Characteristics of Early Interventions for Pain and Function Following Lower Extremity Joint Replacement: Systematic Review

Abstract

Occupational therapy is beneficial among adults with chronic pain; however, occupational therapy interventions addressing earlier phases of pain have not been clearly explicated. This systematic review characterized acute and subacute interventions billable by occupational therapy after hip or knee replacement to improve pain and function. Seven articles met inclusion criteria. Six articles had a low risk of bias. Three intervention types were found: task-oriented exercise, water-based, and modalities. Only task-oriented interventions improved both pain and function one-year after surgery. There are long-term benefits to early task-oriented exercise. Further research is needed to contextualize occupational therapy’s role in early pain interventions.

Introduction

Pain is a major source of disability worldwide (Dahlhamer et al., 2018). The Center for Disease Control and Prevention (CDC) estimates 1-in-12 adults experience pain limiting their daily function (Dahlhamer et al., 2018). In addition to functional deficits, pain costs $560–635 billion annually due to lost productivity and increased health care utilization, which demonstrates a need for pain interventions that consider the biopsychosocial nature of the disease (American Occupational Therapy Association, 2021; National Pain Strategy Task Force, 2015). Occupational therapy practitioners can provide the needed holistic interventions to address both pain and function (National Academies of Sciences, Engineering, and Medicine, 2019). Throughout this study, we refer to occupational therapists and assistants as “practitioners.” Literature on the role of practitioners in pain management has prioritized chronic pain and a dearth of studies have examined early intervention options (Hesselstrand et al., 2015; Lagueux et al., 2018; Pollack et al., 2020; Simon & Collins, 2017). In summary, occupational therapy plays an important role in chronic pain management via patient education, functional goal setting, activity-based training, environmental adaptation, and interdisciplinary referrals (Rochman, 2014).

Expanding the role of practitioners in acute and subacute pain management is a step toward mitigating chronic pain (Treede et al., 2015). The present review considers acute and subacute pain in the common surgical population of total hip or knee replacement. In 2010, over seven million individuals received hip or knee replacements in the United States (Maradit Kremers et al., 2015). By 2030, total hip replacement (THR) and total knee replacement (TKR) are projected to increase by 71% and 85%, respectively (Sloan et al., 2018). Normal healing time for THR and TKR should not exceed 6 months; therefore, we consider pain persisting past 6 months as chronic pain and we define pain prior to post-operative month 6 as acute or subacute (American Association of Hip & Knee Surgeons, 2017; Treede et al., 2015).

Occupational therapy interventions for acute and subacute pain have not been clearly explicated, indicating a gap in the literature. Research is needed to guide early pain interventions because an estimated 49% of THR and TKR patients still use prescription opioids three months after surgery, indicating subacute pain following these surgeries is an issue so prevalent that nearly half of these patients exceeded the recommendations for post-operative opioid utilization (Kim et al., 2017; Michigan Opioid Prescribing Engagement Network [OPEN], 2021). Without early and effective treatment, subacute pain may persist and increase the risk for prolonged opioid use, which is associated with adverse outcomes (Pritchard et al., 2021). Other adverse outcomes like substance use disorder may prevent participation in healthier occupations (Wasmuth et al., 2014). Still, prescription opioid use is only one objective evaluation of pain. A more comprehensive view of pain includes patient-reported outcomes like the subjective evaluation of pain emphasized by the Occupational Therapy Practice Framework: Domain and Process Fourth Edition (Boop et al., 2020). Practitioners consider an individual’s objective and subjective evaluation of pain because these two constructs define performance capacity and give insight into one’s lived experience (Kielhofner, 2008). Therefore, research is needed to study occupational therapy’s impact on outcomes of opioid use and patient-reported pain together.

