The present study was based on the HPI Health Profile Institute cohort (HPI, Stockholm, Sweden. www.hpihealth.se), containing data from Health Profile Assessments (HPAs) carried out by employees at companies connected to occupational or other healthcare services. The HPA consists of a questionnaire including physical activity pattern, lifestyle factors and perceived health, a dialogue with a HPA coach, and a physical examination. All data were subsequently stored in the central database. The test protocol, methods used and education of HPA coaches follows a standardized procedure and has been the same since the start of HPA in the middle of the 1970s. Participation was voluntary and free of charge for the employee. Although data were available in the database since the 1980s, we based the present analyses on data from January 2014 to November 2019 to get a current analysis of the working population (n=107,170). After exclusion of individuals with missing data for occupational group (n=34,294) and individuals <18 and >75 years of age (n=21), the final cross-sectional sample consisted of 72,855 participants (41% women).
Classification of occupational groups
Occupation was reported by the participants and documented into the HPI database coded as a Swedish Standard Classification of Occupation (SSYK) (12) number. SSYK is a categorization of occupations based on the international Standard Classification of Occupation (ISCO) (13). Each occupation is labelled and defined by a four-digit code, which refers to the job performed (defined as the tasks and duties of an employee) and the degree of qualification needed (defined as the knowledge and expertise needed to perform the tasks and duties of an occupation) for each occupation. The four-digit codes contained information on different levels; first digit defines Major group of occupation (e.g. 5=Service, care and shop sales), second digit defines Sub-major group (53=Personal care occupations), third digit refers to Minor group (531=Child minders and teacher aids) and fourth digit Unit group (5311=Child care occupations). Ten major groups of occupations were defined; 1=Managers, 2=Professionals, 3=Associate professionals, 4=Administrative and customer service, 5=Service, care and shop sales, 6=Agricultural and forestry, 7=Building and manufacturing, 8=Mechanical manufacturing and transport, 9=Elementary occupations and 10=Military. The first nine were included in the present analyses and Military were excluded due to low N. As there were further heterogeneity within these nine occupational groups, for example with regard to contact with clients/patients/students (which may induce a different psychosocial working situation) or occupational physical activity pattern, sub-major groups were identified a priory to the analyses based on this heterogeneity. Sub-major groups Healthcare (SSYK 22) and Education (SSYK 23) were defined as contact workers, while Science and engineering (SSYK 21) and All otherprofessionals (SSYK 24) were defined as non-contact workers. Personalcareoccupations (SSYK 53) was defined as contact workers, and Serviceandshopsales (SSYK 51-52) as non-contact workers. Mechanicalmanufacturing (SSYK 81-82) was defined as physically demanding and Transport (SSYK 83) as less physically demanding. In the tables and figures, these were labelled as a sub-code based on the Major group digit (2.1, 2.2 etc), rather than their SSYK-code. Occupational groups were aggregated into white- (Major group 1-5) and blue-collar (major group 6-9) occupations, and further by skill-level within white- and blue-collar occupations; high-skilled white-collar (major group 1-3), low-skilled white-collar (major group 4-5), high-skilled blue-collar (6-7) and low-skilled blur-collar (8-9) (14,15). Description of type and numbers of workers on Minor group level is presented in supplementary 1a and 1b.
Physical activity pattern
Exercise, physical working situation, sitting at work and sitting in leisure were self-reported through the following statements; I exercise for the purpose of maintaining/improving my physical fitness, health and well-being…with the alternatives Never, Sometimes, 1–2 times/week, 3–5 times/week, or At least 6 times/week; My physical work situation… Sitting with some movement, Physically active, Occasionally physically demanding, or Occasionally very physically demanding; I sit at work… and I sit in leisure time… Almost all the time, 75% of the time, 50% of the time, 25% of the time and Almost never.
