Patient Engagement and “Real World Evidence”
The day-long September 2018 workshop, “Medical Devices-Patient Engagement in Real World Evidence: Lessons Learned and Best Practices,” sponsored by the Food and Drug Administration (FDA) and University of Maryland (UM), met on the Baltimore campus, the city where I spent my graduate school years. In contrast to Baltimore’s palpable desperation, UMB’s health campus gleamed with newness, its brick walkways and tastefully planted vegetation viewable through floor-to-ceiling windows. In the well-appointed auditorium, Dr. Jeffrey Shuren, director of the Center for Devices and Radiological Health (CDRH, pronounced ‘cedar’), closed his introduction to the conference with the pronouncement that as the FDA moved toward real world evidence (RWE), “patient engagement” and the data patients may collect are invaluable for RWE.
In the last decade, the FDA has sought to incorporate patients’ involvement in the drug and medical device development cycle. In addition to the Baltimore event, CDRH has facilitated workshops and advisory committees for patient voices, including the Patient Engagement Advisory Committee (PEAC), which I attended in November 2018. New metrics, “Patient Reported Outcomes” (PRO) and “Patient Preference Information” (PPI), support the FDA’s analysis of new devices’ risk-benefit; these metrics seek to incorporate data, and perhaps more importantly, experiences, that standard clinical trial endpoints have traditionally neglected. The only patient speaker, an older white woman with temporomandibular joint disorder (TMJ), described with nuance and conviction how she became a patient advocate after serious harm she’d experienced due to lack of proper FDA oversight of medical devices and related surgery. A few months later, during a phone interview, she explained, “you go [to see your doctor] because you need help. If you’re being kicked out because you’re empowered, you’re not getting help.” Empowerment, she suggested, sucks.
At the Baltimore conference, the FDA attendees asked repeatedly how to turn “data into evidence” and what to do with, what they called, qualitative “anecdotes.” They asked a sociologist, who was there also doing fieldwork, and me about these perceived limits of qualitative data, as though there were not 40+ years of medical anthropology, medical sociology and STS to give contour to their questions. The FDA strategies to incorporate patients as novel data generators juxtapose starkly with open source advances among a highly engaged patient group: members of the type 1 diabetes (T1D) community.
Open Source Patients
Driven by patients and their caregivers, the T1D community coined a hashtag, #WeAreNotWaiting, to proclaim their impatience with commercial/government developments to cure T1D, as well as other limitations in T1D quality of life. Not only does the hashtag reflect a refusal of the promised, yet, perpetually-regressing five-years-away diabetes “cure,” for some it has motivated creative bricolage and hacking medical devices to create tools to improve T1D life. Although those living with T1D are not the only patients who have managed to upend the doctor-as-expert model (for example, Hugo Campos whose determination to gain ownership and control of his medical device data is extraordinary), the communities that have created and use the do-it-yourself artificial pancreas systems (APS), and related DIY T1D data platforms, exemplify a movement that deserves to be reckoned with.
A super-abbreviated history that galvanized these communities is as follows: insulin, a hormone the pancreas produces, was first administered to a T1D patient in 1922; its patent was sold for $1 to ensure universal access; insulin is now outrageously expensive in the U.S. and in many other places where access is limited and/or the quality of the insulin varies (and therefore its effectiveness is unpredictable); home blood glucose monitoring became possible in the late 1970s, partly as an artifact of gestational diabetes, a domestication of clinical blood glucose testing; insulin pumps became available in the early 1980s; the continuous glucose monitor (CGM), that allows for 5-minute intervals of glucose level reading through sensors embedded in the skin became commercially available in the early 2000s. Yet, the practices of managing T1D have remained relatively static in the last 100 years. T1D requires a person to keep her body within a controlled glucose level range, which is impossible without insulin. Glucose highs (hyperglycemia) and lows (hypoglycemia) introduce both immediate and long-term problems, including death.
Living with T1D requires an immense amount of data and self-surveillance, with “approximately three hundred [daily] decisions.” For those who are insulin dependent the CGM, with its constant update of data, is a remarkable technological advance: it allows wearers to understand how eating/exercising/sleeping/hormonal fluctuations affect their glucose levels, which in turn gives them insight into what they might change (or, do nothing). It helps people respond to hypoglycemia when they may not notice dropping glucose levels. Even my interlocutors not using a semi-automated system described their CGM as irreplaceable.
In 2011, two hackers announced a security flaw in a leading manufacturer’s insulin pump. The breach inspired others to capitalize on software vulnerabilities, leading to CGM data sharing “in the cloud,” which became Nightscout. This remote monitoring feature was later incorporated in commercial manufacturers’ CGMs. In the last five years, some T1D communities have surpassed commercial manufacturers’ technologies, developing a customizable hybrid closed-loop system. The community-built systems semi-automate diabetes control through OpenAPS, AndroidAPS, or Loop, with predictive algorithms, allowing the pump to respond to CGM data. They have also assembled novel tools to facilitate interoperability, such as RileyLink and the OpenAPS RaspberryPi HAT.
These technological developments are collaborative, with notable individuals contributing to the progress, including Dana Lewis, Scott Leibrand, Ben West, Nate Racklyeft, Pete Schwamb, Chris Hanneman, Katie DiSimone, Kate Farnsworth, and many others. Their efforts have created meaningful change for life with T1D. Most of the work has occurred remotely and asynchronously, although since 2016 Nightscout has organized five in-person hackathons on the U.S. west coast. In addition, a non-profit diabetes software platform, Tidepool, has added another layer to the assemblages of diabetes-related products developed within T1D communities. Tidepool presents a different direction for patient-organized, open source health interventions but merits a separate discussion. The community’s hyper-engagement has been challenged by commercial companies seeking to protect their “proprietary” software, despite the data being generated by the users’ bodies. The FDA has had to evolve in its own expectations of device manufacturers’ products, approving and facilitating interoperability in diabetes devices in 2018. The T1D open source communities take health care innovation on a radically different trajectory than existing commercial and regulatory imaginations of patient “participation” and “engagement.”
