Project Summary
Low-value care (LVC), defined as care where harms or costs outweigh the benefits, is common and costly. For
patients, LVC can result in physical harm from the service itself, unnecessary follow-up tests/procedures, and
needless out-of-pocket costs. Patients with coronary artery disease (CAD) are at high risk of receiving LVC:
studies estimate that up to 50% of tests and 15% of procedures performed on patients with CAD may be LVC.
To identify and reduce LVC, accurate measurement is critical. Despite widespread adoption of administrative
claims-based measures to assess LVC, most measures have not been tested against the “gold standard” of
chart review (i.e., they do not have good criterion validity). Using inaccurate LVC measures undermines quality
reporting and leads to inaccuracies in understanding the frequency, costs, and trends in use of LVC services; a
critical knowledge gap in understanding which LVC services should be prioritized for reduction.
Bundled payment models (BPMs), which pay organizations a fixed price for a 90-day episode of care such as
elective percutaneous coronary intervention (PCI), could reduce LVC but could also have unintended
consequences. Because BPMs provide a fixed payment to providers independent of patient risk, BPM
providers might perform more PCIs on low-risk patients for whom medical therapy alone is the guideline-
recommended treatment (i.e., they may perform more LVC PCIs). BPM providers might also perform fewer
appropriate PCIs for high-risk patients, or those perceived to be high-risk such as racial/ethnic minorities. To
assess the impacts of the elective PCI BPM, we will use data from the National Cardiovascular Data Registry,
which contains a reliable measure to classify PCIs as LVC or appropriate using clinical chart review data.
We have assembled an experienced team of cardiologists, economists, and outcomes researchers. In prior
and pilot work we have validated 3 claims-based LVC measures, but also found that 1 measure was not valid.
In Aim 1, we will use chart review as a “gold-standard” to assess the validity of commonly-used claims-based
LVC measures relevant for CAD patients. In Aim 2, we will use validated LVC measures to assess LVC
frequency, costs, and trends among a national sample of Medicare beneficiaries with CAD, and provide
policymakers with a priority list of LVC services to target for reduction. In Aim 3, we will conduct a difference-in-
differences analysis to assess the impact of elective PCI BPMs on a) LVC PCIs and healthcare disparities for
appropriate PCIs using detailed clinical data from the NCDR, and b) other validated LVC services using linked
Medicare claims. For patients with CAD, this research will 1) establish valid LVC measures which are needed
for quality improvement and accountability programs, 2) guide efforts to reduce LVC services, and 3) evaluate
the effect (and potential unintended consequences) of a policy intended to reduce LVC.