Search results
Forty-two studies were included from the parallel reviews examining effectiveness and AEs of opioid-tapering interventions and barriers and facilitators to tapering of opioids in adults with chronic non-cancer pain. The PRISMA flow chart is reported elsewhere (see Figure 1).
Study characteristics
Included study characteristics are presented in Table 24. Included studies were conducted across 5 countries (31 USA, 6 Australia, 2 Denmark, 2 Germany and 1 UK). Only two studies105,108 reported on outcomes of tapering initiatives delivered exclusively to a disadvantaged PROGRESS-Plus population subgroup (low income) and one study83 specifically aimed to examine the differential impact of opioid tapering on average daily opioid doses of patients at higher risk for opioid-related adverse outcomes compared with lower-risk patients.
| Author, date, country | Study design | Included in effects/AE/barriers and facilitators review | Aim | PROGRESS-Plus inclusion/exclusion criteria | PROGRESS-Plus sample characteristics and author-reported generalisability issues considerations relating to PROGRESS-PLUS sample characteristics (overall figures unless otherwise stated) | Control variables | Inequalities assessed on effects | Inequalities assessed on access | ||
|---|---|---|---|---|---|---|---|---|---|---|
| Association of PROGRESS-Plus with tapering outcome | Effect of tapering on PROGRESS-Plus outcome | Effect of PROGRESS-Plus on participation | ||||||||
| Austin, 2019,66 USA | Retrospective case review | Effects | To assess the impact of an opioid-tapering curriculum | InclusionAge: ≥ 18 yearsExclusionNo PROGRESS-Plus | N = 707 patientsG: 34.5% male, 65.5% female+A (mean ± SD): 62.5 ± 15.2 years | G, +A | +G, +A | |||
| Benintendi, 2021,93 USA | Qualitative | Barriers and facilitators | To describe the ways in which well-intentioned taper initiatives impacted people living with chronic pain | Inclusion | N = 41 patients | P, R, So, +Co, +St | ||||
| Age: ≥ 18 years | R: 2.44% American Indian/Alaska Native | |||||||||
| Exclusion | 21.95% Black/African American | |||||||||
| Not specified | 65.85% White | |||||||||
| 7.32% Other | ||||||||||
| 2.44% Mixed (> 1 race) | ||||||||||
| G: 44% male, 56% female | ||||||||||
| +A: 31.71% 25–44 years | ||||||||||
| 21.95% 45–54 years | ||||||||||
| 26.83% 55–64 years | ||||||||||
| 19.51% 65 years+ | ||||||||||
| Bienek, 2019,67 Germany | Retrospective cohort | Effects/AEs | To compare the tolerability, efficacy and safety of opioid withdrawal under a fixed dose and individualised dose protocol | InclusionNo PROGRESS-PlusExclusionNot specified | N = 195 patientsG: 57.44 male, 42.56 female+A (mean ± SD): 54 ± 12 years individualisedStarting dose51 ± 13 years fixed starting dose | +Co | ||||
| Capano, 2020,110 USA | Prospective cohort | Effects/AEs | To investigate the impact of full hemp extract cannabidiol on opioid use and quality-of-life indicators among chronic pain patients. | InclusionAge: 30–65 yearsExclusionComorbidities: Any psychotic disorder, any epileptic activity in the last 12 months, incapacitating systemic disorder (cardiac, renal or hepatic) | N = 97 patientsG: 30.07% male, 64.03% female+A (mean): 56.1 years | G | S | |||
| Cunningham, 2016,68 USA | Prospective cohort | Effects | To describe opioid tapering and withdrawal symptoms in fibromyalgia patients on opioids | InclusionNo PROGRESS-PlusExclusionNo PROGRESS-Plus | N = 55 patientsR: 91% ethnicity, CaucasianO: 18% currently workingG: 16% male, 84% femaleE (mean, ± SD): 16.6 ± 2.42 years of education+A (mean, ± SD): 48.6 ± 13.2 years | +Co | ||||
| Firemark, 2021,94 USA | Qualitative | Barriers and facilitators | To identify factors that influence or interfere with referrals by PCPs to a pharmacist-led telephone-based programme to assist patients undergoing opioid tapering | InclusionNo PROGRESS-PlusExclusionNo PROGRESS-Plus | N = 35 providersG: 42.86% male, 57.14% femaleO: For PCPs (N = 20) duration of healthcare system employment (mean, range): 11.9 years, 1.3–26.5 years | So | ||||
| Frank, 2016,96 USA | Qualitative | Barriers and facilitators | To explore patients’ perspectives on opioidtapering | InclusionAge: ≥ 18 yearsR: English speakingExclusionNo PROGRESS-Plus‘This purposive sampling strategy also sought to achieve a diverse sample according to gender and age’. p. 1839 | N = 24 patientsR: 79% White raceG: 46% male, 54% femaleE: 21% high school or general educational development29% some college50% college graduate+A: (mean, ± SD): 52 ± 10 years | So, +St | ||||
| Garland, 2014,69 USAa | RCT | Effects | To evaluate the feasibility of developing a clinical trial comparing acute (pre–post) and longer term (3-month follow-up) efficacy of mindfulness-orientated recovery program MORE with that of a conventional support group (SG) in reducing chronic pain and prescription opioid misuse. | InclusionNo PROGRESS-PlusExclusionNot specified | N = 115 patientsR: 3.48% American Indian18.26% African American65.22% Caucasian3.48% other9.56% no responseO: 25.21% work status full timeG: 32% male, 68% femaleE: 70.43% some collegeS (income level): 23.48% below $20,00023.48% $20,000–39,99998.70% $40,000–59,99997.83% $60,000–79,99996.96% over $80,000025.57% no responseA (mean, ± SD): 49.3 ± 13.68 years MORE Group vs. 47.4 ± 13.56 years Support Group | R, G, E, S, +A, +Co | +Co | G, E, S, +A, +Co | ||
| Gersch, 2021,70 USA | Retrospective matched cohortstudy | Effects | To assess the effectiveness of a PEP, a multidisciplinary tele mentoring service based on the Extension for Community Healthcare Outcomes (ECHO) model to reduce opioid use in the outpatient setting | Inclusion+A: 18 years and above and below 80 yearsExclusionNo PROGRESS-Plus | N = 665 patientsIntervention (N = 125) vs. control (N = 540)R: 78.4% vs. 77.8% White raceG: 33.6% vs. 33.1% male, 66.4% vs. 66.9% female+A (mean, ± SD): 58.9 ± 10.8 years vs. 59.3 ± 10.6 years+Co: 4.8% vs. 1.1% alcohol abuse28% vs. 23.2’ chronic pulmonary disease5.6% vs. 5.6% congestive heart failure22.4% vs. 17.8% diabetes44.8% vs. 33.3% hypertension8.8% vs. 4.1% liver disease4.0% vs. 4.4% myocardial infarction20% vs. 10.6% renal insufficiencyCharlson Comorbidity Index (mean, ± SD): 1.5 ± 1.6 vs. 1.2 ± 1.7Chronic Disease Score (mean, ± SD): 6.4 ± 4.1 vs. 5.7 ± 4.0 | |||||
