Real-world Data From Multiple Sources May Be Needed to Conduct Biosimilar Comparison Studies

Although real-world data (RWD) can be useful when conducting a study comparing the effects of a biosimilar with its reference product, researchers should ensure that RWD sources are comprehensive enough to help answer research questions, according to a poster presented at Academy of Managed Care Pharmacy 2021.

The use of large health insurance claims databases when conducting postmarketing analyses is common. However, certain relevant measures entered into claims databases may be difficult to assess in the oncology space.

To address the insufficiencies of administrative claim data alone, the investigators convened a multistakeholder workgroup of industry sponsors, researchers, and clinicians from the Biologics and Biosimilar Collective Intelligence Consortium to assess the utility of and develop recommendations for RWD sources used for comparison study purposes.

The workgroup had discussions surrounding hypothetical comparative studies for originator trastuzumab and its biosimilars in patients with breast cancer. After the discussions, the workgroup created and distributed a questionnaire to potential RWD sources. The questionnaire asked the sources to report:

• The characteristics of the RWD source
• Their experience with collaboration
• Their availability of select data elements deemed necessary to conduct comparison studies
• Their ability to link to external data sources
• Their counts of patients treated with trastuzumab

The questionnaire was disseminated to 18 potential sources identified from a literature review and suggestions from members of the workgroup. The sources were chosen to reflect a variety of data source types. All chosen sources were contacted in July 2020 and received up to 2 follow-up emails in August 2020.

Overall, 7 (39%) sources responded, 2 of which were primarily based on electronic health records (EHRs), 1 was claims-based, and 4 were a combination of EHR and claims data.

Four (57%) respondents reported having access to biomarkers, such as patients’ estrogen receptor status. Additionally, 71% (n = 5) of respondents had the ability to distinguish between reference products and biosimilars, and 43% (n = 3) had the ability to link to external data sources, including cancer and death registries.

In total, 3 respondents provided feasibility counts of a potential population that could be used in a hypothetical study comparing trastuzumab products. The population of patients with breast cancer in these sources ranged from 66,000 to 128,000. All 3 respondents could report the size of the population who had HER2-positive cancer and who had stage IV or metastatic cancer.

Only 1 respondent was able to differentiate between patients receiving reference trastuzumab or trastuzumab biosimilars and provide numbers of patients treated with a reference product vs a biosimilar. The population of patients receiving trastuzumab ranged from 3000 to 12,000.

The investigators said that lack of biomarker information, information on cancer progression, and mortality data could lead to limitations in researchers’ ability to answer certain oncology-related research questions.

To overcome data limitations, the investigators said that oncology research will likely need to include claims data in combination with links to external registry data or EHR data because they can provide clinical data that may not be captured by administrative claims data alone, such as cancer stage, disease grade, and radiographic and laboratory findings.

“For studies requiring detailed clinical oncology data, the feasibility of potential RWD sources should be directly assessed via a questionnaire similar to that employed in this study,” wrote the investigators.

Reference

Seals RM, Lockhart CM, Lin N, Seeger JD. Characterization of real-world data sources to support observational studies of biologics, including biosimilars in oncology. Presented at: Academy of Managed Care Pharmacy 2021; April 12-16, 2021. Poster C18.