It's 1997, and I'm sprawled on the waiting room floor of a hospital, as bright lights glare down at me. A recent attempt to repair a previous surgical injury to my spinal cord has failed, and I remain unable to sit or stand and in severe pain. When my name is called, my husband gathers me in his arms like a sleepy six-year-old child and carries me to a treatment room. An entire treatment team crowds into the small room, as my primary physician begins, “The thing is—we've put you through painful procedure after painful procedure. We've tried dozens of medications, blocks, infusions. I have a duty to do no further harm. It's time for you to consider prescription opioids.”Footnote 1
INTRODUCTION
The ethical duty to “Do No Harm” is as deeply embedded in modern medicine as any. It is usually attributed to a phrase in the Hippocratic Corpus, from Of the Epidemics: “The physician must be able to tell the antecedents, know the present, and foretell the future—must mediate these things and have two special objects in view with regard to the disease, namely, to do good or to do no harm.”Footnote 2
To “Do No Harm” (“DNH”) translates to the idea that, in certain situations, inaction is better than action: that is, it is preferable to do nothing than to intervene if there is potential to cause more harm than good. In each physician-patient interaction, this duty must be balanced with the physician's obligation to “do good:” to cure disease where possible and, if not, to prolong life, ease suffering, and improve function.
In the context of what has been dubbed the opioid crisis, DNH has proven an elusive precept that has been used to justify opposite sides of the pendulum swing on opioid prescribing.Footnote 3 In the past, it was used to justify the rejection of invasive procedures and in favor of palliation using opioids.Footnote 4 Beginning in the 1990s, physicians, like the doctor in the quote at the start of the Article, who were concerned about undertreated pain, weighed the calculus in favor of expanded prescribing.Footnote 5 Today, DNH is as often a rationale for discouraging doctors from prescribing.Footnote 6
Given this use of the DNH principle to justify opposing positions on opioid prescribing, we set out to examine whether treating pain with opioids raises issues that are distinct when balancing the call for action (to cure disease or ease suffering) with the counsel of inaction (to do no harm). In so doing, we consider opioid prescribing for both acute and chronic pain, but focus on prescribing for chronic pain, as it has garnered the most attention and raises thornier ethical issues. This paper thus contributes to the substantial literature on the opioid crisis by identifying ways in which opioid prescribing differs from the perennial issues clinicians face.
We make three arguments. First, we argue that the fact of pain's contemporaneousness reduces uncertainty about the need for action over inaction, and thus strengthens the reasons to act. Second, we note that while the DNH principle applies to both physicians and policymakers, each has distinct duties: physicians have a duty to the individual patient, while policymakers have a duty to society. We argue that harm to society should weigh little, if at all, in the treating physician's assessment of whether action or inaction is best, whereas policymakers appropriately consider societal harm. We offer guidance to inform policymakers' impulses toward action or inaction aimed at mitigating the ethical burdens placed on physicians and respecting the testimony of patients in pain. In addition, we offer cautions to public health officials gleaned from the context of opioid prescribing about eschewing decisions based on the status of those affected, avoiding policy cycling, and considering how their recommendations are likely to be implemented by subsequent actors. Finally, we address what duties are owed to patients who are currently taking opioid medication. We argue that once treatment has been initiated, health professionals have special duties to these patients.
OPIOID PRESCRIBING IN THE CURRENT ENVIRONMENT
Hippocrates' advice to physicians to heed the antecedents, know the present, and foretell the future is prescient—even chilling—when we consider the medical establishment's failure to anticipate the societal effects of increased opioid prescribing. The gravity of these consequences cannot be overstated: in 2017, over 70,000 people died from a drug overdose.Footnote 7 Although deaths in 2019 seem to be leveling off, in 2018, for the third year in a row, life expectancy in the United States fell, due primarily to overdose deaths and suicide.Footnote 8 Nearly 2.1 million Americans currently have an opioid use disorder related to prescription or illicit opioids.Footnote 9
The question of what constitutes appropriate opioid prescribing stands to affect many people with considerations weighing on both sides. Weighing on one side is the persistence of pain. Relief from suffering is a central duty of physicians, and by all measures, pain remains undertreated.Footnote 10 Some 50 million Americans suffer from persistent, daily, or near daily pain; 40 million report severe pain, and nearly 20 million have pain that is effectively disabling.Footnote 11 Pain is a chief cause of disability,Footnote 12 and the most commonly reported reason for misusing a prescription opioid (62.3%) is to relieve physical pain.Footnote 13
Weighing on the other side are the risks of prescribing opioids for pain. We now know that these risks are higher than once believed and that they increase when opioids are used long-term.Footnote 14 A small but significant portion of people who are prescribed an opioid for pain may develop an opioid use disorder, with risks ranging from .07%Footnote 15 to 8%.Footnote 16 While this danger is important, the substantial majority of people who take opioids do not become addicted to them.Footnote 17 The problem is that, beyond careful practices like using screening tools, providing patient education, and performing ongoing monitoring, we really don't know enough to distinguish in any definitive way for whom the benefits are likely to outweigh the risks.