Post-surgical pain and function are treated with nonpharmacologic and pharmacologic approaches. Systematic reviews have studied the use of nonpharmacologic interventions for pain and function, but did not include occupational therapists and were limited to follow-up periods shorter than six months (Skelly et al., 2020; Tedesco et al., 2017). The role of pre-operative and acute care occupational therapy is growing (Pritchard et al., 2019). To develop occupational therapy’s’ role in pain management, research is needed to determine whether early nonpharmacologic interventions for pain and function could be provided by practitioners with lasting benefits. This is an important area of study because clinical practice guidelines from the CDC encourage using nonpharmacologic interventions instead of high-risk pharmacologic interventions—namely prescription opioids (Dowell et al., 2016). Although opioids are effective in managing acute or subacute pain, evidence shows they do not improve function or quality of life (Sites et al., 2014). Furthermore, adults 65 years and older—the typical age group for joint replacement—are the only age group with an increasing rate of prescription opioid-related overdose mortality (Scholl et al., 2019). Practitioners may play an important role in acute and subacute pain management following THR and TKR because occupational therapy is associated with a reduction in prolonged opioid use and has improved pain and function in chronic pain populations (Pritchard et al., 2021; Simon & Collins, 2017). Furthermore, health management is an occupation that considers pain management integral to maintaining health and supporting participation in other occupations (Boop et al., 2020). To our knowledge, no reviews have analyzed acute and subacute occupational therapy interventions on outcomes of patient function, chronic pain (i.e., persisting ≥6 months post-surgery), and prescription opioid use 6 months or more following THR or TKR.

Objectives

Our objectives were to systematically review randomized controlled trials and evaluate the characteristics and long-term (≥6 months) benefits of nonpharmacologic interventions (billable by practitioners) for adults in the acute and subacute phase of recovery (<6 months post-surgery) after total hip or knee arthroplasty. We proposed to answer the following questions:

First, what are the characteristics of early nonpharmacologic interventions for pain and function after lower extremity joint replacement in adults? Second, which intervention types exhibited long-term efficacy on primary and secondary outcomes? Primary outcomes included any measure of pain and function at least 6 months after surgery. A secondary outcome was prolonged opioid consumption (Dowell et al., 2016).

Materials and methods

The protocol for this systematic review is registered at Prospero at University of York under registration number CRD42020205532. We followed the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guidelines (Page et al., 2021).

Search methods

The following databases were searched on 6/17/2020 and again on 12/27/2021: Ovid Medline, PsychInfo, Cochrane CENTRAL, Cochrane Reviews, Scopus, CINAHL, OTSeeker.

The search strategies were created by a reference librarian using input from the systematic review team, and the search string was peer reviewed by a second librarian. The search strategies were translated to each database depending on the subject heading availability and functionality of each database. The full search strings for each database can be found in the Appendix.

The search in most of the databases included synonyms of hip/knee replacements and a list of occupational therapy interventions. The OTSeeker search only focused on hip/knee replacement terms. A combination of various keywords, subject headings, adjacency, and truncation were the methods used to carry out the search strings in each database. Limiters were used depending on the availability in each database. The first set of limitations on the results included publications from 1995 to current and English language. Then, additional limiters were used for finding randomized controlled trial (RCT) publication types and covering the middle aged and older aged groups from the previous set of results. These were applied to the results on a different line of the search strategy so that any articles that were not picked up with them could be screened through separately for possible inclusion into this review. This was to ensure that no relevant articles on this topic would be missed during the literature search process.

Literature collection and screening

A preliminary screening of the first 100 articles yielded by the search was conducted in blinded pairs to establish the percent of agreement among rater pairs. Each pair screened by title and abstract to determine which full texts should be included, then these were read and assessed for eligibility. A moderator measured agreement within pairs and resolved conflicts. Pairs showed 78% and 88% agreement. The same process was utilized to screen the remaining articles by title and abstract. Articles that met inclusion criteria were read in blinded pairs to assess eligibility, with disagreements resolved by a moderator. This process was conducted using Rayyan (Ouzzani et al., 2016).

Eligibility criteria

Randomized control trials involving adults (18+) who had undergone a hip or knee replacement were eligible for inclusion in the systematic review. Literature published in the English language from 1995 onward was included, as this duration spans the opioid epidemic that prompted modern research on nonpharmacologic pain interventions (Van Zee, 2009). Since the authors could only read English, literature was restricted to English to reduce errors from incorrect translations or misunderstandings. Studies implementing occupational therapy interventions for pain, operationally defined using Current Procedural Terminology codes, were included (see Supplemental Table S1). We emphasized billable occupational therapy interventions because these are reimbursed by national payers such as the Center for Medicare and Medicaid Services and private insurers. Therefore, such interventions could readily translate to all practice settings. Interventions had to be provided within three months after surgery with a follow up period of six months or more to ensure temporality between the intervention and the outcome. These criteria were selected to focus our analysis on long-term outcomes of acute or subacute pain treatment. Continuous passive motion, interventions that focused solely on muscle training without functional activities, and physical agent modalities provided without functional training, were excluded to focus on therapeutic use of occupations and functional activities because occupation and activity are implied in the overall goal of occupational therapy (Boop et al., 2020). Although such interventions may be billable by practitioners and contribute to their intervention plan, we recognize some preparatory interventions are not distinct to occupational therapy and may be delivered by technicians, nurses, or other rehabilitation professionals, such as physical therapy practitioners or chiropractors. Outcomes included any measures of pain, function, and prescription opioid use. This broad definition allowed us to cast a wide net in our search and accounts for the expected variability in measurement standards across practice settings.

