Socioeconomic status and duration and pattern of sickness absence. A 1-year follow-up study of 2331 hospital employees

Table 4 Sickness absence characteristics of persons with any sickness absence (n = 1889).

			Longest absence spell¹⁾			Results from logistic regression analyses, final model
	Number of persons with any absence	Number of absence days per year	Short (1-3 days)	Medium (4 -14 days)	Long (>14 days)	Medium versus short spells²⁾	Long versus short spells³⁾	Long versus medium spells⁴⁾
Occupational group⁵⁾	n	median	n (%)	n (%)	n (%)	OR (95% CI)	OR (95% CI)	OR (95% CI)
Doctors	159	3	116 (73)	34 (21)	9 (6)	1.00	1.00	1.00
Physio- therapists	216	7	117 (54)	78 (36)	21 (10)	2.24 (1.25-4.03)	1.51 (0.58-3.96)	0.95 (0.32-2.79)
Nurses	594	8	265 (45)	267 (45)	62 (10)	3.24 (1.94-5.40)	2.27 (0.98-5.29)	0.78 (0.29-2.07)
Medical secretaries	269	8	111 (41)	135 (50)	23 (9)	4.49 (2.60-7.78)	2.09 (0.83-5.26)	0.58 (0.20-1.67)
Nursing assistants	437	11	131 (30)	235 (54)	71 (16)	5.84 (3.41-9.99)	4.25 (1.83-9.87)	0.84 (0.31-2.27)
Cleaners/ porters	214	12	47 (22)	144 (67)	23 (11)	11.2 (6.08-20.8)	4.71 (1.82-2.19)	0.72 (0.25-2.07)
Total	1889	9	787 (42)	893 (47)	209 (11)
p⁶⁾						<0.0001	0.0003	0.8151
Mean change of estimates compared to start model						6.2%	-11.0%	3.4%

By occupational group. Odds ratios (OR) with 95% confidence intervals (CI) of medium versus short, long versus short and long versus medium spells, adjusted for other significant covariates in multiple logistic regression analysis (see text).
1) Short = only short spells. Medium = any medium but no long spells. Long = any long spells. See figure 1.
2) Adjusted for age, gender, work-place unit, follow-up time, overall demands, control, job strain and general health. Persons with long spells (n = 209) were excluded from the analyses.
3) Adjusted for age, gender, work-place unit, follow-up time,violence and general health. Persons with medium spells (n = 893) were excluded from the analyses
4) Adjusted for age, gender, work-place unit, follow-up time, violence, insecurity at work, duties and general health. Persons with short spells (n = 787) were excluded from the analyses.
5) See table 1.
6) Probability of no difference between occupational groups, χ²-test.

ISSN: 1471-2458