You are here

  • Home
  • Archive
  • Volume 57, Issue 15
  • Delayed timing of physical therapy initiation increases the risk of future opioid use in individuals with knee osteoarthritis: a real-world cohort study

Article Text

Article menu
Original research
Delayed timing of physical therapy initiation increases the risk of future opioid use in individuals with knee osteoarthritis: a real-world cohort study
Free
  1. Deepak Kumar1,2,
  2. Tuhina Neogi2,
  3. Christine Peloquin2,
  4. Lee Marinko1,
  5. James Camarinos1,
  6. Kosaku Aoyagi2,3,
  7. David T Felson2,
  8. Maureen Dubreuil2,4
  1. 1 Department of Physical Therapy, Boston University, Boston, Massachusetts, USA
  2. 2 Section of Rheumatology, Department of Medicine, Boston University School of Medicine, Boston, Massachusetts, USA
  3. 3 Physical Therapy Program, The University of Texas at El Paso, El Paso, Texas, USA
  4. 4 Visiting Scholar, Optum Labs, Eden Prairie, Minnesota, USA
  1. Correspondence to Dr Deepak Kumar, Physical Therapy, Boston University, Boston, MA 02215, USA; krdeepak2pro{at}gmail.com

Abstract

Objective We assessed whether late versus early initiation of physical therapy (PT) was related to greater risk of future opioid use in people with knee osteoarthritis (OA) who receive PT.

Methods We used Commercial and Medicare Advantage claims data from 1999 to 2018 from American adults with incident knee OA referred for PT within 1 year of diagnosis. We categorised people as opioid naïve or opioid experienced based on prior prescriptions. We examined the association of timing of PT initiation with any and chronic opioid use over 1 year.

Results Of the 67 245 individuals with incident knee OA, 35 899 were opioid naïve and 31 346 were opioid experienced. In the opioid naïve group, compared with PT within 1 month, PT 1 to <3, 3 to <6, 6 to <9, 9–12 months from diagnosis was associated with adjusted risk ratio (aRR (95% CIs)) for any opioid use of 1.18 (1.10 to 1.28), 1.49 (1.37 to 1.61), 1.73 (1.58 to 1.89) and 1.93 (1.76 to 2.12), respectively; aRRs (95% CIs) for chronic opioid use were 1.25 (1.01 to 1.54), 1.83 (1.48 to 2.26), 2.29 (1.82 to 2.89) and 2.50 (1.96 to 3.19). Results were similar among opioid experienced; aRRs (95% CIs) for any opioid use were 1.19 (1.14 to 1.24), 1.32 (1.26 to 1.37), 1.39 (1.32 to 1.45) and 1.54 (1.46 to 1.61); aRRs (95% CIs) for chronic opioid use were 1.25 (1.17 to1.34), 1.43 (1.33 to 1.54), 1.53 (1.41 to 1.66) and 1.65 (1.51 to 1.80).

Conclusion Compared with PT initiation within 1 month, delayed PT initiation was associated with higher risk of opioid use in people with incident knee OA. The longer the delay in PT initiation, the greater was the risk.

  • exercise
  • knee
  • cohort studies
  • rehabilitation

Data availability statement

Data may be obtained from a third party and are not publicly available. Data used for this study are owned by Optum Labs.

https://doi.org/10.1136/bjsports-2022-106044

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    WHAT IS ALREADY KNOWN ON THIS TOPIC

    • Exercise and education, commonly delivered as part of physical therapy (PT), are the first line recommended interventions for people with chronic pain due to knee osteoarthritis (OA).

    WHAT THIS STUDY ADDS

    • Delayed initiation of PT (ie, >1 month after diagnosis) versus early initiation (ie, within 1 month of diagnosis) is associated with increased future risk of any opioid use and chronic opioid use in people with incident knee OA. The longer the delay in PT initiation, the greater is the risk.

    HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

    • In people referred to PT after a diagnosis of knee OA, earlier initiation of care could lead to more effective pain management and reduce reliance on opioids.

    Introduction

    The global opioid crisis has been partly fueled by the need to manage chronic musculoskeletal pain.1–4 Osteoarthritis (OA) is a leading cause of chronic pain in middle-age and older-age adults, and knee OA accounts for over half of the global burden of OA.5 6 In the USA, joint pain due to OA is the second most common reason for patients to seek medical care7 and prescriptions for arthritis pain represent over 50% of all opioid prescriptions.8 9 Hence, interventions that could reduce utilisation of opioids in people with knee OA are needed.

    Exercise, usually delivered as part of physical therapy (PT) care, is recommended as first-line intervention for knee OA across the world.10–14 However, in individuals undergoing knee replacement, utilisation of PT has remained variable and suboptimal in countries where such data have been reported.15 16 Contrary to treatment guidelines, intra-articular and oral analgesics are the most common initial interventions for knee OA.10 11 17–19 In fact, there has been an increase in prescription analgesics over time.1 3 20 Early versus late initiation of active exercise-based PT interventions after knee OA diagnosis could improve pain management and function leading to reduced utilisation of prescription opioids, as has been reported for people with low back pain or patellofemoral pain.21 22 These studies may guide clinicians to use PT as an early strategy to reduce or prevent opioid use in people with knee OA.

