A community colleague asked for advice about how an organization can boost the signal for their patient registry. Recruitment levels are not what they expected and nowhere near what they need. I’m sharing what I wrote back so that others can chime in with their advice in the comments.
First, look at what successful registries and learning health networks have done — and what critiques they get, even as they succeed at their mission.
For example, the Cystic Fibrosis Foundation’s patient registry is acknowledged to be a linchpin in the progress made in extending the lives of people living with this chronic condition. Atul Gawande’s 2004 article, “The Bell Curve,” is a must-read on that topic.
Erin Moore is a mom of five kids, one of whom lives with cystic fibrosis. She has also educated me about the power of the registry — and the deep frustration among participants when the holding org doesn’t share access with patients & their family members.
She has pressed the Cystic Fibrosis Foundation to create an enhanced patient registry by adding, for example, a field for antibiotic use. As she put it succinctly in a tweet:
“All great talk about new trials for effect of antibiotic(s) on infections….that wouldn’t be necessary if registry tracked antibiotic use. We would already have our data.”
How might we ask people living with a condition for suggestions about what to track? I bet nobody at your organization is asking anyone other than their partner scientists and clinicians about what should be included in the registry. That’s a loss.
For more on the human right to actively participate in science, check out Article 27.
In a different but related context: ImproveCareNow is a national network of pediatric gastroenterologists who are changing how we treat kids with IBD and ulcerative colitis by harnessing the power of data. Check out their “purpose and success” page (and imagine how you’d create one for your organization, to better tell your story).
NIH All of Us has been wildly successful in recruiting a large, truly diverse cohort from all over the U.S. It may be worth a deeper dive into their website and other public documentation to see why.
Sage Bionetworks is an All of Us partner, helping them craft an informed consent process that conveys deep respect for what they are asking people to hand over. How might your organization create an inviting, engaging, respectful intake that makes people feel the love, feel the buy-in that you want them to feel?
John Wilbanks is one of the architects of this consent approach and he lays out the basics in his TED talk:
John wrote a series of tweets about one of their recent studies that was so good that Sage turned it into a blog post. It — and the full study — are worth reading because it contains insights about why people who signed up for a remote digital health study (like a registry, for example) stuck with it long-term.
Finally, your organization is not effectively tapping into the HUGE reservoir of energy that peer networks represent. Peer-to-peer health is the focus of my work these days. Here’s a 10-minute video (featuring a story about Erin Moore!) that captures what I mean:
How might you find ways to get the word out about the registry through existing networks of patients and caregivers? How might you gain the trust of key influencers in these peer communities, who can then act as your agents?
Now it’s your turn: What advice do you have? If you have joined a patient registry, what motivated you? If you have hesitated or decided not to join, why? What could change your mind?
Image: Donate Your Data to Science, a CareCard by GoInvo.
My takehome after building a lot of registry-like communities on PatientsLikeMe is that you can’t get traction of the type to be transformative unless the registry plugs into care somehow. A standalone black hole of data, no matter how neat the UX or pleading the marketing, will never engage more than 1-5% of the community, and even then only for a time (I’d love to hear stats from the PCORnet projects that tried this too).
The first time a clinician says “So I was looking at your and I think we need to discuss the following things today” is probably going to be the game-changer. Yes that is incredibly difficult, *but* I’m not saying she has to have it ported through on her EHR (or that she has to be a physician). Yes it could be activated patients doing some of the work, but again you’re talking “PAM level 4” activated patients and they are not usually where the most need is.
Those were super-high-level and a bit nerdy – I’m also happy to offer some free war stories to whoever is trying to address this challenge =)
Thanks, Paul!
One footnote for those who may not know about the Patient Activation Measure (PAM) which is a commercial product developed with funding from the Robert Wood Johnson Foundation (which always rankles me — why didn’t the funders ensure that the resulting tool remain open-access? But I digress.)
PAM quick summary:
Level 1 = Disengaged and overwhelmed
Level 2 = Becoming aware but still struggling
Level 3 = Taking action and gaining control
Level 4 = Maintaining behaviors and pushing further
For more, see the Insignia Health site.