Data extraction

Method

Data from studies were extracted using the Oxford center for evidence-based medicine evidence table as a standardized method of retrieving and organizing information (Center for Evidence-Based Medicine, 2011). Reviewers extracted data from the text of the articles, then tables were cross-checked by a second reviewer. Discrepancies about data extraction were resolved through discussion.

Data items

Extracted data included clinical setting, adherence, specific intervention types, frequency, duration, and control or placebo used. Reviewers extracted patient demographics, including age, gender, and type of replacement. Outcomes at the latest follow up point were extracted, including participant retention and significant within and between group findings from measures of pain, function, or prescription opioid use.

Risk of bias in individual studies

Risk of bias was assessed using a standardized table (American Journal of Occupational Therapy, 2020; Shea et al., 2017). Risk of bias at both the study level and outcomes level were evaluated and reported as either “low,” “moderate,” or “high.”

Synthesis of results

Results were synthesized and the overall strength of evidence supporting interventions was evaluated using a strength of evidence table (American Journal of Occupational Therapy, 2020; USPSTF, 2018). This table synthesizes study design and risk of bias, based on intervention type, to rate the quality of evidence for that intervention as high, moderate, or low. We used the Template for Intervention Description and Replication (TIDieR) to characterize and describe interventions (Hoffmann et al., 2014).

Results

While a total of 4794 articles and 1 record in a registry were identified, the full-text screen resulted in seven articles for qualitative synthesis (Figure 1). Table 1 includes the level of evidence and study details. Intervention dosage, comparison groups, outcomes of pain and function, and practice setting varied extensively, preventing our ability to conduct a meta-analysis. Table 2 details the risk of bias for each article. Overall, each study had a low risk of bias except for Lenguerrand et al. (2020), which had a moderate risk of bias.

Figure 1.

PRISMA 2020 flow diagram for new systematic reviews which included searches of databases and registers only

*Consider, if feasible to do so, reporting the number of records identified from each database or register searched (rather than the total number across all databases/registers).

**If automation tools were used, indicate how many records were excluded by a human and how many were excluded by automation tools.

For more information, visit: http://www.prisma-statement.org/

Table 1.

Evidence table: a summary of study quality and design characteristics.