    While there is some evidence from clinical trials that greater number of PT sessions are related with better outcomes, whether this is true in real-world practice is not known.23 24 Further, while passive interventions (eg, heat and cold) should only be provided with active PT interventions that require patient participation (eg, self-management training, exercise, gait retraining), it is not known if active interventions protect against opioid use.10 11 Information on optimal dose and type of PT interventions could guide clinical practice for knee OA.

    Our primary objective was to determine the association of timing of PT initiation with subsequent opioid use in individuals with incident knee OA who receive PT who were opioid naïve and those with prior opioid use. We also assessed the associations of dose and type (ie, active vs passive) of PT with any and chronic opioid use in this population.

    Methods

    Study design

    We conducted a longitudinal cohort study in individuals with incident knee OA who received PT within 1 year of diagnosis (figure 1). We assessed the relation of the PT timing, dose and type after diagnosis with future opioid use over a 1-year period.

    Figure 1
    Figure 1

    Study design. EOC, episode of care; PT, physical therapy. *The PT EOC can start anytime within the 12-month period following the index date. In cases where the PT EOC is initiated close to the end of the 12-month period, it may extend into the outcome assessment period.

    Study sample

    We used data from the Optum Labs Data Warehouse (OLDW, Eden Prairie, USA), which includes deidentified medical and pharmacy claims, PT claims, laboratory results and enrolment records for commercial and Medicare Advantage enrollees. Members in the database had full insurance coverage for physician, hospital and prescription drug services.25 The database contains longitudinal health information on patients representing a diverse mix of ages, ethnicities and US geographical regions.

    We included data from individuals ≥40 with incident knee OA between 2001 and 2016. Incident knee OA was defined by knee OA diagnosis (International Classification of Diseases (ICD), code 715 .x6 (9th revision) or M17.x (10th revision)) with no claims for a diagnosis of knee OA during the prior 24 months. The date of knee OA diagnosis was defined as the index date. To be included in the analyses, individuals were required to have continuous medical and pharmacy claims from 24 months prior to and following the index date. We excluded individuals with history of total knee replacement, other knee surgery, cancer or rheumatoid arthritis. We also excluded individuals who received PT in the 12 months preceding the index date and individuals with missing information for sex, geographical region or insurance type.

    We categorised included individuals as those with any prior opioid use (opioid experienced) and those without prior opioid use (opioid naïve) as done in prior studies.26 27 Since single opioid prescription for acute medical issues is related to greater risk of future use, prior opioid use was defined as 1 or more filled opioid prescriptions within 2 years prior to index date28–32; opioid naïve individuals had no opioid prescription within the 2 years.

    Exposures

    Our primary exposure was timing of PT initiation relative to the index date. Secondary exposures were PT dose and type (described in next paragarph). Exposures were defined by first identifying the initial outpatient PT episode of care (EOC) in the 12-month period following the index date (figure 1). The initial PT visit was identified by a current procedural terminology (CPT) evaluation code for PT (online supplemental eTable 1) from an outpatient facility. A PT EOC was considered to have ended when there were no PT claims ≥12 weeks. To ensure that PT was for knee OA, all ICD codes from the initial PT visit were reviewed (DK) to identify those indicating knee OA. Individuals with ICD codes that did not suggest PT for knee OA were excluded. We also excluded individuals who only underwent PT evaluation but no intervention.

    Supplemental material

    We categorised PT timing as the time (in months) between index date and the initial PT visit (ie, <1 month (reference), 1 to <3 months, 3 to <6 months, 6 to <9 months, 9 to 12 months). Guided by clinical practice, PT dose was categorised as the number of unique sessions on unique dates (ie, 1–5 (reference), 6–12, 13+) during the PT EOC. PT type was defined as active or passive (reference) based on CPT codes during the PT EOC (online supplemental eTable 1). Active PT interventions were defined as ≥50% of codes being active interventions during the PT EOC, as per prior studies.33

    Outcomes

    Outcomes of interest were as follows: (1) any opioid use and (2) chronic opioid use over a 12-month period starting 1 year after index date (ie, outcome assessment period) (figure 1). We included oral/enteral opioid formulations as selected by a rheumatologist investigator (MD) (online supplemental eTable 2). Any opioid use was defined as ≥2 filled prescriptions of an opioid during the outcome assessment period.34 Chronic opioid use was defined as ≥90 days of filled opioid prescriptions as defined previously in non-surgical cohorts.35

    Potential confounders

    We selected confounders based on factors that could influence both our exposures and outcomes. All analyses were adjusted for age, sex, race, obesity, type of insurance (commercial or Medicare Advantage), geographical location, physical and mental health comorbidities. Comorbidities were identified throughout 24 months prior to the index date using the Elixhauser Index.36 37 Physical comorbidities were included as the count of physical comorbidities, except for: obesity, alcohol abuse, drug abuse, psychoses, depression.21 37 Bipolar disorder, post-traumatic stress disorder and schizophrenia disorder that were combined because of the potential collinearity. Other psychological disorders, obesity and sleep disorders were included individually.36–38 Healthcare utilisation was calculated as number of outpatient claims ≥7 days apart within the 2 years prior to index date. Calendar year was added as a covariate for all models to account for secular trends.