Registries often rely on clinician recruitment — for example, a “Take One” display in a specialist’s exam room — and the recent study by Pratap et al. indicates that a clinician’s recommendation does spur action (see my comment here) — but I’d love to see a study that tracks the influence of a peer network. Has anyone done such work yet?
Thanks for this post Susannah, I’ll reply to the larger question of patient registry issues when I’ve had a chance to get my thoughts in order but I find it curious that the Patient Activation Measure was permitted to become private property after funding by RWJF. I’m pretty sure our contract with RWJF required that any work product be open. Maybe this was instituted later? Still, doesn’t seem right.
Paul
Really terrific comments–thanks for sharing!
Lorraine
I appreciate this discussion, it’s one of the few contexts where the non-biomedical/non-entrenched interest perspective can be rationally considered. I’ve talked to many people in the “needs not being met” category, and one of the most frustrating issues with collaborating on discovery arises when there is either no diagnosis or diagnosis is contested. This has affected people with Lyme, obviously, and it affects many others whose symptoms are part of complex chronic conditions. In the top down patient registry approach, the first step is identifying a group of patients, using diagnostic categories. But if your symptoms are not easily and confidently matched in the symptom/disease lookup table, for instance, if you have migraines, allergies, chronic pain, many kinds of mental health concerns, atypical recovery from surgery, etc etc etc you cannot easily join a registry conceived or managed using a top down process, nor can you always use the clinically appropriate descriptors or measurements that would be recorded in the registry. For chronic, complex conditions (at least), the concept of a registry should perhaps be supplemented by something more general and emergent, like a “participant led research” data store. A willingness to support these kind of ground level common resources would be a welcome departure from standard practice.
Gary
I agree re the contested disease etc, but the real issue I believe is the lack of $$ incentives. If pharma has no interest–b/c, for example, the market for developing generic antibiotics is deemed a poor investment choice, then funding becomes a big issue. Our registry was developed by very well informed and empowered patients–based on 10 years of large scale survey research and peer reviewed publishing–and a board of patients, physicians, etc.
That’s fortunate, but it is difficult for patient advocacy groups to self fund this type of project and develop the internal expertise to conduct research. We are doing it, but it’s a very, very heavy lift for a patient organization. Our population is stimatized also. Hence, protecting community interests is a priority as is vetting researchers and ensuring that data is used solely for the benefit of patients (not sold etc).
It may be that the best way treatment approach for our problem (can’t know now) is another drug already on the market. Then you are really about capturing N of 1 trials that arise from clinician innovation to see what works best. And large #’s of patients to subgroup.
I agree with you that you have to first “break out of jail” of traditional thinking and current models of research. You have to build a research community of bio-statisticians, AI etc around the registry. No helicopter research. We have published our first peer reviewed paper on patient treatment response variation https://www.mdpi.com/2227-9032/6/4/124.
This can be done and will improve with new technical tools that provide better inter-operability between the clinicians, the patient registry, and vetted researchers. It does also involve recreating a research culture that is truly patient centered–which I think means both patient funded (power dynamic) and patient driven research models. You have to have patients who are knowledgeable enough and willing to be patient PIs and not simply rely on the experts–who will provide an expert driven model of research.
And then, of course, LymeDisease.org is using its MyLymeData registry to create a “research engine” ala Groft by working with traditional researchers to help drive recruitment in their trials and collaborating with other patient orgs running a Lyme bio-repository, the Lyme Disease Biobank.
Lorraine
Today is a big day in the world of health data thanks to the U.S. federal government release of a new rule related to the interoperability of electronic health information.
In between reading & sharing updates on that work, I’m grateful to people who are helping me to answer the question that I was asked: How might an organization boost the signal for their patient registry?
Comments are open and everyone is welcome to share your perspective.
On Twitter, I’ve seen:
@ALSAdvocacy tweeted: #dingdingdingding From @SusannahFox Pls just read, esp if you think the world’s most expensive registry is the best registry. … Public Q&A: Patient registries…[In a separate tweet, linking to the CF registry] Look at annual data report. Registries that provide meaningful and timely data motivate patient enrollment. It’s not $100 million rocket science.
