C International Epidemiological Association 1998

Printed in Great Britain

InUmalumalJoumal ofEpidemiology 1998:27:1019-1025

Socioeconomic inequalities and disability pension in middle-aged men Nils-Ove Mansson,a Lennart Rastam,a Karl-Fredrik Eriksson b and Bo Israelssonc

Background The issue of inequalities in health has generated much discussion and socioeconomic status is considered an important variable in studies of health. It is frequently used in epidemiological studies, either as a possible risk factor or a confounder and the aim of this study was to analyse the relation between socioeconomic status and risk of disability pension. Methods Five complete birth year cohorts of middle-aged male residents in Malm 6 were invited to a health survey and 5782 with complete data constituted the cohort in this prospective study. Each subject was followed for approximately 11 years and nationwide Swedish data registers were used for surveillance. Results Among the 715 men (12%), granted disability pension during follow-up, three groups were distinguished. The cumulative incidence of disability pension among blue collar workers was 17% and among lower and higher level white collar workers, 11% and 6% respectively. With simultaneous adjustment for biological risk factors and job conditions, the relative risk for being granted a disability pension (using higher level white collar workers as reference) was 2.5 among blue collar workers and 1.6 among lower level white collar workers. Conclusions Socioeconomic status, as defined by occupation, is a risk factor for being granted disability pension even after adjusting for work conditions and other risk factors for disease. Keywords Disability pension, early retirement, occupation, risk factors, screening, socioeconomic status Accepted 27 April 1998

The influence of socioeconomic status on morbidity and mortality is well known and the issue of inequalities in health has been widely discussed.1"10 Social class is considered an important variable in studies of health and is frequently used in epidemiological studies, either as a possible risk factor or a confounder.11"13 In three Swedish reports on disability pension, attempts have been made to describe the socioeconomic background of the pensioners. People with low social status in terms of education and occupation were found to be overrepresented among the disability pensioners.14"16 The issue of confounding is of great interest when various risk factors, e.g. alcohol consumption, obesity and socioeconomic status, are discussed in this context. In previous studies we have reported on aspects of risk of disability pension17'18 and the question has been raised whether or not some of the identified risk relations could be explained by confounding from socioeconomic status.

Departments of "Community Medldne, bAngk>logy and cCardiology, Lund University, Malmd University Hospital, Malm5, Sweden. Reprint requests to: Dr Nils-Ove Mansson, Department of Community Medicine, Malm5 University Hospital, S-205 02 Malm6, Sweden.

It is important to understand the factors determining differences in health and the explanation must include not only the understanding of biological factors but also an insight into the socioeconomic influence on mortality, morbidity and work disability. The aim of this study was, therefore, to analyse the independent relation between socioeconomic status as indicated by occupation and risk of disability pension.

Materials and Methods In Sweden, subjects aged 16-64 years, can be granted disability pension if their working capacity is impaired by at least 50% (from July 1993: 25%) due to disease. A temporary pension is granted if the impairment is not regarded permanent. Full disability pension, including national supplementary pension (ATP), amounts to approximately 65% of the annual income. Supplementary benefits, e.g. housing supplement, constitutes in many cases an additional compensation. Applications for disability pension are made by either the patient or the social insurance office and are handled by special boards at the soda! insurance offices. A dorter's certificate accompanying the application is compulsory.

1019

1020

INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

Subjects In 1974-1978, five complete birth year cohorts of middle-aged male residents of Malmo, Sweden (251 431 inhabitants, 1 January 1974), were invited to participate in a screening programme. 1 9 Of 7697 invited men, born 1926-1930, 5932 (77.1%) participated. The study cohort has been presented in detail elsewhere. 2 The project was approved by the ethics committee of the Medical Faculty, Lund University. The original protocol was to screen men at age 48; men born in 1926 should have been included in 1974, those bom in 1927 in 1975 and so on. However, due to temporary deviations from this protocol, there was a slight variation in age at screening and mean age (±SD) was 48.1 ± 0.7 years. Included in this study were, however, only those who, without a previous retirement with disability pension, participated in the screening (n = 5798). Each subject was followed from inclusion, defined by the date of examination, until death or the end of the calendar year when he turned 58, i.e. a total study period of approximately 11 years. The Nordic Classification of Occupation (Nordisk yrkesklassificering, NYK) was used at screening for classification according to occupational category, NYK-74 from the start of the screening up until 27 September 1978 and NYK-78 thereafter. 21 The NYK code was, prior to the analysis, re-coded in accordance with the occupational classification used in the Population and Housing Census 1980. 22 In this classification, a code for socioeconomic distribution is included, which categorizes the subject in accordance with the qualification level (not only the education required) of his occupational position. The categories used in this study were (a) unskilled blue collar workers; (b) skilled blue collar workers; (c) lower level white collar workers; (d) intermediate level white collar workers; (e) higher level white collar workers; (f) self-employed (farmers and self-employed wholesale and retail traders). Subjects lacking an occupational code were merged to form a separate category.

Screening methods Data on fasting blood glucose (B-Glucose), serum gamma glutamyl transferase (S-GT), serum total cholesterol (S-Chol), body height, body weight, blood pressure, alcohol habits, smoking habits and a number of questions concerning occupational conditions were obtained during the initial health examination. The B-glucose, S-GT and S-Chol were measured in venous blood samples and the analyses were performed using standard methods at the Department of Clinical Chemistry, Malmo University Hospital.

Supine systolic and diastolic blood pressure was measured after 10 min rest in the right arm to the nearest 5 mmHg. Hypertension was defined as systolic blood pressure >160 mmHg and/or diastolic blood pressure 2*95 mmHg and/or current antihypertensive medication. Weight (to the nearest 0.1 kg) and height (to the nearest cm) were measured using calibrated scales. Body mass index (BMI), defined as weight (kg) divided by height squared (m ), was used to estimate the degree of over/ underweight. It was classified as suggested by Bray and as used on the national level by Statistics Sweden in the Surveys of Living Conditions. 23 ' 24 Thus, underweight was defined as BMI