6

Collaborations with Patient Groups

Pat Furlong of Parent Project Muscular Dystrophy (PPMD) highlighted the importance of data sharing and dataset linkage for rare disease research. She discussed PPMD’s work to develop a clinical trial master protocol and network for research on Duchenne muscular dystrophy, a rare progressive neuromuscular disease. She noted that other rare diseases face similar challenges related to research, so PPMD’s goal is to create a framework that could be replicated.

Furlong said that several organizations have conducted research studies and clinical trials to investigate treatments for Duchenne. She also noted that several past clinical trials for drugs to treat the disease have failed. Each of those clinical trials, she said, set up a significant data infrastructure, but when the trial failed, the infrastructure was dismantled, and frequently the data collected during the trial were lost. Furlong also noted that the data from failed clinical trials were not integrated into any other datasets. PPMD determined that this was a major barrier for progress and has been working on finding a mechanism to improve data sharing and linking for research.

Another challenge related to Duchenne research highlighted by Furlong was trial participation. Furlong noted that participation in a clinical trial requires time, money, and travel that are not always accessible to the patients and their families. She explained that any or all of those factors play a role in the decision to participate in such a trial. If a clinical trial master protocol were deployed at multiple sites, she suggested, this could increase the number of patients who could participate.

Furlong noted that PPMD’s first action to improve data sharing for research was to work with the Critical Path Institute. In this work they

developed a disease progression model for Duchenne muscular dystrophy, one that offers a central source for those data to be accessed by those involved in research. She explained that one of the data needs in research on this disease concerns its natural history and the effects of delayed treatment, data that could be culled from the placebo arms of clinical trials. PPMD is seeking to address this by creating a centralized database as part of its clinical trial master protocol. A key part of that infrastructure would be a mechanism for linking non-proprietary data from clinical trials that are using the master protocol. PPMD expects that this would result in larger datasets for analysis than are typically available for rare disease research. PPMD is also working on creating systems for standardizing data collection as part of its master protocol. Furlong concluded by emphasizing the importance of data sharing and improving stakeholder access to data for the benefit of patients.

James Lewis, University of Pennsylvania, spoke about his experiences with patient-centered outcomes research (PCOR) data in the context of his research on inflammatory bowel disease (IBD). He described several barriers related to the patient-centered component of PCOR, including the following:

• High expenses and time commitment needed,
• Low patient engagement,
• Reluctance among patients to be active research subjects, and
• Patient concerns about the safety of their data.

Lewis noted that in regard to the outcomes research component of PCOR, a barrier of note is that while outcomes research routinely measures processes or clinical outcomes, patients may have different priorities for information.

Lewis also discussed his experiences with the IBD Plexus program, which he described as a collection of ongoing and historical cohort studies from which they are working to build a centralized data warehouse. This data warehouse will be designed to link patients’ electronic health records (EHR) data, biosample-derived data, and claims data to assist researchers in developing a full picture of the patient’s experience with their disease. Lewis noted that an important component of this process was developing a SmartForm, one that could be integrated into Epic EHRs, to standardize how clinicians record patient data in the EHR. They have also engaged another tool in Epic EHRs that allows patients to report their symptom data for integration into the EHR in a standardized format. IBD Plexus is in the early stages of a study that involves a partnership with the Food and Drug Administration (FDA) to use the FDA MyStudies app to collect patient-reported data.

Lewis cited low data quality as one of the challenges with using data from EHRs and patient-reported data. He said these sources of data can be inaccurate or incomplete due to patients’ misinterpreting or skipping questions or because of clinician time constraints. He noted that recently implemented Centers for Medicare & Medicaid Services rules around charting requirements may impact the amount of EHR data available for PCOR.

Lewis described possible solutions for filling data gaps. One of these solutions is to link EHR data, claims data, and patient-reported data, which allows researchers to develop a more holistic dataset for a patient, including information beyond what is captured by a single health care system. However, Lewis noted, concerns about confidentiality and data security among stakeholders—including patients, insurance companies, and researchers—are a challenge for data linkage. He also noted that some patients are concerned about whether their data are being used appropriately.

Finally, Lewis discussed natural language processing to convert free text that is entered in EHRs as another solution for filling in data gaps. However, he noted that given the large number of variables, that process often requires a significant time investment. Another potential solution he proposed was the creation of a neutral national organization to standardize, de-identify, and link data from major EHRs and major sources of claims data for use in research. Lewis concluded by highlighting the need to recognize the multidisciplinary nature of PCOR and the need for collaboration.

Marc Natter, Harvard University, said that early in his career he encountered difficulties finding longitudinal data for PCOR focused on childhood arthritis due to inadequate registries and data infrastructure. He noted that the Childhood Arthritis & Rheumatology Research Alliance (CARRA) and the CARRA Registry were developed to address this challenge. He explained that in the beginning, CARRA developed a registry that served as a federal data warehouse for data derived from studies of pediatric rheumatology diseases.

Natter said that over time the CARRA Registry evolved into a common data collection platform in addition to a research network. The platform includes a master data-use agreement and protocol that applies to all participating institutions, and to all identified and de-identified data, which flow bi-directionally between research sites, CARRA’s Data Coordinating Center, and clinical centers. Natter explained that CARRA continues to build on its data infrastructure, which has facilitated a variety of pediatric rheumatology PCOR studies. He noted that he and his colleagues have also incorporated data infrastructure to facilitate collection of patient-reported outcome (PRO) data, including an application-based tool. He noted that they continue to research how best to collect and integrate PRO data.

Natter also highlighted some lessons that were learned as the CARRA registry and data platform have evolved. He noted that at some institutions

they have faced challenges obtaining Institutional Review Board (IRB) approval for studies that include a patient-facing application platform. He suggested that the Department of Health and Human Services work to develop a set of universal guidelines for these patient-facing applications. He said that this would provide IRBs with clear standards to apply when evaluating proposals that include such applications. He also echoed points made by others about the need for researchers to communicate with patient and physician stakeholders to ascertain what information matters to them and to use that to guide how PCOR analytic results are presented.

Natter also noted that methods research needs more attention and the outcomes of that research should be incorporated into PCOR data standards. He said that validated disease metrics are needed to facilitate more effective use of advanced machine learning algorithms. He concluded by highlighting that as other speakers had noted, researchers need better access to large aggregated datasets that include claims data in order to conduct more robust PCOR data analysis.

DISCUSSION

The brief discussion after the presentations focused on the role patient organizations are best positioned to play in advancing PCOR. Participants highlighted patient organizations’ ability to facilitate connections between researchers and people (patients and others) who are invested in the patient-centered aspect of PCOR. Participants also noted a potential role for patient organizations in facilitating patients’ access to their own health data, including claims data. This would enable patients to contribute their information to research studies. Patient groups can also play an important role in providing input on outcome measures that matter most to people. Many of these ideas echoed the committee’s conclusions from the first workshop.¹

CONCLUSIONS

Collaborations with patient organizations can help in addressing patient concerns about participating in research studies and in building patient engagement, which are both important for achieving a patient-centered approach. Disease registries directed by patient groups can be a particularly useful additional source of data, providing information that would not be available to researchers otherwise.

___________________

¹https://www.nap.edu/catalog/26297/building-data-capacity-for-patient-centered-outcomes-research-interim-report.

CONCLUSION 6-1: Patient groups can be helpful partners in all aspects of patient-centered outcomes research, including engaging patients in order to improve research participation and the impact of results.

CONCLUSION 6-2: Patient-directed disease registries can be a source of in-depth, longitudinal, prospective clinical and patient-reported data that are not available from other data sources.

This page intentionally left blank.