Physical examination indicators
Body mass and height were obtained with standard measures in light-weighing clothes, and BMI was subsequently calculated (kg/m2). Systolic and diastolic blood pressure (mmHg) were measured in the right arm using the standard auscultatory method after 20 minutes of seated resting. Cardiorespiratory fitness was assessed as estimated VO2max, expressed in ml⸳min−1⸳kg−1, using the submaximal Åstrand cycle test (16). The Åstrand test has been validated against directly measured VO2max during treadmill running in an adult population with non-significant mean differences on group level (−0.07 L⸳min−1, 95% CI −0.21 to 0.06) and with an absolute error and coefficient of variance similar to other submaximal tests (SEE=0.48 L⸳min−1, CV=18.1%) (17).
Perceived health and other lifestyle-related indicators
Perceived health and diet were self-reported through the statements I perceive my physical and mental health as... and I consider my diet, regarding both meal frequency and nutritional content to be…with the alternatives Very poor, Poor, Neither good or bad, Good, or Very good. Smoking habits and civil status derived by the statements I smoke… with the alternatives At least 20 cig/day, 11–19 cig/day, 1–10 cig/day, Occasionally, or Never; and Civil status… Living alone, Living alone with children, Living together with someone, Living together with someone and with children.
Internal and external validity analyses
Internal validity analysis; missing data was low for exercise (0.3%), civil status (3%), smoking (0.3%), BMI (1.1%), blood pressure (1%), perceived health (0.4%) and diet (0.3%), while it was higher for cardiorespiratory fitness (22.2%), physical work situation (17.5%), sitting at work (22.4%) and sitting in leisure (22.6%). Comparing participants with missing data for the four latter variables and those included in the analyses revealed statistically significant but small differences in other central variables (Supplementary table 2). External validity analysis; sex- and age-distribution in major and sub-major groups in the present study were compared with national register data from 2014 to 2018 (Statistics Sweden: www.scb.se). The proportion of women in different major groups in the HPA data was similar to national register data, with only three occupational groups having more than a 5% difference; Health professionals (79% vs 67%), Personal care workers (83% vs 90%) and Elementary occupations (54% vs 60%) (Supplementary table 3). Only two occupational groups in the present data had a difference in mean age of greater than two years; Service and sales workers (37y vs 43y) and Elementary occupations (38y vs 44y).
Chi-square test was used in the non-response analysis of internal missing data. For external validity, the proportion women and mean age was compared numerically. Independent t-tests were used to test for differences between continuous variables between high and low-skilled occupations (Table 1a). Ten health risk were identified and dichotomized according to alternatives of reply or conventional cut-off values for increased health risk; No regular exercise (Never or Sometimes), Physically demanding work (Occasionally physically demanding or Occasionally very physically demanding), High sitting at work (Almost all the time or 75% of the time), High sitting in leisure (Almost all the time or 75% of the time), Poor perceived health (Very poor or poor), Perceived poor diet (Very poor or poor), Daily smoker (≥1 cig/day), Obesity (BMI>30 kg/m2), Hypertension (diastolic BP≥90 and systolic BP≥140 mmHg or using self-reported blood pressure medicine) and Low cardiorespiratory fitness (estimated VO2max <32 ml⸳min−1⸳kg−1). A total health risk indicator was derived by adding the single risk indicators, ranging from zero to seven possible risk factors for each participant (excluding physically demanding work, high sitting at work and high sitting in leisure due to high internal missing), and further dichotomized into ≥3 risk indicators or fewer for clustered risk analyses. Logistic regression modelling was used to study sex- and age-adjusted odds ratios (OR) with 95% CI confidence intervals (CI) for a) the ten different risk indicators and b) the clustered risk variable, in relation to the major and sub-major occupational groups (Managers set as reference), as well as c) the clustered risk variable in relation to grouping of occupations into blue collar and white collar, high-skilled and low-skilled (white high-skilled set as reference). Data were handled and analysed using R 3.6.3 (R Core Team, 2018) and the Tidyverse (18), the jtools (19) and the finalfit (20) packages.