Clinically focused interventions for people living with pain, such as health professional education, clinical decision support systems, prescription drug monitoring programs, and multidisciplinary care to support opioid tapering, have all been promoted as important solutions to the North American opioid crisis. Yet none have so far delivered substantive beneficial opioid-related population health outcomes. In fact, while total opioid prescribing has leveled off or reduced in many jurisdictions, population-level harms from opioids have continued to increase dramatically.

We attribute this failure partly to a poor recognition of the epistemic and ethical complexities at the interface of clinical and population health. We draw on a framework of knowledge networks in wicked problems to identify 3 strategies to help navigate these complexities: (1) designing and evaluating clinically focused interventions as complex interventions, (2) reformulating evidence to make population health dynamics apparent, and (3) appealing to the inseparability of facts and values to support decision-making in uncertainty.

We advocate that applying these strategies will better equip clinically focused interventions as complements to structural and public health interventions to achieve the desired beneficial population health effects. (Am J Public Health. 2022;112(S1):S56–S65.

Population health challenges are widely understood as wicked problems—they defy definitive formulation, relate to sets of other problems in complex causal chains, lack clear end points, and are changed irreversibly by every intervention, among other notable characteristics.1,2 The ever-increasing harms from opioids in high-income North America have been identified as one such wicked problem.3,4 There are multilevel and multidirectional dynamics at play in this crisis ranging from pain management to substance use to opioid use disorder. Likewise, diverse sets of actors with diverse experience have been called upon to address this wicked problem. This crisis is perhaps unique in that clinical interventions, particularly the prescribing of high-dose, long-acting opioids for pain management, initially fueled the “first wave” of the crisis.5,6 This approach has driven ongoing crisis responses using clinically focused interventions, namely interventions focused on changing clinical practice and clinician behavior. Yet, the particular focus around reducing opioid prescribing has been met with limited success and contributed to subsequent waves of the crisis. Prescribing has fallen or leveled off in many jurisdictions (though absolute levels remain very high compared with previous decades, to other countries, and to levels considered safe and effective for chronic pain management),7–10 but opioid-related harms have continued to escalate, especially because of ongoing criminalization of substance use, which has fostered the conditions for a toxic street drug supply.

Clinical interventions failing to improve population health outcomes may be a general problem. Clinical systems have increasingly been oriented toward population goals, but the investments in high-performing, networked clinical systems have yet to substantially deliver.11–13 Many have appropriately called on the increased prioritization of population health interventions focusing on root causes of the opioid crisis, particularly around intersecting issues relating to the social determinants of health, criminalization of drug use, stigma, and discrimination.14 We very much agree.15 However, we propose here that addressing the complex dynamics at the interface of clinical practice and population health can inform better design and delivery of clinical interventions to achieve the desired population effects and, thus, better complement existing and future population health interventions.16

As Weber and Khademian note in their synthesis of the role of knowledge and networks in addressing wicked problems, “a fundamental challenge to effectively managing any public problem in a networked setting is the transfer, receipt and integration of knowledge across participants.”2(p334) In particular, they identify 3 kinds of knowledge translation processes that can increase the collaborative capacities networks to address wicked problems: knowledge as “syntactic,” as “semantic,” and as “pragmatic.” On the basis of our previous research and experience intervening in the North American opioid crisis, we use this 3-fold strategy to identify how clinical interventions can be more likely to address population health goals for the chronic pain and opioid crises: (1) designing and evaluating clinically focused interventions as complex interventions (knowledge as syntactic), (2) reformulating evidence to make population health dynamics apparent (knowledge as semantic), and (3) appealing to the inseparability of facts and values to support decision-making in uncertainty (knowledge as pragmatic). We focus our discussion here on the subissues of chronic pain and opioid analgesic prescribing, but we aim for the approaches to be generalizable to related crisis interventions, such as the treatment of opioid use disorder.

Weber and Khademian label the first knowledge-based strategy for improving network capacity as “syntactic”—namely, “finding ways to standardize or make compatible methods of communication to facilitate the transfer of knowledge from one participant or organization to the next.”2(p339) We suggest here that applying practices for the design and evaluation of complex interventions to clinical interventions can support appropriate knowledge transfer for the opioid crisis.

Common clinically focused interventions for the opioid crisis have included education programs for clinicians as well as the deployment of prescription drug monitoring programs and clinical decision support systems.17 Chronic pain and opioid prescribing education have been identified as key opioid crisis solutions.18–20 This has been driven by the recognition of the long-standing deficiencies across the educational continuum, including the co-optation of education by commercial interests.21–23

A recent systematic review by Sud et al.24 synthesized the evaluations of 32 opioid analgesic continuing education programs, primarily from the United States and Canada, and demonstrated that 84% of programs used population-level opioid harms to justify their development and design. Only 3 programs25–27 reported population health outcomes, none of which could be related directly to the specific patients or communities of program participants. Generally, program evaluation designs were insufficient for determining population health effects—a challenge endemic to continuing medical education perhaps because of the lack of conceptual frameworks for population health in this field.24

Clinical decision support systems and prescription drug monitoring programs have also been widely deployed as interventions to improve opioid prescribing and chronic pain care. Clinical decision support systems are electronic systems that assist clinical decision-making by providing point-of-care, patient-specific data.28–30 The scoping review of clinical decision support systems for opioid prescribing in primary care by Spithoff et al.31 identified 14 program evaluations between 2009 and 2019. Outcomes typically focused on opioid-prescribing patterns and concordance of clinical practices with guidelines. There was poor utilization of evidence-based design components, minimal assessment of program implementation, no measurement of patient health or population health outcomes, and no assessment for unintended negative consequences. Similar findings have been reported for prescription drug monitoring programs.32 Importantly, the comprehensive systematic review by Furlan et al. of interventions for opioid prescribing and opioid-related harms identified a high rate of unintended negative consequences from clinical interventions, such as overdose, improper prescribing, and increased stigma, even though not all evaluations assessed for these consequences.17 Health interventions of many types have benefited from being conceptualized, delivered, and evaluated as complex interventions.33 Clinically focused interventions for the opioid crisis often meet criteria for complexity—they involve the actions of people and complex chains of steps, are embedded in social systems shaped by context, and are open systems subject to change.34,35 Based on a recent systematic review,24 we will describe the only 2 continuing education programs explicitly conceptualized as complex interventions with formal implementation evaluations to highlight how acknowledging this complexity may help overcome some of the challenges identified previously.

