Climate Related Mortality and Morbidity in Scotland: Modelling Time Series of Counts

Gemmell, Islay M (2000) Climate Related Mortality and Morbidity in Scotland: Modelling Time Series of Counts. PhD thesis, University of Glasgow.

Abstract

Earlier research has demonstrated that excess winter mortality is greater in the countries of the United Kingdom than in those on comparable latitudes elsewhere in Europe. The purpose of this thesis was to provide an up-to-date analysis of excess winter mortality in Scotland. This involved exploring the relationships between mortality, morbidity (as reflected in rates of emergency hospital admissions), climate, influenza epidemics, and socio-demographic variables. The majority of the analysis was concerned with temporal relationships between these variables, however, latterly spatial relationships were also considered. Chapter 1 reviews the literature in support of seasonal patterns in health and assesses the merits of the various statistical techniques that have been used to demonstrate these patterns. Much of the previous analyses have used simple descriptive statistical methods with few acknowledging the Poisson time series nature of the data. In chapter 2 the seasonal pattern of mortality and morbidity from three main disease groups was described using a generalised linear model with Poisson errors incorporating a cosine term. The method was used to analyse the seasonal pattern by sex, age group, social class, deprivation category and health board. In chapter 3 the effect of climate on mortality and morbidity is explored. This chapter is chiefly concerned with the comparison of possible methods of analysis. Firstly the problems with summary methods are demonstrated before the principles of time series methodology are introduced. The final comparison involves three methods, ARIMA time series methods, Poisson regression and Zeger's method. Zeger's method is as a time series regression method for Poisson data. The methods are compared by assessing the effect of temperature on weekly deaths from respiratory disease. Examination of the residuals and the standard errors of the model coefficients reveal that Zeger's method is the most appropriate for this type of analysis. Zeger's method is used in Chapter 4 to assess the relationship between temperature and mortality and morbidity in more detail, by considering the effects of age, socio-economic deprivation and city of residence. This chapter also includes a detailed examination of the effects on mortality of a variety of different temperature patterns. In chapter 5 the spatial aspect of the data is included in the analysis. Space-time variations in emergency admissions for respiratory disease are assessed at various levels of aggregation. Overall there is no clear evidence of space-time patterns in emergency respiratory admissions over the time period, however spatial relationships are demonstrated. Finally, methods which account for spatial autocorrelation are used in an analysis of the relationship between emergency admissions and socio-economic deprivation in Glasgow. This analysis demonstrates, as with the previous temporal analysis, that if autocorrelation exists it is vital to account for this in any modelling procedure. Chapter 6 provides a summary of the main findings of the analysis in terms of both the epidemiological results and the methodological concerns. The limitations of the study concerning problems associated with the use of routinely collected data are also recognised. The thesis has demonstrated that seasonal patterns in mortality and morbidity are still a significant public health problem in Scotland and that Zeger's method is the most appropriate method to use when assessing the direct relationship between climate and ill health.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Additional Information: Adviser: Marian Scott
• Keywords: Statistics
• Date of Award: 2000
• Depositing User: Enlighten Team
• Unique ID: glathesis:2000-76466
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 19 Nov 2019 14:18
• Last Modified: 19 Nov 2019 14:18
• URI: https://theses.gla.ac.uk/id/eprint/76466