Results Summary
PCORI funded the development of PCORnet®, the National Patient-Centered Clinical Research Network, to make research faster, easier, and less costly to conduct. PCORnet is made up of Partner Networks of healthcare systems, patients and communities, and health plans that harness the power of large amounts of health data.
PCORI supports brief, descriptive projects to assess the feasibility of conducting research using data gathered and shared securely through PCORnet. This project is one of several designed to test the network while addressing priorities identified by PCORI and its stakeholders.
What was the project about?
Doctors use molecular tests to get a better understanding of a patient’s cancer. These test results help doctors figure out if the cancer is one that is likely to respond to a treatment called targeted molecular therapy. Molecular tests that are related to targeted therapy are sometimes called companion diagnostic tests.
PCORnet created a shared database system that includes information about test results and treatments from patients’ electronic health records. This shared data system includes information from 11 clinical data research networks that are part of PCORnet. In this project, the team looked at what information was available about these molecular tests and treatments in the database. They wanted to learn whether the database had enough information to be useful for answering patient-centered questions about molecular tests and treatments. Patient-centered research respects patient preferences, needs, and values. When research is patient-centered, the patient’s values guide all healthcare decisions.
What were the results?
- How many patients got molecular tests? Whether patients got testing varied. The amount ranged from 26 percent for patients with endocrine cancers to 76 percent for those with breast cancer.
- How many patients got molecular targeted therapy? Information from the shared database showed that about 5 percent of all patients in the 11 networks received molecular targeted therapy after their tests. Patients age 60 years and older had the lowest use of this therapy. Patients with breast, lung, or head and neck cancer had the highest use of this therapy.
- Did patients get the right kind of therapy? The project team looked at data from two health systems to answer this question. In these systems, fewer than 20 percent of patients with colorectal cancer received molecular targeted therapy. But all patients who had molecular targeted therapy had test results that supported that treatment decision.
- Was the database complete? The project team compared two sets of data. The first set only included data from the database. The second set included data from the database plus Medicare insurance data. Using database data alone found that just 49 percent of patients had molecular tests, Medicare plus database data showed that 60 percent of these patients had the tests. Using database data only showed that about 4 percent of the patients had molecular therapy compared with about 7 percent of patients found using Medicare plus database data.
- Can researchers use the database for future research? The project team did find information on molecular tests and therapy in the database.
Who was in the project?
The project team included electronic medical record data linked to the database for 86,154 patients from 11 research sites in 10 states. Data were from 2013 to 2016. Patients had many types of cancers.
What did the project team do?
The project team used different methods to answer each question. They looked at
- How often doctors used molecular testing for different types of cancer
- How often patients got molecular targeted therapy after getting tested
- Whether patients got the right kind of therapy based on their test results
- Whether the database had enough information by itself, or whether it was more accurate when combined with Medicare data
- Whether the database had enough information to be useful when planning studies on using molecular tests and treatments for different types of cancer
What were the limits of the project?
Some of the information in the database wasn’t specific enough to know exactly which molecular test or therapy patients received. The project team used patient medical records to find the information they needed.
Future research could make data from the database more specific to help research teams find information about molecular tests and treatments.
How can people use the results?
Research teams can use these results when planning studies that use the PCORnet CDM dataset to understand molecular tumor testing and therapy.
Professional Abstract
PCORI funded the development of PCORnet®, the National Patient-Centered Clinical Research Network, to make research faster, easier, and less costly to conduct. PCORnet is made up of Partner Networks of healthcare systems, patients and communities, and health plans that harness the power of large amounts of health data.
PCORI supports brief, descriptive projects to assess the feasibility of conducting research using data gathered and shared securely through PCORnet. This project is one of several designed to test the network while addressing priorities identified by PCORI and its stakeholders.
Objective
To use PCORnet to document the patterns of use of molecular tests and molecular targeted cancer therapies for patients with solid tumors and to test the capacity of PCORnet to describe cancer pathology, test results, cancer treatment, and outcomes
Study Design
Design Elements | Description |
---|---|
Design | Retrospective descriptive analysis project |
Data Sources and Data Sets | Records of patients with a single primary solid tumor diagnosed between 2013 and 2016 from 11 PCORnet sites |
Analytic Approach | Review of EHR, tumor registry, and Medicare claims data |
Aim 1. Frequency of molecular tests and associated cancer treatment Aim 2. Concordance between molecular test results and associated treatment Aim 3. Completeness of data on molecular tests, treatments, and outcomes within EHRs relative to Medicare claims data Aim 4. Capacity of networks to conduct pragmatic outcomes research related to molecular testing in cancer |
The project included tumor registry and electronic health record (EHR) data from 86,154 patients linked using the PCORnet common data model (CDM). The CDM organizes data into a standard structure for use by researchers. The patients had single solid tumors and received care from 1 of 11 research sites in 10 states between 2013 and 2016.
The project had four aims:
Aim 1. The project team examined how often common molecular tests and molecular targeted therapies to treat patients with cancer are recorded in the CDM.
Aim 2. In a subset of patients who had molecular tumor tests for colorectal cancer at two PCORnet research sites, the research team reviewed hospital tumor registry and EHR data to determine whether molecular targeted therapy was in concordance with molecular tests.
Aim 3. In a subset of patients with breast cancer, the project team assessed the completeness of information about molecular tumor tests and molecular-targeted therapy within the PCORnet CDM data by comparing CDM data plus Medicare claims data with CDM data only and Medicare data only.
Aim 4. Based on results from aims 1–3, the project team described how well the CDM captured diagnoses, therapies, tests, test results, and treatment outcomes for research purposes.
Results
Aim 1.
- Molecular tumor tests. The prevalence of having any molecular tumor testing varied by tumor location, ranging from 26% for endocrine tumors to 73% for breast tumors.
- Molecular targeted therapy. Among patients with cancer, 5% had molecular therapy. Use of this treatment was lowest among patients ages 60 years and older and highest for patients with breast, lung, oral cavity, and pharynx cancer.
Aim 2. Between 16% and 18.7% of patients with colorectal cancer at the two sites received molecular targeted therapy. All 36 patients that received molecular targeted therapy received the therapy concordant with their molecular tumor test results.
Aim 3. CDM data plus Medicare claims data showed that 60% of patients had molecular tumor tests compared with 49% of patients when using CDM data only. CDM data plus Medicare claims data showed that 6.7% of patients had molecular targeted therapy compared with 4.4% of patients when using CDM data only.
Aim 4. The project team successfully and rapidly obtained CDM and tumor registry data from many sites in multiple data networks.
Limitations
The CDM data rely on standardized codes that may not be specific enough to identify molecular tests and molecular targeted therapy. The project team had to review underlying patient EHRs to retrieve these data.
Conclusions and Relevance
The project team linked CDM data with both tumor registry and EHR data and used these combined data to describe use of molecular tumor tests and molecular targeted therapy. The team used the combined data to investigate concordance between molecular tumor test results and the targeted therapy patients received. The team also demonstrated the importance of linking Medicare claims data to CDM data to ensure data completeness.
Future Research Needs
Future projects could employ natural language processing or structured pathology data to improve the identification of specific molecular tests and targeted treatments.