Author/Year	Level of Evidence Study Design Risk of Bias	Participants Inclusion Criteria Study Setting	Interventions and Control Groups	Outcome Measures	Results (including significance of findings)
Harmer et al., 2009	Level 1B Study Design Randomized single-blind trial Risk of Bias Low	Participants N=102 −1 patient died from water group −2 patients withdrew -No baseline differences Inclusion Criteria -Rescreened for eligibility following discharge from acute care Intervention Setting Subacute rehabilitation	Intervention: Land and water-based group exercise Control: Land-based group exercise	Function -WOMAC −6MWT -Stair climbing power	Significant Findings -Land-based and water-based exercise programs were associated with comparable improvements in timed walk distance, joint pain, and range of motion immediately following surgery and after 6 months Nonsignificant Findings -Hypothesis of water-based group doing better was not supported -Water medium results did not translate into perceptible clinical gains -Water based therapy elicited greater improvements in stair climbing power and edema resolution at 6 months post op, but not evident immediately after treatment secession.
Ko et al. 2013	Level 2B Study Design Randomized superiority trial Risk of Bias Low	Participants N=249 -lost 16 -participants were similar across three treatment groups Inclusion Criteria -Undergoing primary unilateral or simultaneous bilateral TKRs -Willing to take rehab to one of the participating hospitals Intervention Setting Four metropolitan hospitals	Intervention: Task oriented exercises Controls: Group therapy One-on-One monitored home program	Function -WOMAC -OKS −6MWT Pain -VAS	Significant Findings -One on one therapy is not superior to group or monitored home program in short term or long term -Cohort demonstrated similar self-reported and physical function recovery profiles Nonsignificant Findings -Client satisfaction rates were not significantly different between intervention and control groups.
Lenguerrand et al., 2019	Level 1B Study Design Multi-center, unblinded, RCT Risk of Bias Moderate	Participants N=180 (mean age 69 years) Inclusion Criteria Adults on waiting list for primary TKR due to osteoarthritis Intervention Setting Outpatient rehabilitation	Intervention: Task oriented exercises Control: Usual care by inpatient physical therapy	Function -KOOS -LEFS questionnaire	Significant Findings Intervention group scored higher on the LEFS questionnaire at 3, 6, and 12 months. Highest treatment effect was measured at 3 months. Nonsignificant Findings There were no measured differences in mean total KOOS scores, HADS anxiety or depression subscales, or in Patient Satisfaction Scale scores at 3, 6 or 12 months. No functional differences between groups at 12 months reached a “clinically meaningful level”.
Moffet et al., 2004	Level 1B Study Design Single-blind RCT Risk of Bias Low	Participants N=77 (mean age 66–68 years old) Inclusion Criteria -Diagnosis of OA of the knee, -Waiting for a first TKR -Live in the Quebec metropolitan area -Ambulatory with or without a walking aid Intervention Setting Outpatient and home exercise program	Intervention: Task oriented exercises Control: Usual Care	Outcome: Function −6MWT -WOMAC scores -SF-36 Pain: WOMAC	Significant Findings Participants in the IFR group walked significantly further in 6MWT at all 3 measurements. At implementation and 2 months, the IFR group had significantly less pain, stiffness, and difficulty in performing daily activities (WOMAC subscores). Nonsignificant Findings No difference was measured between groups in the 8 SF-36 Scales, except for role-physical at 2 months. There was no difference between groups in WOMAC scores at 8 months.
Monticone et al., 2013	Level 1B Study Design Randomized, parallel group, controlled superiority trial Risk of Bias Low	Participants N=110 (40 participants were male, mean age= 67 years old) Inclusion Criteria -Primary TKR 7–10 days prior to rehabilitation treatment -understood Italian - >50 years old Intervention Setting Home-based setting and highly specialized rehabilitation center	Intervention: Task oriented exercises Control Group: General discharge education	Function −42-item KOOS −13-item TSK Pain −11-point NRS	Significant Findings -There was a mean adjusted difference of 14.22 (95% confidence interval, 8.35–20.08) at t1, indicating a clinically significant difference between the intervention and control groups. -There was a treatment effect of about 25% -The intervention program of home-based functional exercises aimed at managing kinesiophobia was superior to general advice to stay active in terms of reducing disability, fear-avoidance beliefs and pain, and enhancing the QOL of patients with TKR discharged from a rehabilitation unit. -The results lasted for at least 6 months after the end of intervention. -Clients in the intervention group perceived significantly higher efficacy of treatment.
Monticone et al., 2014	Level 1B Study Design Randomized, parallel-group, controlled, superiority trial Risk of Bias Low	Participants N=100 (40 males, 60 females, mean age= 69 y/o) Inclusion Criteria -Primary THR 4–7 days before admission into rehabilitation hospital -Insufficient home support, additional medical care after surgery - >50 years old -understood Italian Intervention Setting Specialized rehabilitation clinic	Intervention: Task oriented exercises Control: Traditional rote exercise treatment	Function -FIM -WOMAC Pain −11-point NRS	Significant Findings -WOMAC: physical function score decreased by almost 50% in the intervention group and 20% in control. -Decrease in pain intensity (NRS) was more significant in the intervention group than control. -Improvement in FIM score was more significant in the intervention group than control. -That a program of inpatient task-oriented exercises associated with early full weight-bearing is superior to a program of traditional exercises associated with partial weight-bearing for clients with THR. -Clients in the intervention group perceived significantly higher efficacy of treatment.
Stevens-Lapsley et al., 2012	Level 1B Study Design RCT Risk of Bias Low	Participants N=66 (Age 50–58) Inclusion Criteria -Age 50–85 -Underwent tricompartmental TKR Intervention Setting Inpatient, home, outpatient therapy and NMES at home	Intervention: NMES Control: Standard rehabilitation	Function -TUG -SCT -WOMAC -Perception of knee functional ability on GRS -Knee AROM −6MWT Pain −11-point NRS	Significant Findings The addition of NMES treatment to the quadriceps muscles yielded benefits that were most pronounced in the first month after surgery but persisted for the first year. Between group outcomes remained significant for muscle strength, TUG, SCT, 6MWT, GRS, SF-36, MCS, and WOMAC scores.

