Editor's note: Dr. Bell is Director of Patient Safety and Discovery at OpenNotes, Beth Israel Deaconess Medical Center and Associate Professor of Medicine at Harvard Medical School. Her research focuses on transparency in health care delivery systems and partnering with patients to improve health care. We spoke with her about patient engagement and her experience with the OpenNotes project.
Dr. Robert M. Wachter: Tell us a little bit about your background and how it led to your work.
Dr. Sigall K. Bell: Over a decade ago, Tom Delbanco, asked me to help write a facilitator guide for a film about patient and family experiences with medical error. I watched the film and was shocked by how patients and families harmed by health care described their experiences. I started thinking about how to improve communication when things go wrong, cowrote the facilitator's guide, started teaching with it, and then did some research in that space. I started working with Patient and Family Advisory Councils (PFACs), clinicians, and health care systems through research looking at transparency, open and honest communication, and organizational culture to help patients and families when things go wrong. The natural next step was thinking about how to enable effective partnerships with patients and families to ensure that things go right. That was the basis for my work in OpenNotes.
RW: I remember the study on patient and family experiences after errors vividly. I was also shocked by the degree to which patients felt guilty after errors. How did you come to understand that?
SB: I think we all harbor a natural fear that if something bad happens to a loved one—maybe it could have been prevented if only we had done something differently. Patients or family members may hold those feelings for a very long time because health care organizations don't always recognize emotional or psychological harm stemming from errors. Now, in medical error disclosure training, we try to emphasize creating space to explore emotions, such as guilt, fear, and isolation, that patients and families may experience after harmful events and to address them specifically upfront.
RW: As you thought about the gaps in patient communication and patient engagement, a big part of that is the paternalism of medicine and in general how we engage with patients. Then, there are some very specific issues around errors and the malpractice system and how in some ways it promotes silence and a lack of communication. Was your initial interest in one part of that or another or both?
SB: Definitely both. I started off in patient safety thinking about how we could do a better job when things go wrong. But over time, it became very clear to me that we could be applying those same principles—transparency, respect, and open communication—more "upstream" while engaging patients to help ensure that things go right.
RW: If you talked to a hundred doctors or nurses, I'd guess that they'd all say engaging patients is a good idea. As you began in this journey, what did you think the main obstacles were?
SB: Patient engagement has definitely become a buzzword, but it means different things to different people. I think the idea of engaging patients is broadly endorsed, but the devil is in the details. We don't always know how to do it well. Overstretched clinicians understandably worry about change—what will this mean for time demands? So even as safety leaders emphasize patient engagement, it has been challenging to establish tangible, efficient, and sustainable mechanisms. We're trying to understand what patient engagement means to patients. If we apply that principle of looking through the patients' eyes to design a health care system that they can do well in, we could improve health outcomes and ideally make things more efficient for clinicians. Some organizations are doing a great job with patient engagement and certainly getting it right, and some systems are employing user-centered design principles. But we still fall short. In general, medicine still has more of a clinician-centric approach to patient engagement.
RW: As you got into this, did you think that the problem was a cultural problem and the paternalism of medicine and the way we're trained? Or is it a time, logistics, and payment system problem? Or is it the malpractice issue, with everybody going to their own corner after an error? Did you have a sense of which of those you thought were going to be the most important things to tackle?
SB: The short answer is that I think it's all those things. We've ended up in this peculiar place in medicine today. Many of our problems are multifactorial in terms of difficult to change culture, a historically paternalistic system, immense productivity pressures, and fear of being sued. It's not a great formula for meaningful relationships between patients and clinicians or for organizational learning. In some ways, both patients and clinicians are vulnerable to emotional exhaustion and depersonalization in our health care system—two key components of burnout. We hear a lot about the constraints of short visits and pressed clinicians that cannot take just one more thing. When we hear resistance to the concept of OpenNotes, we realize it's often not really resistance to OpenNotes, it's resistance to any one more thing.
