Selective participation and loss to follow-up in population-based studies: novel approaches for dealing with unmeasured causes of selection

Population-based studies are important sources of new knowledge about risk factors, effects of treatments and prevention strategies for chronic and infectious diseases, as well as inequalities in health and disease burden.

However, the validity of such studies is often negatively affected by low participation at the recruitment stage, as well as substantial drop out at follow-up examinations - especially in socioeconomically disadvantaged groups.

Consequently, the general trust in population-based research is at risk as the results may not reflect the diversity that exists in the population.

We have previously demonstrated that detailed population and health register data can be used to improve the validity of population-based studies.

However, we have also encountered situations where the Swedish population and health registers are insufficiently detailed to fully adjust for the effects of selective participation, even if appropriate statistical methods are used.

This methodologically-oriented project therefore aims to develop and evaluate new statistical and epidemiological approaches for improved handling of selective participation and drop-outs in population-based studies. The methods will be tailored towards studies conducted within the register infrastructures in the Nordic countries.

We will apply the methods on data from four on-going population-based studies in Sweden on cancer, cardiopulmonary diseases, stroke, and mental disorders.

We will also address selection in new studies of the Covid-19 pandemic, where we will pay particular attention to variations in testing strategies over time – both for detection of on-going infection and antibodies.

We expect that our tools will provide researchers with a better chance of obtaining results that reflect true conditions and associations in the intended target population, thereby ensuring that research results are as relevant as possible for public health decision-making in the Nordic countries.