Call to Action: Sharing of Patient-Level Data for Parkinson’s Research and Care

Sue Dubman, Patient, Patient Advocate and Researcher
Sue Dubman has more than 20 years of experience in health care and research informatics for academia, industry and the government (NIH). Currently the Senior Director, Informatics Innovation, at University of California San Francisco Cancer Center, Sue is part of a multi-disciplinary team to innovate changes in clinical care and research. Diagnosed with Parkinson's disease (PD) in 2009, Sue has become an active advocate for improving the lives of those living with PD. Sue will have a poster at WPC 2016 (Exhibit Hall B, Level 1, Poster Board Number: P40.09).

Introduction
As a person in cancer research, I never even considered Parkinson’s (PD) as a possibility. Everyone in my family died of cancer. But, there I was in 2009 getting a diagnosis of a chronic and progressive disease, a diagnosis that was devastating as much for me as my family. After a few years of denial, I knew I needed to do something and believed that my experience, as both a patient and as someone in biomedical research and care, could be put to good use so that, in the future, others and their families wouldn't have to suffer like mine.

It was a natural for me to consider becoming a patient advocate. When I was at the National Cancer Institute, I worked with many patient advocates so I was already familiar with what they do and how important they are. Today, I am an active patient advocate with a focus on accelerating treatments to market. As a person in cancer research, I have seen many new innovative ways to bring better and safer treatments to market faster and I want to apply that knowledge and experience to Parkinson’s research. It probably won’t help me but my hope is that someday others will not receive the same message that I got, which is “you have Parkinson’s, a chronic and progressively debilitating disease but we can only treat the symptoms”, but rather the doctor will say “you don’t have to worry, we can treat this”. There's hope in the science and that's what motivates me. 

The problem
Today in Parkinson’s clinical research, the sharing of patient-level data has been very limited in scope. Much of the existing data is almost exclusively considered the property of those that paid for the development of a dataset. This means that others often don’t get access to the data at all or, if they do, they get access many years after the data was collected.

The data that needs to be shared comes from multiple sources including clinical trials, insurance claims, electronic health records (EHRs), patient reported outcomes, web-based patient “social” communities (such as Patients Like Me), registries, wearables, devices and other observational and study datasets. This includes datasets specifically created for Parkinson’s research as well as those datasets that contain valuable Parkinson’s patient information although they were not created for that purpose (such as insurance claims systems or the Veteran’s Administration Million Vets Program knowledgebase). 

For the purposes of this blog, I want to focus on patient data from clinical trials. While recently there have been more attempts to share clinical trial data, the data made available to others is often not in a form that is really useful, is stale and is restricted to a small community of researchers. In addition, it is very difficult and, at times, nearly impossible to combine that data with other data since existing data standards were not used. All of this limits the value of data. 

The current model is based on paper-based tools developed in ancient China during the Han dynasty (206 BC – 220 AD) and disseminated via print publication, the origins of which date back to the time of Guttenberg (1440). If we continue to adhere to this model, our understanding of clinical interventions and our ability to develop new treatments will continue to be limited. In clinical research, the impact is felt in the following ways:

    1. A select number of individuals often decide which analyses to conduct, choosing some at the exclusion of others. An analysis that might have been of great interest to another investigator (and which may have a direct bearing on clinical practice) may not be performed.
    2. Among the many findings generated, only a select number might be included in any peer-reviewed publication. The research community and clinicians may never know about findings generated, but not disseminated.
    3. Among all trials conducted, there may be significant publication delays so the knowledge gained in the research may not be known to other investigators for many years.
    4. Only a limited number of trials are eventually published. The “failed” studies are often archived and never to be used again. Don’t we learn as much or more from our failures as our successes?

The cumulative effect is that patients, neurologists, other healthcare professionals and the research community are placed in the position of making clinical care and research decisions with access to only a fraction of the relevant clinical evidence that might otherwise be available. When neurologists recommend treatment options to Parkinson’s patients, this is routinely done on the basis of information that is biased and seriously incomplete. This standard of practice is tolerated because we are accustomed to it. The care delivery and research communities often only become aware of a treatment’s shortcomings when safety concerns are raised about a drug, device, or other treatment strategy.

One patient’s view
As a researcher turned patient, I understand the importance of clinical research data and results. I want my information to be shared for the greater good. While I certainly have concerns about use and re-use of my data, these can be alleviated by responsible sharing of data that protects patient privacy and security; by incentives for researchers to produce and ensure high quality data for sharing with peers, the broader scientific community and the public; by increased data circulation and by data sharing encouraged or mandated by government. Best practices to ensure better transparency and to enable reproducibility of results in creation of datasets include:

    • Use of common, shared data definitions so data is actually re-usable and shareable.
    • Good, clear documentation so it is easier for others to understand data content and encourage collaboration.
    • Location and access to the data to make it easier to share.
    • Making the data easily discoverable through advertising and easy access. 

The investigator’s dilemma
Data sharing is increasingly common in some areas of medical research, particularly among genomics investigators; however, individual, patient-level clinical trial data sharing is less common because of concerns and challenges with the actual act of data sharing. The principal concern, voiced by investigators, is that a substantial amount of individual time and effort has been invested to design the trial and collect the data and that, in return, they argue that they deserve ample opportunity to conduct their analyses and disseminate their findings.

