Productivity at Work and Quality of Life in Patients With Rheumatoid Arthritis

Myrthe van Vilsteren; Cecile RL Boot; Dirk L Knol; Dirkjan van Schaardenburg; Alexandre E Voskuyl; Romy Steenbeek; Johannes R Anema


BMC Musculoskelet Disord. 2015;16(107) 

In This Article


Baseline measurements were available for 150 participants. Table 1, Table 2 and Table 3 describe the characteristics of the study population. In general, women participated in this study (84.0%). The mean age of the study population was 49.7 years, and on average, our participants had been diagnosed with RA 10.4 years ago. The overall WLQ score was 7.1, meaning that 7.1% of the time during the past two working weeks had been lost due to lost work productivity. This finding led to an average estimate of 4.0 work hours lost due to presenteeism over two weeks, based on an average work week of 28.7 hours. The study population scored low on the RAND 36 scales for vitality, physical role limitations, general health perception, and perceived health change scales, indicating that they had worse health on these subscales compared to the other RAND 36 subscales.

Following univariate regression analyses (Table 1, Table 2 and Table 3), we built multivariate models for internal and external factors. In the internal model, we retained mental health, pain, physical role limitations and biological therapeutic use since diagnosis for further analyses. In the external model, we retained the variables for having a job contract, being a supervisor, limitations in functioning at work, job satisfaction, work instability, heavy and demanding work, and having both physical and mental job demands for further analyses.

The final model (Table 4) shows that participants who experienced worse mental health and more physical role limitations, who had been treated with a biological therapeutic since diagnosis, who were not or only moderately satisfied with their work, and had a higher work instability score experienced more at-work productivity loss. The R2 for this model was 0.50. The R2 of this model in the cross-validation sample was 0.32.

For the second aim of our study, more at-work productivity loss was significantly associated (ß -0.10, 95% CI -0.13; -0.08) with worse mental health (Table 5). The adjusted analysis showed a regression coefficient of -0.06 (95% CI -0.09; -0.03), meaning that more at-work productivity loss was associated with worse mental health. The crude analysis of at-work productivity and the quality of life dimension for physical role limitations had a regression coefficient of -0.03; the adjusted analysis had a regression coefficient of -0.01. This means that on the physical role limitations scale (0–100, where a higher score indicates fewer role limitations) for every 10 points lower on the scale, an additional 0.1 hours are lost due to presenteeism. At-work productivity loss was significantly associated (ß -0.05, 95% CI -0.07; -0.02) with more pain. The adjusted analysis showed a regression coefficient of -0.03, 95% CI -0.06; 0.003), meaning that more at-work productivity loss is associated with more pain.