    Statistical analysis

    We evaluated the relations of the exposures to the outcomes in the opioid experienced and naïve cohorts separately, using Poisson regression with a robust variance.39 We report risk ratios adjusted for the covariates (adjusted risk ratio, aRR). Analyses were conducted using SAS V.9.4 (SAS Institute). We calculated E-values to examine the robustness of the findings to unmeasured confounding.40 The E-value reflects the minimum strength of association needed between an unmeasured confounder and both the exposure and the outcome to nullify the observed effect estimate. Sensitivity analyses were as follows: (A) redefined opioid experienced individuals as those with two or more filled opioid prescriptions within 24 months prior to index date to avoid inclusion of single prescriptions for acute medical or dental issues,34 (B) redefined chronic opioid use outcome as at least 90 days of filled prescriptions without ≥30 days break in supply and (C) additionally included prior non-steroidal anti-inflammatory drugs (NSAIDs) use as a covariate. We also performed sensitivity analyses for PT type by redefining active PT as each PT visit having at least one active code.

    Equity, diversity and inclusion statement

    We used data from OLDW, which includes persons of all genders, race/ethnicities and geographical regions in the USA. The research team included four women (senior author is a woman) and four men, of whom three are early career researchers. The author’s disciplines include PT, rheumatology, epidemiology and biostatistics. We considered sex and race as covariates in our analyses. We also considered geographical region and type of insurance as covariates which may partially capture socioeconomic status. However, since we only included data from persons with insurance, those who are underinsured or uninsured are not captured in this study.

    Patient and public involvement

    Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of our research.

    Results

    We identified 2 194 144 individuals with an incident knee OA diagnosis between 2000 and 2016 of whom 67 245 met the study criteria (figure 2). Of these 35 899 were opioid naïve and 31 346 were opioid experienced (table 1). Opioid experienced individuals had greater prevalence of obesity, NSAIDs use and other health conditions.

    Figure 2
    Figure 2

    Participant flow diagram. OA, osteoarthritis; PT, physical therapy.

    View this table:
    Table 1

    Cohort characteristics

    The mean duration of PT EOC was 6.2±7.5 weeks and the mean intensity of PT was 2.4±1.9 visits/week. Overall, in the opioid naïve cohort, the prevalence of any opioid use and chronic opioid use during the outcome assessment period was 12.3% and 1.9% in the opioid naïve cohort and 35.0% and 15.0% in the opioid experience cohort.

    Timing of PT initiation and opioid use

    In the opioid naïve group, when compared with initiation of PT intervention within 1 month of knee OA diagnosis, initiation of PT 1 to <3 months, 3 to <6 months, 6 to <9 months, 9 to 12 months from diagnosis was associated with greater risk of any opioid use with aRR of 1.18 (95% CI 1.10 to 1.28), 1.49 (95% CI 1.37 to 1.61), 1.73 (95% CI 1.58 to 1.89) and 1.93 (95% CI 1.76 to 2.12), respectively (figure 3). Similarly, initiation of PT 1 to <3 months, 3 to <6 months, 6 to <9 months, 9 to 12 months from diagnosis was associated with increased risk of chronic opioid use with aRR of 1.25 (95% CI 1.01 to 1.54), 1.83 (95%CI 1.48 to 2.26), 2.29 (95% CI 1.82 to 2.89) and 2.50 (95% CI 1.96 to 3.19), respectively (figure 4).

    Figure 3
    Figure 3

    Relation of timing of PT initiation with any opioid use in opioid naïve (grey) and opioid experienced (black) individuals. aRR, adjusted risk ratio; PT, physical therapy.

    Figure 4
    Figure 4

    Relation of timing of PT initiation with chronic opioid use in opioid naïve (grey) and opioid experienced (black) individuals. aRR, adjusted risk ratio; PT, physical therapy.

    Similar to the findings in the opioid naïve group, in the opioid experienced group, initiation of PT 1 to <3 months, 3 to <6 months, 6 to <9 months, 9 to 12 months from diagnosis was associated with increased risk of any opioid use with aRR of 1.19 (95% CI 1.14 to 1.24), 1.32 (95% CI 1.26 to 1.37), 1.39 (95% CI 1.32 to 1.45) and 1.54 (95% CI 1.46 to 1.61), respectively (figure 3). Similarly, initiation of PT 1 to <3 months, 3 to <6 months, 6 to <9 months, 9 to 12 months from diagnosis was associated with increased risk of chronic opioid use with aRR of 1.25 (95% CI 1.17 to 1.34), 1.43 (95% CI 1.33 to 1.54), 1.53 (95% CI 1.41 to 1.66) and 1.65 (95% CI 1.51 to 1.80), respectively (figure 4).

    Results were largely unchanged when additionally adjusted for prior NSAIDs use in both cohorts (online supplemental eTable 3). Results were also unchanged with more stringent criteria were used to define the opioid experienced cohort (online supplemental eTables4 and 5) and to define the chronic opioid use outcome (online supplemental eTable 6).

