g. Stephens et al. 2002). This fact could explain why health status is no longer the primary factor in sick leave after 2 years, which is consistent SB203580 nmr with the observations of the current study as well. Literature shows that some of the factors mentioned by the experts in the present study have also been mentioned in quantitative studies on factors related to sickness absence spells shorter than 1.5 years. It must be noted that most quantitative

studies on these relevant factors are not focused on absence spells of 1.5 years of more. This is concordance with the findings in a systematic review on factors associated with long-term sick leave in sick-listed employees (Dekkers-Sánchez et al. 2008). Quantitative studies on the relevant factors associated with sick leave longer than 1.5 years are needed to confirm our findings. Methodological considerations The electronic Delphi technique we used proved to be a feasible, time- and cost-efficient method. A strength of this study is that we elicited the views of a wide range of

experts that covered a broad representation of views. Although the Delphi method has been widely used in health research, studies using the Delphi technique have some variMS-275 order ability in their methodology (Sinha et al. 2011). In the present study, consensus was defined as an agreement of at least 80 % 3-deazaneplanocin A (Piram et al. 2011). In the last round, we decided that factors selected by a majority of panellists would be included in the final list, and 55 % can thus be accepted as a majority (Slebus et al. 2008). Some authors have suggested that the use of a structured questionnaire in the first round, instead of an open-ended questionnaire, may restrict the ability

of the experts to respond to the original question (Thompson 2009). In the first questionnaire, we used a preliminary list of factors generated in previous studies, but we also encouraged participants to add new factors to the preliminary list. This method ensured that we did not overlook any important factors, and it allowed us to elicit 35 new factors that were incorporated in the subsequent questionnaire. Other studies have also used this pragmatic approach successfully (e.g. Payne et al. 2007; Dionne et al. 2008). This study makes a unique Hydroxychloroquine contribution in several ways. First, the study increased our understanding of important factors that should be considered in the assessment of the work ability of employees on long-term sick leave and that are independent of the diagnosis. Second, it covers, from the physicians’ perspective, a breadth of factors associated with RTW of employees on long-term sick leave. Third, it is based on a large and heterogeneous sample of experts from all geographical regions in the country, with different demographics and varying experience with employees suffering from all types of medical complaints.