6.3.1 Associations between lifestyle risk factors and work ability and sick leave in the general working population (Paper I and Paper III)
The multiple logistic regression analysis showed that individual lifestyle risk factors had a statistically significantly association with low work ability (Figure 7). Further, low work ability was associated with a higher lifestyle risk index score.
Figure 7. Statistically significant associations between lifestyle risk factors and WAS (Paper I)
*CI= Confidence interval
Figure 7 shows that obesity was the factor which was most strongly associated with low work ability, with an OR of 1.5 (95%CI 1.3, 1.7). Low physical activity (OR 1.4; 95%CI 1.2, 1.6), current smoking (OR 1.3; 95%CI 1.2, 1.5), former smoking (OR 1.2; 95%CI 1.1, 1.4)
0 0,5 1 1,5 2 2,5 3 3,5
Healthy Average Unhealthy High physical activity Low physical activity Normal weight Underweight Overweight Obesity Never Former Current Low-risk score Moderate-risk score High-risk score Very high-risk score
Diet Physical
activity BMI (kg/m²) Smoking status Lifestyle risk index
OR with 95% CI*
Paper I
Line indicating 95% CI Odds ratio
___
43
and unhealthy diet (OR 1.3; 95%CI 1.02, 1.5) were also associated with low work ability.
No significant association was observed between average diet and work ability, or between overweight and work ability. Detailed results can be found in Paper I, Table 3.
The results in Figure 7 have been mutually adjusted for the individual lifestyle variables and background variables (sex, age, educational level and occupational group). However, the adjusted OR were not substantially different from the crude OR, indicating that the associations were independent of the background variables.
At follow-up, the multinomial logistic regression analysis showed that unhealthy diet, moderate or low physical activity and former and current smoking were associated with low work ability. In the case of sick leave, overweight and former and current smoking were associated with short-term and long-term sick leave. Obesity was also associated with long-term sick leave.
Furthermore, it was observed a higher odds of low work ability and sick leave by higher lifestyle risk index score.
All of the observed associations were adjusted for age, sex, educational level and dependent variables (work ability score and sick leave) at baseline.
6.3.2 Associations between lifestyle risk factors and work ability and sick leave among persons with physician-diagnosed asthma and other non-communicable diseases (Paper II and Paper III)
The cross-sectional study population used in Paper II comprised a stratified analysis of persons with physician-diagnosed asthma (n=1 110; 11%) and a control group including subjects without physician-diagnosed asthma (n=9 245). The study population characteristics were similar for the two groups, but more persons with physician-diagnosed asthma (41%) reported being on sick leave in the past 12 months than persons without asthma (31%). Also, more persons with physician-diagnosed asthma (18%) reported low work ability, than persons without (13%). The percentages were similar for
___
44
all the lifestyle risk factors with low physical activity (46% for persons with physician-diagnosed asthma and 49% for persons without) and obesity (20% for persons with asthma and 14% for persons without) being the most dissimilar.
Of the individual lifestyle risk factors, only low physical activity had a statistically significant association with low work ability among persons with physician-diagnosed asthma. Further, high and very high lifestyle risk factor scores had a statistically significant association with low work ability among persons with physician-diagnosed asthma.
The multiple logistic regression analysis showed that obesity and former and current smoking were associated with sick leave among persons with physician-diagnosed asthma. Moreover, moderate, high and very high lifestyle risk factor scores were associated with sick leave in persons with physician-diagnosed asthma. The inclusion of the interaction term in the regression model confirmed the association between moderate and high lifestyle risk scores (Table 4), obesity, smoking and sick leave. This may indicate that asthma is a potential effect modifier between multiple lifestyle risk factors and sick leave.
___
45 Table 4. Example of interaction effects when studying the association between lifestyle risk factors and sick leave (Paper II).Model 2 comprises a stratified analysis (Table 2 in Paper II), while Model 3 includes interaction terms. All associations are statistically significant with a p-value <0.05.
†Illustration showing how the interaction terms were incorporated into the multiple logistic regression model.