    The E-values (figures 3 and 4) showed that an unmeasured confounder would have to be associated with both the exposure (timing of PT) and the outcome (any opioid use or chronic opioid use), over and above the measured confounders, by an RR range of 1.64–4.44 (E-value for lower limit of CI ranging from 1.11 to 3.33) in the opioid naïve group. Similarly, the E-value range was 1.67–2.69 (E-value for lower limit of CI ranging from 1.54 to 2.39) for the opioid experienced group.

    PT intervention dose and type, and opioid use

    In the opioid naïve group, when compared with 1–5 PT visits, having 6–12 PT visits did not alter the risk for any opioid use (aRR 0.96, 95% CI 0.90 to 1.02), whereas the risk was increased for >12 PT visits (aRR 1.10, 95% CI 1.02 to 1.18) (online supplemental eTable 7). The risk for chronic opioid use was lower for 6–12 PT visits (aOR 0.81, 95% CI 0.68 to 0.97) and was equivocal for >12 PT visits (aOR 1.07, 95% CI 0.89 to 1.28) compared with 1–5 PT visits (online supplemental eTable 7).

    In the opioid experienced group, 6–12 visits were associated with a lower risk of any opioid use (aRR 0.93, 95% CI 0.90 to 0.97) and chronic opioid use (aRR 0.88, 95% CI 0.83 to 0.94) than those with 1–5 PT visits (online supplemental eTable 7). Having >12 PT visits compared with 1–5 PT visits was not associated with any opioid use (aRR 1.01, 95% CI 0.98 to 1.05) or chronic opioid use (aRR 0.97, 95% CI 0.91 to 1.03) (online supplemental eTable 7).

    For PT type, active PT was associated with lower risk of any opioid use (aRR 0.92, 95% CI 0.86 to 0.98) but not with chronic opioid use (aRR 0.86, 95% CI 0.72 to 1.02) in the opioid naïve group (online supplemental eTable 8). This was also true in the opioid experienced cohort (online supplemental eTable 8). Sensitivity analysis showed similar results except there was a larger protective effect of active PT for chronic opioid use (aRR 0.77, 95% CI 0.62 to 0.95) in the opioid naïve group (online supplemental eTable 9).

    Discussion

    In this large real-world cohort study of persons with incident knee OA who receive PT, delayed PT initiation versus early initiation was associated with greater risk of opioid use, irrespective of prior experience with opioids. These results provide support for initiation of PT within 1 month of diagnosis for individuals with knee OA who have been referred to PT as being associated with lower risk of opioid use in the subsequent 12 months. However, it is important to note our findings are generalisable only to patients with incident knee OA who receive PT and that a majority of the patients typically do not get referred to PT despite clinical practice guidelines.

    The association of timing of PT initiation with risk of opioid use was noted in both opioid naïve and experienced groups, with somewhat larger relative effects in those who were opioid naïve, potentially reflecting their lower background risk than those who are opioid experienced. Our findings were robust to adjustment for prior NSAIDs use which may be an indicator of pain severity. Similar findings have been reported in people with new onset low back or neck pain where early initiation of PT care was associated with lower odds of opioid use (effect sizes between 0.15 and 0.84).21 41–46 In a cohort of people with musculoskeletal pain of different origins, PT initiation within 90 days of the index primary care visit was associated with ~16% lower odds of opioid use in those with knee pain.47 Our findings (ie, 18%–93% increased risk of any opioid use for later PT initiation) are in line with these previously reported data. In a randomised trial of people with low back pain, early PT initiation compared with usual care was associated with improvements in disability which did not meet minimal important difference criteria.48 To our knowledge, no randomised trials have specifically investigated early versus late PT in people with knee OA. Our study provides additional important information about the relation of timing of PT to chronic opioid use with increased risk ranging from 25% to 150% depending on the length of delay and prior opioid experience. Our findings were robust even when we used a more stringent definition of chronic opioid use outcome. For a majority of our cohort, PT was initiated more than 1 month after diagnosis; this trend may reflect referral patterns, access to care or patient preferences.49

    We observed a lower risk of any opioid use with 6–12 PT sessions vs 1–5 PT sessions, but only in the opioid experienced cohort. No further reduction in risk was seen with >12 sessions and in fact, an increased risk of opioid use was seen with >12 sessions in the opioid naïve group. It is possible that requiring >12 PT visits may reflect suboptimal pain management leading to subsequent opioid use. In a meta-analysis of exercise trials in knee OA, more reduction in pain was seen with a greater number of in-person sessions.50 In contrast, no further benefit was observed for >12 in-person sessions in our study. However, our effect estimates overall were small for the relation of PT dose to opioid use, and clinical meaningfulness of these findings remains unclear. The association of 6–12 sessions of PT with lower risk of chronic opioid use was relatively consistent across both opioid naïve and experienced cohorts. It is worth noting that approximately 45% of patients received ≤6 sessions of PT.

    We also observed lower risk of opioid use with active PT interventions in both cohorts. However, the effects observed in our study were small except when active PT was defined as every PT visit having ≥1 active PT code. In this definition of compliant care, active PT was related to a 23% lower risk of chronic opioid use in the opioid naïve cohort. Exercise has beneficial effects in this population.10 11 51 Real-world data from 16 499 patients with knee or hip OA in Denmark who received a standardised 8-week active intervention (2–3 session of patient education, 12 sessions of supervised exercise therapy) across >400 clinics also demonstrated reductions in analgesic use, including opioids (decrease from 7.0% to 4.5%).52 Our data support and extend these previous findings with an association of active PT interventions with a lower risk of opioid use outside of a highly structured setting with selected patients.