Safer Opioid Prescribing is a Canadian continuing education program designed as a multipart, scalable intervention to improve clinical practices around chronic pain and opioid prescribing, aiming for positive population-level impacts (A. S. is director for this program).36 The program was designed using the PRECEDE‒PROCEED model, commonly used in population health initiatives.37,38 This framework allowed program developers to

a) contextualize [the program] within the specific circumstances of the Canadian contemporary opioid epidemic and the range of other policy options for addressing it; b) involve the target audience for the intervention in program planning; and, c) conceptualize and categorize specific implementation and effectiveness outcomes during the initial design stages.36(p3)

Initial evaluation of the program identified an important syntactic convergence between frameworks for the planning and evaluation of complex interventions and frameworks for continuing education, including distinguishing among implementation, effectiveness, and impact outcomes.39,40 Formally evaluating implementation is fundamental in answering questions about how and why programs work, besides answering questions of whether they work.41,42 Likewise, implementation evaluations help explain how programs operate within specific environments and so can be useful in informing transferability to different contexts.43

Barth et al.44 describe the use of the Medical Research Council’s complex intervention framework to develop and evaluate an academic detailing intervention to improve use of the South Carolina prescription drug monitoring program. The key advantage of conceptualizing the program in this way is that it allowed the developers “to identify and clarify potential component parts of the intervention and how the active components might relate to the expected outcome[s].”44(p103) Using the Medical Research Council framework further allowed them to identify and address implementation challenges, iteratively develop the program, and adapt it to different practice settings and evolving legislation.

Although neither of these programs have yet demonstrated population-level impacts, their explicit use of complexity frameworks still accomplishes 3 important objectives. First, it provides the infrastructure for understanding program function and its relation to outcomes, including possible unintended negative outcomes. Second, it provides opportunities for iterative program development and responsiveness to changing context, such as policy and epidemiological trends. Third, it provides a connecting “syntactic” bridge between the fields of population-focused and clinically focused interventions—this can allow for better intercalation between these kinds of interventions, which is essential for responding to a wicked problem in which no single intervention, or intervention type, is going to be sufficient.2

As a further example, chronic pain self-management has been implemented as key population health intervention.45 Interventions that are effective for pain management—even at a population level—are not guaranteed as effective population-level interventions for reducing opioid-related harms. However, we advocate here that a syntactic approach of using complex interventions, especially the concurrent assessment of implementation and effectiveness outcomes,46 is a useful means to design and assess for such effects.47

Interventions and, thus, their effects are shaped by the knowledge drawn upon for their development. This influence is conspicuous in some of the interventions described previously, whose goals were described as “guideline-concordant care.” Thus, attention to the knowledge called upon by opioid crisis interventions, including specifically knowledge synthesized in clinical practice guidelines, is essential to equipping them for population-level impacts. This follows Weber and Khademian’s second knowledge-based strategy—namely a “semantic” strategy of translating and interpreting available knowledge to make it useful for the task at hand.2 We propose here that opioid crisis‒related evidence must be considered in terms of population health dynamics to expect interventions to achieve population health effects.48 In the following sections we explore the knowledge synthesis and translation process especially as it has played out in North American opioid prescribing guidelines. This is relevant because of the prominence afforded to such guidelines despite the absence of evidence as effective policy interventions.

Risk Concentration vs Risk Distribution

Over the past decade, there has been substantial and sustained attention to the harms caused by high-dose prescribed opioids. This concern can be traced to recommendations in 2 national clinical practice guidelines and observational studies referenced therein.49–52 In 1 study, Dunn et al.51 examined a cohort of 9940 US health management organization patients on long-term opioid therapy for pain. They stratified the cohort into 5 groups by opioid dose and determined the overdose hazard ratio. The primary observation was that overdose risk increased with dose. Those who were prescribed more than 100 morphine milligram equivalents (MME) per day had a nearly 9-fold higher overdose risk compared with those prescribed 1 to 20 MME per day (Table 1). This important and compelling finding has influenced innumerable clinical interventions, all aiming to decrease harms in the highest-risk population with doses greater than 100 MME per day.


TABLE 1— Hazard Ratios Between Opioid Doses and Overdose: Washington State, 1997--2005

TABLE 1— Hazard Ratios Between Opioid Doses and Overdose: Washington State, 1997--2005

Opioid Dose Patients Who Overdosed, No. (%) Person-Years Overdose Rate per 100 000 Person-Years (95% CI) All Overdose Events, HR (95% CI)
None 6 (12) 16 780 36 (13, 70) 0.31 (0.12, 0.80)
Any opioid use 45 (88) 17 582 256 (187, 336) 5.16 (2.14, 12.28)
 1 to < 20 MME 22 (43) 13 770 160 (100, 233) 1 (Ref)
 20 to < 50 MME 6 (12) 2 311 260 (95, 505) 1.44 (0.57, 3.62)
 50 to < 100 MME 6 (12) 886 677 (249, 1317) 3.73 (1.47, 9.50)
 ≥ 100 MME 11 (22) 614 1 791 (894, 2995) 8.87 (3.99, 19.72)

Note. CI = confidence interval; HR = hazard ratio; MME = morphine milligram equivalents per day.

