28 Chapter 2 Table 1. Criteria list for assessment of methodological quality of longitudinal studies (6, 24, 26) CriteriaA Participation 1. Adequate description of source population (i.e. clear in- and exclusion criteria) 2. Adequate description of sampling frame, recruitment methods, period, and place of recruitment 3. Participation rate at baseline at least 80% or non-response not selective (i.e. selected population does not significantly differ in key characteristics from source population) Attrition 4. Provision of the response rate (n or %) during follow-up measurements 5. Response at follow-up at least 80% of the n at baseline or non-response during follow-up measurements not selective (i.e. follow-up population does not significantly differ in key characteristics from selected population) Data collection 6. Temporal determination of the work exit B Data analyses 7. Statistical model used appropriate and described with point estimates and measures of precision (i.e. CI or SE) 8. Population size suitable for answering the research question 9. Important confounders or effect modifiers (i.e. age, sex) identified and adjusted for (i.e. stratification and/or interaction term) A: rating of criteria: + = positive; - = negative; n.a. = not applicable; B: Temporal determination of the work exit means how regular this transition was assessed. Studies were rated positive if exit from work was determined on an annual basis. If this was not the case studies were rated negative. Abbreviations: n=sample size; CI=confidence interval; SE=standard error Results Study selection The flow chart, presented in Fig. 1, demonstrates the study selection. The search strategy yielded 8961 articles. After removing duplicates, 4765 articles were screened on title and abstract, and subsequently, 108 articles on full text. The search resulted in 19 articles (17, 27-44). The references of these articles were screened, which resulted in five additional articles (15, 45-48). In total, 22 studies were included in this review. Study characteristics The study characteristics are presented in Tables 2–4. The most remarkable differences are described here. Sample sizes ranged from 186 to 245,082 participants (35, 47). For measuring the effects of exit from work on health, studies mostly used the following datasets cohorts and registers; the Health and
RkJQdWJsaXNoZXIy MTk4NDMw