When you think about what brings clinicians to medicine—the genuine desire to help people—that involves talking to and knowing people. The role of narrative is central in what brings meaning to our work. Yet we're feeling too overwhelmed to find time to talk with patients and families, let alone tell patient stories. Over the years, it seems like that has shrunk more and more, and it's sometimes nearly absent now when you look at a note, using the note as just one example of depersonalization and loss of narrative. There's a lot of data, templates, and imported macros. Patients have even said to us, "I know this note is about me, but I'm really hard-pressed to see myself in this note." These interrelated challenges have affected how clinicians operate within the system. They're constantly time-crunched to generate that data, follow billing rules, or meet the regulations. Somehow, we've lost the very essence of what brought us to medicine in the first place. Needless to say, solutions will likely also need to be multifactorial, emphasizing culture change, patient empowerment, new approaches to medical errors and liability, and restoration of joy and meaning in work.
RW: There's an interesting tension as we get into OpenNotes. You're talking about giving patients access to their notes at a time when the note seems to have less value than it did in the past.
SB: It's true, and I wish we could fast-forward to better notes, but data suggest that patients benefit from notes even as they are. I think putting patients on the other end of notes is a little bit of a social nudge. The simple gesture of putting patients on the other end of the notes we write is humbling and it keeps us honest. It's not new that there's an abuse of copy–paste behavior in notes. It's not new that there's note bloat or all this imported data and difficulty understanding what this note is actually about. It's not new that physicians are so pressed for time that they cannot adequately give feedback to their residents when they're precepting them in clinic, let alone go the extra step of reviewing their note and making sure it is ready for a patient's eyes. Those things are old problems. But once you introduce OpenNotes, you're suddenly forced to look at those problems harder and with a deeper lens. That's a good thing for medicine, because these are problems that we should be addressing before they spiral even further out of control. Patients seem to really appreciate the ability to look at the note after the visit and they say they glean a lot from it. We could certainly be doing a better job, but the benefits have already declared themselves even with current notes.
RW: Summarize what the literature and your own experience have taught us about OpenNotes.
SB: The take-homes so far are that the majority of patients report benefits related to patient engagement, clinicians are not overwhelmed, patients are not overwhelmed, and shared notes can enhance patient–doctor relationships. The original study started in 2010 at three health care organizations across the country, here [Beth Israel Deaconess Medical Center (BIDMC)], at Geisinger Health System in rural Pennsylvania, and at Harborview Medical Center in Seattle, an urban safety net hospital. It initially involved about 100 primary care physicians who volunteered to invite about 20,000 patients to read their visit notes. About three out of four patients reported that they better understood their health conditions, took better care of themselves, remembered the plan of care better, and were more prepared for their visits and more in control of their care. In addition, about two-thirds reported doing a better job taking medications as prescribed, which was itself a very positive finding. Another set of metrics looked at clinician workflow. Although doctors worried about negative effects on time demands before the intervention started, this concern largely did not materialize. OpenNotes really didn't change their day-to-day workflow. Not a single doctor reported ordering more tests as a result of OpenNotes in terms of worries about defensive medicine. Email volume before and after the intervention didn't change at all. Patients also were not overwhelmed. Doctors worried that patients might be confused or worried or find notes offensive. But most patients did not report such outcomes. These results have now been repeated at several organizations across the country.
At the end of the year, there was a single "yes or no" primary outcome: Would patients and doctors want OpenNotes to continue? And 99% of the patients said "Yes." By the same token, none of the doctors (who were then able to discontinue OpenNotes) shut it off, and it still continues today at a growing clip. Starting with 20,000 patients who participated in the initial study, there are now more than 27 million patients across the country with access to their notes. One other illuminating data point from that initial study was that 85% of patients said they would use the availability of notes as a criterion for choosing a health care provider or a health care system. That has been an important mover in influencing a change toward transparency, where there are some market forces at play. Patients are saying they will vote with their feet.
