A research project on health and work focusing on health-promoting factors among farmers was designed in 1989. A study cohort with farmers and rural referent was established. All male farmers born between 1930 and 1949 living in nine selected rural municipalities in Sweden were identified from the Swedish National Farm Register. Occupational activity in farming was thoroughly checked. For each farmer a rural referent, matched by age, sex, and residential area, was identified in the National Population Register. The referents were to be occupationally active, in work other than farming, according to the most recent census.
Altogether 2350 men (1220 farmers and 1130 non-farmers) met the criteria and were included in the cohort. They were invited to an extensive health survey in 1990–91 including questionnaires, interviews, physical examinations, and laboratory test. The participation rate in this baseline survey was 75.8% [24
The entire cohort was invited to a follow-up survey in 2002–03. The participation rate at this examination was 67.6%. In total 1405 men participated in both surveys, that is 59.8% of the initial cohort and 63.6% of those still alive at the time for survey 2.
The health examinations were performed by specially trained teams of physicians and nurses traveling out to the various areas. The surveys were done solely for research purposes and hence were not part of any ordinary or local health program.
The study was approved by the Research Ethics Committee at the Karolinska Institute in Stockholm, Sweden and by the Regional Ethics Board, Uppsala, Sweden. All men who participated in the health surveys gave their informed consent.
Central obesity was defined as waist hip ratio ≥ 1. Waist and hip measurements were taken at both surveys with a tape measure at the level of the umbilicus and at the widest part of the hips with the participants dressed in light wear.
Dairy fat intake
Specific food choices were assessed by a 15-item questionnaire answered at the two surveys. Dairy fat intake was assessed by combining three questions relevant to Swedish eating habits, namely usual spread on sandwiches (butter, low fat margarine, or no fat), type of milk normally consumed (non- homogenized farm milk, full fat milk with 3.0% fat, semi-skimmed with 1.5% fat, or skimmed milk with 0.5% fat) and intake of whipping cream, also in sauces (daily, sometimes during the week, or seldom/never). Low consumption of dairy fat was defined as no butter, low fat milk (1.5% fat or less), and seldom or never intake of cream. High consumption of dairy fat was defined as butter as spread, full fat milk, and intake of whipping cream daily or several times a week. All other combinations of spread, milk, and cream consumption were defined as medium dairy fat intake.
In the food questionnaire one item assessed intake of fruit and berries (nearly every day, several times a week, once a week, or seldom/never) and one item assessed intake of vegetables, legumes or root vegetables, except potatoes, with the same alternatives as for fruit and berries. These items were combined into one variable for fruit and vegetables dichotomized as daily versus less than daily intake.
Weight and height were measured with standard procedures and body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared. Tobacco use, alcohol consumption, and physical activity were assessed in detail in structured interviews. Smoking habits were for these analyses dichotomized as current daily smoking versus no smoking. Alcohol consumption was assessed as frequency of use and amount of alcoholic beverages consumed on each occasion. Average alcohol intake was then computed as grams of pure alcohol consumed per week. Physical activity during leisure time was assessed on a four-grade scale from sedentary to vigorously active. Educational level was classified as mandatory, vocational school, secondary school, college, or university according to self-report in questionnaire.
The internal non-response rate was low for all included variables. Data analyses were performed using SPSS® version 14.0. A significance level of 0.05 was considered to indicate statistical significance and all tests were two-tailed. Analyses of association were performed with logistic regression in multiple analyses adjusting for confounder variables. The results are presented as odds ratios (OR) with 95% confidence intervals (95% CI). Confounder variables were categorized and eliminated from the model in a backward selection of non-significant variables requiring a p-value below 0.10 for the variable to be kept in the model.