The focus of this study is to identify the potential prognostic factors and symptoms that are associated with Cryptococcal Meningitis and to establish a statistical model for the prediction of outcomes (survival and mortality) among in-hospital patients. Materials and Method: The hospital admission books in the medical wards and pharmacy of the East London Hospital Complex were searched to identify the folder numbers of all the patients that were admitted, diagnosed and or treated for Cryptococcal Meningitis at the hospital between the 1st of January 2009 and the 31st of August 2012. 237 folders out of 519 folders reviewed had confirmed cases of Cryptococcal Meningitis. Data on patients’ demographics, In-hospital care, and Concurrent infection/health condition were collected and analysed in a cross-sectional study, using the univariable and multiple logistic regression. Analysis of data was done with SAS version 9.1.3 and NCSS version 2007 software. Results: In a multivariable logistic analysis of variables found to be significantly associated with Cryptococcal Meningitis in a univariable logistic regression, Being Conscious (i.e. a Glasgow Coma score of 15) (OR= 5.34,CI=2.85-9.99 p =<0.000); Having no history of TB infection (OR=28.91, CI= 3.68-226.89, p = <0.001); No Fits (OR = 2.59, CI= 1.18-5.68 p = 0.017); Being a non-smoker (OR =2.22,CI=1.13-4.34 p = 0.020); and Adhering to treatment instruction and guidelines (OR = 2.38, 1.15-4.89, p = 0.019) were the variables found to be significantly associated with the survival of a Cryptococcal Meningitis patient. The uninterrupted use of Amphotericin B (OR=3.04, CI=1.06-8.72, P=0.038) as a first line regimen was also found to be significantly associated with survival.On the other hand, being unconscious (i.e. Glasgow Coma score <15) (OR =5.34, CI=2.85-9.99, p = < 0.000), Currently having a TB infection (OR = 9.20, CI=2.77-30.57, p = < 0.000), Not adhering to treatment guidelines (OR=2.38, CI=1.15-4.89, p=0.019 ); Being a smoker (OR = 2.22, CI=1.13-4.34, p = 0.020) and having Fits (OR=2.59 CI=1.18-5.68 p=0.017 ) were found to be significantly associated with mortality. Headache (p= 0.505) was found not to be a significant predictor of survival contrary to the findings in many publications on Cryptococcal Meningitis. Owing to time constraint, testing data was not collected to validate the prognostic models. However, model diagnostics was done and the relevant statistics confirmed the goodness of fit and the predictive ability of the model Conclusion: It has been established in this study that certain baseline variables can be helpful in the prognosis of Cryptococcal Meningitis infection. It is therefore believed that these variables will help in improving the prognosis of the infection especially at the East London Hospital Complex. Though the statistical models will work well in predicting the outcome of Cryptococcal Meningitis infection for patients admitted at the East London Hospital Complex, adequate precaution must be exercised while attempting to apply it in other geographical areas.
| Identifer | oai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:ufh/vital:11782 |
| Date | January 2012 |
| Creators | Okorie, Ikechukwu Obinna |
| Publisher | University of Fort Hare, Faculty of Science & Agriculture |
| Source Sets | South African National ETD Portal |
| Language | English |
| Detected Language | English |
| Type | Thesis, Masters, MSc (Biostatistics and Epidemiology) |
| Format | 91 leaves; 30 cm, pdf |
| Rights | University of Fort Hare |
Page generated in 0.0021 seconds