| Giannitrapani, 2018,97,98 USA | Qualitative | Barriers and facilitators | To understand providers’ perceptions of barriers to reducing opioid use and improving the use of non-pharmacological pain management therapies for chronic pain | InclusionNo PROGRESS-PlusExclusionNot specified | N = 60 providersNo PROGRESS-Plus characteristics specified | P, S, +Co | ||||
| Henry, 2019a,99 USA | Qualitative | Barriers and facilitators | To characterise patients’ experiences with opioid tapering and identify communication strategies that are likely to foster productive, patient-centred discussions of opioid tapering | InclusionA: 35–85 yearsExclusionNo PROGRESS-Plus | N = 21 patientsR: 10% African American5% Asian/Pacific Islander76% Caucasian5% Native American5% American/Mexican/Indian10% HispanicO: 19% full-time employment5% Out of work57% Not able to work19% RetiredG: 48% male, 52% femaleE: 14% high school or less57% some college19% bachelor’s degree10% master’s degreeS (annual household income): 24% < $40,00048% $40,000–80,00029% > $80,000+A (mean, ± SD): 58.2 ± 8.3 years | O, So, +Co, +St | ||||
| Henry, 2019b,100,101 USA | Qualitative | Barriers and facilitators | To identify patient statements about opioids that indicate potential openness to tapering opioids or trying non-opioid pain treatments | InclusionA: AdultsExclusionR: Spoke a language other than English | N = 86 patients, 49 providersPatients:R: 14% Hispanic65.1% White27.9% Black0% Asian2.3% Native American4.7% multirace/otherO: 12.8% working full time or part time1.2% not working34.9% retired51.2% disabled/unable to workG: 36.1% male, 43.9% femaleE: 17.4% did not finish high school18.6% high school graduate34.9% some college12.8% associate or technical degree16.3% bachelor’s degree or moreS (annual household income): 24.7% ≤ $10,00035.3% $10,001–20,00012.9% $20,001–40,0008.2% $40,001–60,00010.6% $60,001–80,0008.2% > $80,000+A (mean, ± SD): 59.6 ± 10.5 years ≤ $10, 000Providers:R: 2% Hispanic49% White4.1% Black42.9% Asian0% Native American4.1% multirace/otherG: 24.5% male, 75.5% female+A (mean, ± SD): 29.6 ± 3.6 years | |||||
| Hudak, 2021,71 USA | RCT | Effects | To test whether participation in MORE vs. supportive group (SG) psychotherapy would occasion increased alpha and theta power as well as frontal midline theta (FMT) coherence during a laboratory-based mindfulness-meditation session. | InclusionNo PROGRESS-PlusExclusionNo PROGRESS-Plus | N = 62 patientsR: 3.23% African American4.84% Hispanic/Latino82.26% White4.84% Native American/American Indian3.23% otherG: 85.5% male, 14.5% female | |||||
| Huffman, 2017,72 USAb | Retrospective cohort | Effects/barriers and facilitators | To examine outcomes of patients weaned from opioids in an interdisciplinary chronic pain rehabilitation programme | InclusionNo PROGRESS-PlusExclusionNot specified | N = 941 patientsG: 36.68% male, 61.32% femaleSo (marital status):57.49% married20.5% single11.8% divorced7.86% other+A: (mean, ± SD): 59.3 ± 9.9 years | G, So, +A | ||||
| Jacobs, 2015,73 USA | Prospective cohort pilot study | Effects/AEs | To describe the development and implementation of the Chronic Opioid Assessment Program (COAP), and provide an analysis of initial findings and the impact of COAP on opioid prescribing practices. | InclusionNo PROGRESS-PlusExclusionNo PROGRESS-Plus | N = 148 patientsG: 98% male, 2% female+A (mean): 64 years+Co: 33.8% depression19.6% substance use disorder32% post-traumatic stress disorder (PTSD) | |||||
| Jackson, 2021,111 USA | Pilot RCT | Effects/AEs | To integrate acupuncture within the standard of care during outpatient opioid tapering and assess impact of this treatment on cumulative withdrawal symptoms, psychological distress, and pain. | InclusionG: males or femalesR: English speaking+A; 18–65 yearsExclusionSignificant psychological disease requiring ongoing treatment | N = 15 patientsR: 93.3% White6.7% Black0% Hispanic100% not HispanicG: 53% male, 47% female+A (mean, ± SD): 56.5 ± 17.3 years | +Co | ||||
| Kennedy, 2018,102 USA | Qualitative | Barriers and facilitators | To explore PCPs’ experiences discussing and implementing opioid tapering with patients on long-term opioid therapy | InclusionNo PROGRESS-PlusExclusionNot specified | N = 40 providersR: 70% non-Hispanic White3% non-Hispanic Black10% Hispanic18% otherG: 53% male, 47% female+A: (mean, ± SD): 44 ± 20 years | So, +Co | ||||
| Krumova, 2013,74 Germanyc | Prospective cohort | Effects | To evaluate the relation between the pain intensity shortly after OW and the probability for long-term opioid non-use. | InclusionNo PROGRESS-PlusExclusionPsychiatric diseases with increased risk of withdrawal complication | N = 102 patientsG: 53.9% male, 46.1% femaleSo (marital status): 55.9% married+A (mean, ± SD): 51 ± 13 years | +Co | G | |||
| Kuntz, 2021,95 USA | Mixed methods | Barriers and facilitators | To identify factors from patients and pharmacists, that were cited as hindering or contributing to successful opioid tapering | InclusionA: ≥ 21 yearsExclusionNo PROGRESS-Plus | N = 25 patients, 12 providersPatientsR: 91% non-Hispanic WhiteG: 37% male, 63% femaleS (Neighbourhood Deprivation Index):25% quintile 1 (lowest)25% quintile 225% quintile 325% quintile 4 (highest)So:+A: 24% 21–49 years50% 50–65 years26% > 65 years+Co:Charlson Comorbidity Index:41% 023% 114% 223% ≥ 311% substance use disorder53% depression34% anxiety6% PTSD29% fibromyalgiaChronic pain diagnoses, mean (SD): 3.8 (1.7)ProvidersG: 16.7% male, 83.3% female | R, G, S, +A, +Co | O, S, So, +Co, +St | |||
| Kurita, 2018,75 Denmarkd | RCT | Effects | To evaluate the efficacy of stabilising opioid therapy followed by a sequential opioid tapering off programme in chronic non-cancer pain patients | InclusionE: At least 7 years of schoolingA: ≥ 18 yearsExclusionR: Not fluent in Danish language+Co: Poor general health condition, dementia, encephalopathy, braindamage, cranial base trauma, renal or hepatic failure, other metabolic disturbances (medical diagnose) | N = 35 patientsO: 22.9% workingG: 40% male, 60% femaleE (mean, ± SD): 11.5 ± 3.6 years education in yearsS (income in DKK): 11.8% < 200,00064.7% 200,000–499,00023.5% ≥ 500,000So: 68.6% co-habitation+A (mean, ± SD): 53.0 ± 12.6 years | +Co | G, +A, +Co | , | ||
