Health Promotion, Work Participation in Musculoskeletal Pain
Health Promotion, Work Participation in Musculoskeletal Pain
This study aimed to investigate the relationship between health promotion resources and work participation in a population reporting musculoskeletal pain. The work group represented 32% (n=2161) and was fairly equally distributed with 50.6% female (n=1094) and 49.4% men (n=1067), and with a mean age of 51.29 (SD= 9.68). The sick leave group represented 67% (n=4511) and consisted of 65.1% (n=2935) female and 34.9% (n=1576) men, with a mean age of 49.9 (SD= 9.95).
The distribution of MSP showed an equal pattern between the groups with shoulder pain (60%) in the working group and in the sick leave group (64%); neck pain was reported by 56% versus 60%; pain in lumbar regions was reported by 50% versus 54%; and pain in hips was reported by 35% versus 41%. The K-S test for normality was (D(2129) =.082, p<.001) for the work group and (D(4427) =.078, p<.001) for the sick leave group, which indicated deviation from normality. Descriptive results were presented in Table 1 with grouping the variable "Have you been on sick leave in the past 12 months?" (n=6672).
A Mann- Whitney U test was conducted to evaluate nine resources (physical exercise, EPQ-E and N, feeling strong, social activities, work support, work load and work control) if there were any differences between the work group and the sick leave group. A significant difference was found in levels of SRH between the work group and the sick leave group, p <.001. No differences were found between the two groups in levels of physical exercise, social activities or levels of EPQ-E. EPQ-N levels were significantly higher in the sick leave group than in the work group, p = <.001. Further, the work group, reported significantly higher levels of feeling strong than the sick leave group, p = <.001. The work group reported higher levels of work support than the sick leave group, p = <.001, and the work group reported higher levels of work load than the sick leave group, p = <.001. Moreover, the work group reported higher levels of work control than the sick leave group, p = <.001. This effect, however, was below the .03 criterion for a medium effect and could be considered a small effect. The total results were presented in Table 2 .
Variable relationships were assessed thorough chi-square statistics (phi) and showed significant gender difference between the groups, x(6672) = −.138, p = < .001. Only the significant variables that were seen as health promoting resources as SRH, feeling strong, work support, work load and work control were included for further analysis.
Three categorical variables (meaning, friends' support and friends' cohesion) were assessed with chi-square statistics (Phi for 2 × 2 Tables and Pearson). There was significant gender difference between the two groups, with more men in the work group, x(6672) = −.138, p = < .001. No significant difference was found between the work and sick leave groups in the meaning variable, x(3) = 2.935, ns.
The work group reported slightly more support at work (95.4%) than the sick leave group (95.1%) (x(1) = −.008, ns). However, participants in the sick leave group reported significantly more friends' cohesion (91%) than participants in the work group (89%)(x(1) = −.29, p <.005).
For further analysis, the variables EPQ-E, meaning, and friends support were not included because of non significance differences between groups. Friends' cohesion was not included because of the preliminary findings and a small difference. For constructing a model of prediction for the work group, only variables with statistically significant group and gender differences in means were presented in Figure 1.
(Enlarge Image)
Figure 1.
Work and sick leave group, female and men composite scores of means.
According to the aim of the study and preliminary findings, a logistic regression analysis was conducted to predict belonging to the work group using five variables (SRH, feeling strong, work support, work load, and work control) as independent associations. Age and gender were included. A test of the full model (forward LG method) against a constant only model was statistically significant, indicating that the independent associations are a set, reliably distinguished between the work group and sick leave group (chi-square = 408.264, p=000 with (df = 7)). Nagelkerke's (Rof .092) indicated a weak relationship between prediction and grouping. Prediction success overall was 68%, which was more than by chance. The model fit is acceptable (x(8) =10.973, p=.203). The Wald criterion (with acceptable S.E) demonstrated that only SRH made a significant contribution to prediction (p=.0001) of the work group. Collinarity statistics were satisfied, and Leverage values were satisfyingly low.
Due to the subtle differences between each step, the main outcomes were reported in the last step. EXP(B) indicated that when SRH raised with one unit, the odds to be in the work group increased by 66%. EXP(B) for work support, work control, and work load indicated that when these variables increased by one unit each, the odds to be in the work group increased by 29%, 21%, and 16%, respectively. EXP(B) for the feeling of being strong indicated that when it increased by one unit, the odds to be in the work group increased by 12%. According to gender, we see that females had a statistically significant lower probability (B= −.53) to be in the work group than men, with an odds of 41%. The total results were presented in Table 3 .
