Shared Decision-Making and Discontinuation of Opioid Therapy for Chronic Pain

Discussion

Our findings of less frequent discontinuation of opioid therapy at 3 months and more frequent opioid prescribing over 12 months in the greater SDM group, coupled with no apparent benefits from opioid therapy, support the view that SDM is a necessary but insufficient step to improve opioid prescribing for patients with chronic pain.²⁴ To promote better opioid prescribing decisions, the National Institute for Health and Care Excellence (NICE) in the United Kingdom issued guidelines that more specifically advised physicians on dealing with patients already using opioids for chronic pain. In this setting, the guidelines recommended that physicians describe the lack of evidence for continued opioid use in general, explain the individual patient risks of continuing to use opioids if they report little benefit or significant harm, and encourage and support them to reduce opioid dosage or discontinue opioid use if possible.²⁵ Subsequently, NICE issued a more comprehensive guideline for safe prescribing and withdrawal management for patients using opioids for chronic pain. During SDM with a given patient, the guideline recommends consideration of issues such as their changing opioid benefit-to-risk ratio over time, avoidance of abrupt discontinuation of opioids, dose reduction and taper speed, and risks potentially associated with opioid withdrawal and available support options.²⁶ Relatedly, if a physician believes that continued opioid therapy is not in the patient’s best interest but SDM cannot resolve the issue, the General Medical Council in the United Kingdom recommends that the physician not prescribe the opioid, explain the reasons for this decision, document all discussions carefully and provide a copy to the patient, and offer a second opinion.²⁷

There are complex challenges in engaging in SDM during medical encounters for chronic pain management when long-term opioid therapy is being considered.²⁴ These may involve patient issues such as irrational or short-sighted expectations, overdependence on pain intensity in guiding treatment progress, and flawed decision-making owing to long-term opioid use. In response, a model of supplemented SDM has been advocated.²⁴ This may include the participation of spouses, family, and friends, greater focus on physical function and HRQOL in monitoring treatment progress, use of appropriate decision aids, offering evidence-based nonpharmacological treatment options, and clarifying treatment goals through motivational interviewing.

Beyond knowing and wishing to adhere to clinical practice guidelines for prescribing opioid therapy for patients with chronic pain, presenting unbiased opioid risk information to be understood, synthesized, and acted on by patients can be a demanding task for physicians. Prospect theory was originally developed as a model within the field of economics to explain how persons make decisions involving monetary outcomes in the face of risk and uncertainty.²⁸ However, the model may be extended to apply to health care decisions. In this setting, the model posits that patients will choose riskier treatments, such as opioid therapy rather than nonpharmacological or nonopioid alternatives, when faced with the likelihood of ongoing or worsening pain. Alternatively, patients may choose more conservative treatments when physicians use less technical medical terminology, which is perceived to indicate a less severe condition.²⁹ Patients may also be susceptible to availability bias by believing that personal experiences with opioid therapy outweigh evidence, or optimism bias by believing themselves to be less susceptible to opioid harms than others.³⁰

Although the President’s Commission for the Study of Ethical Problems in Medicine and Biomedical and Behavioral Research introduced the concept of SDM in 1982, ethicists worry that its common application is premature in some ways.³¹ In particular, patients and physicians both rate performance on weighing the benefits and risks of treatment options lower than all other aspects of SDM.³² When SDM involves maintaining, tapering, or discontinuing opioid therapy it is particularly difficult, as patients may exhibit compromised decisional capacity owing to dependence, abuse, and diversion.²⁴ Engaging in SDM may also be time-consuming and costly in primary care settings.³³ While conversation aids may encourage and improve the quality of SDM within medical encounters, they compete for time with other tasks such as documentation for reimbursement.³⁴ Decision aids generally have focused on improving care and reducing costs for medical or surgical procedures, rather than for common ongoing care issues.³⁵ Thus, many are concerned that policy initiatives promoting widespread implementation of SDM, such as recommendations involving chronic disease management in millions of patients, carry the risk of unintended consequences if relational aspects of SDM are not maintained over time.³¹ Although informing patients about benefits and risks of opioid therapy and seeking input about their values is necessary,² SDM requires a fundamental shift in the patient-physician relationship in which both parties are considered partners. Less frequent discontinuation of opioid therapy and more frequent opioid prescribing with greater SDM in our study suggest that shared decisions to initiate opioid therapy are more resistant to reversal than unilateral decisions stemming from more paternalistic care. Physicians engaged in greater SDM may be more reluctant to taper or discontinue opioid therapy to avoid the perception of patient abandonment.²