Without question, investigators do deserve extra time during which they can prioritize their analyses and publish their work; however, when dissemination delays are two, three or more years, this inevitably slows and diminishes the impact of any research. As a researcher, I understand the concern of investigators being “beaten to the punch” with their own data. As a patient, however, I ask that investigators weigh the time and effort they have invested (often with government/taxpayer funds) against the value of earlier data sharing, if for no other reason than to avoid duplication of effort. 

Another investigator objection to data sharing includes concerns that multiple analyses by various independent research groups will produce analyses with differing results, either because of human error or because external investigators conduct inappropriate analyses; that clinical trials are designed with pre-specified study protocols and that additional analyses amount to “data-dredging”; and that data ownership belongs by right to the original investigative team. As a researcher, I know that the scientific community is well-positioned to review and put into context differing results from the same (semantically equivalent) trial data, as well as to judge whether data has been “dredged” or appropriately analyzed. As a patient, I ask: “whose data set is it anyway?” While I understand that data legally belongs to the investigators, medical science is essentially an enterprise conducted for moral reasons.

The scientific community must reach a consensus about several critical points before the promise of sharing clinical research data can be fully realized. These include:

    1. What are the responsibilities of the original investigator team? To share data effectively, they must produce a clean, well-described, and accurate data file that can be used by others and protects patient confidentiality.
    2. Who supports the investigators’ efforts to create these data sources?
    3. What if there are subsequent questions and inquiries? Who bears responsibility for the shared data?

Patient privacy and security
Another broad issue is the question of who should be allowed to access the data and how access should be provided.

The scientific community began the process of developing standards and solutions to these data sharing problems as early as 2009. Some examples include:

    • The Institute of Medicine (IOM) set forth recommendations on managing research data in the Information Age in 2009.
    • The Wellcome Trust convened a number of research funders, including the World Bank, the National Institutes of Health, and the Bill and Melinda Gates Foundation, to develop a coherent vision, principles and goals to promote the sharing of research data to improve public health in 2012.
    • Journal editors have developed guidance for the preparation of raw clinical data for publication.

In Parkinson’s and other areas of neurological research there are several prominent examples of data sharing currently underway that are illustrative and can inform our expectations for open scientific and data exchange.

    • Critical Path Institute’s Critical Path for Parkinson’s (CPP) consortium is focused on sharing precompetitive patient-level data from observational cohorts and legacy clinical trials and implementing consensus data standards. See: https://c-path.org/programs/cpp.
    • Parkinson’s UK, the strategic partner for CPP, has been a leader in bringing together many participants in the Parkinson’s research enterprise to find ways to share data and work collaboratively.
    • The Parkinson’s Progression Markers Initiative (PPMI) funded by the Michael J. Fox Foundation for Parkinson’s Research is aimed at identification and validation of biomarkers foster open data sharing of patient level data.
    • The National Institute of Neurological Disorders and Stroke (NINDS), Critical Path Institute and the Clinical Data Interchange Standards Consortium (CDISC) have invested in development of clinical data standards for Parkinson’s research, a necessary step in improving combinability of data.
    • The mPower App is a digital health smartphone app aimed at tracking Parkinson’s symptoms where >10,000 patients with PD agreed to share their data at the time of downloading the app.

Summary of key data sharing benefits
Data sharing, especially for Parkinson’s where there are no effective treatments, is critical. Science is a community, continually building on each other’s ideas. In the era of electronic knowledge exchange, only when data sharing becomes the norm, can we derive its full benefits, including:

    • Increased speed of quality, structured and re-usable information from care delivery to research and back to care delivery tohelp create a true learning health care system so that evidence is available when and where it is needed, resulting in more effective and more efficient care and research.
    • More connection and collaboration between patients, researchers, industry and government, which can result in important new findings within the field.
    • Ability to leverage the data and build upon the work of others rather than repeating already existing research.
    • More widely disseminated information for more informed decision-making for planning and policy.

Increased open science and information exchange through data sharing will further the value of all clinical trial research. It is in the public's interest to have access to comprehensive clinical trial data to ensure a complete understanding of drug or device safety and effectiveness.

Call to action
It is clear that increased sharing of patient-level data holds promise for scientific advancement in Parkinson’s and other research but we need to overcome the barriers that limit data sharing today. This is a call to action. Parkinson’s patients and their advocates should insist and expect that:

    • Government continues to mandate guidelines to ensure responsible sharing of clinical trial data that balances the interests of stakeholders with the public demand for information on the effectiveness and safety of therapies. 
    • All stakeholders, from drug and device manufacturers to medical researchers to medical journals and regulatory bodies must understand the importance of data sharing from a patient perspective and make data sharing the norm rather than the exception.
    • Data sharing plans are mandated by government to be a part of all clinical research protocols including detail on what data will be available to others for the purpose of advancing science.
    • Clinical research data are made available for sharing much more quickly than they are today. 
    • Groups like CPP create new information technology platforms to support data sharing. 21st century problems require 21st century solutions. Leverage solutions that have been used in other industries to speed implementation of new drugs and devices.
    • Government and industry implement strategies to encourage use of clinical data standards and development of incentives for data sharing, both “carrot” and “stick”.

While data sharing of clinical trial data is not a panacea, it is an important piece of the puzzle to bring new safe and effective products to market much faster. For those who question, “Why now?”, I say “Why not sooner?”. I and patients like me can’t wait 10 to 20 years for new drugs or devices to be approved.