6MWT=Six Minute Walk Test, ADL= Activities of Daily Living, CST=30-second chair stand test (Rikli & Jones, 1999), FIM= Functional Independence Measure, GRS=Knee Functional Ability Global Rating Scale, (Irrgang et al., 1998), HOOS = Hip Dysfunction and Osteoarthritis Outcome Score (Nilsdotter et al., 2003), KOOS=Knee Injury and Osteoarthritis Outcome Score, LEFS=Lower Extremity Functional Scale (Binkley et al., 1999), NMES= neuromuscular electrical stimulation, NRS=Numeric Rating Scale, OKS=Oxford Knee Score (Dawson et al., 1996), QOL= Quality of Life, RCT= randomized control trial, ROM= range of motion, SCT=Stair Climb Test (Shields et al., 1995), SF-36=36 item Short-Form Health Survey, SF-12=12 item health survey (Ware et al., 1996), THR= total hip replacement, TKR=total knee replacement, TSK=Tampa Scale for Kinesiophobia, TUG=Timed Up and Go, WOMAC=Western Ontario and McMaster Universities Osteoarthritis Index,

Table 2.

Risk of bias: a summary of each study’s risk for major types of methodological bias.

	Selection bias			Performance bias		Detection bias			Reporting bias
Citation	Random Sequence Generation	Allocation concealment (until participants enrolled and assigned)	Baseline differences between intervention groups (suggest problem with randomization?)	Blinding of participants during the trial	Blinding of personnel during the trial	Blinding of outcome assessment Self-reported outcomes	Blinding of outcome assessment Objective outcomes (assessors aware of intervention received?)	Incomplete Outcome data (data for all or nearly all participants	Selective reporting (results being reported selected on the basis of the results?)	Overall risk of bias assessment (low, moderate, high risk)
Harmer et al., 2009	+	+	+	+	-	+	-	+	-	Low
Ko et al. 2013	+	+	+	+	-	+	+	+	+	Low
Lenguerrand et al., 2019	+	+	+	-	-	-	-	+	+	Moderate
Moffet et al., 2004	+	+	+	-	-	+	+	+	+	Low
Monticone et al., 2013	+	+	+	+	-	-	+	+	+	Low
Monticone et al., 2014	+	+	+	+	-	-	+	+	+	Low
Stevens-Lapsley et al., 2012	+	+	+	-	-	+	-	+	+	Low

Note: “+” = low risk of bias; “-“ = high risk of bias

To answer our first research question, the characteristics of interventions fell into three categories: task-oriented exercises, water-based exercises, and physical agent modalities. The TIDieR descriptive overview of the procedures for each intervention type is presented in Supplemental Table S2–S4. Task-oriented exercises (S2) varied, but all could be delivered by an occupational therapist and were defined as a goal-directed therapeutic activity or functional exercise. For example, interventions included knee-specific and lower limb functional exercises (Moffet et al., 2004), kinesiophobia management education (Monticone et al., 2013), and strategies for performing activities of daily living with full weight-bearing on the affected leg (Monticone et al., 2014). Water exercise (S3) included walking, step-ups, jogging, jumping, kicking, knee range-of-motion exercises, lunges, and combined squats and upper extremity exercises. Modalities intervention (S4) included neuromuscular stimulation (NMES) at home in conjunction with therapeutic exercises and functional mobility training. No studies meeting our inclusion criteria were provided by occupational therapy practitioners.

Improvements in function and pain

Following TKR, only Monticone et al. (2013) found significantly better scores for the task-oriented intervention group after 1 year, compared to the control group, using the Knee Injury and Osteoarthritis Outcome Score (KOOS) (Roos et al., 1998), Tampa Scale for Kinesiophobia (TSK) (Miller et al., 1991), Numerical Rating Scale (NRS) (Jensen et al., 1986), and Short-Form Health Survey (SF-36) (Stewart & Ware, 1992).

Following THR, Monticone et al. (2014) found significant one-year improvements in the task-oriented intervention group’s Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) (Bellamy, 2005), NRS of pain intensity, and Functional Independence Measure score (Granger et al., 1986) compared to the control group.