DoubleTrojan tweeted: Patients can’t query the @CF_Foundation‘s registry – they aren’t allowed access despite giving their information to it. Incredibly paternalistic and a slap on the face to those with #cysticfibrosis . The CFF is sidelining some great learnings due to this
Uppity Cancer Pt tweeted: The study @MBC_Project has been so successful in recruiting because patients helped design the registry and did a great deal of recruiting. The investigators also built trust by promising the data would be shared with other researchers for public good not profit #bcsm
Gunnar Esiason tweeted, in reply to my tweet linking to the CF registry: The only knock against the registry is that as of yet, patients are completely reliant on provider interpretation of the registry to inform care decisions and recommendations. We need to open it up to the patient community so that we can be better informed to make choices.
Is it possible that the architecture of patient registries is not adequate for making the discoveries we need? Where the data model is determined in advance, and especially where it is entirely dependent on clinical measures, it won’t be possible to address the many additional questions that develop over time or from communities of patients that weren’t anticipated by the registry designers. Erin’s critique about lack of information about antibiotic use in the CF registry is an illustration of this problem. Another approach to enabling collaboration on discovery would be support the development of “ground up” registries with maximally flexible data storage schemas (including mere file uploads of non-standard data), with data integration functions added on top. This makes some clinically oriented researchers I know roll their eyes, as they see it as pure chaos. But I’m not sure it’s correct to dismiss this alternative approach. By creating small, community driven data commons, such as the kind stewarded by Open Humans for the OpenAPS community, or the kind we created for our Quantified Self Bloodtesters project, we set the stage for research done with smaller cohorts using much more high density datasets, and allow the common data collection and analytical workflows to emerge out of these tests. So the order of registry development would be: Small cohort, tests, best practices, expansion of registries, workflow standards, data standards. Data standards come at the final stage, not the first stage.
Here’s another insight shared on Twitter by Kristen Honey: Lyme #MyData patient registry created by @lymepolicywonk w/ 12K+ ePatients trusting @Lymenews https://www.lymedisease.org/mylymedata-lyme-disease-research/
Lyme disease is an example of a condition that falls into the “needs not being met” and “people with resources/power don’t know how to help*” quadrant of a 2×2 diagram I’ve been working with recently. The asterisk is in honor of this Upton Sinclair quote: “It is difficult to get a man to understand something, when his salary depends upon his not understanding it!”
One might argue that ALS is another example of a health condition in that category since there are no known treatments. Cystic fibrosis *was* until their patient registry began to yield insights and now research that has led to the development of drugs promising to help 90% of those living with CF.
These are groups of people who are HIGHLY motivated to join a registry and donate their data. Is this a prerequisite for a vibrant patient registry? That is, if you or a loved one is facing a death sentence, you are motivated to join. But if you consider your health condition to be manageable, does that mean an organization recruiting you for a patient registry has to meet a higher standard than, say, the conditions I list above?
I’d love to hear from other people whose experience resonates with that description of the “needs not met” and “people with resources don’t know how to help” group.
I’d also love to hear from people whose experience is more along the lines of a “needs are being met” (i.e., there are known treatments, for example) group. Is there a patient registry for your condition? If it’s been successfully implemented, what motivated people to join?
Note: The organization that asked me this question is following the discussion, so please know that you are helping them immensely.
Thanks for this terrific informative post and comments. As I read them, I should point out that not all patient registries are alike. For example, the Cystic Fibrosis registry, the ImproveCareNow registry, and our patient driven registry, MyLymeData are very different. They are driven by the incentives and key strengths of their funders.
Clinical care registries like ImproveCareNow are different from stand-alone patient registries that don’t include a clinical care component. Patient foundation registries like CFF may actually be variations on clinical care registries using centers of excellence. Patient foundations may or may not be patient centered. Often they are clinician or researcher centered. Advocacy or grass roots organizations use different means of recruiting and retaining patients. What type of organization are you working with? What are its assets? Engagement with the community, $$, researchers, clinicians or???
Our registry, MyLymeData addresses a “research disadvantaged disease” using rare disease research strategies. Pharma is not interested in antibiotic development as a profit model. So there we sat. Except “We are not waiting” (@danalewis?)
When no one’s helping, roll up your sleeves. 30+ years of patient grass roots engagement & earned trust. Patient driven research w patient driven data stewardship. “No data about us w/o us”. (@Bravebosom?) So we built a grass roots community for 30 years where we were a trusted third party to amplify the patients voice with legislation and advocacy before we put on our research hat, about 10 years ago and started conducting large scale surveys that evolved into a patient registry.