Source. Dunn et al.51

In the influential formulation of population health by Rose et al.,48 focusing interventions on those prescribed high doses is a classic example of the “high-risk” strategies often favored by clinicians—it focuses intervention on the subpopulation with the highest concentration of risk. Such a high-risk orientation is made explicit in the 2017 Canadian Opioid Guidelines, which recommended tapering opioids to the lowest effective dose for patients currently using 90 MME of opioids per day or more.50 This guideline targets “those who prescribe opioids or create policy regarding this issue,”50(p.e660) clearly signaling its health policy‒informing intentions.

Yet, one of Rose’s key injunctions is that population health interventions not only must address this high-risk element of the population but also must consider risk distribution within a population.48 In a typical population, the majority of cases will be contained in the low-risk subpopulation, apparent when we reinterpret the observations of Dunn et al.51 from a risk distribution perspective (Figure 1).

As made visible in this representation, more than twice as many overdoses were in the 1 to 20 MME compared with the 100 MME or higher group. We advocate here that clinically focused interventions aiming to have population-level effects should “semantically” reinterpret their evidence to identify such distributive effects and, thus, inform the nature of their interventions toward populations.

Context and Complexity in Evidence Synthesis

One clinical strategy promoted as an opioid crisis response has been multidisciplinary care (MDC).53 Indeed, there is long-standing evidence for the efficacy of MDC for chronic pain management, including some evidence that MDC can constrain opioid prescribing.54 Despite this evidence and the high burden of complex chronic pain, even high-resource health systems provide persistently poor access to MDC.55,56 In response, as 1 example, the Ontario Ministry of Health initiated CAD $17 million in annual funding to support the development and operation of 17 MDC pain clinics. Notably, the funding announcement came not as part of a provincial response to a crisis of chronic pain but specifically as a response to the province’s opioid crisis.57 While the mechanisms and effects of MDC for the management of chronic pain have been well-studied and understood for decades, the operation of MDC as an opioid dose‒reduction strategy aiming for population-level effects is much less clear. Given complex dynamics, it is not necessary that an effective chronic pain management intervention will be effective for opioid-related harms.

A recent systematic review of 21 studies using MDC strategies for opioid tapering identified a mean opioid discontinuation rate of 87% (range = 29%–100%). The significant heterogeneity across studies with respect to program components, personnel, philosophical approaches, duration, and settings did not allow any further synthesis using traditional aggregative synthesis methods. Despite this heterogeneity, the 2017 Canadian Opioid Prescribing Guideline made a strong recommendation that patients using opioids and experiencing serious challenges in tapering should be referred to formal multidisciplinary program.50 In addition to the lack of access to MDC, an important challenge of this recommendation is that the guideline did not clearly define MDC for opioid tapering. Without a clear definition, it is difficult to interpret and operationalize this recommendation as either a clinical or opioid crisis intervention.

The appropriate mobilization of evidence for complex problems or interventions is not a new challenge. The methodological and epistemological diversity of evidence available to inform such interventions requires synthesis methods distinct from those used for the aggregation of outcomes from clinically focused trials in highly controlled conditions.58 Methods such as realist synthesis instead acknowledge the complexity of context and the relevance of context to determining intervention effects.

Sud et al.59 conducted a realist synthesis of 95 MDC program evaluations across 5 decades, which reached distinct findings compared with previous systematic reviews and guideline recommendations, and suggests specific population health lessons not previously apparent. First, this review identified 3 necessary but insufficient components related to opioid-dose reductions: pain relief, behavior change, and active medication management. This defies the conventional understandings of MDC mechanisms, which suggest that substituting pharmacological analgesia with nonpharmacological analgesia is sufficient for achieving opioid dose reductions. Second, the review identified that context very much mattered. The national orientation toward opioids directly influenced MDC program design and effects. While Northern European and American programs had similar pain and function outcomes, Northern European programs typically did not include active medication management and, thus, did not reduce opioid doses. Finally, the rate of return to opioid use after achieving opioid dose reduction was as high as 20% to 40%. This could be an acceptable and harm-reducing outcome, especially as the return to use was often at a lower dose than at program outset. However, in the contemporary context of a highly toxic street drug supply and restricted access to pharmaceutical opioids, even a small return to use involving street drugs could easily undo any modest population benefits of prescribed dose reductions.

In summary, because of heterogeneity in how interventions are defined, implemented, and then also evaluated, an intervention that is effective for improving chronic pain carries no guarantee of being an effective intervention for the opioid crisis. To expect clinically focused interventions to have specific population effects, the knowledge base informing these interventions must be “semantically” reinterpreted and synthesized from the perspectives of population health. Besides the complex nature of the interventions, this includes the complex nature of the contexts in which they will be deployed.60

While “syntactic” processes can be developed to help speak a common language within a network, and “semantic” interpretation may help adapt different kinds of knowledge available in a network, knowledge is always incomplete.61 Knowledge is also always provisional and defeasible—new knowledge may arise to supplant or otherwise change existing knowledge.62 Yet, the duty to care and act, on the parts of both clinicians and policymakers, is not removed when knowledge is incomplete, provisional, and defeasible.63 Building on Weber and Khademian’s identification of knowledge as localized and tied to practice,2 we identify here that knowledge as the basis for decision-making in response to wicked problems, besides being constituted by facts, is also constituted by values. While some may despair at this acknowledgment that values are inherent to knowledge and, thus, to clinical and policy decision-making, lest it undermine evidence-based practice, we offer an alternative view that making such values explicit can improve the capacity to address wicked population health problems.64

