Why Do Doctors Overprescribe?
How to get doctors to change their behavior
Posted Feb 05, 2018
Are doctors losing the public’s trust? Some recent news coverage seems to suggest that the idea of not trusting one’s doctor is relatively widespread. While this may or may not be true, it is certainly the case that there has been increasing discussion in recent years about the various ways in which physicians’ judgment can falter.
In recent months, there has been special attention paid to the role that doctors play in the current opioid addiction and overdose crisis in the U.S. Many people, including us, have noted that the opioid crisis is extremely complex and that a single, simple cause cannot be identified. This is certainly the case, but it’s also true that physician behavior has played an important role in the development of this crisis and will continue to play an essential role in how it continues to evolve.
While doctors are not to solely blame for the opioid crisis, it is true that there is a pattern of overprescription of these medications. Before we proceed with examining why this behavior persists and what to do about it, we would like to point out that the notion of placing blame, even for overprescribing behaviors, is not helpful here. These behaviors also have complex, systemic causes and the idea of whose “fault” it is should not be the focus of these discussions.
Leaving “fault” aside, we can see some parallels between overprescription of opioids and another relatively common physician error, overprescription of antibiotics. So why do doctors overprescribe? The reasons are varied and complex. In the case of antibiotics, we know that doctors often feel pressure from patients to provide antibiotics, and they identify this pressure as a major reason why they frequently prescribe them for illnesses that do not respond to antibiotics. The same phenomenon, pressure from patients, is likely to also be part of the motivation for unnecessary or excessive opioid prescribing as well.
We also know that doctors, just like anyone else, are sensitive to the behaviors of others in their social networks. For this reason, we often see very similar physicians decisions and behaviors within specific social networks. This phenomenon was brought to light as a result of some brilliant studies done in the 1970s as part of the Dartmouth Atlas of Healthcare project. These studies identified a phenomenon called “small-area variation” in healthcare practice, which basically refers to the fact that doctors in similar communities make treatment decisions simply based on the habits and practices of those in their immediate vicinity. These studies suggest that physician behavior is therefore influenced by social signals and not always governed by evidence, best practice, or guidelines. It’s highly likely that patterns of both antibiotic and opioid prescription are similarly subject to this same kind of social network effect.
In the case of opioid overprescription in particular, there are also a few unique systemic and cultural reasons why doctors persist in this behavior. Some have argued that a cultural shift in the way medicine conceives of pain over the past few decades has contributed to the opioid epidemic and in particular to doctors’ over-dependence on opioids to manage chronic pain. At some point, organized medicine began espousing the view that “all pain is treatable.” In this view, the patient’s subjective experience of pain governs the doctor’s decision to utilize conventional medical modes of treatment, most notably pills. It came to be considered a routine part of a doctor’s duty to assess and ultimately to completely eliminate any kind of significant pain, as determined by the patient. This kind of approach has likely led to a situation in which it becomes the expectation that a patient experiencing more than mild discomfort should be treated with pills and that the goal is to be completely pain-free. In fact, it is rarely possible to completely eliminate pain; trying to do so often results in excessive administration of opioids.
At the same time, a variety of structural factors have created a situation in which doctors are poorly educated about pain management and do not have the time or incentive to fully evaluate whether an individual patient should be prescribed pain medication with highly addictive properties. Doctors receive very little training on pain management, including alternatives to prescribing opioids, in medical school and in continuing education programs. In addition, poor communication in our extremely fragmented healthcare system can easily lead to situations in which patients who should not be receiving opioids are able to obtain them.
In the emergency department (ED), which is where many opioid prescriptions originate, doctors often have little time to assess patients’ pain and understand their risk of dependency and addiction. In fact, ED doctors are usually under enormous pressure to see as many patients as possible and are often evaluated based on how many patients they see rather than on the quality of care they provide. Patients are also asked to fill out satisfaction surveys as they depart the ED, including questions about whether they feel their pain has been addressed and resolved. Because doctors are evaluated based in part on the scores they get on these surveys, those questions provide an incentive to overprescribe potent analgesics. All of these factors - a medical culture that emphasizes the need for pills to treat pain, lack of training on pain management for physicians, a disjointed healthcare system, and time pressure on doctors - have come together to create a situation ripe with opportunities for vast overprescription of opioids.
