The Ethical Implications of Rapid Learning Systems in Oncology
Health care providers and research scientists frequently utilize emerging technology to enhance their ability to study diseases and develop effective treatment. A major area of growth in medicine has surrounded the use of health information technology and electronic health records (EHRs), which can be easily accessed and transferred between individuals and institutions. The rapid adoption of EHRs has allowed doctors to electronically accrue the entire medical history of their patients, allowing for better coordination of care and dissemination of information about patients and providers.1 A 2013 report issued by the Institute of Medicine asserted that “professional organizations and the Department of Health and Human Services [should] deliver the necessary components of a learning health care information technology system.”2
Throughout the oncology world, this has taken place through the development of rapid learning systems (RLS). These programs allow data to be culled from patient care and clinical research, which are then centrally banked and analyzed.3 The American Society of Clinical Oncology (ASCO) entered the RLS word with CancerLinQ,4 which allows oncologists and cancer treatment centers to track treatments and progress using data extracted from EHRs. Introduced at the 2015 ASCO Annual Meeting,5 CancerLinQ has been adopted by 70 oncology practices in 39 US states, representing data from more than 1 million patients and 1500 practitioners.6
Oncology practices considering the implementation of an RLS must also consider the ethical implications of such a step. Although a study by Grande and colleagues found that patients with cancer are often more willing to share data regarding their inherited genetic information than patients without cancer,7 questions have remained about the proper measures needed to explain and obtain informed consent in this context. Some researchers have argued that not all studies using these data may require patient consent, and that a streamlined consent process may appropriately replace written informed consent8; others have worried that relaxing rules regarding informed consent may bring about another Henrietta Lacks situation.9
In order to better understand how clinicians and researchers are approaching this debate, Journal of Clinical Pathways spoke with Reshma Jagsi, MD, DPhil, professor and deputy chair of the department of radiation oncology at University of Michigan Health System, as well as a research investigator at the Center for Bioethics and Social Sciences in Medicine (Ann Arbor, MI). Dr Jagsi and colleagues recently published a paper studying the ethical implications of RLS implementation from a patient perspective in Journal of Oncology Practice, which showed that patients whose medical providers establish trust are more likely to consent to including their data.10 We asked Dr Jagsi to detail strategies for doctors to gain their patients’ trust and increase their comfort level, and how best to approach the subject of electronic health information in older patients and those less familiar with technology.
What are the potential benefits that an RLS system could hold for the oncology world?
The idea of a system that harnesses data that are collected in routine clinical care to promote research and quality improvement is particularly useful in a complex area like oncology care. It is a wonderful idea to create a virtuous cycle where the information we would gather anyway, as part of health care delivery, can be used to actually answer important research questions and then feed back to the care we deliver to patients. We know that certain cancers are rare, and that certain patients are different from one another—by gathering the data generated from routine clinical settings, we can learn the best ways to manage patients and make sure that the quality of care is optimized.
From where do the ethical questions arise in designing this kind of a system?
The ethical questions center around insuring that we are appropriately respectful to our patients. This is their information, their data. What we don’t want to do is end up in a situation like Henrietta Lacks, where there was inadequate respect shown to an individual patient whose biological specimen contributed so much to medical research. In that specific case, neither she nor her family were told that the specimen had been used, and it really serves as an example of how the medical community didn’t adequately respect a patient’s dignity. In these cases, the patients are not providing their information to providers to enhance research or medical knowledge, or to improve the quality of care for future generations of patients. That said, they might be very happy to do so, and actually enjoy the benefits of engaging in that form of altruism. But we rob them of that if we don’t actually present them with the choice.
One of the big questions is who we should appropriately engage and notify patients that this idea is taking route, and explaining how we are creating these learning systems from which they benefit. How do we get their buy-in to a system like this? Another important question centers around data use and governance. We need to understand what the appropriate limits are on use of these data. Both of these issues need to be informed specifically from the patient’s perspective.
Your research found that there were differing levels of comfort among patients regarding the use of their data, based on factors like personal knowledge and level of education. Based on that information, what are some of the best ways for medical professionals to go about starting these conversations, especially with patients who may be more skeptical?
We found that trust between the physician and patient is very important. The wonderful thing about these conversations between physicians and patients is that they can start from a place of trust, and the physician can explain how the data are protected, and how they have the potential to help other patients. They can also explain how a system like this benefits everyone within it. By doing this, they empower the patient by making them realize that they are contributing something vitally important. They also engage the patient in a joint endeavor, and make them feel like they have contributed an essential element to the research.