Facts and Values Are Not Separate

A common assumption is that scientific evidence presents “just the facts.” Indeed, the logic of evidence-based medicine is that empirically derived facts can reduce uncertainty around intervention effects and mechanisms and thus inform what interventions ought to be implemented. Facts, or knowledge as generated through scientific methods, are considered as separate from values, which cannot be derived empirically or logically. Values (e.g., normative ethics analysis about how the world ought to be, including what counts as an “effective” intervention) are subjective and part of personal or communal ethics. This separation of facts and values is called the fact‒value distinction.65–67

Closely examining the practice of opioid tapering as a harm-reducing opioid crisis intervention provides a useful example of the inseparability of facts and values. Retrospective observational studies have demonstrated that stopping opioid therapy can actually substantially increase overdose or suicide mortality.68,69 Inappropriate tapering can also lead to poor pain control and loss of functional abilities as well as a sense of medical abandonment.63 This presents a challenging situation: the facts about the harm-reducing effects of lower doses of opioids are provisional and defeasible, and so seem insufficient to guide appropriate intervention.67 Instead, values, implicitly or explicitly, guide decision-making in this context.

Identifying a crisis of chronic pain and epistemically valuing its treatment versus identifying a crisis of opioid-related mortality and valuing avoidance of iatrogenic harm justify different kinds of interventions. For example, Juurlink70 construes the (subjective, unprovable) benefits of long-term opioid therapy as illusory and, thus, less relevant in decision-making compared with the (objectively) demonstrated risk of harms. This differential valuing of objective over subjective knowledge entails a particular mode of action—in this example, reducing opioid doses over maintaining them. Decisions about whether one ought to pursue opioid tapering reflects an interrelationship between facts and values that is often overlooked. Chronic pain, as an inherently subjective condition, frustrates the core epistemology of clinical biomedicine that relies on evidence of objective pathology. This introduces an epistemic hierarchy within clinical care and policy, prioritizing objective over subjective evidence. As a result, pain sufferers and people who use drugs continue to have their credibility undermined and their testimonies about their individual experiences, values, or priorities marginalized.71

Clinical‒Population Health Values

When discussions of values regarding the concurrent chronic pain and opioid overdose crises have been made explicit, questions have been raised about whether ethical analysis should focus on clinical or population health levels.72 From a clinical perspective, opioid tapering raises ethical issues of respect for patient autonomy, voluntary free and informed consent, safety, individual risk‒benefit profiles, and the patient‒clinician relationship.73,74 Alternatively, hazards attributable to commercial involvement in aggressive marketing, sponsorship of medical education, and interference in policy processes are well established and have been raised as prominent population health ethics issues. So have the issues of discrimination and stigmatization, which have intensified the inequities of people living with pain, alongside people who use drugs.75

The concept of structural iatrogenesis can help make explicit the values at the intersection of these 2 levels, specifically to support decision-making when facts are insufficient. Stonington and Coffa76 define structural iatrogenesis as the harm to patients caused by bureaucratic systems within health care, including systems intending to benefit patients.76 They provide a case example of how a clinician could notice how aspects of her clinic’s opioid prescribing policy (e.g., urine screens, opioid contracts) created frequent gaps in medication coverage and created harms for patients within larger-scale social forces (e.g., lack of transportation, manual labor for economic survival). Other examples of structural iatrogenesis abound at the intersection of pain and opioid use in the form of requiring prescriptions for naloxone, requiring special authorizations to prescribe agonist therapies,77 or the use of (often punitive) opioid treatment agreements.78 Such decisions are not only individual actions but they are also the result of structural processes reflecting intersecting concerns, interests, and actions of people in pain, people who use drugs, clinicians, professional and political organizations, insurance companies, pain and opioid policies, and laws.

The values inherent to structural iatrogenesis can provide guidance for the improvement of health at the clinical‒population health interface (Figure 2). Such values include solidarity (a shared interest in survival, safety, and security)80; distributive and social justice, represented by equitable access to life-saving interventions such as pain management, naloxone, and opioid-agonist therapies; and epistemic humility, which is of particular relevance when considering knowledge-based processes. Making these values explicit in decision-making can help shift the gaze from individual patients as “dysfunctional” or “high-risk” to a focus on their needs and the responsibilities of the broader public to help meet those needs.81

Specifically, epistemic humility is a disposition and a commitment to engage in collaborative effort that arises out of recognizing the limits of one’s knowledge.71,78,82 One’s picture of a clinical scenario, or one’s facts about how structural processes might disproportionately harm certain patients, are incomplete. Identifying the limits of facts, and thereby making apparent the active role of implicit values, is a key strategy that Stonington and Coffa identify for addressing structural iatrogenic harms.76

Embracing epistemic humility under conditions of uncertainty challenges the presumed knowledge hierarchy between decision-makers and people with lived experience, recognizes that people living with pain have unique epistemic access to their lives, and emphasizes that inclusion of their voices in treatment decision-making and policymaking is necessary. Decision-makers can avoid committing epistemic injustice by not presuming “objective facts” to be the only relevant considerations in treatment and intervention decisions.83 A commitment to epistemic humility when making decisions in the context of uncertainty means balancing the patient’s experiences, priorities, and values with the specialized content knowledge and inseparable values of the decision-maker.

