Medications and Weight Gain in PCORnet: The MedWeight Study - 7 Project Summary/Abstract
Several prior studies have assessed whether individual prescription medications are associated with weight
gain and increased risk for conditions such as diabetes and high cholesterol. These negative health effects
from medications are a particular concern for medications used to treat chronic disease, which often require
very long term or lifelong treatment. Patients also identify weight gain as a common reason for medication
nonadherence, leading to worse health outcomes. Limitations of prior studies include limited long-term follow-
up, lack of comparison of all subclasses or commonly-prescribed individual medications within a class, and
lack of use of causal methodologies that can help manage baseline and time-varying differences in patients
prescribed different medications. Prior studies have typically not been large enough to determine whether
associations differ by age when medications are initiated or by other patient characteristics, such as race,
gender, and weight status. Using data on patients in a large network of healthcare institutions, this study will be
a comprehensive assessment of weight gain and metabolic risk up to 10.5 years after initiating medications
used to treat diabetes, high blood pressure, epilepsy, depression, psychoses, as well as contraceptives. These
are all classes used to treat chronic diseases and have been associated with weight gain and metabolic risk;
each of these classes also have several choices that can be used for treatment. The study will include children
and adults and will involve separate evaluations of the 6 medication classes, comparing effects between
subclasses and commonly-prescribed individual medications within each class. The setting for the study will be
19 diverse healthcare institutions participating in the National Patient-Centered Clinical Research Network
(PCORnet). PCORnet facilitates cross-institutional, comparative effectiveness studies though use of a
Common Data Model, in which each institution organizes its clinical information in the same format allowing for
straightforward merging of data across sites. The first study Aim will be a detailed data quality assessment that
will measure deviations from expected patterns of weight and height and rates of medication prescriptions in
each class. This Aim will adapt existing data quality methods to allow for troubleshooting and remediating data
issues and will produce an approach that can be scaled across PCORnet and other healthcare networks. The
second Aim will be an examination of medication initiation and weight outcomes, incident diabetes, and change
in cholesterol, and the third Aim will assess more sustained use of medications. In these latter Aims, we will
incorporate advanced statistical methods that can account for differences in patients in novel ways, both at
baseline and over the course of follow-up. This large-scale study will be the most comprehensive study to date
of medication-induced weight gain and metabolic risk and will provide important, actionable information to help
patients and clinicians make the most informed choices about medications.