previously researched rather each variable has been studied independently. The model utilised in the Geographical Information Systems (GIS) offers an opportunity to graphically overlay these individual variables and then examine the potential effect. It was only possible to explore sociodemographic variables within this Masters research project. 3. Determining the geographical accessibility ofthe public provincial hospitals offering psychiatric services to the population, (particularly to persons with schizophrenia). This is relevant to the provision of maintenance programmes and early intervention as encouraged by recent research (Hodges, Byme, Grant & Johnstone, 1999; Yung et al, 1998). The socio-demographic profile of the certified patients at FNH correlated with most of the international findings. The presentation of schizophrenia remained constant in South Afiica in comparison to international countries and cultures (Warner & de Girolamo, 1995). Overall, the pattern produced by the schizophrenia sensitivity model was unclear other than high and low schizophrenia sensitivity ranges being evenly spread throughout KZN. The model has yet to be further refined to increase its sensitivity to schizophrenia risk. Although statistical significance was not established, the accessibility model suggested that large parts ofKZN fall within the range of easy accessibility in relation to the 19 provincial hospitals that provide a psychiatric service. The potential of the schizophrenia sensitivity model and accessibility model, in spite of the crude and incomplete variables utilised was evident. The GIS was a powerful tool and utilised in this study to collect (from surveys and other databases), store (retrieval and query), manipulate (transforming data, analysis and modelling) and produce data output (data reporting, such as maps and reports) (Foote & Lynch, 1995). The GIS has been a very effective tool in working with epidemiological information and it would be of great use in service planning. It is strongly recommended that there be greater collaboration between mental health services and the GIS Unit, Department of Health. / Thesis (M.Soc.Sc.)-University of Natal, Pietermaritzburg, 2002.

The impact of the emotional climate in the home of the schizophrenic on relapse has been researched extensively through the construct of Expressed Emotion (EE). Most often patients from high EE homes have a higher relapse rate than those from low EE homes. This is a robust finding throughout many western and some non-western countries. However, no published research documents EE status and relapse in African countries. This study reports on the EE status and rehospitalisation rates of 29 Zulu-speaking schizophrenic patients in a South African sample. EE was assessed using a translated version of the Level of Expressed Emotion (LEE) scale, a 60 - item, self-report measure developed in Canada (Cole & Kazarian, 1988). A multistage translation procedure, comprising back-translation, the committee approach and decentering was employed. The Zulu SCL-90-R was administered as an indicator of psychological distress. Follow-up data on rehospitalisation was collected nine months after index admission. Results indicated somewhat unsatisfactory internal reliabilities on some of the subscales of the Zulu LEE scale. High scores on the Zulu LEE scale were not significantly predictive of rehospitalisation at follow-up; however, they were significantly predictive of greater previous admissions. The psychometric properties of the Zulu SCL-90-R were found to be satisfactory, indicating the validity of using this instrument for the purpose of screening for psychiatric illness. Zulu schizophrenics were found to have a lower rehospitalisation rate (17% at nine month follow-up) than found in international studies. The course for female schizophrenics was better than that for male schizophrenics. Findings are inconclusive regarding the impact of EE on the course of schizophrenia in a Zulu sample. / Thesis (M.Soc.Sc.)-University of Natal, Pietermaritzburg, 2000.