For the intersecting crises of chronic pain and opioids, the potential for conflict between the knowledge and values of clinical practice with those of population health is high. Yet, wicked population health problems necessarily call on the knowledge and values of diverse system actors to identify tractable solutions. Using Weber and Khademian’s 3-fold conceptualization of “knowledge in networks”2 to increase collaborative capacity for addressing wicked problems, we have identified 3 means for action at this complex interface of clinical and population health: Knowledge as “syntactic” facilitates knowledge transfer and use by allowing network actors to “speak the same language,” as we have identified around designing and evaluating clinically focused interventions as complex interventions. Knowledge as “semantic” identifies the importance of interpretation across network actors, as we have suggested in reformulating evidence in terms of both risk concentration and distribution, as well as in terms of synthesizing evidence that is context-responsive. Finally, a “pragmatic” view of knowledge understands that knowledge is changing, highly contextual, socially situated and embedded in practice, and, thus, irreducibly connected to values. Clinically focused measures continue to be essential for addressing chronic pain and the opioid crisis. We advocate that applying this more nuanced understanding of knowledge at the clinical‒population health interface will better equip clinically focused interventions to have the desired beneficial population health effects.

See also Doctor and Sullivan, p. S15, and Nicholson, p. S18.


A. Sud receives an academic stipend as director of Safer Opioid Prescribing. D. Z. Buchman received salary support from the Centre for Addiction and Mental Health. A. D. Furlan is the inventor of the Opioid Manager, a point-of-care tool and app that is sold in iTunes; all proceeds belong to the hospital where she works. A. D. Furlan receives advertising revenues from Google Inc. for her YouTube channel. P. Selby reports receiving grants, salary, and research support from the Centre for Addiction and Mental Health, Health Canada, Ontario Ministry of Health and Long-Term Care, Canadian Institutes of Health Research, Canadian Centre on Substance Use and Addiction, Public Health Agency of Canada, Ontario Lung Association, Medical Psychiatry Alliance, Extensions for Community Healthcare Outcomes, Canadian Cancer Society Research Institute, Cancer Care Ontario, Ontario Institute for Cancer Research, Ontario Brain Institute, McLaughlin Centre, Academic Health Sciences Centre, Workplace Safety and Insurance Board, National Institutes of Health, and the Association of Faculties of Medicine of Canada. P. Selby also reports receiving funding, honoraria, or both from the following commercial organizations: Pfizer Inc/Canada, Shoppers Drug Mart, Bhasin Consulting Fund Inc, Patient-Centered Outcomes Research Institute, ABBVie, and Bristol-Myers Squibb. Furthermore, P. Selby reports receiving consulting fees from Pfizer Inc/Canada, Evidera Inc, Johnson & Johnson Group of Companies, Medcan Clinic, Inflexxion Inc, V-CC Systems Inc, MedPlan Communications, Kataka Medical Communications, Miller Medical Communications, Nvision Insight Group, and Sun Life Financial. R. E. G. Upshur receives salary support from the Lunenfeld Tanenbaum Research Institute and holds an endowed chair from the Dalla Lana School of Public Health at the University of Toronto.


The development of this article did not involve research including human participants.