RW: One concern has always been that physicians would feel uncomfortable writing certain things in the note, for better or worse. Was there evidence about whether they changed the way they documented?
SB: The original OpenNotes study addressed the question about changes in documentation as perceived by providers about specific sensitive issues. Things like obesity, substance abuse, and cancer discussions. About 20% to 30% of clinicians said they have given more thought or changed the way that they document those topics. We don't yet know in what way. That research needs to be done. Have they changed their documentation of sensitive topics in a way that is more user-friendly to patients? Did they say less; did they say more? A lot of people worry about the note changing, which I find interesting because people also have a lot to say about how bad the notes are. So change may not necessarily be a bad thing. In fact, one classic example from the University of Washington involved a clever researcher who showed an ophthalmology note to general medicine clinicians and asked, "How many of you understand this note?" On a scale of 0 to 10, nearly every clinician plotted in the 0 or 1 category. As a result, ophthalmologists have changed the way they are documenting their notes so not only are they more meaningful to patients, but also more meaningful to clinicians. Some changes may be helpful for both patients and clinicians. But further studies are needed.
RW: Let's switch to the implications and the evidence regarding patient safety. The theory here was that errors would potentially be caught by patients, clinicians would document more carefully, or it would open conversations between clinicians and patients that would ultimately improve safety. How is that playing out and is there any evidence regarding whether it has made care safer?
SB: The OpenNotes and patient safety story is exciting. First, transparent notes can help with simple things, like access. Patients who go to see another doctor in a different system and the doctor cannot access their notes. That's a well-recognized problem affecting efficiency, cost, unnecessary test duplication, and timely diagnosis. Patients who have OpenNotes can access their own notes. Anecdotally, an elderly patient receiving care at our hospital was travelling with family, got sick in Singapore, and was able to access and download his notes there, preventing what could have resulted in delays or complications in care.
RW: A workaround solution for interoperability.
SB: That's right. It raises some questions about where the interoperability should happen. At the health care organization level? At the patient level? While not a universal fix, could this at least help until more comprehensive solutions are available? It raises some interesting philosophical questions.
The second thing we've seen with safety is what patients find in notes, as you were alluding to. Patients are finding medication errors, family histories that are not updated, and other documentation errors, such as incorrect description of symptoms—things that could directly affect the diagnostic process. At BIDMC, we built an add-on to OpenNotes using a patient-reporting tool. The idea was rather than just invite patients to read their notes passively, we ask them to actively give us feedback. Did we get it right? Were there things you perceived as possible mistakes? What are they? These were the kinds of mistakes patients reported. OpenNotes puts more eyes on the chart.
In addition to what patients find in the notes, a third category of safety effects is what they do with that information. For example, they're more likely to share their information with care partners. Vulnerable, elderly, chronically ill patients who have complex care are at risk for safety hazards. They can share information with their care partners or during transitions of care to close a common information gap. Patients also report that notes help them to better remember next steps. It's astounding that we don't have a better solution for that. Across the country every day, earnest doctors see patient after patient, doing their best to deliver the right information at the right time. But patients leave the office and a substantial portion of that information, some experts say 40% to 80%, is misremembered or forgotten. Notes can serve as a reminder. Is there a test I'm supposed to do? Is there a referral I'm supposed to go to? Survey data from patients, families, and caregivers tell us that reading the notes helps them to close the loop on that next step. About 75% of 10,000 surveyed patients, parents, and caregivers said reading notes helps them understand the rationale for tests and referrals, and about half reported that it helps them to complete those steps. So there are important implications for improving the diagnostic process by tackling common ambulatory risk areas through patient engagement.
The fourth category of benefits with safety implications has to do with the relational effects of reading notes. Studies demonstrate enhanced trust in the provider. Patients talk about positive emotions from reading their notes, greater appreciation for the provider (e.g., "She really knows me" or "He really cares"), and a greater sense of activation in their care. Patients describe understanding more about how the doctor thinks, which may even lead to being able to help the doctor if there's a next step that has fallen through the cracks. These stronger relationships may empower and encourage patients to help keep the treatment or diagnostic plan on track.