| Laigaard, 2020,76 Denmark | Prospective cohort | Effects | To investigate cognitive function and health-related quality of life in patients with chronic non-cancer pain during opioid reduction | InclusionA: > 18 yearsExclusionE: Patients not able to speak and understand Danish | N = 51 patientsG: 33.3% male, 66.7% femaleO (any employment): 22%E (highest education):35% primary/secondary school49% vocational16% universitySo (living alone): 41%A (mean, ± SD): 55.9 ± 13 years | Co | O, G, E, So, +A | |||
| Langford, 2020,103 Australia | Qualitative | Barriers and facilitators | To explore the perspectives of healthcare professional stakeholders on the challenges associated with opioid deprescribing and factors to be considered in the development of opioiddeprescribing guidelines | InclusionNo PROGRESS-PlusExclusionNot specified | N = 31 providersG: 48.4% male, 51.6% femaleO (profession):4 GPs1 anaesthetist1 general physician3 geriatricians1 addiction specialist1 rheumatologist3 clinical pharmacologists14 clinical pharmacists3 registered nurses | P, S, +Co, +St | ||||
| McNeilage, 2021,106 Australia | Qualitative (longitudinal study) | Barriers and facilitators | To identify patterns in individuals’ experience of tapering and explore whether patient characteristics, readiness to taper, opioid-tapering self-efficacy, or psychosocial context were related to tapering trajectory | InclusionNo PROGRESS-PlusExclusionA: < 18 yearsR: Insufficient English+Co: major psychiatric conditions | N = 21 patientsP (geographic location):10 cities, 11 ruralsO (employment status):4 working17 not workingG: 10 male, 11 femaleE (educational level):6 vocational training3 high school7 bachelor’s degree1 master’s degree1 PhD3 NRS (Neighbourhood Deprivation Index):So (in a relationship):13 yes8 no+A (mean, range): 55 years, 29–83 years+St (opioid therapy duration, range): 4 months to 25 years | P, O, E, So, +A | P,O, S, So, +Co, +St | |||
| Magee, 2021,104 Australia | Mixed methods (qualitative data only) | Barriers and facilitators | To explore patients’ access to and use of mobile technology, interest in mHealth support for opioid tapering, preferences for the form and content of mHealth support, and potential barriers to and facilitators of engagement with mHealth support for opioid tapering | InclusionA: > 18 yearsExclusionR: insufficient English+CoM: Major psychiatric conditions, comorbid opioid use disorder | N = 21 patientsO (employment status):19% employed78% unemployed or not working5% NR/data missingG: 48% male, 52% femaleE: 14% high school graduate29% vocational training33% bachelor’s degree10% postgraduate degree14% NR/data missingS (Neighbourhood Deprivation Index):So (relationship status):10% single48% married14% in a relationship10% widowed19% NR/data missing+A (mean, ± SD): 55 ± 12.26 years+St duration of opioid treatment, years [mean (SD)]: 9.3 ± 7.5 years | P, S, So, +Co | ||||
| Matthais, 2017,105 USA | Qualitative | Barriers and facilitators | To understand communication processes related to opioid tapering, to identify best practices and opportunities for improvement. | InclusionA: ≥ 18 yearsR: Speaks EnglishExclusionNot specified | N = 37 patients, 9 providersPatientsR: 41% White46% Black2.5% American Indian8% other2.5% don’t know/refusedO (employment):16% employed8% not employed19% retired54% unable to work3% other/refusedG: 32% male, 25% femaleE: 38% some high school or less30% high school/GED27% some college/technical or business school5% college degree or aboveS (income): 11% comfortable43% just enough43% not enough3% refused+A (mean): 58 yearsProvidersR: 78% White11% Black1% AsianO (employment):100% employedE: 89% MD11% physician assistant+A (mean): 45 years | So, Co | ||||
| Montgomery, 2020,77 USA | Retrospective cohort | Effects | To analyse the effectiveness of battlefield acupuncture to decrease chronic pain immediately and 6 months after treatment and to decrease the number of opioids needed to manage chronic pain | InclusionNo PROGRESS-PlusExclusionNo PROGRESS-Plus | N = 47 patientsIntervention vs. controlR: 96% vs. 96% White race/ethnicityG: 83.3% vs. 95.6% male, 16.7% vs. 0.04% femaleS (Neighbourhood Deprivation Index):+A (mean, ± SD): 61 ± 8.8 years vs. 65 ± 8.5 years | |||||
| Murphy, 2013,78 USAe | Retrospective cohort | Effects | To evaluate the associations between opioid cessation and subsequent multidomain treatment outcomes | InclusionNo PROGRESS-PlusExclusion+Co-M: Current medical or psychiatric status precluded them from full benefit and engagement | N = 221 patientsO (employment):13.6% full time4.1% part time7.3% unemployed75.1% disabled/retiredG: 82.4% male, 17.6% femaleR: 66.1% White18.1% African American9.5% Hispanic6.4% otherE (mean, ± SD): 13.86 ± 2.45So (marital status):55.2% married12.2% never married28.1 divorced or separated2.7% cohabiting not married1.8% widowed+ A (mean, ± SD): 49.08 ± 10.97 years | |||||
| Panicker, 2022,79 USA | A retrospective quality improvement (QI) project | Effects | To evaluate chronic non-cancer pain management of veterans using an advanced practice registered nurse (APRN)-led multidisciplinary team approach to incorporate non-opioid and non-pharmacological modalities to affect self-reported pain and use of prescribed opioids | Inclusion+A: > 19 yearsExclusionNot specified | N = 34 patientsG: 91.8% male, 8.2% femaleO (employment status):36.4% retired27.3% disability18.9% unemployed18.9% employedSo (marital status):52.9% married+ A (mean, ± SD): 63.18 ± 5.39 years+St: 21% receiving prescription opioids < 12 months and similarly for patients receiving opioids 1–5 years.32% were prescribed opioids for 5–10 years26% prescribed beyond 10 years+Co: Of the 34 veterans' only 8 did not have a mental health diagnosis. Depression and PTSD were found equally at 35% and anxiety among 1%. | |||||