The size of our registry, 12,000+ reflects that engagement, our “disadvantaged research” status, along with the fact that we have the largest communications platform for Lyme disease attracting over 3 million unique visitors a year. We have a group forum where we are considering a MyLymeData subgroup, but have not done this yet. It seems smart, if, but only if, you have an engaged community.
Gary Wolf wrote “Is it possible that the architecture of patient registries is not adequate for making the discoveries we need? Where the data model is determined in advance, and especially where it is entirely dependent on clinical measures, it won’t be possible to address the many additional questions that develop over time or from communities of patients that weren’t anticipated by the registry designers.”
It’s a great question. As Susannah knows in Oct 2009, at the “E-patients, social media, and cancer” workshop organized by the National Cancer Policy Forum at the Institute of Medicine to examine the elements of a rapid learning health care system for cancer that includes patient-centered quality, value, safety, and innovation I made fun of the co-founder of Patients Like Me for exactly that reason. Engineers feel too often they know the answers but in reality it’s patients and caregivers who end up knowing what needs to be collected. In 2009, when I allowed myself to be humorous in this very serious meeting, telling Jamie Heywood he had become too conservative, I had been working hard to create such a patient-defined registry. IOW, a system where any patient could come and add a field to the existing database. It was a concept, has remained a concept but whoever will succeed in making it a reality will greatly improve the value of patient registries. Gary’s idea for such a system is exactly what I had in mind then. I now believe it could be implemented as a hybrid system, where all the validated data is held into an IRB approved data silo while the patient-driven additional data collection held by a different entity who may not need to have either IRB approval or even be HIPAA compliant. Hybrid systems offer so much greater flexibility.
Gilles, thank you for this comment and for all you’ve done in this domain. I feel sometimes like a broken record on the topic of more democratic and participatory approaches to knowledge making; and I think we owe it to ourselves to confront the barriers that have prevented these ideas – ideas that were evident to you already more than a decade ago – from influencing biomedical practice. Again and again, conventional top down designs fail to deliver what they promise, and again and again new proposals based on the same failed models get funded as if they are promising revolutionary resources, only to fail in the same ways. What is preventing the institutional actors from learning from their failures? I’ve concluded in my own work that it’s a mistake to ask too much of biomedical collaborators. Practically, this turns me against the hybrid approach, because the resources expended trying work through incoherent biomedical requirements doom the projects from the start. Instead, I prefer an unapologetic bottom up approach, where self-collected observations accumulate value at each stage: individual, individual + caregivers/family/community, small group research, larger cohorts. Curious about your take on this since you have seen the same failures I have.
Gary, I am more optimistic than you about the future of participatory research 🙂
I think it is essential to work with biomedical collaborators that perceive the possibilities of this kind of collaboration. That means collaborators who don’t rely on the status quo for their income (I know that takes out a majority of researchers). Those who show interest in collaborating don’t need to fully understand what can be achieved, because that’s usually asking too much of them. If I had worked in academia or doing any form of classical/validated research for decades I am not sure my mind would be pliable enough to understand the extraordinary value of patient-driven research.
Given that you have found some biomedical collaborators willing to take a risk with you, the hybrid system could deliver amazing results IF the research conducted by them uses some form of agile protocol that can evolve over time based on current results. Then the patient-driven part of the research can feed into the “established” research and help them evolve their theories and hypotheses. It’s what I see as the first effective level of this kind of collaboration.
For the purpose of this discussion I didn’t touch on the complexity of interoperability of the 2 systems used to collect data. It’s obviously a non-trivial barrier for collaboration.
As an internet geologist, let me dig down into the substrate and point everyone to the post I wrote to capture that intense, amazing symposium: Superheroes and rock stars at the Institute of Medicine. A more formal capture was printed as a book, happily available for free online. Here’s a link to Chapter 5: Patient-Driven Rapid Learning Systems.
The good news is that Gilles, Jamie, Amy Abernethy, Simone Sommer, Josh Sommer, and others who were there that day have been consistent in our call to involve patients as equal partners in research — and we have made progress. The bad news is that we have a long way to go to achieve the vision we laid out in 2009.