Whatever the risk-benefit calculus may be for individual patients, however, the public health consequences of opioid prescribing have increased substantially in recent years. Beginning in the 1990s, we saw a significant increase in non-medical use of prescription opioids and the developing public health crisis that arose partly from unduly discounting the risks of addiction.Footnote 18 The largest risk of increased opioid prescribing is diversion—meaning the use of a prescription opioids by someone other than the person to whom it was prescribed.Footnote 19 The overwhelming weight of the data suggests that the vast majority of people who misuse prescription opioids did not receive them from a doctor but from a family member, friend, or on the street.Footnote 20 Among those who misuse—either not using them as prescribed or using opioids prescribed to another—an estimated 8 to 12% may develop a use disorder.Footnote 21
Presently, the efficacy of opioids for relieving long-term pain is being re-evaluated,Footnote 22 and the U.S. Food and Drug Administration (“FDA”) and the National Academies of Science, Engineering and MedicineFootnote 23 are in the process of providing guidance on prescribing for acute pain. While prescribing medication is but one aspect of effective pain management, all current guidelines continue to provide a role for prescribed opioids in both acute and chronic pain management.Footnote 24 Thus, assessing the ethical obligations of prescribing remains a pressing concern.
Below, we examine the DNH principle in the general case, before turning to distinct ethical problems posed by pain treatment.
THE PHYSICIAN'S TWIN AIMS: THE GENERAL CASE
The correct balance between the twin duties of action and inaction in medicine is not easy to determine. First, despite Hippocrates' wise counsel, the physician rarely knows for certain what the effects of her action or inaction will be. Instead, she operates under conditions of uncertainty, making evidence-based predictions about the likelihood that various outcomes will occur.Footnote 25 This first challenge leads to another: how to combine the physician's degrees of confidence regarding each outcome with a principle for balancing the twin duties.
Suppose Jane is ill with a disease that has a 20% chance of killing her, a 30% chance of significantly reducing her quality of life and a 50% chance of causing no harm. A therapy is available, but both its efficacy and its side-effects come with uncertainty. Perhaps it cures the disease 80% of the time, fails to work in 15% of cases, and leads to significant long-term harm for the patient 5% of the time. To say that the injunction to “do good or to do no harm” will be complicated to operationalize is an understatement.
In addition, the DNH mandate appears to rest on an empirical supposition about clinicians themselves. The DNH principle speaks in a cautionary voice and is offered as a corrective for the physician's assumed tendency to respond to the patient's distress with action. It counsels restraint when the desire to “do something” is insistent. A further complication, then, arises from the fact that it is hard to know how large this bias is in order to appropriately calibrate a principle to correct for it.
With this general background in view, we pose a new question: when doctors treat pain, is anything different? In our view, it is.
HOW PAIN AND OPIOID PRESCRIBING COMPLICATE THE ANALYSIS
Opioid prescribing for pain differs from the standard case in two important respects. First, unlike the standard case in which patient prognosis, treatment effectiveness, and occurrence of side-effects are future events, the presence of patient pain is occurrent. This fact about pain matters because it affects the need for action. Second, in the standard case, the harms to be avoided by the caution underlying DNH are harms to the patient. With opioid prescribing for pain, these harms include harms to third parties and to society as a whole.
