The following is a guest post from Bill Gardner, a psychologist who studies the mental health service system for children. Bill is an American living in Canada and a professor of pediatrics at Dalhousie University (Nova Scotia) and the Ohio State University. Bill blogs at Inequalities, and you can follow him on Twitter at @Bill_Gardner.
As Obamacare becomes settled law, we must address a more fundamental health policy task: reducing the cost of medical care while making it more effective. This will require changes through out medicine, possibly including medical ethics.
One proposal for improving the efficiency and value of medical care comes from from the Institute of Medicine: the Learning Health System.
A learning health care system generates and applies the best evidence for the collaborative health care choices of each patient and provider; drives the process of discovery as a natural outgrowth of patient care; and ensures innovation, quality, safety, and value in health care. In such a system, knowledge flows seamlessly between and among patients, providers, diagnostic facilities, and related community services. The best knowledge about treatments, diagnostics, and care delivery is naturally embedded in the delivery process, and new knowledge is captured as an integral by-product of the delivery experience.
The idea is to reengineer the health care system so that it gets better at learning which treatments work best and how to deliver them most efficiently. We need a better learning process because there probably won’t be a single awesome discovery that suddenly drops medical prices (while maybe raising the dead). However, we can get a more effective and efficient system by accumulating lots of incremental changes, if the health care system were rebuilt to find and implement them quickly.
This kind of rapid evolution is a matter of deliberate practice at firms like Google. Experiments are running continuously and every click you make contributes data to studies that drive incremental product improvements. This isn’t the case in medicine. Experiments to improve service delivery occur, but they are uncommon, expensive, and are not part of routine care. The Learning Health System proponents want to change that and they have proposals to transform the procedures, incentives, and data infrastructure of medicine.
Ruth Faden and her colleagues propose that transformation of the US health care system to a Learning Health System will also require changes in our clinical and research ethics. In An Ethics Framework for a Learning Health Care System they argue that research is not just a matter for university professors or health industrialists. Instead, all health care providers have an obligation to be continuously engaged in learning activities. More controversially, they argue that
Just as health professionals and organizations have an obligation to learn, patients have an obligation to contribute to, participate in, and otherwise facilitate learning.
To that end, they open a discussion of modifying current practices in research ethics. Current regulations require patients to give informed consent to participation in experiments, with very few exceptions. The requirement to obtain consent significantly raises the cost of carrying out an experiment. This may seem like mean-spirited complaint: so what if data costs more? The problem is that data are valuable en masse — that’s the point of the buzzword ‘big data’ — and even small differences in the cost of acquiring and using a data point can make research too expensive to undertake and reduces the size and representativeness of the patient populations who can be studied.
It is important that Faden et al. do not propose that we relax the requirement of consent to participate in the trial of a new drug, surgical procedure, or medical device. These activities “could proceed only with patients’ express, affirmative agreement, obtained through a valid informed consent process.” In other cases, where the care process is varied in ways that do not expose patients to risks exceeding those of the care they would routinely receive, there would not be a requirement to obtain prior consent.
Faden et al. do not provide an example of a study that would not require consent, but I believe they have something like the following in mind. Suppose a clinic believes that patients would benefit from closer monitoring of the side effects of a risky medication. A possible way to monitor those effects would be to regularly call patients taking the drug and ask them questions about side effects they might be experiencing. This intervention appears to have minimal risk and a clinic could simply implement such a policy for all patients without obtaining prior consent from those patients, or even warning them that the calls would occur.
However, suppose the clinic decided instead to randomly call half of the patients taking the drug and ask them about their symptoms, while the other half of the patients did not receive calls (which was, please recall, the usual standard of care). After six months, say, the clinic could check which patients were experiencing the feared side effect. This randomized controlled trial would provide the clinic with far more information about whether it’s worth making the calls. Interestingly, if the clinic did this trial and just used the data to improve their own practice, there would still be no requirement to obtain informed consent from the patients for participation in the experiment. Trials like this involving minimal risk are considered to be “quality improvement” studies and do not require prior consent.
However — and this is the critical point — if the clinic also wanted to publish the results of this randomized trial so that other providers could learn from their experience, then suddenly the “quality improvement study” becomes “research” and it would require informed consent. This is an arbitrary restriction that impedes system learning and I believe that it is an example of the impediments to learning that Faden and her colleagues have in mind.
Faden et al. do not spell out how higher-risk research participation requiring consent would be discriminated from low-risk procedures not requiring patient consent, or whether or how persons who object to any research use of their medical data might opt out. They are simply raising the question. This important proposal challenges decades of received doctrine in medical ethics and deserves widespread public discussion.
