Advocates & Other Big Data Stakeholders Module 4 This training was partially funded through a Patient-Centered Outcomes Research Institute (PCORI) Eugene Washington PCORI Engagement Award (1492-RUF).
The following information has been developed with assistance and input from Jane Perlmutter, PhD – Founder of Gemini Group Consulting, patient advocate, and member of the BD4P Steering Committee. This information does not constitute an endorsement by Gemini Group. Disclosure
Module Objectives Review and refine the definition of “big data” Identify stakeholders in healthcare big data Introduce some key big data issues for patients and their advocates Discuss how patients and advocates are involved in big data Compare and contrast modes of stakeholder involvement in same big data projects Module Objectives
What is Big Data?
What are the 4 Vs of Big Data? Velocity – the speed of the data. Volume – the size of the data set. Variety – the different types of data being used. Veracity – the measure of how well the data reflects reality. What are the 4 Vs of Big Data?
Modes of Data Collection Data can be collected everywhere… During the regular course of care – EHRs, medical records, doctors’ notes, diagnostic tests, etc. After regular course of care – insurance claims, treatment summaries, etc. From patient-reported data – wearables/fitness tracker information, patient-reported registries, patient-reported outcomes, health surveys, etc. From research – clinical trials, observational studies, molecular data, etc. Modes of Data Collection
Ethical Concerns of Data Collection There are ethical concerns about non-profit vs. for- profit research collection: Who has my data? Who are they sharing it with? Am I okay with a company making a profit from my data? Is/where is my data being stored? Who can access my data? Are there any other ethical concerns you can think of regarding non-profit vs. for-profit collection? Ethical Concerns of Data Collection
Types of Data/Examples Quantity/ Quality of Data Inference Method Interventional – Randomized Clinical Trials Small/ Generally Excellent Analysis of Variance Observational – Registries, Surveys, Passive Mobile Data Medium/ Very Variable Regression Unstructured – Social media, research articles Large/ Often challenging to analyze Artificial Intelligence, Natural Language Processing Big Data – Combining Multiple Types and/or Data Sources Often difficult to combine Multi-dimensional Analytics and Visualization tools What IS Big Data? This chart is one view of how big data relates to other types of data, each of which have advantages and disadvantages, and have unique approaches to drawing inferences
Types of Big Data Research Hybrids Real World Evidence Rigor Clinical Trials Molecular Databases Observational Studies Types of Big Data Hybrids Pragmatic Trials Registries Rigor Generalizability Real World Evidence Electronic Media Records (EMR) Billing Data Patient Blogs, Discussion Groups, Etc.
Data is not information, information is not knowledge, knowledge is not understanding, understanding is not wisdom. - Clifford Stoll Seek Wisdom
Why Big Data Now? One theory… Greater demand for data Across all sectors, including healthcare Increased supply of data EMRs and clinical data Advances in technology Greater functionality and ability to synthesize and analyze data More government support and utilization Focus on increasing data availability and interoperability Source: McKinsey & Company, “The Big Data Revolution in Healthcare” Why Big Data Now?
What and Who are Big Data Stakeholders? A stakeholder is an individual, business, or organization that has an interest in an particular issue, or one who is involved in or affected by a course of action Every stakeholder is going to have different interests in being involved in big data, and different roles in the data process What and Who are Big Data Stakeholders? Big Data Patients & Caregivers Advocates & Advocacy Organizations Payers Providers & Health Systems Professional Organizations Government Researchers Pharmaceutical Companies
Stakeholder Goals in Big Data Diverse stakeholders in biomedicine and health research have different needs and objectives for collecting and utilizing big data: Patients, Caregivers, and their Advocates – using the results to learn more about their conditions and benefitting from better treatments. Payers – determining which treatments are cost- effective Providers & Health Systems – informing better treatments and working in partnership with patients in patient-centered and evidence-based care Researchers (including governments and pharmaceutical companies) – analyzing results to make advancements in treatments and therapies Stakeholder Goals in Big Data
Outcomes of Big Data Use How do innovations and the use of big data affect diverse stakeholders? Patients & Caregivers More effective and efficient treatments Opportunities to participate in initiatives Better understanding of disease and treatment options Advocates & Advocacy Organizations Better understanding of disease and how to advocate Promotion of patient-centered treatments and therapies Payers Finding new methods to increase value and quality while decreasing costs Changing payment structures based on insights from big data Outcomes of Big Data Use
Outcomes of Big Data Use (cont.) Providers & Health Systems Increase in use of evidence-based medicine and personalized treatments Researchers (including governments and pharmaceutical organizations) A move towards new forms of research and gain more comprehensive information Using new tools and innovations to change policies and standards Obtaining better, more effective, and cost-efficient therapies Outcomes of Big Data Use (cont.)