Although SDM may be challenging for young physicians navigating the challenges of acquiring medical experience while also learning to communicate and collaborate with patients, experienced physicians are also often unaware of their deficiencies in SDM by falsely perceiving that they involve patients in decision-making.³² Such physicians may be familiar with guidelines for opioid prescribing among patients with chronic pain but they are not adequately versed in implementing SDM. Consequently, more medical education and better physician training seem warranted to deal with opioid prescribing issues during SDM with patients having chronic pain. Linking SDM curricula with evidence-based medicine is a logical step for educating physicians to implement opioid prescribing guidelines.³⁶ Another strategy involves integrating online SDM courses with existing continuing medical education requirements.³⁷ Two related state-level initiatives are noteworthy. Oregon developed opioid prescribing guidelines to address challenging patient conversations that may arise during SDM, including discussing updated guidelines, compassionate refusal of opioid initiation or dose escalation, and tapering or discontinuing opioid therapy.³⁸ Minnesota has a Shared Decision-Making Collaborative that promotes routine SDM during ongoing CLBP management.³⁹ However, research is needed on the impact of voluntary programs such as those developed in Oregon and Minnesota. Some believe that mandatory physician education may be required to address the crisis of prescription opioid misuse, potentially linking it to medical licensure.⁴⁰

Study strengths included selecting patients to mirror those with CLBP throughout the United States on characteristics such as age, sex, educational level, cigarette smoking, and medical comorbidities⁴¹; a validated measure to assign patients to SDM treatment strategies¹¹; propensity-score matching of SDM groups to control for potential confounders; double blinding of patients and investigators; and follow-up for 12 months using intention-to-treat analysis. Nevertheless, there were study limitations. First, because PRECISION is not a population-based registry, the results may not be generalizable to the United States population. Second, because the patients assigned to treatment strategies had pre-existing relationships with their physicians at baseline, it is unknown if patient ratings of physician SDM may have been influenced by prior opioid prescribing frequency or opioid dosage. It is possible that patients on longstanding opioid therapy at baseline (ie, those less likely to discontinue opioids) may have rated their physicians more highly on SDM than patients who initiated opioid therapy more recently. It is also possible that physicians may have attempted greater SDM with patients having poor adherence to recommended CLBP treatments in the past.⁴² Third, although there were no statistically significant differences between SDM groups in any of 21 variables measured at baseline, standardized between-group differences in sex, prior lumbar spine surgery, and pain self-efficacy exceeded the 0.1 threshold generally accepted for negligible differences.^43,44 Fourth, data were self-reported by patients and not otherwise corroborated, including that opioid therapy may have been prescribed by physicians other than those designated as the usual CLBP providers. Finally, only opioid prescribing frequency was routinely measured, not opioid dosage. Some physicians may prescribe opioids more frequently, but at lower doses, to appease patients.⁴⁵ Nevertheless, an unplanned post hoc analysis using available PRECISION data on a subset of 142 patients reported in a prior study⁴⁶ found no difference in baseline morphine milligram equivalent dosages between those in the greater or lesser SDM groups (mean, 40.7 vs 34.0, respectively; P = .12). Thus, it seems unlikely that physicians treating patients in the greater SDM group prescribed opioids at lower doses, or that differences in baseline opioid dosages had a material effect on opioid tapering or discontinuation over time based on recommended taper speed.⁴⁷