1. Rittel HWJ, Webber MM. Dilemmas in a general theory of planning. Policy Sci. 1973;4(2):155169. CrossrefGoogle Scholar
2. Weber EP, Khademian AM. Wicked problems, knowledge challenges, and collaborative capacity builders in network settings. Public Adm Rev. 2008;68(2):334349. CrossrefGoogle Scholar
3. Lee JC. The opioid crisis is a wicked problem. Am J Addict. 2018;27(1):51. Crossref, MedlineGoogle Scholar
4. McCrea AM. Can administrative capacity address wicked problems? Evidence from the frontlines of the American opioid crisis. Adm Soc. 2020;52(7):9831008. CrossrefGoogle Scholar
5. Alam A, Juurlink DN. The prescription opioid epidemic: an overview for anesthesiologists. Can J Anaesth. 2016;63(1):6168. Crossref, MedlineGoogle Scholar
6. Centers for Disease Control and Prevention, National Center for Injury Prevention and Control. Understanding the epidemic. March 17, 2021. Available at: Accessed June 25, 2021. Google Scholar
7. Centers for Disease Control and Prevention. US opioid dispensing rate maps. December 7, 2020. Available at: Accessed June 22, 2021. Google Scholar
8. Jones W, Vojtila L, Kurdyak P, Fischer B. Prescription opioid dispensing in Canada: an update on recent developments to 2018. J Pharm Policy Pract. 2020;13(1):68. Crossref, MedlineGoogle Scholar
9. Scholten WK, Christensen A-E, Olesen AE, Drewes AM. Quantifying the adequacy of opioid analgesic consumption globally: an updated method and early findings. Am J Public Health. 2019;109(1):5257. LinkGoogle Scholar
10. Richards GC, Aronson JK, Mahtani KR, Heneghan C. Global, regional, and national consumption of controlled opioids: a cross-sectional study of 214 countries and non-metropolitan territories. Br J Pain. 2021; Epub ahead of print. Google Scholar
11. Farmanova E, Kirvan C, Verma J, et al. Triple aim in Canada: developing capacity to lead to better health, care and cost. Int J Qual Health Care. 2016;28(6):830837. MedlineGoogle Scholar
12. Government of Ontario, Ministry of Health, Ministry of Long-Term Care. Patients first: action plan for health care. January 31, 2009. Available at: Accessed February 19, 2021. Google Scholar
13. Washington AE, Coye MJ, Boulware LE. Academic health systems’ third curve: population health improvement. JAMA. 2016;315(5):459460. Crossref, MedlineGoogle Scholar
14. Dasgupta N, Beletsky L, Ciccarone D. Opioid crisis: no easy fix to its social and economic determinants. Am J Public Health. 2018;108(2):182186. LinkGoogle Scholar
15. Webster F, Connoy L, Sud A, Pinto AD, Katz J. Grappling with chronic pain and poverty during the COVID-19 pandemic. Can J Pain. 2020;4(1):125128. Crossref, MedlineGoogle Scholar
16. Orkin AM, Bharmal A, Cram J, Kouyoumdjian FG, Pinto AD, Upshur R. Clinical population medicine: integrating clinical medicine and population health in practice. Ann Fam Med. 2017;15(5):405409. Crossref, MedlineGoogle Scholar
17. Furlan AD, Carnide N, Irvin E, et al. A systematic review of strategies to improve appropriate use of opioids and to reduce opioid use disorder and deaths from prescription opioids. Can J Pain. 2018;2(1):218235. Crossref, MedlineGoogle Scholar
18. Alford DP. Opioid prescribing for chronic pain—achieving the right balance through education. N Engl J Med. 2016;374(4):301303. Crossref, MedlineGoogle Scholar
19. Webster F, Rice K, Sud A. A critical content analysis of media reporting on opioids: the social construction of an epidemic. Soc Sci Med. 2020;244:112642. Crossref, MedlineGoogle Scholar
20. Kahn N, Chappell K, Regnier K, Travlos DV, Auth D. A collaboration between government and the continuing education community tackles the opioid crisis: lessons learned and future opportunities. J Contin Educ Health Prof. 2019;39(1):5863. Crossref, MedlineGoogle Scholar
21. Webster F, Bremner S, Oosenbrug E, Durant S, McCartney CJ, Katz J. From opiophobia to overprescribing: a critical scoping review of medical education training for chronic pain. Pain Med. 2017;18(8):14671475. Crossref, MedlineGoogle Scholar
22. Van Zee A. The promotion and marketing of oxycontin: commercial triumph, public health tragedy. Am J Public Health. 2009;99(2):221227. LinkGoogle Scholar
23. Spithoff S. Industry involvement in continuing medical education: time to say no. Can Fam Physician. 2014;60(8):694696, 700–703. MedlineGoogle Scholar
24. Sud A, Molska GR, Buentello-Salamanca F. Evaluations of continuing health provider education for opioid prescribing: a systematic scoping review. Acad Med. 2021; Epub ahead of print. Google Scholar
25. Cochella S, Bateman K. Provider detailing: an intervention to decrease prescription opioid deaths in Utah. Pain Med. 2011;12(suppl 2):S73S76. Crossref, MedlineGoogle Scholar
26. Morton KJ, Harrand B, Floyd CC, et al. Pharmacy-based statewide naloxone distribution: a novel “top-down, bottom-up” approach. J Am Pharm Assoc. 2017;57(2 suppl):S99S106.e5. CrossrefGoogle Scholar
27. Katzman JG, Comerci GD, Landen M, et al. Rules and values: a coordinated regulatory and educational approach to the public health crises of chronic pain and addiction. Am J Public Health. 2014;104(8):13561362. LinkGoogle Scholar
28. Kawamoto K, Houlihan CA, Balas EA, Lobach DF. Improving clinical practice using clinical decision support systems: a systematic review of trials to identify features critical to success. BMJ. 2005;330(7494):765. Crossref, MedlineGoogle Scholar
29. Sim I, Gorman P, Greenes RA, et al. Clinical decision support systems for the practice of evidence-based medicine. J Am Med Inform Assoc. 2001;8(6):527534. Crossref, MedlineGoogle Scholar
30. Haynes RB, Wilczynski NL, Computerized Clinical Decision Support System Systematic Review Team. Effects of computerized clinical decision support systems on practitioner performance and patient outcomes: methods of a decision-maker‒researcher partnership systematic review. Implement Sci. 2010;5(1):12. Crossref, MedlineGoogle Scholar
31. Spithoff S, Mathieson S, Sullivan F, et al. Clinical decision support systems for opioid prescribing for chronic non-cancer pain in primary care: a scoping review. J Am Board Fam Med. 2020;33(4):529540. Crossref, MedlineGoogle Scholar
32. Moride Y, Lemieux-Uresandi D, Castillon G, et al. A systematic review of interventions and programs targeting appropriate prescribing of opioids. Pain Physician. 2019;22(3):229240. Crossref, MedlineGoogle Scholar
33. Craig P, Dieppe P, Macintyre S, et al. Developing and evaluating complex interventions: the new Medical Research Council guidance. BMJ. 2008;337:a1655. Crossref, MedlineGoogle Scholar
34. Pawson R, Greenhalgh T, Harvey G, Walshe K. Realist synthesis: an introduction. Manchester, England: ESRC Research Methods Programme, University of Manchester; 2004:155. Google Scholar
35. Sridharan S, Nakaima A. Ten steps to making evaluation matter. Eval Program Plann. 2011;34(2):135146. Crossref, MedlineGoogle Scholar
36. Sud A, Doukas K, Hodgson K, et al. A retrospective quantitative implementation evaluation of Safer Opioid Prescribing, a Canadian continuing education program. BMC Med Educ. 2021;21(1):101. Crossref, MedlineGoogle Scholar
37. Gielen AC, McDonald EM, Gary TL, Bone LR. Using the PRECEDE‒PROCEED model to apply health behavior theories. In: Health Behavior and Health Education: Theory, Research, and Practice. San Francisco, CA: Jossey-Bass; 2008;4:407‒429. Google Scholar
38. Green LW, Kreuter MW. Health Promotion Planning: An Educational and Ecological Approach. Mountain View, CA: Mayfield Publishing Company; 1999. Google Scholar
39. Moore DEJr, Green JS, Gallis HA. Achieving desired results and improved outcomes: integrating planning and assessment throughout learning activities. J Contin Educ Health Prof. 2009;29(1):115. Crossref, MedlineGoogle Scholar
40. Moore DEJr, Chappell K, Sherman L, Vinayaga-Pavan M. A conceptual framework for planning and assessing learning in continuing education activities designed for clinicians in one profession and/or clinical teams. Med Teach. 2018;40(9):904913. Crossref, MedlineGoogle Scholar
41. Astbury B, Leeuw FL. Unpacking black boxes: mechanisms and theory building in evaluation. Am J Eval. 2010;31(3):363381. CrossrefGoogle Scholar