Pharmaceutical companies also influence what doctors prescribe. Many physicians receive gifts and payments from manufacturers of opioids and this increases the likelihood that they will prescribe opioids to their patients.
While we may not know all we need to know about how to deal with the opioid crisis and patterns of problematic physician decision-making, we do know a bit about what doesn’t work so well. In general, approaches that are very narrow, addressing only one piece of the puzzle, tend to be pretty inefficient. These kinds of approaches tend to involve short continuing education courses for practicing physicians that review guidelines and laws around opioid prescribing and evidence for alternative forms of pain management. While these programs tend to have positive effects on knowledge and sometimes behavior in the short-term, in general education-only approaches tend to have limited impact. This is true across a broad range of behaviors in an array of different populations. In general, simply educating people on its own, especially if it’s a short-term activity, does not go far in the way of permanent behavior change. While education is often necessary for changing behavior, it is almost never sufficient.
Similarly, changing medical guidelines is often necessary but not sufficient. Changes in guidelines sometimes capture physicians’ attention, at least in the short term, and can make them aware of a broader body of evidence on a particular issue. In a limited number of cases, guidelines may be enough to generate changes in practice, but for the most part, instituting new guidelines in and of itself will not create sweeping, large-scale changes in physicians’ behavior. Guidelines can be an important piece of the puzzle, but they are most useful as enablers of change rather than a catalyst to long-term behavior change in and of themselves.
Broader behavior changes generally require a more in-depth, comprehensive kind of intervention that, in addition to providing new guidelines and more education, also addresses underlying incentives and motivations that lead to specific kinds of decisions and behaviors. For curbing overprescription of antibiotics, it does seem to be the case that social benchmarking approaches work relatively well. Some studies have shown that when high prescribers are told that they are outliers compared to their peers, their rates of prescription drop significantly. In addition, requiring some kind of public accountability also seems to work. In one study, prescribers were forced to enter a justification for any antibiotic prescription that did not comply with guidelines in a medical record that would be available to other physicians. Being required to enter this kind of “public” justification also led to a considerable decrease in doctors’ prescribing behavior.
Some similar interventions would most likely work for opioid prescribing, although more testing of this is still needed. In addition, it seems that “nudge” approaches implemented in electronic medical records, including pre-populating with lower starting doses or tapering doses for patients for whom opioids have not relieved pain over long periods of time, would also work to curb rates of opioid prescriptions.
While behavioral interventions are extremely important, they are not, however, sufficient. Behavioral interventions will not affect the opioid crisis significantly unless they are accompanied by important structural and policy-level changes. These include things like patient education programs administered by nurses aimed at reducing the “pills for pain” culture and providing counseling to help patients manage their pain in other ways. Communication systems within the healthcare system also need to be improved. For example, mandating use of state-level prescription drug monitoring programs (PDMPs) before any new prescription of opioids would also be an essential safeguard against inadvertent prescribing of opioids to patients with histories of abuse. Although 49 states have PDMPs, checking them is voluntary in all but a few. Changing incentives and reimbursing more time with patients, especially in the emergency department, would also help prevent patients with a history of abuse from getting new prescriptions and could also stop patients from becoming addicted. Removing questions about pain treatment from patient satisfaction surveys may also help reduce unnecessary opioid prescriptions. Finally, gifts and payments to prescribers from pharmaceutical companies in connection with opioids must be permanently banned.
With a problem as complex and pervasive as the opioid crisis, the solution will of course not be simple. Nonetheless, there are some very promising behavioral interventions for physicians that, combined with appropriate structural and policy changes to support these physicians, could make a huge difference in this national tragedy.