| Quinlan, 2021,107 UK | Mixed methods (qualitative and quantitative data) | Barriers and facilitators | To investigate the extent of the psychological comorbidity of patients with chronic pain and long-term opioid use and their hopes and fears before embarking on opioid tapering | InclusionNo PROGRESS-PlusExclusionNot specified | N = 49 patientsG: 48% male, 52% femaleO: few were in employment+A (mean, range): 52.5 years, 23–76 years+Co: Generalised Anxiety Disorder (mean, median, IQR): 12.41–13.50, 8.00–17.00+St: many having been on these doses for over 10 years. | So, +Co, +St | ||||
| Rivich, 2018,80 USA | Retrospective cohort | Effects | To identify the impact of OSI chart reviews on total daily opioid dose, prescribing practices of opioids and benzodiazepines, and adherence to VA/VISN policy and monitoring parameters. | InclusionNo PROGRESS-PlusExclusionNo PROGRESS-Plus | N = 147 patientsG: 90% male, 10% female+A (median, range): 61 years, 33–84 years | |||||
| Seal, 2020,81 USAf | PMC | Effects | To evaluate the effectiveness of an Integrated Pain Team (IPT) clinic in decreasing opioid dose and mitigating opioid risk. | Inclusion (patients matched on)G: male or female+A: < 55 or ≥ 55+Co: number and type of psychiatric diagnoses (PTSD, depression, bipolar disorder, and/or other psychiatric diagnosis)ExclusionNo PROGRESS-Plus | N = 294 patientsIntervention vs. controlR: 6.8% vs. 6.8% Hispanic or Latino81% vs. 82.3% not Hispanic or Latino59.9% vs. 70.1% White17% vs. 6.8% Black or African American6.8% vs. 4.1% American Indian or Alaska Native1.4% vs. 1.4% Native Hawaiian or other Pacific Islander, or Asian multiraceG: 89.8% vs. 89.8% male, 10.2% vs. 10.2% femaleSo (marital status): 37.4% vs. 44.2% married35.4% vs. 32.7% divorced or separated21.1% vs. 18.4% never married4.8% vs. 4.1% widowed/widow/widower+A (mean, ± SD): 62.1 ± 12.4 years vs. 62.9 ± 11.4 years+CoM: psychiatric and substance use disorders:34% vs. 38.1% depressive disorder | |||||
| 26.5% vs. 25.9% PTSD12.2% vs. 9.5% alcohol use disorder17.0% vs. 6.1% opioid use disorder7.5% vs. 2.7% other drug use disorderChronic medical conditions:12.9% vs. 12.2% sleep apnoea22.5% vs. 15.7% chronic pulmonary disease2.7% vs. 1.4% cognitive impairment8.8% vs. 6.1% chronic kidney disease10.2% vs. 8.8% hepatic disease+St (duration of VA opioid therapy (years): 14.1 (5.3) vs. 13.6 (5.1) | ||||||||||
| Sharp, 2018,82 USAd | Retrospective cohort | Effects | Whether reducing opioid prescriptions for patients without cancer to safer levels is associated with low satisfaction scores among patients with high opioid use in a real-world managed care setting. | InclusionNo PROGRESS-PlusExclusion+A: younger than 18 years | N = 2492 patientsR:72% White9.3% African American15.1% Hispanic2.2% Asian/PI or Native American1.4% unknownG: 43.2% male, 56.8% female+A: 54.5% < 65 years45.6% ≥ 65 years+Co (Elixhauser score):70.26% < 329.7% ≥ 3 | R, G, +A, +Co | ||||
| Sullivan, 2017,113,114 USA | RCT | Effects/AEs | Does a prescription opioid taper support intervention for patients receiving moderate or higher dose long-term opioid therapy for chronic non-cancer pain work? | InclusionNo PROGRESS-PlusExclusion+Co: considered medically unstable (as judged by the referring physician); report of suicide attempt or psychiatric hospitalisation in the past 10 years or current suicidal ideation with specific plan or intent; report of psychotic symptoms on the modified Mini International Neuropsychiatric Interview; report of current abuse of substances other than nicotine or marijuana according to the National Institute on Drug Abuse Alcohol, Smoking, and Substance Involvement Screening Test | N = 35 patientsIntervention vs. controlR: 72.2% vs. 100% white5.6% vs. 0% black11% vs. 5.6% Asian5.6% vs. 0% Native American, Alaska Native, or Pacific Islander5.6% vs. 0% HispanicG: % male, % femaleE: 11.1% vs. 23.5% high school44.4% vs. 47.1% some college27.8% vs. 23.5% college graduate16.7% vs. 5.9% graduate or professional school+Co Score of > 10 on the PHQ-9 (moderate or greater depressive symptomseverity): 61% vs. 53%Overall+A (mean, ± SD): 54.4 ± 10.1 years+St (duration of their current use of opioid medication, mean, ± SD): 10.2 ± 4.3 years | +Co | +Co | |||
| Thakral, 2018,83,84,85,86,87 USAd | Non-patient cohort Interrupted time series | Effects/AEs | To determine if opioid risk reduction initiatives including dose reduction and risk mitigation strategies for chronic non-cancer pain patients receiving COT had a differential impact on average daily opioid doses of COT patients at higher risk for opioid-related adverse outcomes compared with lower-risk patients. | InclusionS: At least 12 months of continuous enrolment in the health insurance plan prior to and including the current quarter+A: ≥ 26 yearsExclusionNo PROGRESS-Plus | N = 23,809 patientsP: 87.6% Western WashingtonG: 36.2% male, 63.8% female+A: 24.3% 26–45 years47.1% 46–64 years28.7% 65 +years+Co:9.9% substance use disorders56.2% mental disordersComorbidity score:57.4% 08.6% 114.8% 219.3% 3 + | G, +A, +Co | ||||
| Townsend, 2008,88 USAg | Prospective cohort | Effects | We test the hypothesis that patients with chronic pain and long-standing opioid use who undergo opioid withdrawal in the course of rehabilitative treatment will experience significant and sustained improvement in pain and functioning similar to patients who were not taking opioids. | InclusionNo PROGRESS-PlusExclusionNo PROGRESS-Plus | N = 213 patientsR: 96.7% whiteO (% working for a wage): 26.3%G: 21.2% male, 79.8% femaleE (mean no. of years ± SD):14.6 ± 2.9 years92.5% completed high schoolSo (marital status): 64.8% married+ A (mean, ± SD): 44.5 ± 14.2 years | P, +Co, | +Co, So | |||
| Twillman, 2018,89 USA | Retrospective cohort | Effects/AEs | Patients with chronic pain receiving long-term opioid therapy were surveyed to assess the incidence and impact of opioid dose reduction following this guideline’s promulgation. | InclusionNo PROGRESS-PlusExclusionNo PROGRESS-Plus | N = 362 patients | O, So, +Co | ||||