RW: A lot of clinicians worried about patients disputing something in the note. What is the feedback loop for a patient to get that corrected?
SB: Having a system for patients to provide feedback about their notes is really important. For one thing, there's a lot you can learn from patients. I learned that lesson from you a few years ago. You showed a slide from JAMA at a talk. It was a picture that a 7-year-old girl had drawn of her visit to the doctor's office. She's sitting on the exam table and her mom is there, then off to the side there's the doctor with his back to the family typing on the computer. It's called "The Cost of Technology"—at least the way we often use it now. But it begs the question: is there a better way? Can we bring patients, families, and doctors together around health information transparency so we're all looking at the same thing together? This kind of patient feedback, even from a 7-year-old girl, is so powerful. Doctors can see themselves through the patients' eyes. The same is true for enabling feedback from patients about what is in the record. It's an underutilized opportunity to improve documentation and to improve safety. Right now there is often no feedback loop. A note gets signed and is part of the medical record. Patients don't routinely see what is there. Busy clinicians don't always get to check or correct the dictation system errors. We know that EHRs are not always accurate. As health information transparency grows, patients will need a clear way to report errors. Health care systems will need ways to create a "learning EHR."
We've harmonized the patient-reporting tool with our QI [quality improvement] workflow. The reports are generated through the patient portal. When patients report a possible error or problem, the report is automatically flagged as a possible safety concern. They are sent directly to Patient Relations, where every safety concern report is reviewed and the information is vetted. If there is a confirmed safety or care concern, we can close the loop with the clinician. We've found that the majority of patient reports that flag a concern represent legitimate safety concerns. The majority of the time when we contact clinicians about these issues, there's a change in the record of care. This is not an isolated finding. It resonates with the literature we're seeing now that patients and families can be active partners in care and in safety, and they can recognize errors that are unapparent to clinicians.
RW: That sounds complicated and expensive. If you have 100,000 patients out there reading their records and finding things that they don't quite agree with, sending all those emails to your understaffed Patient Relations department and getting back to the doctor. Does that work okay in real life?
SB: Well, you'd be surprised. Patients are thoughtful about feedback. They don't use the tool indiscriminately, and some even comment about how busy clinicians are and not wanting to bother them. If anything, we should worry as a profession about why so few patients use reporting tools. What are the barriers? When we ask those questions, we find that there's still a considerable proportion of patients and family members who are hesitant to speak up. That is a bigger concern for us to be thinking about from the safety standpoint. The way our system has worked, the Patient Relations staff has not felt overwhelmed by the reports that are coming in. In fact, we also asked patients about positive feedback about their health care provider. The Patient Relations staff has enjoyed forwarding this on to clinicians. Being recognized by patients for "getting it right" can generate positive culture change.
At the end of the day, some patients are going to call or email their doctors about a perceived problem in their notes anyway. The reporting tool adds a formal flow so that it's landing in the same box and going through a step-by-step process. It also offloads the clinicians, because the messages come through a central mechanism rather than going directly to their inbox. And there's a distinction to be made between a disagreement with the clinician's judgment and a factual error, which is part of the patient and clinician education process. As far as cost goes, it's amortized. There are some upfront costs to training the Patient Relations staff, and education for patients or clinicians, which is important for a system like this to work, and to help patients become safety partners. But after that, it's integrated into the fabric of the QI system. At Boston Children's they have implemented the patient/family reporting tool across nearly the entire ambulatory system without additional staff or FTE [full-time equivalent].
RW: I remember Tom Delbanco, cofounder of OpenNotes, telling me the story of the patient who was keeping his own notebook and said to the surgeon, "I'll write my notes, you write yours, and we'll compare." You've done some work on collaborative note-writing rather than the provider's note being the note of record and the patient making requests to correct it. More Wikipedia kind of models or different ways of thinking about this. Can you talk a little bit about that?
