U.S. flag An official website of the United States government
  1. Home
  2. Science & Research
  3. Science and Research Special Topics
  4. Advancing Regulatory Science
  5. Improving the Integrity of Novel Digital Health Technology-derived Endpoints through Rigorous Simulation Studies of Multiple Imputation Techniques
  1. Advancing Regulatory Science

Improving the Integrity of Novel Digital Health Technology-derived Endpoints through Rigorous Simulation Studies of Multiple Imputation Techniques

CERSI Collaborators: Manisha Desai, PhD; Thomas Robinson, MD, MPH; Marco Perez, MD; Bryan Bunning

FDA Collaborators: Maria Matilde Kam, PhD; Paul Schuette, PhD; Andrew Potter, PhD; Lili Garrard, PhD

Project Start: September 1, 2023

Regulatory Science Framework

Primary Charge: I. Modernize development and evaluation of FDA-regulated products: J. Methods to Assess Real-World Data to serve as Real-World Evidence
Secondary Charge: I. Modernize development and evaluation of FDA-regulated products: E. Clinical Outcome Assessment

Regulatory Science Challenge

Wearables such as smart watches that track your movement or heartbeat are used frequently in research to track patient health. These devices generate a large amount of data, and new types of statistical methods are needed to properly analyze them and address challenges such as when patients do not wear the device. This collaboration will help create standards for handling non-wear periods that can be used to advance medical product development and improve patient health.

Project Description and Goals

The goal of this project is to improve understanding of data from wearable devices. To that end, investigators will

  • Develop an open-source tool to help simulate patient data from wearables including non-wear periods using databases from completed clinical trials
  • Create and test strategies for handling non-wear periods
  • Establish open source code for recommended strategies

Overall, advancement of science will help take data from patient wearables and translate them into evidence that can be used to get better treatments for all patients.

Anticipated Outcomes/Impact

The anticipated outcomes from this CERSI project include the following:

  • Development of an open-source simulation tool (R package) to help simulate patient data from wearables including non-wear periods using databases from completed clinical trials.
  • Creation of methods for identifying and handling non-wear periods
  • Establishment of open-source code for the developed multiple imputation methods.
  • Development of manuscripts:
    1. a paper describing the gaps in handling missing data for digital health device studies;
    2. a paper to identify non-wear using ensemble methods; and
    3. a paper to investigate various multiple imputation methods for handling missing data from digital health devices.
 
Back to Top