    Clinical implications

    Guidelines recommend exercise and education, typically delivered as part of PT care, as the first-line intervention for managing pain due to knee OA. However, utilisation of PT for people with knee OA remains low and healthcare providers should consider referral to PT as an early strategy for these patients. In addition, our findings from this large, representative cohort of American adults with incident knee OA suggest that in people referred for PT, strategies to ensure early initiation of PT care are needed to reduce reliance on opioids. Expanded reimbursement of self-referral, telerehabilitation, stepped care treatment protocols and implementation of clinical practice guidelines, are some approaches to improve early access and referral to PT. Also, physicians, insurance providers and PT practices may consider ensuring that 6–12 sessions of PT are provided to people with knee OA. More work is also needed to standardise PT interventions for knee OA and to determine the relation of type of PT interventions and subsequent opioid use.

    Limitations

    This study was performed using claims data. As such we did not have the ability to validate knee OA definition against diagnostic criteria; however, we combined the claims-based diagnosis with a treatment for the same condition (PT), which would be expected to increase specificity. Also, individuals may have sought care for symptoms related to knee OA prior to the index date. We adjusted for prior healthcare utilisation and NSAIDs use to partly account for this. Given the observational nature of our data, residual confounding is a possibility, though the reported E-values and adjustments for a large number of confounders raise confidence that our finding are unlikely to largely be the result of unmeasured confounding. We did not compare individuals who did and did not receive PT due to concerns with confounding by indication. We did not adjust for access to PT care across different states because regulations regarding access to PT have changed over time and it was not possible to adjust for this variation in our study comprising data from a 15-year period, though we accounted for calendar year to address secular trends. We did not have information on OA severity or clinical outcomes (eg, pain, function); hence, opioid use only serves as a proxy for efficacy of pain management. We used claims data through 2018; lack of substantive changes in OA management guidelines over the recent years and continued evidence of opioid use in this population suggest that our findings are relevant to current practice. We modelled PT timing as a categorical variable and the effect sizes for future opioid use may be different if PT timing were modelled as a continuous variable. While our cohort was reflective of the racial/ethnic and geographical diversity in the USA, it is unclear if our findings would be generalisable to other countries with different demographic and healthcare utilisation patterns. We also did not have access to data from underinsured or uninsured individuals who typically have a lower socioeconomic status.

    In conclusion, we observed that delayed versus early PT initiation after knee OA diagnosis was associated with greater risk of opioid use irrespective of prior experience with opioids. For people with knee OA who are referred for PT, these real-world data provide support for strategies to initiate PT care early. These findings on timing, dose and type of PT interventions may be used by payers and providers for implementing optimal PT interventions to reduce utilisation of opioids in these patients.

    Data availability statement

    Data may be obtained from a third party and are not publicly available. Data used for this study are owned by Optum Labs.

    Ethics statements

    Patient consent for publication

    Ethics approval

    Since this study involved analysis of pre-existing, deidentified data, it was exempt from institutional review board review.