Pain's Presence
A clinician applying the DNH principle must weigh the costs of action versus those of inaction, discounted by the probability that each cost will materialize. For the physician treating a patient for a condition other than pain, many of the outcomes she considers will occur in the future. What distinguishes the patient in pain is that a patient in pain is currently in pain. The pain is present, certain, occurring. For this reason, the doctor treating a patient in pain has more reason to act than does a doctor treating a condition whose harm to the patient is largely in the future and is thus uncertain. This is our first claim.
We must be careful not to overstate this claim. Non-pain patients also seek treatment for presently occurring conditions: loss of function, depression, etc. And when doctors treat pain, they must also make predictions about the future. Is the pain likely to lessen or worsen if untreated? Will untreated pain cause other problems, both physiological and psychological? Still, in the case of the patient in pain the greatest part of why she seeks treatment relates to a presently occurring state for which no prediction is necessary. This matters.
Pain's presence thus functions as a thumb on the scale in favor of action. This counsel for action does not suggest any specific treatment, however. The clinician should make treatment decisions guided by the evidence and clinical experience and consider opioid prescribing to the extent that both support its use.Footnote 26 Our point is simply that the presence of pain reduces uncertainty about the need for action and so alters the calculus in the clinician's assessment of the twin aims of medicine.
A critic of this point might object that the presence of pain is not certain from the perspective of the doctor. Only the patient knows whether the pain is currently present. Because pain is not objectively verifiable, the doctor might worry that the patient is exaggerating, or even lying in order to get drugs.Footnote 27 One might think that the doctor should discount the harm of the patient's current pain by her certainty that the patient is reporting accurately and honestly. We disagree. Discounting the weight accorded to a future harm by its predictive likelihood of occurring is morally different than discounting the weight accorded to a current harm by the doctor's assessment that the patient is reporting correctly. The first involves a prediction about the future, which the doctor cannot be expected to know. The second involves an assessment about the patient's reliability, which—we suggest—the doctor is not entitled to doubt, unless doing so would benefit the patient.
In order to treat the patient with the respect she is owed, we believe that the doctor must trust her patient and accept her testimony about her pain.Footnote 28 This duty is grounded in two ways. First, distrusting the patient may have negative consequences for the patient. Second, distrusting the patient itself violates the duties inherent in that relationship irrespective of its consequences.
Dismissing patients' accounts harms patients because pain can be a signal that something else is wrong. A physician may miss a serious disease. Also, where trust is absent, patients may experience fear or shame, which may lead them not to share symptoms out of fear of being disbelieved or to cease seeking medical attention altogether.Footnote 29 Lastly, the experience of being distrusted itself causes humiliation and distress.Footnote 30
Even if distrust did not cause negative consequences for the patient, the doctor's duty of loyalty to the patient includes an obligation to show her respect as an agent to be believed. To ask doctors to adopt skeptical attitudes toward their patients undermines that relationship in a way that is inherently morally troubling. There is one exception to this principle. If the doctor thinks the patient may be misusing the drugs and is likely to harm herself by doing so, the doctor must balance her concern for her patient's interests against her obligation to treat the patient respectfully by believing her. But in cases where this exception does not apply, the doctor wrongs the patient by adopting a skeptical attitude toward her report.
One further consideration counts in favor of clinicians trusting patient reports of pain. If the testimony or reports of some groups—people of color, women, individuals with disabilities, and others who are marginalized—are discounted or challenged more often than are the accounts of others, this tendency to disbelieve these groups wrongs these patients and amounts to what the philosopher Miranda Fricker terms “testimonial injustice.”Footnote 31 The “Me Too” movement aims to confront precisely this phenomenon of testimonial injustice and it is for this reason that their slogan is “Believe Women.”Footnote 32 To the extent that chronic pain patients are disproportionately from groups who tend to be disbelieved,Footnote 33 the physician has an additional reason, both epistemic and moral, to correct this injustice and “Believe Pain.”Footnote 34
In sum, in balancing the conflicting calls to action and inaction, the fact that a patient reports current pain weighs distinctively in favor of action. The presence of the pain obviates the need for future prediction about the imperative to treat.
Harm to third parties and to society
Opioid prescribing for chronic pain differs from the standard medical context in another important respect, one that has garnered significantly more attention than the contemporaneousness of pain. Communities around the country have been devasted by the “opioid epidemic,” words that convey the scope of the problem.Footnote 35 In what follows, we argue that while this problem is real, serious, and warrants governmental attention, the individual doctor treating an individual patient ought not to consider broader societal effects when making a treatment recommendation.
