Just a moment, the page is loading...

Functional Estimation of Interventional Effects

Functional Estimation of Interventional Effects

Joon Park

Professor, Department of Economics
Indiana University
Bloomington, IN, US



26 March 2014

The purpose of our research is to develop new statistical tools for the analysis of results data, particularly in understanding the distribution effects of a treatment. We are interested in developing and applying a novel method to analyze the consequence of a policy effect (in this case, a drug intervention) across a heterogeneous sample of individuals. The heterogeneity in this case will arise from the bio-marker profiles/values of the patients in the study. Such markers may include factors such as fasting blood glucose, blood pressure, or other numerical assays/profiling tools.

Standard methods aim to communicate the random average effect of a medication, such as the change in the mean value of a bio-marker value. Our method will, in addition to this information, communicate the random distributional effect of a medication, i.e. the entire probability profile of potential bio-marker results of a medical treatment. Secondary measures such as bio-marker value dispersion, or increases or decreases in extreme bio-marker values, may be equally as important as the average effect of a drug and may have significant impact on overall patient outcomes. Other unusual distributional properties of medical treatments are also observable under our methodology (including multi-modality, skewness, or any result that can be computed from a complete probability distribution).

The aim of the research is strictly statistical in nature, and does not attempt to extend in any way medical procedure (with which we are not familiar), and strives only to provide an additional statistical tool. Core findings will consist of the distribution data of drug effects. In general, we may make use of our results to comment upon efficacy of the drugs based upon commonly agreed upon 'normal' or healthy values of the bio-markers in question. In essence, this will provide empirical evidence for the utility of the statistical technique in question, which has a theoretical framework which we have developed. Findings, consisting of both code to execute the method and references to the specific datasets involved will be freely available at publication. General methodology of the technique will be free for public use.

[{ "PostingID": 475, "Title": "GSK-49653/048", "Description": "A Randomized, Double-Blind Study to Compare the Durability of Glucose Lowering and Preservation of Pancreatic Beta-Cell Function of Rosiglitazone Monotherapy Compared to Metformin or Glyburide/Glibenclamide in Patients with Drug-Naive, Recently Diagnosed Type 2 Diabetes Mellitus

Medicine: rosiglitazone, Condition: Diabetes Mellitus, Type 2, Phase: 3, Clinical Study ID: 49653/048, Sponsor: GSK" },{ "PostingID": 476, "Title": "GSK-49653/137", "Description": "A Study to Evaluate the Efficacy of Rosiglitazone (BRL-049653) on Reduction of Microalbuminuria in Subjects with Type 2 Diabetes Mellitus

Medicine: rosiglitazone, Condition: Diabetes Mellitus, Type 2, Phase: 3, Clinical Study ID: 49653/137, Sponsor: GSK" },{ "PostingID": 485, "Title": "GSK-49653/347", "Description": "A Multicenter, Randomized, Double-Blind, Parallel-Group, Placebo-Control, Clinical Evaluation of Insulin Plus Rosiglitazone (2mg and 4mg) Compared to Insulin Plus Placebo for 24 Weeks in Subjects with Type 2 Diabetes Mellitus Who are Inadequately Controlled on Insulin

Medicine: rosiglitazone, Condition: Diabetes Mellitus, Type 2, Phase: 4, Clinical Study ID: 49653/347, Sponsor: GSK" },{ "PostingID": 488, "Title": "GSK-49653/374", "Description": "Medicine: rosiglitazone, Condition: Diabetes Mellitus, Type 2, Phase: 3, GSK Clinical Study ID: 49653/374, Sponsor: GSK ." }]

Statistical Analysis Plan

The publication citation will be added after the research is published.