Intervention effects (job rotation) on the number of working hours lost due to sick leave (mean and standard deviation) during 12-month follow-up of participants (n=491). Results of the linear mixed-effects models (MIXED). [95% CI=95% confidence interval.]
| Primary outcome (working hours lost due to sick leave per month)* | Unadjusted within-group mean differences (baseline minus follow-up) | Adjusted between-group mean differences (intervention minus control) | |||
|---|---|---|---|---|---|
| Intervention (n=266)† | Control (n=225)† | Intervention (n=266)† | Control (n=225)† | ||
| Intention-to-treat analysis | |||||
| Baseline | 6.7 (18.2) | 11.4 (26.5) | |||
| 3 months | 9.0 (19.1) | 11.3 (20.8) | 2.3 (−0.5 to 5.0) | −0.1 (−3.9 to 3.8) | 2.3 (−44.2 to 48.9) |
| 6 months | 83.7 (256.6) | 85.5 (254.5) | 77.0‡ (45.8 to 108.1) | 74.1‡ (40.4 to 108.0) | 2.8 (−43.7 to 49.4) |
| 9 months | 105.3 (284.2) | 123.0 (305.1) | 98.6‡ (64.2 to 133.0) | 111.6‡ (71.1 to 152.0) | −13.0 (−59.5 to 33.5) |
| 12 months | 143.9 (329.9) | 191.1 (368.6) | 137.1‡ (97.2 to 177.0) | 179.8‡ (131.1 to 228.5) | −42.6 (−89.2 to 3.9) |
| Per-protocol analysis | |||||
| Baseline | 6.4 (19.2) | 12.1 (28.9) | |||
| 3 months | 7.3 (15.2) | 10.3 (17.3) | 0.9 (−2.0 to 3.9) | − 1.8 (−6.2 to 2.6) | −3.0 (−6.3 to 0.3) |
| 6 months | 10.8 (29.6) | 13.7 (23.7) | 4.4‡ (0.4 to 8.5) | −1.6 (−3.2 to 6.4) | −2.8 (−8.5 to 2.7) |
| 9 months | 15.7 (37.8) | 16.5 (33.0) | 9.3‡ (4.1 to 14.6) | 4.4 (−1.2 to 10.0) | −0.76 (−8.2 to 6.6) |
| 12 months | 17.3 (40.3) | 25.6 (58.6) | 10.9‡ (5.5 to 16.3) | 13.5‡ (5.3 to 21.7) | −8.3 (−18.5 to 1.9) |
Results of the linear mixed-effects models.
*The working hours lost due to sick leave for musculoskeletal disease were recorded daily and the amount per month for each individual, and this was used to calculate the mean working hours lost for each individual and per group (intervention and control) every 3 months. Values are unadjusted mean (SD).
†The number of participants assessed at follow-up was n=487 (n=246 intervention, n=209 control) at 3 months, n=442 (n=241 and n=201) at 6 months, n=418 (n=231 and n=187) at 9 months, and n=368 (n=200 and n=168) at 12 months. The participant sample (n) was used in the per-protocol analysis.