42. Haji F, Morin M-P, Parker K. Rethinking programme evaluation in health professions education: beyond “did it work?” Med Educ. 2013;47(4):342351. Crossref, MedlineGoogle Scholar
43. Damschroder LJ, Aron DC, Keith RE, Kirsh SR, Alexander JA, Lowery JC. Fostering implementation of health services research findings into practice: a consolidated framework for advancing implementation science. Implement Sci. 2009;4(1):50. Crossref, MedlineGoogle Scholar
44. Barth KS, Ball S, Adams RS, et al. Development and feasibility of an academic detailing intervention to improve prescription drug monitoring program use among physicians. J Contin Educ Health Prof. 2017;37(2):98105. Crossref, MedlineGoogle Scholar
45. Kerns RD, Philip EJ, Lee AW, Rosenberger PH. Implementation of the Veterans Health Administration national pain management strategy. Transl Behav Med. 2011;1(4):635643. Crossref, MedlineGoogle Scholar
46. Curran GM, Bauer M, Mittman B, Pyne JM, Stetler C. Effectiveness‒implementation hybrid designs: combining elements of clinical effectiveness and implementation research to enhance public health impact. Med Care. 2012;50(3):217226. Crossref, MedlineGoogle Scholar
47. Sud A, Nelson MLA, Cheng DK, et al. Sahaj Samadhi Meditation versus a health enhancement program for depression in chronic pain: protocol for a randomized controlled trial and implementation evaluation. Trials. 2020;21(1):319. Crossref, MedlineGoogle Scholar
48. Rose GA, Khaw K-T, Marmot M. Rose’s Strategy of Preventive Medicine: The Complete Original Text. Oxford, England: Oxford University Press; 2008. CrossrefGoogle Scholar
49. Dowell D, Haegerich TM, Chou R. CDC guideline for prescribing opioids for chronic pain—United States, 2016. MMWR Recomm Rep. 2016;65(1):149. Crossref, MedlineGoogle Scholar
50. Busse JW, Craigie S, Juurlink DN, et al. Guideline for opioid therapy and chronic noncancer pain. CMAJ. 2017;189(18):E659E666. Crossref, MedlineGoogle Scholar
51. Dunn KM, Saunders KW, Rutter CM, et al. Opioid prescriptions for chronic pain and overdose: a cohort study. Ann Intern Med. 2010;152(2):8592. Crossref, MedlineGoogle Scholar
52. Gomes T, Mamdani MM, Dhalla IA, Paterson JM, Juurlink DN. Opioid dose and drug-related mortality in patients with nonmalignant pain. Arch Intern Med. 2011;171(7):686691. Crossref, MedlineGoogle Scholar
53. Frank JW, Bohnert ASB, Sandbrink F, McGuire M, Drexler K. Implementation and policy recommendations from the VHA State-of-the-Art Conference on Strategies to Improve Opioid Safety. J Gen Intern Med. 2020;35(suppl 3):983987. Crossref, MedlineGoogle Scholar
54. Flor H, Fydrich T, Turk DC. Efficacy of multidisciplinary pain treatment centers: a meta-analytic review. Pain. 1992;49(2):221230. Crossref, MedlineGoogle Scholar
55. Jeffery MM, Butler M, Stark A, Kane RL. Multidisciplinary Pain Programs for Chronic Noncancer Pain. Rockville, MD: Agency for Healthcare Research and Quality; 2011. Google Scholar
56. Choinière M, Peng P, Gilron I, et al. Accessing care in multidisciplinary pain treatment facilities continues to be a challenge in Canada. Reg Anesth Pain Med. 2020;45(12):943948. Crossref, MedlineGoogle Scholar
57. Ontario Newsroom. Ontario taking action to prevent opioid abuse. Available at: Accessed February 19, 2021. Google Scholar
58. Pawson R, Greenhalgh T, Harvey G, Walshe K. Realist review—a new method of systematic review designed for complex policy interventions. J Health Serv Res Policy. 2005;10(suppl 1):2134. Crossref, MedlineGoogle Scholar
59. Sud A, Armas A, Cunningham H, et al. Multidisciplinary care for opioid dose reduction in patients with chronic non-cancer pain: a systematic realist review. PLoS One. 2020;15(7):e0236419. Crossref, MedlineGoogle Scholar
60. Moore GF, Evans RE, Hawkins J, et al. From complex social interventions to interventions in complex social systems: future directions and unresolved questions for intervention development and evaluation. Evaluation. 2019;25(1):2345. Crossref, MedlineGoogle Scholar
61. Greenhalgh T, Papoutsi C. Studying complexity in health services research: desperately seeking an overdue paradigm shift. BMC Med. 2018;16(1):95. Crossref, MedlineGoogle Scholar
62. Upshur RE. Seven characteristics of medical evidence. J Eval Clin Pract. 2000;6(2):9397. Crossref, MedlineGoogle Scholar
63. Sandhu HK, Abraham C, Alleyne S, et al. Testing a support programme for opioid reduction for people with chronic non-malignant pain: the I-WOTCH randomised controlled trial protocol. BMJ Open. 2019;9(8):e028937. MedlineGoogle Scholar
64. Bensimon CM, Upshur REG. Evidence and effectiveness in decisionmaking for quarantine. Am J Public Health. 2007;97(suppl 1):S44S48. LinkGoogle Scholar
65. Väyrynen P. Thick ethical concepts. In: Zalta EN, ed. The Stanford Encyclopedia of Philosophy. Metaphysics Research Lab, Stanford University. 2021. Available at: Accessed June 19, 2021. Google Scholar
66. Kelly MP, Heath I, Howick J, Greenhalgh T. The importance of values in evidence-based medicine. BMC Med Ethics. 2015;16(1):69. Crossref, MedlineGoogle Scholar
67. Upshur R, Goldenberg MJ. Countering medical nihilism by reconnecting facts and values. Stud Hist Philos Sci. 2020;84:7583. Crossref, MedlineGoogle Scholar
68. Oliva EM, Bowe T, Manhapra A, et al. Associations between stopping prescriptions for opioids, length of opioid treatment, and overdose or suicide deaths in US veterans: observational evaluation. BMJ. 2020;368:m283.; Crossref, MedlineGoogle Scholar
69. James JR, Scott JM, Klein JW, et al. Mortality after discontinuation of primary care‒based chronic opioid therapy for pain: a retrospective cohort study. J Gen Intern Med. 2019;34(12):27492755. Crossref, MedlineGoogle Scholar
70. Juurlink DN. Rethinking “doing well” on chronic opioid therapy. CMAJ. 2017;189(39):E1222E1223. Crossref, MedlineGoogle Scholar
71. Buchman DZ, Ho A, Illes J. You present like a drug addict: patient and clinician perspectives on trust and trustworthiness in chronic pain management. Pain Med. 2016;17(8):13941406. Crossref, MedlineGoogle Scholar
72. Kofi LS, Sud A, Buchman DZ. Ethics at the intersection of chronic pain and substance use. In: Buchman DZ, Davis K, eds. Developments in Neuroethics and Bioethics. San Diego, CA: Elsevier Science and Technology; 2018:127150. Google Scholar
73. Kertesz SG, Manhapra A, Gordon AJ. Nonconsensual dose reduction mandates are not justified clinically or ethically: an analysis. J Law Med Ethics. 2020;48(2):259267. Crossref, MedlineGoogle Scholar
74. Rieder TN. Is nonconsensual tapering of high-dose opioid therapy justifiable? AMA J Ethics. 2020;22(1):E651E657. MedlineGoogle Scholar
75. Tyndall M, Dodd Z. How structural violence, prohibition, and stigma have paralyzed North American responses to opioid overdose. AMA J Ethics. 2020;22(1):E723E728. MedlineGoogle Scholar
76. Stonington S, Coffa D. Structural iatrogenesis—a 43-year-old man with “opioid misuse.” N Engl J Med. 2019;380(8):701704. Crossref, MedlineGoogle Scholar
77. Fiscella K, Wakeman SE, Beletsky L. Buprenorphine deregulation and mainstreaming treatment for opioid use disorder: X the X waiver. JAMA Psychiatry. 2019;76(3):229230. Crossref, MedlineGoogle Scholar
78. Buchman DZ, Ho A. What’s trust got to do with it? Revisiting opioid contracts. J Med Ethics. 2014;40(10):673677. Crossref, MedlineGoogle Scholar
79. Levesque J-F, Breton M, Senn N, Levesque P, Bergeron P, Roy DA. The interaction of public health and primary care: functional roles and organizational models that bridge individual and population perspectives. Public Health Rev. 2013;35(1):14. CrossrefGoogle Scholar
80. Baylis F, Kenny NP, Sherwin S. A relational account of public health ethics. Public Health Ethics. 2008;1(3):196209. CrossrefGoogle Scholar
81. Valentine K. Complex needs and wicked problems: how social disadvantage became multiple. Soc Policy Soc. 2016;15(2):237249. CrossrefGoogle Scholar
82. Valles SA. Philosophy of Population Health: Philosophy for a New Public Health Era. Milton Park, England: Routledge; 2018. CrossrefGoogle Scholar
83. Carel H, Kidd IJ. Epistemic injustice in healthcare: a philosophical analysis. Med Health Care Philos. 2014;17(4):529540. Crossref, MedlineGoogle Scholar


