Leveraging seasonality of 25-hydroxyvitamin D to control for unmeasured confounders in vitamin D epidemiology
Thursday, September 15, 2016 — Poster Session II
- MD Shardell
- L Ferrucci
The body produces 25-hydroxyvitamin D [25(OH)D] in response to sunlight exposure, therefore measured concentrations of 25(OH)D typically exhibit seasonal variation. Researchers often address this phenomenon in epidemiology studies by statistically adjusting for season of blood collection. However, if the study outcome does not vary seasonally then this adjustment may increase rather than decrease bias. On the other hand, if season of blood collection is not confounded with the study endpoint, then seasonality may be leveraged to control for unmeasured confounders by adapting instrumental variable methods. We present and illustrate this approach by examining the impact of 25(OH)D, as measured by radioimmunoassay, on depressive symptoms, as measured by the Center for Epidemiology Studies Depression Scale (CES-D), in participants aged at least 65 years enrolled in the Aging in Chianti study (n=3,147 person-visits). We found that participants with 25(OH)D < 20 ng/ml had 4.62 times greater risk of severe depressive symptoms (CES-D > 16) than did participants with 25(OH)D ≥ 20 ng/ml (95% confidence interval 3.25 to 6.56). We also present results from a Monte Carlo simulation experiment showing that the proposed method performs well with respect to bias and efficiency relative to conventional instrumental variable methods. Lastly, we present R software code to enhance the accessibility of the method for applied researchers.