Improvements in function

Following TKR, Stevens-Lapsley et al. (2012) reported that the modality-based NMES intervention produced significantly better outcomes on the WOMAC, Timed Up and Go (Podsiadlo & Richardson, 1991), and 6 Minute Walk Test (6MWT) (Butland et al., 1982) compared to the control. An 11-point NRS scale was used to measure pain, but these results were not reported.

Non-significant findings

Three studies on task-oriented interventions found no statistically significant improvements in pain or function one year after TKR (Ko et al., 2013; Lenguerrand et al., 2020; Moffet et al., 2004). Similarly, water-based exercise after TKR was not associated with six-month improvements in the WOMAC, 6MWT, or stair climbing power. As a secondary outcome, we considered whether complementary occupational therapy interventions could mitigate the use of high-risk prescription opioids. No studies measured prescription opioid use as an outcome.

Discussion

This review characterized types of nonpharmacologic interventions for pain and function. We also identified which intervention types demonstrated efficacy during a follow-up of at least six months. Finally, we assessed whether any interventions were used to reduce prolonged opioid use. Three types of potential occupational therapy interventions were identified for the early phases of THR or TKR recovery: (1) task-oriented exercise, (2) water-based therapy, and (3) physical agent modalities. Overall, the strength of evidence for each intervention was low to moderate and the risk of bias was low. These findings build on prior occupational therapy literature that studied chronic pain, because we identified interventions used for acute and subacute pain that produced long-term benefits (Hesselstrand et al., 2015; Lagueux et al., 2018; Simon & Collins, 2017).

Regarding types of interventions, a surprising result was the lack of interventions using behavioral approaches. Prior occupational therapy literature on chronic pain interventions emphasized behavioral and self-management strategies (Simon & Collins, 2017). Our finding could be attributed to the fact that no occupational therapy practitioners were used in the delivery of interventions. Still, the growing need for psychological and behavioral pain interventions is well-documented (National Pain Strategy Task Force, 2015; National Academies of Sciences, Engineering, and Medicine, 2019).

The efficacy of interventions varied. Many intermediate and long-term outcomes from the intervention groups were not statistically different from control groups. The presence of previously reported short-term benefits (Skelly et al., 2020) indicates a possible plateau in recovery after participants reach a certain level of pain and/or function (Rimmer, 2012). The possibility of plateaued recovery should encourage health care providers to shift pain interventions toward the acute and subacute stages of post-surgical rehabilitation. Although earlier interventions may not overcome this plateau, future research may consider whether early interventions help patients recover sooner. In the context of policy and practice implications, providing long-term pain interventions may be a developing role for acute care practitioners because bundled payment initiatives are decreasing post-acute care utilization following THR and TKR (Pritchard et al., 2019).

Long-term benefits of task-oriented interventions were found in two studies measuring outcomes at 1-year follow-up (Monticone et al., 2013; 2014). These findings indicate long-term benefits from task-oriented interventions used in home-based and outpatient settings. In the context of theory-based clinical practice implications, it is possible that task-oriented interventions are readily integrated into patients’ occupational performance. Using a multi-model approach, practitioners can rely on patients’ values to integrate interventions into familiar performance patterns within a variety of environmental contexts. For example, a multi-model approach may use the Model of Human Occupation (Kielhofner, 2008) to guide a biomechanically focused task-oriented intervention (Wong & Fisher, 2015) in acute care or home health. Evidence was lacking to support water-based interventions or physical agent modalities as methods to improve long-term function and pain.

In the context of occupational therapy research, an important finding was the lack of research that considered prescription opioid use as an outcome. This finding was surprising as more than 22% of adults with pain use prescription opioids despite policy and practice initiatives aimed at reducing this percentage (Dahlhamer et al., 2021). In the United States, skilled therapy is underutilized compared to prescription opioids in treating pain (Neprash & Barnett, 2019). Underutilization of occupational therapy is concerning because adults with post-surgical pain may have limited opportunities to access nonpharmacologic alternatives such as occupational therapy as inpatient post-acute care becomes less common following joint replacement (Navathe et al., 2017). In addition to barriers to accessing occupational therapy interventions for pain, there is limited evidence demonstrating the effectiveness of occupational therapy as an intervention for acute or subacute pain capable of producing lasting benefits. Small trials and large observational studies show that nonpharmacologic pain interventions, such as occupational and physical therapy, improve quality of life and reduce high-risk opioid consumption (Frogner et al., 2018; Pritchard et al., 2021; Simon & Collins, 2017). However, randomized control trials are needed to demonstrate the role of early occupational therapy in mitigating pain while improving function, particularly for adults over 65 years, the only age group with an increasing rate (10.5%) of prescription opioid-related mortality (Scholl et al., 2019).