The Fantastical Empowered Patient
What stands out, as I barely capture the significance of the DIY open source T1D efforts, are these patients’ alternate paths to improving their health; they’d prefer not to be the ones to create what they need. Their “super engagement,” which in theory, the FDA might fantasize about in other patient populations, however, have not always been received enthusiastically by clinicians, regulatory, or commercial entities. The FDA’s promises to include patients draw on existing clinical models of research and development, but they do not trouble the fundamental categories of what is possible. Perhaps it is unrealistic to expect the U.S. drug regulatory agency to restructure entirely, and, thus, the slow progress to include patients is still notable. But it is not just the anthropologist’s fantasy that there are alternative worlds possible. The interventions the open source T1D communities have accomplished reveal concrete methods for patient expertise and knowledge to lead health technology. Few disease conditions present as many decision-making moments for patients, and thus, T1D is unique; yet, this community also introduces a paradigm shift. Their enactment of “engaged” patients illuminates how regulatory fantasies of empowered patients neglect patients’ own versions of participation. The FDA imagined fantastical empowered patient is a compliant patient, rather than the patient-as-actor-creator and disrupter.
This work was funded by the National Science Foundation Science, Technology, and Society, grant # 1632716.
Works Cited/Thought With
Blume, S. S., & Rose, D. (2005). Citizens as Users of Technology: An Exploratory Study of Vaccines and Vaccination. In How Users Matter: The Co-Construction of Users and Technology. MIT Press.
Blume, S. S. (1997). The Rhetoric and Counter-Rhetoric of a “Bionic” Technology. Science, Technology & Human Values, 22(1), 31-56.
Epstein, S. (1995). The Construction of Lay Expertise: AIDS Activism and the Forging of Credibility in the Reform of Clinical Trials. Science, Technology & Human Values, 20(4), 408-437.
Epstein, S. (2008). Patient Groups and Health Movements. In J. Wajcman, M. Lynch, & O. Amsterdamska (Eds.), The Handbook of Science and Technology Studies. The MIT Press.
Jain, S. L. (2013). Malignant: How Cancer Becomes Us. Univ of California Press.
Kelty, C., & Panofsky, A. (2014). Disentangling Public Participation in Science and Biomedicine. Genome Med, 6(1), 8.
Klawiter, M. (2002). Risk, Prevention and the Breast Cancer Continuum: The NCI, the FDA, Health Activism and the Pharmaceutical Industry. History and Technology, 18(4), 309-353.
Klawiter, M. (2004). Breast Cancer in Two Regimes: The Impact of Social Movements on Illness Experience. Sociology of Health & Illness, 26(6), 845-874.
Langstrup, H. (2011). Interpellating Patients as Users: Patient Associations and the Project-Ness of Stem Cell Research. Science, Technology, & Human Values, 36(4), 573-594.
Pols, J. (2013). The Patient 2.Many: About Diseases that Remain and the Different Forms of Knowledge to Live with Them. Science & Technology Studies, 2, 80-97.
Pols, J. (2014). Knowing Patients. Science, Technology, & Human Values, 39(1), 73-97.
Rabeharisoa, V., & Callon, M. (2002). The Involvement of Patients’ Associations in Research. International Social Science Journal, 54(171), 57-63.
Rabeharisoa, V., Moreira, T., & Akrich, M. (2013). Evidence-Based Activism: Patients’ Organisations, Users’ and Activist’s Groups in Knowledge Society. CSI Working Papers Series 033.
 I am especially thinking here of the work on patients and their diverse advocacy and participation practices. The literature is obviously quite rich; in addition to Steven Epstein’s work on HIV/AIDS advocates (1995; 2004), Maren Klawiter’s (2002; 2004), and Lochlann Jain’s (2013) distinct works on breast cancer advocacy and patients, there are of course, also Blume and Rose (2005), Blume (1997), Callon and Rabeharisoa (2008), Rabeharisoa et al. (2013), Langstrup (2011), Pols (2013; 2014), Kelty and Panofsky (2014) and many others who have thought about what patients do, what constitutes expertise, and how citizens (not just patients) engage with scientific inquiry and data.
 These communities have received a fair bit of public coverage, including a Prognosis podcast, Wall Street Journal, and other main stream media, only some of which I’ve listed here.
 One diabetes-related site identifies the origin of this term to a 2013 conference, attributing “we are not waiting” to Lane Desborough, a co-founder of Nightscout.
 The high cost of insulin in the U.S. has led to some people rationing their insulin and subsequently dying or having serious complications. In April 2019 insulin makers’ testified before the U.S. Congress to explain the high cost of insulin.
 I use their names because they have public online personas. I have interviewed some of these individuals, as their narratives are important to the history of their work, but I use pseudonyms when I use non-identifiable parts of those narratives in my writing. There are others who have contributed social and community aspects of these projects that are no less significant. I want to acknowledge the generosity of the community in allowing me to observe and to occasionally participate in their hackathons and other activities.
 Tidepool has engaged with the FDA in another novel fashion, as a member of a pilot program to restructure regulation of software as medical devices. A brief history of the FDA’s process of pre-certification can be found here. Tidepool is the only non-profit company among nine companies involved.
Your blog is really comprehensive. More power