| Summary: | Viðskiptafræðideild hefur samþykkt að aðgangi að þessari ritgerð skuli lokað í eitt ár. Background: Instrumental variable analysis has been used in health economics, e.g. to find causal effects of health on labor-market outcomes. With advanced knowledge in genetics there has been a growing interest of using genetic information as instruments. The aim of this paper is to examine the quality of molecular genetic variants as instruments for body mass index (BMI). Data and Methods: The data used is from the Icelandic Heart Association, the Reykjavik Study (RS) established in 1967 and the Age, Gene/Environment Susceptibility–Reykjavik Study (AGES-RS) initiated in 2002. Participants are men and women born 1907-1935 in Reykjavík. Genetic variants or single nucleotide polymorphisms (SNPs) found to be robustly associated with BMI in a Genome Wide Association Study (GWAS) was used as instruments in the analysis. We made regressions where the SNPs were both used as a set of instruments and as a weighted genetic risk score (GRS). Results: First stage regressions show the instruments to be too weak to serve as instruments for BMI. The F-statistics result in a value of around 2, far below the minnimum of 10 that is often used as a threshold. Conclusion: This paper supports and further reinforces the literature in that far stronger genetic instruments are needed for BMI than are available to date. Therefore, results with such instruments need to be cautiously interpreted.
|
|---|