Doctors treating individual patients, like lawyers with individual clients, have obligations that run directly to the patients and that prohibit the sort of divided loyalties that give rise to conflicts of interest.Footnote 36 It is for this reason that ethical issues arise when physicians gain financially by prescribing the drugs made by particular companies,Footnote 37 or when physicians are financially rewarded for offering less care.Footnote 38
While the doctor is not prohibited from ever considering how her treatment of her patient affects others, it is important to minimize conflicting loyalties as much as is practicable. Ethical tension is lessened when the obligation to consider others does not compromise the medical care provided to the patient. For example, the law requires that doctors report suspected child abuse and wounds that appear to be caused by gun violence.Footnote 39 While such reporting is justified by the interests of others, it does not impair the medical care provided. Were a doctor to decline to prescribe opioid medication because she worries that unused drugs may be stolen or sold, the doctor compromises patient care on behalf of community welfare. Instead, she should mitigate potential harms by educating the patient regarding the risk to others if medication falls into the wrong hands, urging safe medication storage and disposal practices and providing the smallest effective supply of medication for her patient.
This analysis implies that the public health official has different obligations that may yield a different recommendation about how best to balance the twin goals of medical practice. We address the public health perspective in the next section.
THE PUBLIC HEALTH PERSPECTIVE
The public health perspective is different. The official charged with developing recommendations for the use of opioid medication must ask herself what is best for everyone. For her, it is appropriate to balance patients' need for treatment against the harm to others that may be caused by the diversion of prescribed drugs. We offer two recommendations and two warnings to guide the public health official.
Before doing so, we first address whether the DNH principle applies to the public health official at all, as the directive in the Hippocratic Corpus focuses on the treating physician. While this is accurate, we argue that the public health official faces analogous ethical imperatives to do good and to avoid harm. As a result, she too must grapple with how best to balance these twin aims. Indeed, the number of people potentially affected by action and inaction is far greater in the case of the public health official, a fact that complicates rather than obviates the balancing the public health official must consider.
First, the policies adopted by the public health official should minimize as far as possible the ethical strain they place on clinicians treating pain patients. Second, public health officials should also keep in mind the issue of testimonial injustice and calibrate their recommendations in light of the fact that disbelieving pain patients may exacerbate these injustices.
We now turn to the warnings. In the context of the individual patient and her provider, the DNH principle served as a counterweight to the supposed physician bias toward action. In the public health context, there is no uniform bias toward either action or inaction.Footnote 40 Instead, what we see are two important dynamics that public health officials must be aware of and accommodate.
First, whether there is a bias toward action or inaction may depend upon who it is that is suffering. When the group suffering enjoys high social status, there is likely to be a bias toward action. When the group suffering lacks social status, there is likely a bias toward inaction. The argument has been made that drug overdose deaths were ignored for far too long owing to the low social status of people who use drugs or have addiction, and that policymakers became responsive to the opioid crisis and began to treat addiction as a public health matter because of the perception of its impact on white Americans.Footnote 41
Second, officials should recognize that policy is often responsive. We see one problem, react; another develops, react. Public health officials should be cautious not to over-react in addressing one problem lest they create a different problem in the policy's wake.
Indeed, each of these dynamics can be seen in our policy on opioid prescribing. As noted earlier, the problem of addiction likely attracted more sympathy when the affected population was perceived to be largely white.Footnote 42 Regarding cycling, we have whipsawed from over-emphasizing the benefits of opioid medication and focusing narrowly on treating pain to caring dramatically more about the risks of addiction.Footnote 43 Our first warning to public health officials is to be attentive to these two dynamics: the tendency of action or inaction to track status and the dangers of cycling.
Our second warning relates to the manner in which classic trade-offs between rules and standards are complicated in the context of opioid prescribing. As lawyers know well, there are benefits and drawbacks to each approach.Footnote 44 Consider two examples. We have speed limits to ensure that motorists drive at a safe speed. We could simply say, “Drive safely,” but we don't, because we believe that we will achieve more safety by telling drivers that they cannot exceed 65 mph. Of course, that rule sometimes produces errors. Some drivers could drive safely at 75 mph and some drivers should be driving 50 mph (given their skill, the weather, or other factors). But the choice of a rule (65mph) versus a standard (drive safely) reflects the judgment that there will be fewer errors with the rule. In other contexts, the standard is the better choice. We direct family court judges to make custody determinations in accordance with “best interests of the child” because we believe these judges will reach better decisions guided by this goal than they would if constrained by a rule that limits discretion and sets specific criteria.Footnote 45 The point to stress is that rules and standards each produce errors. The choice between them requires we ask: will there be more errors with a rule or more errors with a standard, in the particular context?
