Background Epidemiological methods in public health are used to study ill-health and hazardous behaviours in order to take health-promoting measures, e.g. intervention. Data can, as in this dissertation, be based on general purpose sample surveys. Alternative ways of collecting data are registers or medical surveys. One advantage of survey data compared to registers is the large amount of background variables that can be collected with good quality. Another advantage compared to many medical surveys is that the collection can be made before the event and that the survey can be made prospectively. Moreover, there are few possibilities of gathering information about the distribution of risk factors in the population or in subgroups of the population without using large surveys. Many risk factors, such as smoking, lack of physical activity, overweight/obesity and hypertension, play an intermediate role and can be considered both as outcomes and as risk factors, but in different studies with different perspectives. For example, one may study whether low socio-economic status (SES) is related to smoking and hypertension, but one may also study whether smoking and hypertension are related to ill-health or increased mortality. This thesis contains studies with both perspectives. As an indicator of SES, we use both attained level of education and occupational class. It is important to take confounding factors into account when trying to establish a relation between risk factors and outcome variables. It is wellknown that sex, age and country of birth are significant confounders. It has also been established that the foreign-born population generally has poorer health than native Swedes, so that this is a crucial factor in its own right. The foreign-born population is large, and consists primarily of young and middle-aged people, today about 11%. As they are ageing their need for health care will gradually increase.
Objective The main purpose of the five studies is to analyse the relationship between certain risk factors - lifestyle, overweight/obesity, country of birth and socio-economic status - and the outcome variables, long-term illness, medical treatment, poor health status, mortality, suicide and body mass index (BMI), after adjusting for some important background variables.
Data The studies are based on data from the Swedish Survey of Living Conditions (ULF). Each study consists of several (2-7) merged random samples and is limited to the ages 16-74 at the time of the interview. The samples have been followed concerning mortality up to 31 December 1996. Country of birth is defined as born in Sweden or born abroad. The latter group is considered either as a single group or is divided into subgroups.
Method The statistical methods used are logistic regression, proportional hazards regression and generalised estimating equations. Many of the methods are closely related to each other and will be compared empirically. For completeness, Poisson regression and intensity regression are also treated. All but one of the outcome variables are dichotomous and the independent variables are generally categorised. The SAS™ software was used in all studies.
Results and Discussion The studied self-reported risk factors are associated with ill-health irrespective of the outcome factor. The results are unambiguous, with some differences between men and women. A lack of physical activity is associated with a high risk of morbidity/mortality. Therefore, a change in physical activity behaviour should result in smaller health care expenses for society and a large health benefit for the individual at a small cost. Foreign-born people consistently report more morbidity than Swedes, but only Finnish men in Sweden have an excess mortality. These facts must be taken into consideration when planning health care for the future. In addition, there are clear educational differences in morbidity and mortality in Sweden, the lower education, the poorer the health. Furthermore, the risk ratio (RR) in mortality for unskilled/skilled blue-collar workers is also significantly increased (RR = 1.31 and RR = 1.44 for men and women, respectively) compared with middle/upper class employees (RR = 1). An even sharper demarcation regarding ill-health goes, as expected, between those who are employed and those who are unemployed or have a sickness pension. Those with a low SES, defined as unskilled/skilled blue-collar workers or those with only a compulsory school education, also have a more unfavourable risk factor profile than those with a high SES or higher education.

 Key words: self-reported data; morbidity; mortality; risk factors; BMI; lifestyle; hypertension; country of birth; SES; education; survey; cross-sectional; longitudinal. 

Close this Window

Last update: 1998-12-07 / MC