    References

    1. ↵
      2. Song I-A ,
      3. Choi H-R ,
      4. Oh TK
      . Long-Term opioid use and mortality in patients with chronic non-cancer pain: ten-year follow-up study in South Korea from 2010 through 2019. EClinicalMedicine 2022;51:101558. doi:10.1016/j.eclinm.2022.101558
    2. ↵
      2. Yu D ,
      3. Hellberg C ,
      4. Appleyard T , et al
      . Opioid use prior to total knee replacement: comparative analysis of trends in england and sweden. Osteoarthritis Cartilage 2022;30:815–22. doi:10.1016/j.joca.2022.02.621
      OpenUrl
    3. ↵
      2. Rejas-Gutierrez J ,
      3. Sicras-Mainar A ,
      4. Darbà J
      . Future projections of opioid use and cost in patients with chronic osteoarthritis pain in spain. Ther Adv Musculoskelet Dis 2021;13. doi:10.1177/1759720X211010599
    4. ↵
      2. Thorlund JB ,
      3. Turkiewicz A ,
      4. Prieto-Alhambra D , et al
      . Opioid use in knee or hip osteoarthritis: a region-wide population-based cohort study. Osteoarthritis Cartilage 2019;27:871–7. doi:10.1016/j.joca.2019.01.005
      OpenUrlCrossRefPubMed
    5. ↵
      1. GBD 2019 Diseases and Injuries Collaborators
      . Global burden of 369 diseases and injuries in 204 countries and territories, 1990-2019: a systematic analysis for the global burden of disease study 2019. Lancet 2020;396:1204–22. doi:10.1016/S0140-6736(20)30925-9
      OpenUrlCrossRefPubMed
    6. ↵
      1. GBD 2015 Disease and Injury Incidence and Prevalence Collaborators
      . Global, regional, and national incidence, prevalence, and years lived with disability for 310 diseases and injuries, 1990-2015: a systematic analysis for the global burden of disease study 2015. Lancet 2016;388:1545–602. doi:10.1016/S0140-6736(16)31678-6
      OpenUrlCrossRefPubMed
    7. ↵
      2. St Sauver JL ,
      3. Warner DO ,
      4. Yawn BP , et al
      . Why patients visit their doctors: assessing the most prevalent conditions in a defined american population. Mayo Clin Proc 2013;88:56–67. doi:10.1016/j.mayocp.2012.08.020
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      2. Stokes A ,
      3. Berry KM ,
      4. Hempstead K , et al
      . Trends in prescription analgesic use among adults with musculoskeletal conditions in the united states, 1999-2016. JAMA Netw Open 2019;2:e1917228. doi:10.1001/jamanetworkopen.2019.17228
    9. ↵
      Prevalence and trends in prescribed opioid use among US adults with arthritis, 2008-2013, medical expenditure panel survey. ACR/ARHP Annual Meeting; 2016
    10. ↵
      2. Bannuru RR ,
      3. Osani MC ,
      4. Vaysbrot EE , et al
      . OARSI guidelines for the non-surgical management of knee, hip, and polyarticular osteoarthritis. Osteoarthritis Cartilage 2019;27:1578–89. doi:10.1016/j.joca.2019.06.011
      OpenUrlCrossRefPubMed
    11. ↵
      2. Kolasinski SL ,
      3. Neogi T ,
      4. Hochberg MC , et al
      . 2019 American College of rheumatology/arthritis Foundation guideline for the management of osteoarthritis of the hand, hip, and knee. Arthritis Rheumatol 2020;72:220–33. doi:10.1002/art.41142
      OpenUrlPubMed
    12. ↵
      2. Basedow M ,
      3. Williams H ,
      4. Shanahan EM , et al
      . Australian GP management of osteoarthritis following the release of the RACGP guideline for the non-surgical management of hip and knee osteoarthritis. BMC Res Notes 2015;8:536.:536. doi:10.1186/s13104-015-1531-z
      OpenUrlPubMed
    13. ↵
      1. National Institute for Health and Care Excellence
      . Guidelines. osteoarthritis in over 16s: diagnosis and management. London National Institute for Health and Care Excellence (NICE); 2022.
    14. ↵
      2. Overton C ,
      3. Nelson AE ,
      4. Neogi T
      . Osteoarthritis treatment guidelines from six professional societies: similarities and differences. Rheum Dis Clin North Am 2022;48:637–57. doi:10.1016/j.rdc.2022.03.009
      OpenUrl
    15. ↵
      2. Dhawan A ,
      3. Mather RC ,
      4. Karas V , et al
      . An epidemiologic analysis of clinical practice guidelines for non-arthroplasty treatment of osteoarthritis of the knee. Arthroscopy 2014;30:65–71. doi:10.1016/j.arthro.2013.09.002
      OpenUrlCrossRefPubMed
    16. ↵
      2. Grenier JP ,
      3. Rothmund M ,
      4. Missmann M
      . Variation in the utilisation of physiotherapy in patients with advanced knee osteoarthritis prior to total knee arthroplasty a systematic review. Musculoskeletal Care 20, 2022. doi:10.1002/msc.1726
    17. ↵
      2. DeMik DE ,
      3. Bedard NA ,
      4. Dowdle SB , et al
      . Are we still prescribing opioids for osteoarthritis? J Arthroplasty 2017;32:3578–82. doi:10.1016/j.arth.2017.07.030
      OpenUrlCrossRef
    18. ↵
      2. Gwam CU ,
      3. Emara AK ,
      4. Ogbonnaya IA , et al
      . Addressing national opioid prescribing practices for knee osteoarthritis: an analysis of an estimated 41,389,332 patients with knee arthritis. J Am Acad Orthop Surg 2021;29:e337–44. doi:10.5435/JAAOS-D-20-00924
      OpenUrlPubMed
    19. ↵
      2. Dysart S ,
      3. Utkina K ,
      4. Stong L , et al
      . Insights from real-world analysis of treatment patterns in patients with newly diagnosed knee osteoarthritis. Am Health Drug Benefits 2021;14:56–62.
      OpenUrlPubMed
    20. ↵
      2. Khoja SS ,
      3. Almeida GJ ,
      4. Freburger JK
      . Recommendation rates for physical therapy, lifestyle counseling, and pain medications for managing knee osteoarthritis in ambulatory care settings: a cross-sectional analysis of the National ambulatory care survey (2007-2015). Arthritis Care Res (Hoboken) 2020;72:184–92. doi:10.1002/acr.24064
      OpenUrl
    21. ↵
      2. Kazis LE ,
      3. Ameli O ,
      4. Rothendler J , et al
      . Observational retrospective study of the association of initial healthcare provider for new-onset low back pain with early and long-term opioid use. BMJ Open 2019;9:e028633. doi:10.1136/bmjopen-2018-028633
    22. ↵
      2. Young JL ,
      3. Snodgrass SJ ,
      4. Cleland JA , et al
      . Timing of physical therapy for individuals with patellofemoral pain and the influence on healthcare use, costs and recurrence rates: an observational study. BMC Health Serv Res 2021;21:751. doi:10.1186/s12913-021-06768-8
    23. ↵
      2. Lee AC ,
      3. Harvey WF ,
      4. Price LL , et al
      . Dose-Response effects of tai chi and physical therapy exercise interventions in symptomatic knee osteoarthritis. PM R 2018;10:712–23. doi:10.1016/j.pmrj.2018.01.003
      OpenUrl
    24. ↵
      2. Pignato M ,
      3. Arbeeva L ,
      4. Schwartz TA , et al
      . Level of participation in physical therapy or an Internet-based exercise training program: associations with outcomes for patients with knee osteoarthritis. BMC Musculoskelet Disord 2018;19:238. doi:10.1186/s12891-018-2139-y
    25. ↵
      1. OptumLabs
      . OptumLabs and optumlabs data warehouse (OLDW) descriptions and citation. reproduced with permission from optumlabs. 2020. Available: https://www.optumlabs.com/
    26. ↵
      2. Politzer CS ,
      3. Kildow BJ ,
      4. Goltz DE , et al