SB: One possible future direction for OpenNotes is the OurNotes model. OurNotes, led by Jan Walker and team, enables interested patients or proxies to contribute to the note and to agenda-setting before the visit. Rather than have doctors solely responsible for the note, what about making it collaborative? Would that help clinicians by having some of the note written ahead of time by patients or their caregivers? And in turn, would it help patients and caregivers to ensure their story is reflected and their priorities are recognized? It's being piloted in several sites right now. At one site (University of Washington) where it has been previously piloted in a smaller scope, both patients and clinicians really liked the collaborative agenda-setting before the visit. It helped bring the most pressing issues to the table, prepare clinicians for the visit, and improve communication. The doctors who participated also found that visits were more efficient.
RW: Anything else I didn't ask about that you wish I did?
SB: There have been several surprises we've found in doing this work over the last 7 years; I'll mention three. The first is how important it is to challenge assumptions. When people hear about OpenNotes, many say, "Wow, this is a great intervention for tech-savvy patients." We've been surprised and humbled by how universal the interest is. It's not just educated patients who benefit. It's elderly patients; it's vulnerable populations. There's a universal interest and perhaps a particular benefit for the 40 million informal caregivers in our country taking care of vulnerable patients. We have been surprised by data showing that non-Caucasian patients and those with a high school education or less are more likely to rate notes as extremely important to engage in their care.
The second thing that I've been intrigued by is the idea of OpenNotes as a visit extender. Patients have said, "[OpenNotes] is amazing. It's like having the visit all over again." It occurred to me how powerful that is for many things that we do in medicine, take informed consent as just one example. It's helpful for a patient to be able to read again the rationale for why a certain intervention is recommended or not, and to sit with that in the patient's own time and leisure when they're not nervous and can take things in a little bit better. OpenNotes may enhance shared decision making—informing patients so they can ask questions and participate in care decisions. It also has implications for trust and stronger patient–doctor relationships, because time spent with patients is such an important factor, and patients "see" and appreciate the additional time clinicians spend thinking about them. A resident once likened OpenNotes to bedside rounding: rather than spend a minute in the room with the patient and 30 minutes outside the room discussing the case, transparent notes make the doctor's thought process visible. In the future, notes could even be part of the treatment, if patients could go back to read the clinician's reaffirming or encouraging guidance again and again in the space between visits. There's no replacement for face-to-face time, of course, but clinicians can make the time they spend caring for patients count even more.
A patient with Parkinson disease in Sweden, Sara Riggare, came to the realization that for every hour that she spends with her neurologist, she spends 8765 hours on her own—her calculation. It's an astounding statistic. The truth is, we would go miles farther if we could empower and equip patients to do what would improve their health in those 8765 hours in between their visits. OpenNotes offers an amazing way of connecting patients to their care in that space between visits, and in that space where we know safety hazards can arise in the ambulatory sector. Maybe OpenNotes can help close the gap and help us move our emphasis not just from what's happening in the clinic but what's happening between the visits. It's like the flipped classroom. The idea of creating a flipped clinic has a lot of potential but is really just in its infancy. We need to learn how to use clinic time for meaningful discussions and time between visits for education, support, and adherence-building. I am really excited about some of the emerging innovations related to social incentives and community activation in health care.
Finally, another surprising finding is the power of invitation. With OpenNotes, a lot of benefits come from better remembering next steps. If we were told as clinicians that we could have the visit but never look at the note again, it would be hard to come back to the patient at the next visit and remember all the details of what happened. Yet that's what we're asking patients to do. But it's beyond what's on paper. It's about the relational effects of extending the invitation to say "Please read this. You're part of the team and I want you to read it. I want you to give me feedback." The invitation can be empowering. It can help build trust that stems from transparency. Even if patients don't read notes, the offer may open the door to stronger patient–clinician relationships.