Results
This study aimed to investigate the relationship between health promotion resources and work participation in a population reporting musculoskeletal pain. The work group represented 32% (n=2161) and was fairly equally distributed with 50.6% female (n=1094) and 49.4% men (n=1067), and with a mean age of 51.29 (SD= 9.68). The sick leave group represented 67% (n=4511) and consisted of 65.1% (n=2935) female and 34.9% (n=1576) men, with a mean age of 49.9 (SD= 9.95).
The distribution of MSP showed an equal pattern between the groups with shoulder pain (60%) in the working group and in the sick leave group (64%); neck pain was reported by 56% versus 60%; pain in lumbar regions was reported by 50% versus 54%; and pain in hips was reported by 35% versus 41%. The K-S test for normality was (D(2129) =.082, p<.001) for the work group and (D(4427) =.078, p<.001) for the sick leave group, which indicated deviation from normality. Descriptive results were presented in Table 1 with grouping the variable "Have you been on sick leave in the past 12 months?" (n=6672).
Work Group and Sick Leave Group Comparisons
A Mann- Whitney U test was conducted to evaluate nine resources (physical exercise, EPQ-E and N, feeling strong, social activities, work support, work load and work control) if there were any differences between the work group and the sick leave group. A significant difference was found in levels of SRH between the work group and the sick leave group, p <.001. No differences were found between the two groups in levels of physical exercise, social activities or levels of EPQ-E. EPQ-N levels were significantly higher in the sick leave group than in the work group, p = <.001. Further, the work group, reported significantly higher levels of feeling strong than the sick leave group, p = <.001. The work group reported higher levels of work support than the sick leave group, p = <.001, and the work group reported higher levels of work load than the sick leave group, p = <.001. Moreover, the work group reported higher levels of work control than the sick leave group, p = <.001. This effect, however, was below the .03 criterion for a medium effect and could be considered a small effect. The total results were presented in Table 2 .
Variable relationships were assessed thorough chi-square statistics (phi) and showed significant gender difference between the groups, x(6672) = −.138, p = < .001. Only the significant variables that were seen as health promoting resources as SRH, feeling strong, work support, work load and work control were included for further analysis.
Meaning, Friends' Support and Cohesion
Three categorical variables (meaning, friends' support and friends' cohesion) were assessed with chi-square statistics (Phi for 2 × 2 Tables and Pearson). There was significant gender difference between the two groups, with more men in the work group, x(6672) = −.138, p = < .001. No significant difference was found between the work and sick leave groups in the meaning variable, x(3) = 2.935, ns.
The work group reported slightly more support at work (95.4%) than the sick leave group (95.1%) (x(1) = −.008, ns). However, participants in the sick leave group reported significantly more friends' cohesion (91%) than participants in the work group (89%)(x(1) = −.29, p <.005).
For further analysis, the variables EPQ-E, meaning, and friends support were not included because of non significance differences between groups. Friends' cohesion was not included because of the preliminary findings and a small difference. For constructing a model of prediction for the work group, only variables with statistically significant group and gender differences in means were presented in Figure 1.
(Enlarge Image)
Figure 1.
Work and sick leave group, female and men composite scores of means.
Model of Work With Logistic Regression Analysis
According to the aim of the study and preliminary findings, a logistic regression analysis was conducted to predict belonging to the work group using five variables (SRH, feeling strong, work support, work load, and work control) as independent associations. Age and gender were included. A test of the full model (forward LG method) against a constant only model was statistically significant, indicating that the independent associations are a set, reliably distinguished between the work group and sick leave group (chi-square = 408.264, p=000 with (df = 7)). Nagelkerke's (Rof .092) indicated a weak relationship between prediction and grouping. Prediction success overall was 68%, which was more than by chance. The model fit is acceptable (x(8) =10.973, p=.203). The Wald criterion (with acceptable S.E) demonstrated that only SRH made a significant contribution to prediction (p=.0001) of the work group. Collinarity statistics were satisfied, and Leverage values were satisfyingly low.
Due to the subtle differences between each step, the main outcomes were reported in the last step. EXP(B) indicated that when SRH raised with one unit, the odds to be in the work group increased by 66%. EXP(B) for work support, work control, and work load indicated that when these variables increased by one unit each, the odds to be in the work group increased by 29%, 21%, and 16%, respectively. EXP(B) for the feeling of being strong indicated that when it increased by one unit, the odds to be in the work group increased by 12%. According to gender, we see that females had a statistically significant lower probability (B= −.53) to be in the work group than men, with an odds of 41%. The total results were presented in Table 3 .