Concerns Common to All Big Data Stakeholders Interoperability and data standardization – will the data be useable? Exchange of data – will the data be able to be read by those who need to read it? Access of information – will the data be accessible to those who need to obtain it? Ensuring security and privacy – will the data be securely obtained, held, and disseminated? Concerns Common to All Big Data Stakeholders
Sponsors and investigators share many of the same concerns that patients/advocates do, though they sometimes take a backseat to technical and scientific concerns. Patient advocates are more focused on patient issues. Patient advocates have a responsibility to bring the patient issues to the forefront. Patients can help to identify incorrect information that could lead to inaccurate data sets (i.e. wrong procedure/billing codes) Advocacy Reminders
Big Data Issues Especially Important to Patients/ Advocates For research projects & big data initiatives: How representative are the data of the population? Who pays for the data and infrastructure and how is it sustained? Who profits from the project? Does the informed consent meet best ethical and patient-friendly practices? What plans are in place to return results to those who contribute data? What are the research priorities? Big Data Issues Especially Important to Patients/ Advocates
For research projects & big data initiatives: What data controls have been established? How are the data cleaned and/or curated? Is there a data monitoring plan? How will missing data be handled? Are there other checks and balances? What data sharing plans have been established? Are common data elements and other standards used, when feasible? Are the data readily available to all comers? Are guidelines in place to vet requests for data? Is there adequate funding and other support to facilitate sharing? Big Data Issues Especially Important to Patients/ Advocates – QUESTIONS TO ASK
For research projects & big data initiatives: What security and privacy measures have been established? Do the security and privacy practices meet minimal legal and regulatory requirements? What are the trade-offs for establishing more or less stringent requirements? What’s behind the models, algorithms, and parameters? What assumptions are built into the model? What is the evidence to support them? Have you done a sensitivity analysis of the sensitivity to the assumptions, algorithms, and parameters in the model? What is your evaluation and improvement plan? Big Data Issues Especially Important to Patients/ Advocates – QUESTIONS TO ASK (cont.)
From Patient to Research Advocate Patients Advocates Research Advocates From Patient to Research Advocate Have somewhat longer term perspective Want to prevent others from going through what they have Have a great diversity of knowledge, opinions, and approaches On average, more interested and knowledgeable about science Want to ensure that research is efficient, effective, and patient-focused Cannot wait Often cannot advocate for themselves Often willing to take great risk for low probability of gain
Different Roles for Different Folks Type of Advocate Most Effectively Involved Survivor with a Personal Compelling Story Clarify (naïve) concerns among diverse patients Generate media attention Impact legislators and regulators Staff Member of Advocacy Organization Identify potential contributors of data to big data projects Identify potential advocates to inform big data projects Clarify and advocate about big data policy issues Inform and activate community about big data research efforts and policy Individual Research Advocate Provide (informed) input on patient needs, concerns, and priorities Influence big data research efforts and policy issues Review informed consent documents and big data contributor educational material Different Roles for Different Folks
How are Patients and the Public Involved in Big Data? Contribute their data, knowingly and unknowingly Fundraise, directly and indirectly Lobby for health research Prioritize questions and research efforts that are important to patients Be active members of big data steering committees and advisory boards Review informed consent documents and other educational materials Help inform the public about big data Provide patient perspective on big data panels How are Patients and the Public Involved in Big Data?
Summary: Advocate View of Big Data The promise is huge… ...But there are risks Promise Risk Better predictions about health risks Immature methodology erroneous inferences Faster development of treatments Breach of security loss of privacy More rapid progress toward precision medicine Hoarding of data slowing progress More efficient use of health resources Summary: Advocate View of Big Data
How to Make a Difference? Ensure patients/advocates are “at the table” and heard when decisions are made about big data projects Inform patients/advocates and the public about potential benefits and concerns associated with health data (messaging, training) Be discriminating in providing support to excellent projects and encouraging patients and researchers to share data Learn more How to Make a Difference?
Big data has lots of potential, but it’s more complicated than we can imagine Technical issues Political/economic issues Patient ethical issues Including patients/advocates from the beginning will lead to better, faster, and more acceptable results Take Home Messages
Modes of Stakeholder Involvement in Sample Big Data Projects Goals Modes of data collection Potential outcomes Funding Stakeholder involvement Modes of Stakeholder Involvement in Sample Big Data Projects
What are Similarities & Differences Among these Initiatives? Cancer Big Data Examples ‘Omics – Visualization, analysis and download of large-scale cancer genomics data sets for research Example: NCI’s The Cancer Genome Atlas, cBioPortal for Cancer Genomics Learning Systems – Real-world data for quality improvement and research Example: ASCO’s CancerLinQ, Flatiron Health, The Athena Breast Health Network Clinical Trials – Data sharing from clinical trials for research Example: Project Data Sphere What are Similarities & Differences Among these Initiatives?
What are Similarities & Differences Among these Initiatives? Other Health Examples Surveillance – multi-source data to monitor unidentified toxicities and drug interactions Example: FDA’s Sentinel Initiative Patient-Entered Data – patient support and information sharing Example: PatientsLikeMe Artificial Intelligence Processing – analyzing data and providing patient support Example: IBM Watson Health What are Similarities & Differences Among these Initiatives?