Hypothetical Intervention for Health Related Quality of Life in Children with Juvenile Idiopathic Arthritis during the First Year of Disease
*Bin Huang, Cincinnat Children's Hospital Med Ctr
Keywords: Casual inference, G-formula, Bayesian modeling, Health Related Quality of Life, Patient report outcome, cohort study
Recent development in statistics causal inference proposes to empirically test for and estimate potential intervention effect by applying parametric G-formula to an existing observational data without actually conducting an intervention study. This allows for strategically planning and optimizing effective intervention programs utilizing statistics modeling, and can be easily accomplished by linear regression modeling techniques. Here, we apply and extend parametric G-formula to estimate a set of hypothetical intervention effect on health related quality of life (HRQOL) in a cohort of patients newly diagnosed for Juvenile Idiopathic Arthritis (JIA) during the first year of their disease. Patients were followed up for 12 month in 6 month interval. By applying Bayesian autoregressive modeling in conjunction with G-formula, we demonstrate that Bayesian autoregressive modeling can provide a flexible and power approach that allows for incorporating existing knowledge into the evaluation of intervention effect, and allows for model fit and selection. The proposed approach can be easily implemented using WinBUGS. We present results analyzing the JIA cohort data, and discuss considerations in planning future intervention programs to ensure its maximum impact to improve HRQOL in children with JIA.
Important Dates & Deadlines
- October 9 - 11, 2013