𝑥𝑥1=𝑚𝑚𝑚𝑚𝑐𝑐𝑐𝑐𝑐𝑐𝑎𝑎𝑎𝑎𝑐𝑐 𝑐𝑐𝑟𝑟𝑎𝑎𝑟𝑟 𝑎𝑎𝑐𝑐𝑚𝑚𝑐𝑐𝑐𝑐; 𝑥𝑥2=ℎ𝑟𝑟𝑖𝑖ℎ 𝑐𝑐𝑟𝑟𝑎𝑎𝑟𝑟 𝑎𝑎𝑐𝑐𝑚𝑚𝑐𝑐𝑐𝑐; 𝑥𝑥3=𝑣𝑣𝑐𝑐𝑐𝑐𝑣𝑣 ℎ𝑟𝑟𝑖𝑖ℎ 𝑐𝑐𝑟𝑟𝑎𝑎𝑟𝑟 𝑎𝑎𝑐𝑐𝑚𝑚𝑐𝑐𝑐𝑐; 𝑥𝑥4=𝑎𝑎𝑎𝑎𝑎𝑎ℎ𝑚𝑚𝑎𝑎; 𝑥𝑥5 𝑎𝑎𝑚𝑚 𝑥𝑥10
=𝑎𝑎𝑖𝑖𝑐𝑐,𝑎𝑎𝑐𝑐𝑥𝑥,𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑎𝑎𝑎𝑎𝑟𝑟𝑚𝑚𝑒𝑒𝑎𝑎𝑒𝑒 𝑒𝑒𝑐𝑐𝑣𝑣𝑐𝑐𝑒𝑒 𝑎𝑎𝑒𝑒𝑐𝑐 𝑚𝑚𝑎𝑎ℎ𝑐𝑐𝑐𝑐 𝑐𝑐ℎ𝑐𝑐𝑚𝑚𝑒𝑒𝑟𝑟𝑐𝑐 𝑒𝑒𝑐𝑐𝑒𝑒𝑖𝑖 𝑐𝑐𝑟𝑟𝑎𝑎𝑐𝑐𝑎𝑎𝑎𝑎𝑐𝑐
𝑒𝑒𝑚𝑚𝑖𝑖 �𝑝𝑝−1𝑝𝑝 �=𝛽𝛽̂0+𝛽𝛽̂1∗ 𝑥𝑥1+𝛽𝛽̂2∗ 𝑥𝑥2+𝛽𝛽̂3∗ 𝑥𝑥3+𝛽𝛽̂4∗ 𝑥𝑥4+𝛽𝛽̂5∗ 𝑥𝑥1∗ 𝑥𝑥4 +𝛽𝛽̂6∗ 𝑥𝑥2∗ 𝑥𝑥4+𝛽𝛽̂7∗ 𝑥𝑥3∗ 𝑥𝑥4+ 𝛽𝛽̂8∗ 𝑥𝑥5… . +𝛽𝛽̂13∗ 𝑥𝑥10
*Model 2 and 3 have been adjusted for age, sex, educational level and other chronic lung diseases.
**The reference group consisted of those who responded negative to the question of whether they had physician-diagnosed asthma
In Paper III, the sample was stratified into two groups with persons with NCD groups or illnesses and persons not reporting having any of the disease groups or illnesses. Initially, each disease group had few cases, so we decided to expand the groups to increase statistical strength. The respiratory disease group was expanded to include asthma, COPD and other chronic respiratory diseases. Further, the CVD group was expanded to include persons with diabetes. The multiple logistic regression analysis revealed few associations,
Model 1-Crude
Model 2* – Stratified
analysis Model 3** – Interaction
terms included† All participants
(n 10 355) No physician- diagnosed asthma OR (95% CI)
Physician- diagnosed asthma OR (95% CI)
Physician-diagnosed asthma*lifestyle risk factors
OR (95% CI) Lifestyle risk index
score
Moderate risk score 1.3 (1.2, 1.4) 1.2 (1.1, 1.4) 1.7 (1.2, 2.4) 1.4 (1.02, 2.01) High risk score 1.5 (1.3, 1.6) 1.4 (1.2, 1.6) 2.1 (1.4, 3.0) 1.6 (1.1, 2.3) Very high risk score 1.8 (1.5, 2.1) 1.8 (1.5, 2.1) 2.6 (1.6, 4.2) 1.6 (0.97, 2.7)
___
46
and only former smoking had a statistically significant association with low work ability among persons with mental illness (OR 0.57; 95%CI 0.37, 0.88). Moreover, current smoking was associated with sick leave among persons with CVD, diabetes or mental illness. Finally, a moderate lifestyle risk score was associated with sick leave among persons with respiratory diseases (OR 1.51; 95%CI 1.01, 2.24). The other analyses did not reach statistically significant levels.
___
47