Table 3.

Within-person effect of temporal fluctuations in neck-shoulder pain intensity (scale 0–10) on sick leave (days/month) and work ability (scale 0–10) over 1 year

GEE models	N	Sick leave				Work ability
		P value	RR	95% CI		P value	OR	95% CI
				Lower	Upper			Lower	Upper
Model 1	248	0.008	1.12	1.03	1.21	0.005	1.11	1.03	1.19
Model 2	242	0.008	1.12	1.03	1.21	0.009	1.11	1.03	1.19
Model 3	231	0.011	1.11	1.02	1.21	0.002	1.13	1.04	1.21

RR estimates, indicating the relative increase in the number of days on sick leave per month, were obtained using GEE with a Poisson distribution for days on sick-leave (measured at 4-week intervals). ORs indicating the likelihood of a 1-unit reduction in work ability, were obtained using GEE with a multinomial distribution for work ability (measured at 12-week intervals). Estimates indicate the within-person effect of change in pain intensity on change in sick leave and work ability per month.

Model 1: adjusted for the person mean pain intensity across time points.

Model 2: additionally, adjusted for age, gender and body mass index.

Model 3: additionally adjusted for occupational sector and physical work load.

GEE, generalised estimation equation; RR, relative risk.