This is familiar terrain. What is sometimes neglected is the manner in which the structures and systems involved in the enforcement of a policy affect how rule-like or standard-like a policy actually turns out to be. Our second warning to public health officials is to be aware of how a standard can be transformed into a rule by the actors who play a role in its implementation.
Prescribing guidelines for opioid-based medication provide an apt example of this phenomenon. In 2016, the Centers for Disease Control and Prevention (“CDC”) issued its “Guideline for Prescribing Opioids for Chronic Pain” (“Guideline”) designed to inform primary care physicians about safer opioid prescribing.Footnote 46 This Guideline encourages doctors first to use alternative medications or treatments for chronic pain as opposed to opioids, and, when opioids are used, to prescribe at the lowest effective dose for the shortest effective duration.Footnote 47 The Guideline provides a standard that is non-mandatory guidance, though coming from the CDC, it carries significant weight.Footnote 48
Nevertheless, two of the more concrete provisions in the Guideline have been widely adopted into laws and policies by state legislatures, major insurance companies, major pharmacy chains, pharmacy benefit managers, and others in ways that have had reverberating implications throughout the healthcare system.Footnote 49 These progeny of the CDC's Guideline take a stricter, more rule-like form. One provision that is included in the laws enacted in over half the states limits the number of days for which a physician may prescribe opioids for acute pain.Footnote 50 Another uses dosage guidance that was designed to assist physicians when starting a new patient on opioids but has become, in effect, a benchmark for safe prescribing.Footnote 51
A final actor involved in implementation of the Guideline is law enforcement, which is using the dosage guidance in the CDC's Guideline to make decisions about who to watch.Footnote 52 Prescribing over the dosage suggested in the Guideline may subject a clinician to oversight.Footnote 53 This development has had downstream effects on patients. A landmark report issued by the international watchdog group, Human Rights Watch, for example, found that doctors were tapering patients off of opioids or denying them care—and doing so even against their medical judgment—in an effort to protect their licenses or stay under the radar of law enforcement.Footnote 54
Each of the subsequent actors—pharmacists, insurers, law enforcement officers, and physicians themselves—have, in effect, treated a guideline as a rule. Public health officials who draft policies must take account of these predictable dynamics. While guidelines may be written in a manner that leaves discretion to clinicians, other actors may react in ways that subtly but powerfully transform the policy into more of a rule.
To be fair, the CDC this year addressed concerns that the proliferation of laws and policies stemming from its Guideline risked harm to patients.Footnote 55 The CDC issued clarifications that its guideline was being misapplied by policymakers, and underscored the danger of misapplication in view of the limitations of available evidence.Footnote 56 In a coordinated effort, the FDA issued a warning and announced a label change for opioid medication that cautions against abrupt cessation of opioids, advising that doing so can result in withdrawal symptoms, uncontrolled pain, and suicide.Footnote 57 This is a positive step but one that may not be as effective as hoped: the response has been so wide-reaching that course correction of the previous directive may be hard to dislodge.Footnote 58
At least with respect to one group of people, the between 8 and 13 millionFootnote 59 Americans who currently take opioids for pain, there is reason for concern. We turn to the special obligations to these patients in the next section.