No related items




Abhimanyu Sud, MD, CCFP , Daniel Z. Buchman, PhD, MSW , Andrea D. Furlan, MD, PhD , Peter Selby, MBBS, CCFP, MHSc, Dip ABAM , Sheryl M. Spithoff, MD, MSc , and Ross E. G. Upshur, MD, MA, MSc, CCFP Abhimanyu Sud is with the Department of Family and Community Medicine and Institute for Health Policy, Management, and Evaluation, University of Toronto, Canada. Daniel Z. Buchman is with Centre for Addiction and Mental Health and Dalla Lana School of Public Health, Toronto. Andrea D. Furlan is with the Institute for Work and Health and Department of Medicine, University of Toronto. Peter Selby is with the Centre for Addiction and Mental Health and Department of Family and Community Medicine, Department of Psychiatry, University of Toronto. Sheryl M. Spithoff is with Department of Family and Community Medicine, University of Toronto. Ross E. G. Upshur is with the Bridgepoint Collaboratory for Research and Innovation, Department of Family and Community Medicine, Dalla Lana School of Public Health, University of Toronto. “Chronic Pain and Opioid Prescribing: Three Ways for Navigating Complexity at the Clinical‒Population Health Interface”, American Journal of Public Health 112, no. S1 (February 1, 2022): pp. S56-S65.

PMID: 35143271