Accurate and efficient assessment of outcomes is central to medical device innovation, evaluation and surveillance. Administrative claims could facilitate prospective device evaluations in older individuals if health outcomes of interest (HOIs) are reliable. Various HOI ascertainment tools, including site-based reporting and patient-reporting are commonly used without rigorous evaluation. Each increases resource utilization vs.
cl aims -based HOI assessment without a clear demonstration of added value, particularly in older cohorts. Existing predicates using procedure-oriented national registries for evaluation of coronary stents, defibrillators and prosthetic heart valves already demonstrate efficiencies with high impact potential for device innovation over the total product life cycle. Extension of these efforts to long-term outcomes databases such as claims data, if reliable, would integrate these efficiencies for a far broader scope of device applications. Recently, both the Observational Medical Outcomes Partnership (OMOP) and the Mini-Sentinel groups have highlighted limitations of existing claims-based HOI ascertainment algorithms. Previous HOI validation efforts for myocardial infarction, stroke, heart failure, and bleeding have had limited ability to distinguish between incident and prevalent disease, to establish accuracy estimates within subgroups, or to assess adverse events absent re-hospitalization. In addition, existing claims-based HOI ascertainment algorithms rely solely on diagnostic codes, and the extent to which procedural billing codes would enhance their precision and accuracy is unknown. Enhanced efficiency cannot serve device evaluation without validated accuracy. This proposal applies a novel linkage of existing in-house clinical and claims data to provide a cost- and time-efficient evaluation of claims-based HOI ascertainment versus 'gold standard'CEC adjudicated outcomes. This study proposal is structured in 3 phases to progressively leverage deliverables through 3 potential years of funding. In PHASE 1 of this study (year 1), we will refine and validate administrative claims-based HOI coding algorithms using a novel linkage between Medicare claims data and existing clinical trial and registry records. In PHASES 2 and 3 (years 2 and 3, respectively), we will apply these refined mechanisms of HOI ascertainment to the conduct of both 1) the longitudinal surveillance of transcatheter aortic valve replacement (TAVR) prostheses using available unique device identifiers (UDI), and 2) a pragmatic registry-based clinical trial that builds on the successful piloting of the National Cardiovascular Research Infrastructure (NCRI) by the SAFE-PCI for Women Trial. This proposal expands and strengthens existing research collaborations and infrastructure for the performance of medical device surveillance and registry-based randomized clinical trials.

Public Health Relevance

The accurate and efficient assessment of health outcomes of interest (HOIs) is the foundation of medical device testing and surveillance. This study will evaluate and refine efficient methods for HOI ascertainment; then, it will apply these methods in the development of 1) a national surveillance system for transcatheter valve prostheses, and 2) a pragmatic registry-based IDE randomized clinical trial

Agency
National Institute of Health (NIH)
Type
Research Project--Cooperative Agreements (U01)
Project #
5U01FD004591-03
Application #
8731134
Study Section
Special Emphasis Panel (ZFD1)
Project Start
Project End
Budget Start
Budget End
Support Year
3
Fiscal Year
2014
Total Cost
Indirect Cost
Name
Duke University
Department
Type
DUNS #
City
Durham
State
NC
Country
United States
Zip Code
27705