| Westanmo, 2015,117 USA | Non-patient cohort before and after | Effects | To describe processes and outcomes of a health system quality improvement initiative designed to reduce opioid-related harms | InclusionNo PROGRESS-PlusExclusionNot specified | N = 6942/50,749 patients (April 2011) and 5981/54,636 (October 2014) patients received opioid prescriptions | |||||
| White, 2020,108 Australia | Mixed methods (qualitative data only) | Barriers and facilitators | To evaluate patientand provider perspectives regarding the opioid-tapering intervention Assess, Inform, Manage and Monitor. | InclusionR: English speaking+A: AdultsExclusion+Co: Serious underlying pathology, physically or mentally unable to complete survey; current abuse of illicit substances; unable to use a telephone due to cognitive or hearing impairmentS: in receipt of workers compensation benefits | N = 6 patients, 4 providersPatientNot reported separately for qualitative dataProviderO: 25% GP25% practice nurse25% dietitian25% pharmacist | So | ||||
| Wu, 2019,109 USA | Mixed methods (case series design, qualitative data only) | Barriers and facilitators | To evaluate the effect of this prescribing policy, with secondary aims to gain insight into the patient experience of undergoing opioid tapering and to generate hypotheses for further study | InclusionR: Speak English+A: older than 18 yearsExclusionNo PROGRESS-Plus | N = 6 patientsR: 2 white2 African American2 HispanicG: 4 males, 2 females+A (range): 53–61 years | So, +St, +Co | ||||
| Zheng, 2008,112 Australia | RCT | Effects/AEs | Does real electroacupuncture reduce the use of opioid-like medications consumption and associated side effects in patients with chronic pain when compared to sham electroacupuncture | Inclusion+A: 18–80 inclusiveExclusionNot specified | N = 35 patientsIntervention vs. controlG: 53% vs. 50% male, 47% vs. 50% female+ A (mean, ± SD): 51.1 ± 13 years vs. 48.4 ± 10.5 years | So | ||||
| Zheng, 2019,115,116 Australia | RCT | Effects/AEs | To assess if electroacupuncture is an effective adjunct therapy to standard pain and medication management in reducing opioids use by patients with chronic musculoskeletal pain. | InclusionR: confident in conversational and reading English+A: aged between 18 and 85 yearsExclusion+Co: suffering from chronic musculoskeletal pain; severely depressed with suicidal tendency as judged by pain specialists, unstable heart condition, brain tumour, current cancer, haemophiliaor wearing cardiac pacemakers | N = 108 patientsIntervention vs. control vs. pain managementG: 41.7% vs. 37.9% vs. 51.6% male, 58.3% vs. 62.1% vs. 48.4% femaleE (education level): 20.8% vs. 27.6% vs. 19.4% University of higher68.8% vs. 55.2% vs. 64.5% ≥ 9 year of formal education4.2% vs. 13.8% vs. 9.7% < 9 years6.3% vs. 3.4% vs. 6.5% missingSo (marital status): 44.7% vs. 57.1% vs. 53.3% partnered55.3% vs. 42.9% vs. 46.7% single | +Co | P, G, E, So, +A, +Co | |||
| 1 vs. 1 vs. 1 missing+A (mean, ± SD): 55.9 ± 11.3 years vs. 54.8 ± 13.7 years vs. 58.4 ± 12.8 years+Co [Beck Depression Inventory, mean (SD)]:11.5% vs. 7.8% vs. 14.7% | ||||||||||
| Zhou, 2017,90 USA | Retrospective cohort | Effects | To explore the success rate and important adjuvant medications in the medication-assisted treatment with temporary use of methadone for opioid discontinuation in patients with prescription opioid use disorder. | Inclusion+A: ≥ 18 years at first clinic visitExclusionNo PROGRESS-Plus | N = 140 patientsR: %O (employment status):16.5% employed83.5% not employedG: 37.9%% male, 62.1%% femaleS (Neighbourhood Deprivation Index):So (marriage status): 49.5% single27.8% married22.7% otherLiving status:33% alone67% with other+ A (mean, ± SD): 46.7 ± 12.7 years+Co (anxiety, depression, or both): 90% yes, 10% no+St (unit: years): 7.7 ± 6.1 | O, G, So, +A, +Co | ||||
Figure 6 displays the PROGRESS-Plus characteristics collected at baseline versus analysed. Thirty-nine (93%)66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,88,90,93,94,95,96,99,100,102,103,104,105,106,107,108,109,110,111,112,113,116 collected baseline data on PROGRESS-Plus characteristics (see Figure 6) and 30 (71%)66,68,69,72,74,75,76,82,83,88,89,90,93,94,95,96,97,99,102,103,104,105,106,107,108,109,110,111,113,116 undertook analysis involving PROGRESS-Plus factors.
FIGURE 6 PROGRESS-Plus characteristics collected at baseline vs. analysed (Review 4). E, education; G, gender; O, occupation; P, place of residence; R, race, ethnicity, language; R, religion; S, social capital; SES, socioeconomic status. (Figure not reproduced here; see the original article.)
In the 30 studies that analysed PROGRESS-Plus factors, the most frequent factor analysed was comorbidity (24 studies), followed by social capital (17 studies) and gender (11 studies). No studies were located that examined the impact of ‘Religion’ on inequalities. Five studies analysed 1 factor, 6 studies analysed 2 and 18 studies analysed 3 or more factors.
Altogether, 17 studies66,68,69,74,75,76,82,83,88,89,90,95,106,110,111,113,116 examined whether opioid-tapering initiatives were more or less effective across disadvantaged populations (see Table 3) and 15 studies93,94,95,96,97,99,102,103,104,105,106,107,108,109,110 reported on whether opioid-tapering initiatives were more or less accessible across disadvantaged populations. The quality of the included studies is reported elsewhere (see Figure 3, Tables 13–15, 22 and 23). In brief, the quality of the 13 studies reporting qualitative data in the barriers and facilitators review was moderate to high and 2 studies had major limitations. The overall quality of the studies included in the effectiveness review was moderate.
The effects of PROGRESS-Plus characteristics were analysed further in three ways: (1) by stratifying effect analyses by different categories of a PROGRESS-Plus characteristic; (2) by exploring associations between PROGRESS-Plus characteristics and changes in opioid use or tapering trajectory following a tapering intervention (including differential associations, e.g. male vs. female; low vs. high SES); and (3) by assessing the impact of opioid-tapering initiatives on PROGRESS-Plus-related outcomes (e.g. ability to work).