SPECIAL OBLIGATIONS TO PATIENTS ALREADY ON OPIOID TREATMENT
Opioid prescribing policy should also consider what duty is owed to those who already take opioids long-term for chronic pain. Some of these patients have functioned on opioids for years or even decades.Footnote 60 Some, though certainly not all, are on higher dosages than is now considered ideal, and may do better at lower doses.Footnote 61 Yet, while risks rise with higher doses, there is no set dosage threshold that will be appropriate for all patients.Footnote 62
Recently, there has been an uptick in reports of long-term pain patients being abruptly tapered off their medication or abandoned in care altogether.Footnote 63 As a result, patients report damage to their mental and physical health.Footnote 64 Some are no longer able to work or function and have suffered financial devastation as a result.Footnote 65 Others have resorted to the illegal market and even suicide when their medication or care is denied.Footnote 66 The studies that have emerged just this year are harrowing. One study of Medicaid patients who had been on high dosages for more than 90 days found that the average time to discontinuation of opioids was 24 hours, and that it was often followed by an opioid-related hospital or emergency room visit as a result.Footnote 67 Another showed a three-fold increase of overdose death, just from destabilizing dosage.Footnote 68 Yet another showed that opioid tapering was associated with dissolution in care relationships between providers and patients.Footnote 69 And a recent study found that tapering was happening too abruptly and disproportionately to women and people of color.Footnote 70
Moreover, additional studies suggest that chronic pain patients who use opioids to manage pain are facing increasing barriers to getting access to health care, with one study finding that approximately 40% of primary care physicians are unwilling to take on a new patient who uses opioids to manage pain,Footnote 71 and another concluding that 81% are reluctant to.Footnote 72
Some of these issues are practical in nature and require attention and resources. For example, 70% of physicians report that they need help understanding how tapering can be done in a safe manner.Footnote 73 This data suggests a need for better education. But in an imperfect world in which optimal physician training is unlikely to permeate to all clinicians and thus one in which we can predict that some number of patients may be abandoned or inappropriately titrated by their physicians, what public policy response is appropriate?
Given that public health agencies allowed doctors to prescribe opioids, even at high doses, we believe that society and physicians have special duties to these patients. This contention draws support from the existence of a similar duty in other contexts. For example, although a person has no legal duty to rescue a person in need, if a good Samaritan attempts a rescue, she incurs duties to continue and to carry out the rescue competently.Footnote 74 In a similar vein, Henry Richardson argues that medical researchers incur obligations to research subjects to provide ancillary care and that these duties arise from the fact that the researcher has involved himself in the treatment of the research subject.Footnote 75
Indeed, we already apply these principles to the one in three cancer patients who no longer have active disease but continue to experience chronic pain either as a consequence of the disease or its treatment.Footnote 76 The CDC has stated that its guideline was not intended to apply to anyone who has ever had cancer.Footnote 77 This recognition of the duty to on-going patients supports our contention.
Along these lines, we posit that physicians and society have special obligations that add extra normative heft to the claims of patients currently on opioid-based drugs. These patients did not put themselves on opioids. The duty of the physician who initiated opioid prescribing is the clearer case. But we argue that given the recalibration in views of the medical and public health community on opioid prescribing, this duty extends to the broader medical and public health community. As a result, the harm to existing patients who are inappropriately treated in the wake of more restrictive prescribing policies is not merely a cost to be weighed alongside other costs and benefits, but is instead a wrong that society inflicts on these patients.
One possible solution that retains the benefits of current policy choices, while mitigating harms, might be to provide legal amnesty to clinicians who are willing to take on medication management of patients on long-term opioid therapy a sort of safe haven from heightened oversight. Given misapplication of the CDC's dosage guidance, clinicians who are willing to care for patients on higher doses face elevated oversight risk, and this creates an incentive to discard or fail to treat some of the most vulnerable patients. Any such provision must be drawn carefully and narrowly to avoid backsliding into liberal prescribing. Guidance that is just now emerging from public health authorities might provide sufficient guardrails against such concerns,Footnote 78 while enabling the protection and care of this vulnerable and not insubstantial group of patients.
CONCLUSION
We have highlighted the factors that should inform physician and policy decisions regarding opioid prescribing for pain, examining how both the dangers of opioids and the presence of pain complicate the delicate balance between the twin goals of medicine. We conclude with a final, more general, concern.
At least with respect to patients presently taking opioids, current policy may be turning the ordinary ethical and legal obligations of clinicians on their head. Physicians have an ethical duty to relieve suffering and both an ethical and legal duty not to abandon their patients, especially where serious harm might result from such abandonment.Footnote 79 Refusing to see a patient on the basis of her medical conditionFootnote 80 or the medication she takes for that conditionFootnote 81 also violates that patient's civil rights. To the extent that current policies may be encouraging physicians to do so, these are serious considerations, which raise deeper concerns about undermining both trust in the doctor-patient relationship and the helping nature of the profession.
