There is large debate in the literature about the relationship between job mobility and health. This thesis contributes to this topic by estimating the impact of work mobility on either cardiovascular, musculoskeletal or neuropsychological diseases in a longitudinal study on a sample of Belgian workers followed-up for seven-years. The occurrence of such diseases was assessed through medication use as proxy. In the first part of this thesis, the focus was on CVD and MSD diseases. To this aim, a logistic regression model for autocorrelated data with repeated measures was applied (while controlling for the time-variant and time-invariant confounders). This longitudinal model was adequate to take into account time-dependent covariates and included a Markov chain mechanism that regulates serial dependence. The main characteristic of this model is the ability to handle a series of different lengths of observations across individuals. In the second part, the focus was on the impact of work mobility on the onset of neuropsychological diseases. Therefore, a quasi-experimental approach was used to evaluate the causal effect of a time-varying treatment through the propensity score matching with time-dependent covariates. The optimal (sequential) matching algorithm was used to balance the distribution of the time-dependent covariates at every time point. The hazard of receiving the treatment was estimated using the Cox hazard model with time-fixed and time-varying covariates for a patient receiving the treatment at different times. Data was obtained from IDEWE, the largest Belgian occupational service for well-being at work (external service for prevention and protection at work). Its database includes data on medical history, work conditions, biometrics, vaccinations, medication use and sickness absence of more than 150.000 Belgian workers divided by employment sectors. The association with external job mobility, identified by a change in the 2-digit ISCO code of the employee in the IDEWE database, was statistically ...