Several patients brought up their level of comfort with technology in general, with some older patients skeptical because of lack of understanding. How can that be mitigated?
One of the concerns we had was that we needed particularly to focus on sensitive ways of communicating to older patients who might not have as much familiarity with technology. We actually found that some of the patients who were not familiar with technology couldn’t even imagine what kinds of things could go wrong, and they had such strong trust in the health care system—and their physicians in general—that they were not worried. On the other hand, you sometimes see the fear of the unknown. The challenge is to explain exactly how these systems work, then reassure the patients that we are taking every available precaution to protect the security and privacy of their data.
With that in mind, what do you think the next steps are for implementation research in this area?
What we really need is more information from what I would call the informed and considered opinions of patients. We used traditional social scientific methods in our study: we interviewed patients, we did surveys. Those kind of static interactions sometimes only capture the gut instinct initial response of a patient, but these opinions can evolve, even over the course of an hour. There is an idea in empirical bioethics that we should engage citizens in a jury where people are brought together for an entire day, to think about what the issues really are, to ask questions, and to talk amongst other participants, with the ultimate goal of coming to a truly informed opinion. Obviously you cannot do this with every single patient, but you can bring together a diverse group of patients who can generate a wide range of opinions. This kind of work could be really informative for those who are implementing these kinds of systems, because they might actually be more sophisticated than what we draw out from a simple one-on-one interview.
My colleagues and I are currently funded by the National Cancer Institute to conduct exactly this sort of broader investigation, where patients will be invited to come together for a full day and talk about these issues. We will use the information that we gather here to help form the implementation. We are hoping that bringing patients and experts together, having them sit together at small group discussion tables, and having experts rove the room as the tables deliberate over specific questions will help. We tape record and transcribe all of these discussions and perform qualitative analysis of these data, in addition to having the participants generate individual consensus opinion statements. We use individual surveys, given at the baseline and again several weeks later, to measure how individual opinions evolve over the process. There is a lot to be learned, but the key is that we need to make the implementation process patient-centered. We need to be maximally sensitive to the concerns that only patients can voice for themselves. υ
1. What is an electronic health record (EHR)? Health IT website. https://www.healthit.gov/providers-professionals/faqs/what-electronic-health-record-ehr. Accessed January 31, 2017.
2. Delivering high-quality cancer care: charting a new course for a system in crisis. Institute of Medicine website. http://www.nationalacademies.org/hmd/Reports/2013/Delivering-High-Quality-Cancer-Care-Charting-a-New-Course-for-a-System-in-Crisis.aspx. Published September 10, 2013. Accessed January 31, 2017.
3. Abernethy AP, Etheredge LM, Ganz PA, et al. Rapid-learning system for cancer care. J Clin Oncol. 2010;28(27):4268-4274.
4. ASCO CancerLinQ website. https://cancerlinq.org/. Accessed January 31, 2017.
5. CancerLingQ takes a big leap forward. ASCO website. https://am.asco.org/cancerlinq%E2%84%A2-takes-big-leap-forward. Published May 29, 2015. Accessed January 31, 2017.
6. CancerLinQ platform hits milestone of having more than one million patient records in system. ASCO website. https://www.asco.org/about-asco/press-center/news-releases/cancerlinq-platform-hits-milestone-having-more-one-million. Published October 20, 2016. Accessed January 31, 2017.
7. Grande D, Asch DA, Wan F, et al. Are patients with cancer less willing to share their health information? Privacy, sensitivity, and social purpose. J Oncol Pract. 2015;11(5):378-383.
8. Faden RR, Beauchamp TL, Kass NE. Informed consent, comparative effectiveness, and learning health care. N Engl J Med. 2014;370(8):766-768.
9. Zielinski S. Henrietta Lacks’ ‘Immortal’ Cells. Smithsonian magazine website. http://www.smithsonianmag.com/science-nature/henrietta-lacks-immortal-cells-6421299/. Published January 22, 2010. Accessed January 31, 2017.
10. Jones RD, Sabolch AN, Aakhus E, Spence RA, Bradbury AR, Jagsi R. Patient perspectives on the ethical implementation of a rapid learning system for oncology care [published online January 24, 2017]. doi:10.1200/JOP.2016.016782