      . Trends in opioid utilization before and after total knee arthroplasty. J Arthroplasty 2018;33:S147–53. doi:10.1016/j.arth.2017.10.060
      OpenUrlPubMed
    27. ↵
      2. Bedard NA ,
      3. Pugely AJ ,
      4. Westermann RW , et al
      . Opioid use after total knee arthroplasty: trends and risk factors for prolonged use. J Arthroplasty 2017;32:2390–4. doi:10.1016/j.arth.2017.03.014
      OpenUrlPubMed
    28. ↵
      2. Alam A ,
      3. Gomes T ,
      4. Zheng H , et al
      . Long-Term analgesic use after low-risk surgery: a retrospective cohort study. Arch Intern Med 2012;172:425–30. doi:10.1001/archinternmed.2011.1827
      OpenUrlCrossRefPubMedWeb of Science
    29. ↵
      2. Brummett CM ,
      3. Waljee JF ,
      4. Goesling J , et al
      . New persistent opioid use after minor and major surgical procedures in US adults. JAMA Surg 2017;152:e170504. doi:10.1001/jamasurg.2017.0504
    30. ↵
      2. Calcaterra SL ,
      3. Yamashita TE ,
      4. Min S-J , et al
      . Opioid prescribing at hospital discharge contributes to chronic opioid use. J Gen Intern Med 2016;31:478–85. doi:10.1007/s11606-015-3539-4
      OpenUrlPubMed
    31. ↵
      2. Sun EC ,
      3. Darnall BD ,
      4. Baker LC , et al
      . Incidence of and risk factors for chronic opioid use among opioid-naive patients in the postoperative period. JAMA Intern Med 2016;176:1286–93. doi:10.1001/jamainternmed.2016.3298
      OpenUrl
    32. ↵
      2. Schroeder AR ,
      3. Dehghan M ,
      4. Newman TB , et al
      . Association of opioid prescriptions from dental clinicians for us adolescents and young adults with subsequent opioid use and abuse. JAMA Intern Med 2019;179:145–52. doi:10.1001/jamainternmed.2018.5419
      OpenUrl
    33. ↵
      2. Fritz JM ,
      3. Cleland JA ,
      4. Brennan GP
      . Does adherence to the Guideline recommendation for active treatments improve the quality of care for patients with acute low back pain delivered by physical therapists? Med Care 2007;45:973–80. doi:10.1097/MLR.0b013e318070c6cd
      OpenUrlCrossRefPubMedWeb of Science
    34. ↵
      2. Johnston SS ,
      3. Ammann E ,
      4. Scamuffa R , et al
      . Association of body mass index and osteoarthritis with healthcare expenditures and utilization. Obes Sci Pract 2020;6:139–51. doi:10.1002/osp4.398
      OpenUrl
    35. ↵
      2. Martin BC ,
      3. Fan M-Y ,
      4. Edlund MJ , et al
      . Long-Term chronic opioid therapy discontinuation rates from the TROUP study. J Gen Intern Med 2011;26:1450–7. doi:10.1007/s11606-011-1771-0
      OpenUrlCrossRefPubMed
    36. ↵
      1. Center for Medicare and Medicaid Services
      . Chronic condition data warehouse (CCW) algorithm; 2020.
    37. ↵
      2. Elixhauser A ,
      3. Steiner C ,
      4. Harris DR , et al
      . Comorbidity measures for use with administrative data. Med Care 1998;36:8–27. doi:10.1097/00005650-199801000-00004
      OpenUrlCrossRefPubMedWeb of Science
    38. ↵
      2. Jolley RJ ,
      3. Liang Z ,
      4. Peng M , et al
      . Identifying cases of sleep disorders through International classification of diseases (ICD) codes in administrative data. Int J Popul Data Sci 2018;3:448. doi:10.23889/ijpds.v3i1.448
    39. ↵
      2. Spiegelman D ,
      3. Hertzmark E
      . Easy SAS calculations for risk or prevalence ratios and differences. Am J Epidemiol 2005;162:199–200. doi:10.1093/aje/kwi188
      OpenUrlCrossRefPubMedWeb of Science
    40. ↵
      2. VanderWeele TJ ,
      3. Ding P
      . Sensitivity analysis in observational research: introducing the E-value. Ann Intern Med 2017;167:268–74. doi:10.7326/M16-2607
      OpenUrlCrossRefPubMed
    41. ↵
      2. Fritz JM ,
      3. Childs JD ,
      4. Wainner RS , et al
      . Primary care referral of patients with low back pain to physical therapy: impact on future health care utilization and costs. Spine (Phila Pa 1976) 2012;37:2114–21. doi:10.1097/BRS.0b013e31825d32f5
      OpenUrl
    42. ↵
      2. Childs JD ,
      3. Fritz JM ,
      4. Wu SS , et al
      . Implications of early and guideline adherent physical therapy for low back pain on utilization and costs. BMC Health Serv Res 2015;15:150. doi:10.1186/s12913-015-0830-3
      OpenUrlCrossRefPubMed
    43. ↵
      2. Fritz JM ,
      3. King JB ,
      4. McAdams-Marx C
      . Associations between early care decisions and the risk for long-term opioid use for patients with low back pain with a new physician consultation and initiation of opioid therapy. Clin J Pain 2018;34:552–8. doi:10.1097/AJP.0000000000000571
      OpenUrl
    44. ↵
      2. Horn ME ,
      3. Fritz JM
      . Timing of physical therapy consultation on 1-year healthcare utilization and costs in patients seeking care for neck pain: a retrospective cohort. BMC Health Serv Res 2018;18:887. doi:10.1186/s12913-018-3699-0
    45. ↵
      2. Magel J ,
      3. Kim J ,
      4. Fritz JM , et al
      . Time between an emergency department visit and initiation of physical therapist intervention: health care utilization and costs. Phys Ther 2020;100:1782–92. doi:10.1093/ptj/pzaa100
      OpenUrlCrossRef
    46. ↵
      2. Brown-Taylor L ,
      3. Beckner A ,
      4. Scaff KE , et al
      . Relationships between physical therapy intervention and opioid use: a scoping review. PM R 2022;14:837–54. doi:10.1002/pmrj.12654
      OpenUrl
    47. ↵
      2. Sun E ,
      3. Moshfegh J ,
      4. Rishel CA , et al
      . Association of early physical therapy with long-term opioid use among opioid-naive patients with musculoskeletal pain. JAMA Netw Open 2018;1:e185909. doi:10.1001/jamanetworkopen.2018.5909
    48. ↵
      2. Fritz JM ,
      3. Magel JS ,
      4. McFadden M , et al
      . Early physical therapy vs usual care in patients with recent-onset low back pain: a randomized clinical trial. JAMA 2015;314:1459–67. doi:10.1001/jama.2015.11648
      OpenUrlCrossRefPubMed
    49. ↵
      2. Christiansen MB ,
      3. Dix C ,
      4. Master H , et al
      . “I’ve been to physical therapy before, but not for the knees.” A qualitative study exploring barriers and facilitators to physical therapy utilization for knee osteoarthritis. Musculoskeletal Care 2020;18:477–86. doi:10.1002/msc.1491
      OpenUrl
    50. ↵
      2. Fransen M ,
      3. McConnell S ,
      4. Harmer AR , et al
      . Exercise for osteoarthritis of the knee: a cochrane systematic review. Br J Sports Med 2015;49:1554–7. doi:10.1136/bjsports-2015-095424
      OpenUrlAbstract/FREE Full Text
    51. ↵
      2. Verhagen AP ,
      3. Ferreira M ,
      4. Reijneveld-van de Vendel EAE , et al
      . Do we need another trial on exercise in patients with knee osteoarthritis?: no new trials on exercise in knee OA. Osteoarthritis Cartilage 2019;27:1266–9. doi:10.1016/j.joca.2019.04.020
      OpenUrlPubMed
    52. ↵
      2. Thorlund JB ,
      3. Roos EM ,
      4. Goro P , et al
      . Patients use fewer analgesics following supervised exercise therapy and patient education: an observational study of 16 499 patients with knee or hip osteoarthritis. Br J Sports Med 2020. doi:10.1136/bjsports-2019-101265

