Generalizing breast cancer screening to more representative populations
Principal Investigator
Name
Sarah Robertson
Degrees
PhD
Institution
Dartmouth College
Position Title
Assistant Professor
About this CDAS Project
Study
HIPB
(Learn more about this study)
Project ID
HIPB-18
Initial CDAS Request Approval
Jan 5, 2026
Title
Generalizing breast cancer screening to more representative populations
Summary
The HIP Breast Cancer Screening Trial found that screening reduced mortality, mainly in participants over age 50. However, in earlier reports it noted substantial differences between participants and non-participants, in terms of demographic data, so it is unclear if the results of the HIP Breast Cancer Screening Trial would generalize to more representative participants. We will attempt to answer this question, using transportability methods and more representative data, such as NHIS data. This analysis will serve as a methodological check to validate transportability estimators. We will also explore heterogeneity in the benefit found in participants over age 50 vs under age 50.
Aims
Our primary aims in using the HIP Breast Cancer Screening Trial are as follows:
Aim 1: Compare characteristics of the HIP Breast Cancer Screening trial vs a corresponding representative target populations (e.g., NHIS data) that broadly meet the eligibility criteria for the trial. For the NHIS data we will form different definitions of the target population, from most strict (e.g., insured women meeting the age and enrollment criteria) to the most relaxed (e.g., any women meeting the age and enrollment criteria). This will require lining up baseline covariates (e.g., age, education, smoking, health history) from the NHIS data to the HIP Breast Cancer Screening trial.
Furthermore, an advantage of the HIP trial is that we will also compare participants who were not screened.
Aim 2: Estimate the effects of the screening intervention in these more representative populations, and compare the effects to that estimated in the trial. We will use established transportability methods that only require baseline data from NHIS. We will use the same primary outcome and follow-up time as the HIP trial.
Collaborators
Sarah Robertson Dartmouth College