Abstract Statins are a widely used class of lipid-lowering medication, yet little is known about their effect on the risk of Alzheimer?s disease and related dementias (ADRD). Findings from two large randomized controlled trials of statins found no effect on cognitive outcomes over the short-term (4 to 5 years). Results from observational studies have been mixed. There could be several explanations for these inconsistent findings, including limited ability to account for confounding in observational studies, insufficient follow-up for cognitive events, and the inclusion of selected populations. Furthermore, most trials do not have adequate statistical power to examine heterogeneous treatment effects ? for example, whether the effect of statin on ADRD risk differs by subgroups of age, sex, race/ethnicity, or comorbid conditions. Trials may also not capture or enroll those who are less tolerant and more likely to discontinue statins, a proportion that approaches 70% in some populations. It is logistically challenging and expensive to design a trial that will address all these gaps in our knowledge. Large, administrative observational data sources can provide sufficient sample sizes and diversity to address these limitations, but traditional analytical techniques are insufficient in the presence of strong confounding. In this study, we propose to leverage a widespread clinical guideline and established statistical methods in economics, specifically a regression discontinuity (RD) design, to address confounding and approximate a randomized trial. Our data (N=189,682) will come from the Health Improvement Network (THIN) database which includes general practices in the UK covering about 5% of the 2014 total population. In 2008, the National Institute for Health and Care Excellence in the UK passed a guideline that statins are to be prescribed if a patient?s 10-year cardiovascular disease risk score exceeds 20%. Since treatment is given or withheld according to this guideline, we assume that patients who are ?near? the 20% cutoff will be similar except for the treatment received. Under certain feasible assumptions, this creates a quasi-experiment that will enable us to estimate the average treatment effect of statin on ADRD risk. The proposed research will (Aim 1) calibrate the RD model design in the setting of the effect of statin on myocardial infarction (positive control outcome), and on motor vehicle accidents (negative control outcome). We will apply the same RD methodology to establish the treatment effect of statin on ADRD risk (Aim 2). Heterogeneous treatment effects remain underexplored, so we will (Aim 3) identify treatment effects for relevant subgroups defined by age, sex, race/ethnicity, and comorbid conditions. Using the RD design, we will also examine ADRD risk across patterns of statins, including those who discontinue (Aim 4). This study will provide critical insight into the effect of statin on ADRD risk by leveraging methods from medicine, epidemiology and economics. By also examining heterogeneous treatment effects across key patient characteristics, this study will allow us to move towards targeted ?precision medicine? ADRD prevention, as well as identify those in whom the evidence is equivocal and an RCT is necessary.
|Effective start/end date||8/15/19 → 7/31/20|
- National Institutes of Health: $705,777.00
Hydroxymethylglutaryl-CoA Reductase Inhibitors
Information Storage and Retrieval
National Institutes of Health (U.S.)
Randomized Controlled Trials