Use this URL to cite or link to this record in EThOS:
Title: Modelling of risk factors, case-fatalities, survival and functional health status for stroke in Kelantan, Malaysia
Author: Musa, Kamarul Imran
Awarding Body: Lancaster University
Current Institution: Lancaster University
Date of Award: 2017
Availability of Full Text:
Access through EThOS:
Access through Institution:
Stroke is an important public health problem worldwide. It is a non-communicable disease of increasing importance in the ageing population. There are four major types of stroke: a) ischaemic stroke, b) primary intracerebral haemorrhage, c) subarachnoid haemorrhage and d) undetermined stroke (no computed tomography [CT], magnetic resonance imaging [MRI], autopsy or cerebral angiography). The risk factors and fatalities for stroke vary worldwide and stroke accounts for about 9.7% of all deaths worldwide. Unfortunately, most stroke deaths occur in the less developed countries, where stroke research is scanty. Patients who survive stroke will have a wide range of functional limitations that affect their daily activities. To date, only few reliable data are available for identifying the risk factors and understanding the stroke fatalities in the low- and middle-income countries. The effect of stroke on daily activities is also understudied. Measuring the effect of stroke on daily activities is important to evaluate the recovery process. Understanding the factors affecting daily activities post-stroke helps identify areas where intervention may benefit stroke survivors the most. In this thesis, I have worked on the questions that will improve my knowledge and understanding of the natural history of stroke in the state of Kelantan, Malaysia. In this study, we posed questions to subjects from the population of interest to reflect our results across the greater Kelantan population to develop a model of risk factors for stroke, models of case-fatalities and survival to compare the characteristics and outcomes of two main types of stroke, i.e. ischaemic stroke and haemorrhagic stroke, and lastly to assess the longitudinal change in functional health status using the Barthel Index post-stroke. This thesis includes four draft papers, in which several modelling data collection and data analysis strategies were applied to four datasets: one was provided by the hospitals, two were extracted by us and the final one was based on personal interviews with stroke survivors. The first draft paper is based on an observational study using data from the records offices of two major hospitals in Kelantan. In this paper, I analysed and modelled the risk factors for stroke using a case–control study design. This dataset contained individual-level variables (patient variables from hospitals) and area-level variables supplied by the Department of Statistics, Malaysia, and we utilised logistic regression to model the risk factors for stroke. In the results, we showed the non-linear relationship between age and odds for stroke and the interaction of age with sex in the model. In the second draft paper, we explored the important prognostic factors for in-hospital stroke fatalities. Using Cox proportional hazard regression, we found that the only two independent prognostic factors for stroke fatality in the hospitals were: a) stroke subtype and b) age. To further investigate the different prognostic effect of stroke subtype on admission and on fatality, we recruited consecutive in-hospital stroke patients. In the analysis, we performed Cox proportional hazard regression to quantify the odds of stroke fatality for: a) ischaemic stroke and b) haemorrhagic stroke. In this third draft paper, we showed the prognostic effect of stroke subtype on stroke fatality. In the fourth draft paper, we recruited stroke patients and interviewed them on three occasions. In this longitudinal assessment, we assessed the functional health status of stroke patients until 3 months after hospital discharge. I conducted all interviews and assessed the functional outcome using the well-known Barthel Index. Considering the longitudinal format of the data, we used the linear mixed effect model to model the rate of change of the Barthel Index at the three measurement occasions. We have identified several limitations in this PhD project and have taken several measures to minimize the biases caused by those limitations. The limitations include the need for us to do handsearching for data abstraction, potential informative censoring due to our study design and short follow-up times, limited generalizability of results, small sample sizes, missing observations, missing important variables (to be modelled as covariates), absence of residential coordinates and using data on arrival to Emergency department (no pre-arrival data). If the censoring mechanism provides significant information with time (T), numerical estimates from Kaplan-Meier and Cox proportional hazard regression are biased. The new knowledge stemming from the stroke modelling and outcome assessment developed and analysed in this thesis could help improve our understanding of stroke in Kelantan. The thesis will also improve our understanding and knowledge of the natural history of the disease, i.e. the progression from risk factors to outcome (fatality or functional residuals) after stroke. In conclusion, our data and the four draft papers written based on this PhD project have added new stroke data and knowledge on the progression of stroke, which is understated in the Malaysian and Asian population in general and in Kelantan specifically.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available