Skip to main content
. 2019 Oct 31;9(10):e025967. doi: 10.1136/bmjopen-2018-025967

Table 5.

Predicting the odds for any sickness absence (SA) (logistic model) and the duration of SA, if any (zero-truncated negative binomial (NB) part) in the negative binomial Hurdle model: crude analysis

OG Gender HRA result category N Logistic model (0 vs >0) Zero-truncated NB (>0)
OR 95% CI RR 95% CI
BC
Female (Intercept) 1.08 (0.82 to 1.43) 10.28 (7.81 to 13.53)
No findings 198 1.00 Ref 1.00 Ref
Some symptoms 1471 1.78 (1.32 to 2.40) 1.07 (0.80 to 1.42)
Health risk 843 2.05 (1.49 to 2.80) 1.36 (1.01 to 1.82)
WD risk, 1 risk factor 840 3.13 (2.27 to 4.33) 1.58 (1.18 to 2.12)
WD risk, 2 risk factors 335 4.91 (3.27 to 7.36) 2.36 (1.72 to 3.25)
WD risk, 3–5 risk factors 219 5.59 (3.49 to 8.96) 3.43 (2.45 to 4.82)
Male (Intercept) 1.23 (0.97 to 1.56) 6.64 (5.25 to 8.41)
No findings 274 1.00 Ref 1.00 Ref
Some symptoms 1060 1.37 (1.04 to 1.79) 1.27 (0.98 to 1.63)
Health risk 691 1.44 (1.08 to 1.91) 1.47 (1.13 to 1.92)
WD risk, 1 risk factor 456 2.04 (1.49 to 2.79) 2.89 (2.19 to 3.82)
WD risk, 2 risk factors 147 2.34 (1.51 to 3.63) 3.86 (2.70 to 5.51)
WD risk, 3–5 risk factors 69 3.51 (1.83 to 6.71) 5.32 (3.41 to 8.30)
C Female (Intercept) 0.62 (0.51 to 0.76) 8.05 (6.35 to 10.20)
No findings 425 1.00 Ref 1.00 Ref
Some symptoms 3390 1.58 (1.28 to 1.94) 0.87 (0.68 to 1.11)
Health risk 1344 1.94 (1.55 to 2.42) 1.28 (0.99 to 1.65)
WD risk, 1 risk factor 1403 2.81 (2.24 to 3.51) 1.48 (1.15 to 1.91)
WD risk, 2 risk factors 500 3.32 (2.54 to 4.36) 1.73 (1.30 to 2.30)
WD risk, 3–5 risk factors 272 3.54 (2.56 to 4.88) 3.05 (2.21 to 4.21)
Male (Intercept) 0.34 (0.28 to 0.42) 5.96 (4.51 to 7.86)
No findings 532 1.00 Ref 1.00 Ref
Some symptoms 2198 1.36 (1.10 to 1.69) 1.03 (0.77 to 1.37)
Health risk 953 1.46 (1.15 to 1.85) 1.36 (0.99 to 1.86)
WD risk, 1 risk factor 590 2.59 (2.01 to 3.34) 1.84 (1.34 to 2.54)
WD risk, 2 risk factors 197 3.19 (2.27 to 4.49) 3.22 (2.15 to 4.82)
WD risk, 3–5 risk factors 85 4.16 (2.59 to 6.68) 3.83 (2.29 to 6.41)
P/M

Female (Intercept) 0.58 (0.38 to 0.88) 2.77 (1.59 to 4.83)
No findings 95 1.00 Ref 1.00 Ref
Some symptoms 774 1.23 (0.79 to 1.92) 2.18 (1.26 to 3.78)
Health risk 220 1.45 (0.89 to 2.38) 3.03 (1.65 to 5.58)
WD risk, 1 risk factor 254 1.71 (1.06 to 2.78) 3.09 (1.71 to 5.60)
WD risk, 2 risk factors 69 2.36 (1.25 to 4.46) 4.31 (2.09 to 8.90)
WD risk, 3–5 risk factors 35 2.57 (1.16 to 5.69) 5.00 (2.10 to 11.9)
Male (Intercept) 0.39 (0.29 to 0.52) 2.95 (1.80 to 4.81)
No findings 232 1.00 Ref 1.00 Ref
Some symptoms 913 1.06 (0.77 to 1.46) 1.76 (1.12 to 2.78)
Health risk 292 1.38 (0.95 to 2.01) 2.08 (1.23 to 3.51)
WD risk, 1 risk factor 176 1.43 (0.94 to 2.18) 3.18 (1.76 to 5.73)
WD risk, 2 risk factors 36 2.06 (1.00 to 4.21) 1.35 (0.54 to 3.28)
WD risk, 3–5 risk factors 20 2.57 (1.02 to 6.46) 2.51 (0.80 to 7.82)

Logistic model refers to the model component for predicting membership to the subpopulation A with high propensity to zero absence, and zero-truncated NB to the component predicting the days on sick leave among the susceptible subpopulation B. To facilitate interpretation, for the zero-inflation part we have shown the ORs associated with the complementary propensity to having any sickness absence, that is, inclusion in subpopulation B.

Bold values denote statistical significance at the p < 0.05 level.

BC, blue collar; C, clerical; HRA, health risk appraisal; OG, occupational group; P/M, professional/manager; RR, risk ratio; WD, work disability.