Now I’m going to hold our own “feet to the fire” because this is a good, knowledgeable group, let’s take advantage of the moment to try to articulate what’s really gone wrong. By “gone wrong” I mean: Bottom up, agile approaches have been a thoroughly obvious alternative to top down recruitment-based approaches for more than a decade. But there is there still now generally useful architecture of collecting, sharing, and collaborating to make discoveries based on our own self-observations. If our ideas are so good, why have they not in fact produced real progress? The listserv architecture of ACOR worked well for its time. (For people to whom this is jargon/ancient history, ACOR is the Association of Cancer Online Resources that Gilles started in 1995.) But I believe that this remains a paradigmatic success, subsequent real achievements like Patients Like Me are praiseworthy, but ultimately the biomedical establishment, with its focus on outputting clinical products and standards of care, dominated the development path, at the expense of patient driven/participatory knowledge making. I don’t want to take this discussion too insider dominated, but I think it’s important in noting issues with real projects to acknowledge that I attribute this direction not to any lack of insight by Patients Like Me leadership. If there had been a lack of insight, the conversation Gilles reports from the National Cancer Policy Forum would have corrected it. NO, the domination of the agenda by biomedical development models comes from the relative institutional strength of the parties. Biomedicine has highly credentialed paid staff backed by widespread intellectual consensus to argue their development path is correct, and a clear target to aim for: New commercially viable therapies! On our side, we are talking about wellbeing for individual patients, hope for families, and boosting the knowledge making power of communities, even where there is no commercial product or standard of care (yet) as a target for development. This means that in arguments framed generally by the goals of biomedicine, we will always lose. We just seem vague, idealistic, non-rigorous, etc etc.
At the same time, if you lift your head up and look around above the borders of biomedicine, you see participatory peer-to-peer technology and culture radically transforming other domains. There is so much hope available for innovation based on our increased capacity for empirical observation, data gathering, quantitative analysis, even self-experiment, that we shouldn’t be discouraged just because biomedicine can’t see it (yet).
In the hopes of cross-pollination, I’d like to share a Twitter thread I wrote today on COVID-19 patient registries
When people with a health condition are asked to share their data or experiences, I tune in and start taking notes.
First, my biases: The more access people have to data, information, and tools, the better. The more participation by the people who have the problem, the better the solutions will be.
Second, some background: “In brief, a patient registry is a collection—for one or more purposes—of standardized information about a group of patients who share a condition or experience.” (Workman, 2013) https://www.ncbi.nlm.nih.gov/books/NBK164514/
“When collecting primary data, conduct multistakeholder engagement with potential participants and data users to prioritize data collection needs.” – @PCORI https://www.pcori.org/research-results/about-our-research/research-methodology/pcori-methodology-standards#Data%20Registries
(I think that means: Ask people who have the health condition to help define what is important to include in the survey. @PCORI please correct or refine my interpretation.)
In a global pandemic, EVERYONE is a potential respondent to a COVID-19 survey. How are people responding?
Here’s what I am observing in the U.S. (please, those of you in other countries, share your observations):
– Academic institutions are setting up COVID-19 patient registries, as are local public health agencies.
– Individuals who have had COVID-19 are setting up peer communities.
– Private companies are tracking COVID-19 symptoms.
My questions:
– Are organizations asking people who have (or recovered from) COVID-19 for input?
– Is anyone coordinating the questions being asked?
– Is anyone coordinating the potential pooling of data?
Here’s the post where I’m directing COVID-19 patient registry discussion (hopefully): https://susannahfox.com/2020/05/13/tracking-the-trackers/
Please join me if this sparks questions, answers, complaints, or compliments!
“Registry” has become a confusing term for many. While generally referring to retrospective clinical data, the term is used in a very generic way that does not make clear the separation of four models across two key questions:
* WHO – Which stakeholder is being engaged to participate and share … the patient or the provider?
* HOW – Is data being manually captured/re-entered or is the model based upon sharing of trusted electronic data?
A “patient-facing” registry based on self-reporting data brings very different operating challenges with a “provider-facing” registry based upon sharing EHR data.
A simple 2×2 across these two dimensions answers a great deal of ambuity.