Of the 17 studies that examined whether opioid-tapering initiatives were more or less effective across disadvantaged populations, 7 studies66,82,83,90,95,106,110 examined the association of PROGRESS-Plus characteristics with tapering outcomes, of which, 4 studies83,90,95,110 assessed differential associations (e.g. male vs. female; younger vs. older patients). Four studies66,82,90,95 assessed associations of PROGRESS-Plus characteristics with tapering outcomes and 10 studies68,69,74,75,76,88,89,111,113,116 assessed the impact of tapering on PROGRESS-Plus-related outcomes. One study106 assessed descriptive associations with tapering outcomes. Ten studies66,67,69,74,75,76,88,112,113,116 reported on the number of dropouts or reasons for dropout in relation to PROGRESS-Plus characteristics
PROGnosis RESearch Strategy partnership-Plus characteristics were analysed on 8 tapering outcomes: (1) mental health/depression/anxiety (10 studies68,69,74,75,76,88,89,111,113,116); (2) reduction or change in opioid dose/use/withdrawal (3 studies82,83,95); (3) opioid discontinuation (1 study90); (4) quality of life (1 study110); (5) tapering trajectory (1 study106); (6) ability to work (1 study89); (7) relationship with medical professionals/family/friends (1 study89); and (8) remaining on chronic opioid therapy (1 study66).
Impact of PROGRESS-Plus characteristics on inequalities in effectiveness of, and access to, opioid tapering
Factors relating to inequalities in effectiveness of, and access to, opioid tapering were reported across nine PROGRESS-Plus characteristics (see Tables 40–42 in Appendix 8) and related to individual (pertaining to the patient), interpersonal (interaction with provider), organisational and wider environmental contexts. Factors associated with inequalities in effectiveness of opioid tapering operated at an individual level only. Inequalities in access were found that may arise from dimensions of accessibility of services or abilities of healthcare workforce and/or from the abilities of patients to interact with the dimensions of accessibility to generate access (Figures 7 and 8).
FIGURE 7 Inequalities in effectiveness of, and access to, opioid tapering in patients with chronic non-cancer pain according to the SEM (Review 4). (Figure not reproduced here; see the original article.)
FIGURE 8 Conceptual model of healthcare access of tapering opioids in adults with chronic pain (Review 4). (Figure not reproduced here; see the original article.)
Impact of ‘Place of Residence’ on inequalities in opioid tapering
Factors relating to ‘Place of Residence’ that may potentially widen inequalities were reported for both effectiveness of and access to opioid tapering across seven studies.88,93,97,103,104,106,116 The evidence suggested that inequalities in access due to ‘Place of Residence’ may widen as a result of individual, organisational and environmental factors and may affect a patients’ ability to reach an opioid taper. At an individual level, inequalities may widen when a patient lives in a rural area106 or in areas with poor mobile phone reception.104 Organisational factors relating to the way in which opioid tapering is implemented and wider environmental factors may help to explain why these individual factors widen inequalities, for example, when patients are forced to travel long distances to attend face-to-face opioid-tapering appointments,93,97,103 or when tapering support is delivered by mobile phone apps.104 Wider environmental factors external to tapering that may increase inequality in access include a lack of or poor transportation,93 or living in an area prone to natural disaster occurrences.106 In addition, the way in which resources were distributed appeared to increase inequalities when tapering resources or alternative therapies were not available to providers in certain locations.97,102,103 Three studies,97,103,106 highlighted factors relating to ‘Place of Residence’ that may potentially reduce inequalities in access at the organisational level, including the availability of tapering resources and delivering tapering initiatives in home or community settings.
Impact of ‘Race, Ethnicity and Language’ on inequalities in opioid tapering
Three studies69,82,95 reported on factors relating to ‘Race, Ethnicity, Language’ that may potentially widen inequalities on the effectiveness of opioid tapering at the individual level. Two studies82,95 did not find an association between race or ethnicity and opioid reduction. In addition, one RCT69 did not find a significant difference in race among patients who dropped out of the tapering trial compared to those who completed. Interpersonal factors relating to race that may widen inequalities in patients seeking treatment to taper were noted in one study. For example, Benintendi et al.93 found that fear of racial discrimination may compound existing stigma around opioid use, and that this may prevent patients from seeking help to taper. No studies were found that examined the impact of ‘language’ on effectiveness or access to opioid tapering; however, six studies76,96,105,108,109,111 included language as an inclusion/exclusion criteria.
Impact of ‘Occupation’ on inequalities in opioid tapering
‘Occupation’ was considered a factor that could widen inequalities in both the effectiveness of and access to opioid tapering in three studies.89,95,106 Individual-level factors relating to ‘occupation’ that may widen inequalities in tapering effectiveness were found in two studies. For example, McNeilage et al.106 reported a descriptive association between ‘not working (all receiving pension)’ and a distressing tapering trajectory and Twillman et al.89 assessed the impact of tapering on an occupation-related outcome and found that opioid tapering significantly worsened a patient’s ability to work. One study90 did not find a significant association between employment status (yes/no) and opioid discontinuation and there was no significant impact of occupational status found in two further studies69,76 among patients who dropped out or were lost to follow-up compared to those who completed the study. Factors relating to ‘occupation’ that could widen inequalities in access included individual factors relating to work stress or patient beliefs that opioids were needed in order to work.95,99 Wider environmental factors related to busy work schedules.106 These factors may also act as a barrier to access at the point at which the patient perceives a need to taper, or seeks to taper.
Impact of ‘Gender’ on inequalities in opioid tapering
Factors relating to the impact of ‘Gender’ on inequalities were only reported for tapering effectiveness. At an individual level, a widening of inequalities in opioid-tapering effectiveness due to ‘Gender’ was noted in four studies,82,83,95,110 all of which reported significantly worse tapering outcome for males. Of these, one study83 found a lower dose reduction in males compared to females, one study82 found that patients in the dose reduction group were more likely to be female and one study95 found that the odds of a 50% reduction in opioid use at 12 months were decreased among men compared to women. One further study110 found that males reported a worse PHQ-4 score at week 8 post taper compared to females (though not at baseline or 4 weeks). There were no significant associations between ‘Gender’ and tapering outcomes reported in four studies.66,83,90,110 No significant difference was found between males and females with regards to study participation in six studies.66,69,74,75,76,116
Impact of ‘Religion’ on inequalities in opioid tapering
No evidence of impact of ‘Religion’ on either tapering effectiveness or access was found.
Impact of ‘Education’ on inequalities in opioid tapering
One study106 reported an individual-level factor association of ‘Education’ with tapering effectiveness, finding that patients with low levels of education were associated with a ‘surviving’ opioid-tapering trajectory, while patients with a university education were associated with a ‘resilient’ tapering trajectory. Three studies69,76,116 found no significant difference in study participation due to ‘Education’.