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • Twitter @ProfDeepakKumar

    • Contributors CP and DK had full access to all of the data in the study and take responsibility for the integrity of the data and accuracy of the data analysis. Concept and design: DK, TN, CP, MD, LM, JC and DTF. Acquisition, analysis or interpretation of data: DK, TN, CP, KA, MD, JC, LM and DTF. Drafting of the manuscript: DK, CP and KA. Critical revision of the manuscript for important intellectual content: DK, TN, CP, KA, LM, JC, DTF and MD. Statistical analysis: CP, DK, TN, MD. Obtained funding: DK, TN and DTF. Administrative, technical or material support: DK, MD, DTF and TN. Supervision: DK, TN, DTF and MD.

    • Funding This work is supported by NIH-NIAMS K01AR069720 (DK), K24AR070892 (TN), P30 AR0702571 (DTF) and K23AR069127 (MD).

    • Disclaimer The sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. Author DK accepts full responsibility for the work, had access to the data, and controlled the decision to publish

    • Competing interests TN served as a consultant for Pfizer/Lilly, Regeneron, and Novartis outside the submitted work. DK received grants from the National Institutes of Health during the conduct of the study and grants from Pfizer Inc for unrelated projects outside the submitted work. MD and DTF received grants from the National Institutes of Health.

    • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.