Rapid solutions to the opioid crisis remain elusive. Prescriptions for opioids have decreased. Yet, supply limits have not reduced fatalities. Demand-side interventions have not fared any better. British Columbia, a place hit hard by the crisis, saw harm reduction at an all-time high in 2020. People there received 34 000 naloxone kits and filled 23 735 opioid agonist therapy prescriptions. Both are new records.¹ However, there was a rapid rise in opioid-related deaths in 2020 compared with 2019 (1725 vs 985). A fentanyl-laced drug supply overshadows population health efforts. Because of this, changes in practice metrics have not guaranteed improvement in outcomes.

To address these concerns, Sud et al. (p. S56) discuss three types of translations of knowledge to improve outcomes. The first is to favor complex clinical interventions over simple population-based ones. They criticize educational programs, electronic health record prompts, and prescription drug–monitoring programs for focusing on population metrics without tracking clinical outcomes to ensure better success across implementations. Sud et al. view these interventions through a clinical lens. Yet, there is much more to them than can be measured from the standpoint of the identified patient. Although these interventions happen in a clinical setting, affect change in clinical care, and encourage best practices, patient outcomes will not capture the total social impact of opioid deprescribing. The primary benefit of fewer opioid prescriptions may be to families and communities. We know, for instance, that opioid prescriptions spread through households.² One family member receiving an opioid prescription makes it more likely that another will get one. Judicious prescribing discourages household spread, a benefit not measurable at the patient level.

Opioid use predicts loss of pleasure and motivation even after controlling for pain severity, depression, and dosage.³ Through a loss of motivation, a community’s rate of opioid prescriptions may affect labor force participation: the rate of working-age people having or seeking employment. In early 2000, the labor force participation rate reached a maximum of 67.3%, and it has declined at a steady pace since then, reaching a 40-year low in September 2015 at 62.4%.⁴ Opioid prescriptions per capita increased by a factor of 3.5 nationwide between 1999 and 2015. A careful temporal analysis of county-level labor force participation between 2014 and 2016⁴ linked to county-level opioid prescriptions nationally in 2015 found that county-level opioid prescriptions⁵ in 2015 accounted for 43% of the observed labor force decline.⁶

About half of prime-age men not in the labor force are on pain medication, and two thirds of these take prescription pain medications; these men report low levels of subjective well-being.⁶ It is likely that persons in prime age who are not in the workforce could benefit more from visiting a social worker for counseling than a physician offering pills. Other data indicate that county-level opioid marketing is associated with elevated overdose mortality one year later.⁷ Deprescribing interventions countervail community advertisements, impede the spread of prescribing through households, and may have positive effects on labor and the well-being of community members. Yet, these effects cannot be measured in a clinical setting.

Sud et al. would also like clinicians to factor population knowledge into smaller units that can translate into clinical effects. There are many clear examples. At clinic visits, it is safer for 10 people to receive 5 milligrams fewer than for one person to receive 50 milligrams fewer of an opioid. System rollout of psychological treatment of pain may fail to curtail opioid use because opioid use is appetitive. Buprenorphine orders may not rise when regulators remove administrative and legal burdens. This is because poor care coordination, clinician stigma, and lack of peer support are also barriers.⁸ In each case, population metrics may not do justice to the complexities in different local environments. The authors’ concerns seem right. And, although social benefits may be larger than clinical ones, without a careful implementation strategy, population-based interventions may fail altogether. With behavioral interventions, “details matter.”⁹ Interventions require a clinician’s attention and appropriate interpretation of information. For an intervention to work in a new clinic or system, each feature of the intervention must map in some proper way to the new environment. In the context of widespread drug stigma, how clinicians make sense of intervention is important. There needs to be careful attention to knowledge translation. An opioid taper should not be interpreted as an act of enforcement but as a collaboration with the patient.

Finally, Sud et al. describe a tendency in medicine to rank medical facts above patient values when approaching opioid tapering. They review studies that have documented the harms of opioid tapering but neglect to mention that most studies show that most patients improve with opioid tapering. They accuse Juurlink of construing the subjective benefits of long-term opioid therapy as illusory and less relevant than objective harms. Juurlink argues that “the goal of pain medication is not simpl[e] pain relief. Like any therapy, the goal is to confer more benefit than harm.”^10(pe1222) He does not devalue the patient perspective. He simply does not want to limit the patient’s perspective to pain intensity ratings (a perspective that doctors have largely forced on patients). Patients’ ability to function at work, at home, and in relationships is also relevant.

Sud et al. also call for “epistemic humility”: “Chronic pain, as an inherently subjective condition, frustrates the core epistemology of clinical biomedicine that relies on evidence to objective pathology” (p. S61). We agree with this statement. Modern biomedicine seeks to explain subjective symptoms with objective tissue damage. It has also tended to dismiss “medically unexplained symptoms,” including pain, as not legitimately medical and likely psychological in origin. Yet modern pain research has revealed pain, especially chronic pain, to be a complex human experience with a loose and variable relationship to tissue damage. We agree that we should begin the clinical encounter by believing our patients’ reports of pain. But this is not because, as Sud et al. say, “[P]eople living with pain have unique epistemic access to their lives” (p. S63). The 17th-century French philosopher René Descartes argued that pain experience was incorrigible, but modern philosophy sees pain as interpersonal. We know a mother’s reaction has a direct effect on her child’s experience of pain.¹¹ And, when clinicians help patients dispute chronic pain’s causes and threat value, both pain experience and brain processing change.¹² This offers us new and effective means to treat the clinical and population problems of chronic pain.

Believing our patients’ reports of pain is only the beginning of our clinical responsibility to the patient with pain. Chronic pain has many causes—physical, psychological, and social—that must be investigated and addressed. Short-cutting this process by using a simple quantitative measure of pain intensity and seeking to reduce this number with opioid medication in accord with a claimed right to pain relief has brought us an opioid epidemic that has still not ended.

In sum, liberal opioid prescribing imposed a heavy cost on US communities and spawned a robust and toxic illicit market. Whether population-based interventions impose a cost on individuals, in the long run, is a worthy concern. However, this concern will be solved best by expanding mental health care, community support, and social service rather than turning our gaze back to increasing opioid prescriptions.

REFERENCES

Section:

ChooseTop of pageREFERENCES <<CITING ARTICLES

References