Analyzing sickness absence, disability, and withdrawal from the labour market: The need for causal inference
The research group for causal inference at Department of biostatistics at the University of Oslo is involved in a project on analyzing sickness absence, disability, and withdrawal from the labour market.
About the project
Causal inference is becoming a major international field in statistics with applications in public health, epidemiology, welfare and many other fields. In these areas an important issue is how statistical data can be used for drawing conclusions about useful interventions and for understanding mechanisms, that is, for making assessments of causality. This is a notoriously difficult matter, but the new tools of causal inference can be of help in clarifying the questions and coming closer to causal conclusions. Norway has large registries in the areas of public health and welfare and the unique personal identification number allows linkage of various registries. This gives unique possibilities for analysis. In order to exploit these possibilities close attention must be made to using the most efficient methodologies. The emphasis on methods should match the amount, the richness and the complexity of the data. Detailed individual follow up data are available from Statistics Norway, The Norwegian Labour and Welfare Administration (NAV) and other sources. We shall assume a life-course perspective in our analysis because detailed follow-up of individuals may throw a clearer light on causal issues than and allow insight into underlying mechanisms. We apply methods from event history and survival analysis to analyze dynamics and mediation (direct and indirect effects) and methods like propensity scores and marginal structural models to study effects of interventions. Sensitivity analysis of unmeasured confounders will be an important supplementary tool.
Examples of issues we focus on in this project:
- Study detailed transitions between sick-listed, return to work or disability pension in order to understand the dynamics in these processes. The aim is evaluate possible interventions.
- Analyzing how parental family background could be mediated through individual factors such as health, intellectual ability, education and skill.
- Analyzing resilience.
The project is funded by the Research Council of Norway, through the SYKEFRAVAER (sickness absence, work and health) program.