Impact of ‘Socioeconomic status’ on inequalities in opioid tapering
Six studies95,97,103,104,106,110 reported on individual-, organisational- or wider environmental-level factors that have the potential to widen inequalities in access to opioid tapering. Individual-level factors across three studies were related to changes in individual financial circumstances, including financial stress,95,106 changes in/loss/alteration of patient insurance plans/policies,110 and having their phone service cancelled.110 Organisational and environmental factors examined in five studies included the cost to the patient to taper97,103,104,106,110 (which in one instance prevented a patient from participating in the study110) and the availability of funding provision for patients to access tapering support.97
Based on the findings of one study,95 there was no evidence of an impact of socioeconomic status on tapering effectiveness, which resulted in a widening of inequalities at an individual level. Another study69 also found no significant difference in socioeconomic status between participants who dropped out of the study compared to those who completed.
Impact of ‘Social capital’ on inequalities in opioid tapering
‘Social capital’ was the most reported factor that may widen or reduce inequalities in access to opioid tapering. With regards to tapering effectiveness, ‘social capital’ was assessed at the individual level in three studies,89,90,106 which reported mixed effects of social capital on tapering effectiveness which may potentially widen or reduce inequalities. One study89 reported a negative impact of tapering on social capital-related outcomes which may widen inequalities, including observing a significant worsening of relationships with medical professionals and friends/family among patients whose dose decreased compared to patients whose dose was unchanged. On the other hand, characteristics relating to social capital had a positive association with a tapering outcome, with one study106 reporting a descriptive association between patients in a relationship and a ‘thriving’ tapering trajectory. One study90 did not find a significant association between marriage or living status and opioid discontinuation. Further, while no significant differences were found in two studies76,116 on dropouts/loss to follow-up due to social status, two other studies88,112 reported family issues as reasons for why patients dropped out of tapering.
Eleven studies88,93,94,95,99,102,105,106,108,109,112 highlighted characteristics relating to both the existence of, or lack of, social support at an individual and interpersonal level that may widen inequalities at different points along a tapering trajectory. At an individual level, one study106 found that a lack of social support from family or friends may widen inequalities if increased feelings of social isolation resulted in a lack of motivation among patients to initiate and sustain a taper. Conversely, two studies95,106 found that patients who had strong family relationships experienced barriers when external family stresses and responsibilities, such as caring for sick or dying relatives/friends or marital problems, prevented them from tapering. On an interpersonal level, eight studies93,94,99,102,105,106,108,109 highlighted that a poor patient–provider relationship may widen inequalities in tapering provision when an unsupportive provider or lack of trust between patients and providers leads to unintended consequences such as an ongoing poor relationship with providers on other treatments and, in some instances, patients changing healthcare providers, or providers discontinuing healthcare provision with patients.102
The existence of ‘social capital’ may also help reduce inequalities in access to opioid tapering on an individual, interpersonal and organisational level. Individual-level factors were examined across seven studies.95,96,99,102,104,106,107 Patients with strong family bonds or family responsibilities were more likely to initiate and sustain a taper than patients who felt isolated. Among these patients, either the patients themselves recognised the impact opioids were having on their relationships or family members raised concerns about their opioid use, thereby motivating them to initiate a taper. At the interpersonal level, across seven studies95,96,99,104,105,106,108 it was found that a trusting patient–provider relationship was helpful in reducing inequalities in access to opioid tapering. Findings from four studies94,95,99,108 also suggested that mutual honesty between patient and provider and regular access to the same provider were key to building a trusting patient–provider relationship. At the organisation level, across two studies,96,104 it was found that patient support forums96,104 and out-of-hours support104 have the potential to reduce inequalities associated with social capital. These findings show that ‘social capital’ has the potential to impact on every stage of the patient’s tapering pathway.
Impact of ‘Age’ on inequalities in opioid tapering
A widening of inequalities in opioid-tapering effectiveness due to ‘Age’ at the individual level was observed in three studies82,83,95 which found significant associations of age in favour of younger patients and one study106 reported a descriptive association between being aged 50 years and older and a ‘resilient’ tapering trajectory. No significant difference on opioid use was found on mean age in one study,90 between younger populations (26–45 vs. 45–64 years) in another,83 or for those ≥ 65 years compared to 50–64 years.95 With regards to inequalities in participation, one study66 found a significant association of younger age with remaining an active patient but no association between age and remaining on chronic opioid therapy. Three further studies69,75,76 found no significant difference in study participation due to ‘Age’.
Impact of ‘Comorbidities’ on inequalities in opioid tapering
Thirteen studies68,69,74,75,76,82,83,88,89,95,111,113,116 assessed the impact of ‘Comorbidities’ and tapering effectiveness at an individual level. Of these, three studies68,88,89 measured the impact of tapering on comorbidity outcomes. Mixed results were noted with one study89 reporting that mental health worsened among patients whose dose decreased compared to patients whose dose was unchanged, whereas two other studies found a statistically significant improvement on depression at 3 weeks post taper68 and immediately following opioid withdrawal and at 6 months post treatment.88 Mixed results were also noted for studies examining an association between ‘Comorbidities’ and tapering outcomes. A differential impact which may lead to a widening of inequalities was reported in one study,95 which found significantly greater odds of a 50% reduction in opioid use at 12 months in patients with no Charlson Comorbidities compared to one. Conversely, three studies82,83,95 reported better tapering outcomes for patients with existing comorbidities, including mental disorders,83 substance use disorders,83,95 anxiety95 and an Elixhauser comorbidity score of ≥ 3.82 There was no association of tapering with comorbidities related to depression and anxiety in seven studies69,74,75,76,111,113,116 and in three studies83,90,95 which assessed an association between comorbidities and tapering outcomes.
With regards to participation, five studies67,69,75,88,113 reported issues in participation relating to comorbidities which may widen inequalities. Two studies69,88 reported dropouts due to ‘comorbidity issues’, and one study67 reported late cessation of opioids due to aggravated depressive symptoms and escalating levels of pain, anxiety and depression.113 In addition, one study75 found that patients who dropped out presented significantly worse scores of anxiety, depression, general health as assessed by SF-36 and a higher risk for opioid misuse as assessed by the Pain Medication Questionnaire (PMQ). Two studies69,116 found no significant differences between participants who dropped out compared to those who completed.
