Association between adjustment latitude and sickness presence - A panel study of Swedish employees.

OBJECTIVES This study is analyzing the unexpected reversed or lacking association between high adjustment latitude and sickness presence by examining whether it is due to confounding. MATERIAL AND METHODS Questionnaires were sent in 2004 and 2005 to a cohort of individuals aged 25-50 years, selected from the Statistics Sweden's register of the Swedish population. Information from 2397 individuals who answered both questionnaires was analyzed by the use of logistic regression analysis. RESULTS The odds ratio for sickness presence among the individuals with a low adjustment latitude compared to those with high adjustment latitude was 1.7 (range: 1.4-2.2). This increased likelihood was almost entirely unaffected in the analysis of potential confounders. CONCLUSIONS If the reversed association between adjustment latitude and sickness presence does not reflect confounding, it may be due to reporting bias, which may cause problems in research on sickness presence. We argue that more detailed studies are needed to explore the different sources of possible reporting problems.