Strengths

This review examined a wide range of occupational therapy interventions used to treat pain, function, and prescription opioid use. Our methodology followed PRISMA guidelines, and we followed the methodology in our registered proposal that was pre-published to minimize bias. A deviation from our original proposal was the addition of the TIDieR template, which summarized qualities of each intervention type (Supplemental Table S2–S4). Included studies showed minimal risk of bias and contained high levels of evidence. These findings are highly relevant to practitioners, educators, and researchers because the use of nonpharmacologic pain interventions is a priority research area for major medical institutions, such as the National Academies of Science and CDC.

Limitations

This systematic review was limited to studies in the English language. Gray literature was not included, and it is possible that studies with null results were never fully published, which may present publication bias (Franco et al., 2014). Excluding studies without the therapeutic use of occupations or activities increased feasibility and internal validity, but this exclusion criteria may have overlooked beneficial preparatory interventions that could be provided by interdisciplinary peers. It is possible that we may have excluded relevant articles published prior to 1995. Our sample of only seven studies—coupled with heterogeneity of surgery type, intervention design and dosage, practice setting, and outcome measures—limited our ability to conduct a meta-analysis. This heterogeneity indicates a need for stratified analysis to account for confounding, which was not possible given our small sample. Only one study analyzed non-steroidal anti-inflammatory drug consumption as an outcome measure and none analyzed opioid consumption (Moretti et al., 2012). Although occupational therapy practitioners can bill for CPT codes used in our inclusion criteria (Supplemental Table S1), information bias is possible, as other professions including physicians, advanced practice providers, and physical therapists may bill for these interventions as well.

Future research

Future systematic reviews may build on this work by expanding the inclusion criteria to include high quality observational studies in realworld settings. Expanding inclusion criteria may allow meta-analyses to be more selective in examining specific outcomes based on psychometrics, conceptual definitions, and applicability to practice while controlling for study heterogeneity. We identified a critical research gap as we found no studies with long-term follow-up that relied on occupational therapy practitioners to deliver acute and subacute pain interventions. Furthermore, no studies considered prescription opioid use as an outcome. Comparative effectiveness and cost-effectiveness research will be needed to further characterize the role of occupational therapy practitioners in pain management.

Conclusions

In conclusion, we found that three types of acute and subacute interventions—task-oriented, water-based, and modalities—were used to improve long-term outcomes of pain and function in adults after THR or TKR. Task-oriented interventions improved function and pain at 1-year follow-up. Task-oriented interventions are goal-directed and readily integrated into patients’ performance patterns across any environmental context, which may increase motivation, improve adherence to treatment plans, and instill behavior change, as seen in prior chronic pain literature (Hesselstrand et al., 2015; Lagueux et al., 2018; Simon & Collins, 2017). One modality intervention improved function but not pain. Water-based interventions were not associated with improvements in either pain or function. While the findings identified interventions that occupational therapy practitioners could use for acute and subacute pain, none of the evidence demonstrated occupational therapy practitioner’s distinct role in providing these interventions. This finding is concerning as occupational therapy practitioners regularly evaluate and provide interventions for adults with acute and subacute pain; however, there is a lack of evidence-based research to guide practice.

Abbreviations:

KOOS

Knee injury and Osteoarthritis Outcome Score

NMES

Neuromuscular stimulation

NRS

Numerical Rating Scale

OT

Occupational Therapy

OTP

Occupational Therapy practitioners

RCTs

Randomized control trials

PRISMA

Preferred Reporting items for Systematic Reviews and Meta-Analyses

SF-36

Short-Form Health Survey

TSK

Tampa Scale for Kinesiophobia

CDC

The Centers for Disease Control and Prevention

TiDieR

Template for intervention Description and Replication

THR

Total Hip Replacement

TKR

Total Knee Replacement

WOMAC

western Ontario and McMaster Universities Osteoarthritis index

6MwT

6 Minute walk Test