Factors relating to ‘comorbidities’ that could widen inequalities in access were highlighted at the individual, organisational and environmental levels across 11 studies.93,95,97,99,102,103,104,105,106,107,109 Five studies95,99,104,106,107 found that a worsening of a patient’s health (due to factors external to tapering) or emotional status was more likely to lead to a disruption or discontinuation of tapering, resulting in a widening of inequalities relating to access. Evidence from four studies97,102,103,106 indicated that such inequalities at the individual level were further compounded by environmental-level factors relating to availability of comorbidity treatments, including a lack of timely access to alternative therapies for comorbidities or the lack of recognition for an individualised approach to tapering, thus resulting in patients slowing down or stopping a taper. In addition, two studies93,104 indicated that organisational factors relating to intervention design and setting may inadvertently widen inequalities when a patient’s health status (visual impairment and ability to be mobile enough to attend tapering appointments) prevented them from accessing tapering support. At an individual level, three studies99,104,106 found that improvements in a patients’ emotional state were also seen as characteristics enabling a successful taper. Based on findings from across three studies,99,103,105 a flexible, tailored, tapering approach which recognises changes in patient’s health and emotional status may help reduce inequalities relating to ‘Comorbidities’. Four studies97,103,106,109 highlighted environmental factors around increased availability and access to multidisciplinary care (e.g. pain specialists, psychologists) and alternative therapies to help patients manage their health and emotional issues while tapering.
Impact of ‘Status as a long-term opioid user’ on inequalities
Seven studies,93,95,99,103,106,107,109 which assessed the impact of a patient’s status as a long-term opioid user on tapering, found evidence that tapering interventions may widen inequalities in access. Across these studies, stigma associated with ‘status as a long-term opioid user’ was found to arise directly as a result of opioid tapering on a number of levels (individual, interpersonal and environmental), and sometimes occurring across multiple levels simultaneously. On an individual level (relating to self-stigma), one study93 found that some patients felt that their pain was invalidated or disbelieved, resulting in feelings of shame, guilt and humiliation. Findings across seven studies93,95,99,103,106,107,109 suggested that this self-stigma could be perpetuated on an interpersonal level by discriminatory behaviours and negative attitudes of providers resulting in patients fearing, or experiencing, being labelled a ‘drug addict’ by their providers (relating to social stigma). On a wider environmental level, one study93 reported that some patients felt unfairly targeted by wider government policies to reduce opioid use without consideration for their individual circumstances. The same study93 also highlighted that an unintended consequence of the design of opioid-tapering interventions (e.g. monitoring of opioid use, urine toxicology screening) was that they may reinforce stigma about addiction. As a result of stigmatisation, patients may be reluctant to seek help to taper or engage with the tapering process, thus widening inequalities in access. Conversely, the study93 indicated that monitoring of opioid use was also seen as an opportunity to distance oneself from social stigma. In one study,96 a non-judgemental provider was seen as an interpersonal factor noted as having the potential to reduce inequalities in access.
Summary
To our knowledge, this is the first systematic review to synthesise evidence on whether and how inequalities (as defined by PROGRESS-Plus) impact on the effectiveness of, and access to, opioid tapering in adults with chronic non-cancer pain. Forty-two studies included across the three parallel reviews were of varying quality. We found that most studies had collected and reported baseline data with regards to PROGRESS-Plus categories, but less than half considered the impact of them on effectiveness, and just over a third considered them in relation to inequalities in access to opioid tapering. Only two studies105,108 focused specifically on a disadvantaged group in their study population and only one study83 aimed to assess differential effects of opioid tapering albeit, comparing patients at high risk of adverse effects with lower-risk patients. Of those that assessed PROGRESS-Plus characteristics in relation to effectiveness, no studies reported on differential effectiveness of opioid-tapering interventions across the selected PROGRESS-Plus characteristics. Instead, inequalities in effectiveness were explored by: (1) stratifying effect analyses by different categories of a PROGRESS-Plus characteristic; (2) exploring associations between PROGRESS-Plus characteristics and changes in opioid use or tapering trajectory following a tapering intervention; or (3) assessing the impact of opioid-tapering initiatives on PROGRESS-Plus-related outcomes (e.g. ability to work).
Only four studies assessed the differential effects of opioid tapering across PROGRESS-Plus characteristics. Our findings agree with those of Lehne and Bolte,137 who highlight that many studies frequently measure PROGRESS-Plus factors such as race/ethnicity, social status and education, but that few actually consider them in the analysis. This is likely due to researchers not considering the impact of their research on inequalities when planning their studies.138 Differential effect analyses were more frequently considered for gender, age and comorbidities, with significant findings of a worse tapering outcome for patients who are male and older. Mixed evidence for differential effects was found for comorbidities. For race/ethnicity, occupation and socioeconomic status, no significant differential associations were observed on opioid dose reduction. No differential associations were measured for place of residence, education or social capital. There is also evidence suggesting significant associations (not differential) between gender, age, occupation, social capital and comorbidities on tapering outcomes. Ultimately, any widening of inequalities in access to opioid tapering will likely impact on a widening of inequalities in the effectiveness of opioid tapering. Factors that may potentially widen inequalities in relation to access were highlighted at the individual, interpersonal and wider structural levels and may arise at any point across the entire spectrum of the tapering process, from the point of recognition of a need to taper, initiation of and engagement with a taper. Of note is the influence of ‘social capital’ which is clearly visible at every stage in a patient’s opioid-tapering pathway.
Given that the findings of this review indicate significant associations of some PROGRESS-Plus characteristics, in particular, ‘Gender’, ‘Age’ and ‘Comorbidities’ with changes in opioid use or tapering trajectory, it is imperative that researchers consider including demographic representation of disadvantaged populations in research on opioid tapering. The review findings on inequalities in participation, in particular, reasons for patients dropping out/not completing an opioid taper also have important implications for the way in which interventions are designed and implemented to avoid excluding underserved populations. The NIHR ‘Improving inclusion of underserved groups in clinical research: Guidance from the NIHR INCLUDE project139 supports this view, stating that the ‘principle of “no decision about me, without me” provides the moral justification for ensuring that under-served groups are always included in research’ regardless of the subject matter being researched. In addition, one of the review findings highlighted inequalities in access to alternative tapering options depending upon where patients access tapering support (place of residence). However, in a survey of patient engagement in RCTs of tapering chronic opioid therapy, James et al.140 found that in order to improve engagement, patients wanted access to alternative pain management options.
While some PROGRESS-Plus characteristics were well reported (e.g. age, gender), other PROGRESS-Plus characteristics that may potentially widen inequalities were less well reported and less frequently analysed. This may indicate a lack of diversity in the sampling approach within the included studies, resulting in an under-representative sample. For example, mental health comorbidities are common among patients prescribed long-term opioid use for chronic non-cancer pain,95 and the results of this review indicate an association of mental health comorbidities with tapering outcomes. However, only 12 studies (31%) in this review reported on the comorbidity status of participants. Important differences in how comorbid populations respond to or engage with opioid tapering compared to other groups may not therefore have been captured. The current state of research on opioid tapering may not, therefore, accurately reflect the high healthcare burden of opioid tapering within this group.139