Disability and Rehabilitation

ISSN: 0963-8288 (Print) 1464-5165 (Online) Journal homepage: http://www.tandfonline.com/loi/idre20

Rehabilitation after total joint replacement: a

scoping study

Deborah L. Snell, Julia Hipango, K. Anne Sinnott, Jennifer A. Dunn, Alastair

Rothwell, C. Jean Hsieh, Gerben DeJong & Gary Hooper

To cite this article: Deborah L. Snell, Julia Hipango, K. Anne Sinnott, Jennifer A. Dunn, Alastair
Rothwell, C. Jean Hsieh, Gerben DeJong & Gary Hooper (2017): Rehabilitation after total joint
replacement: a scoping study, Disability and Rehabilitation, DOI: 10.1080/09638288.2017.1300947

To link to this article: http://dx.doi.org/10.1080/09638288.2017.1300947

Published online: 23 Mar 2017.

Submit your article to this journal

Article views: 8

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=idre20

Download by: [Fudan University] Date: 25 March 2017, At: 05:45

DISABILITY AND REHABILITATION, 2017
http://dx.doi.org/10.1080/09638288.2017.1300947

REVIEW ARTICLE

Rehabilitation after total joint replacement: a scoping study

Deborah L. Snella,b, Julia Hipangoa, K. Anne Sinnottb, Jennifer A. Dunna, Alastair Rothwella, C. Jean Hsiehc,d,
Gerben DeJongd,e and Gary Hoopera
a
Department of Orthopaedic Surgery and Musculoskeletal Medicine, University of Otago Christchurch, Christchurch, New Zealand; bBurwood
Academy of Independent Living (BAIL), Burwood Hospital, Christchurch, New Zealand; cMedStar Health Research Institute, Washington, DC,
USA; dMedStar National Rehabilitation Hospital, Washington, DC, USA; eDepartment of Rehabilitation Medicine, Georgetown University School
of Medicine, Washington, DC, USA

ABSTRACT ARTICLE HISTORY

Purpose: The evidence supporting rehabilitation after joint replacement, while vast, is of variable quality Received 16 August 2016
making it difficult for clinicians to apply the best evidence to their practice. We aimed to map key issues Revised 16 February 2017
for rehabilitation following joint replacement, highlighting potential avenues for new research. Accepted 26 February 2017
Materials and methods: We conducted a scoping study including research published between January
2013 and December 2016, evaluating effectiveness of rehabilitation following hip and knee total joint KEYWORDS
replacement. We reviewed this work in the context of outcomes described from previously published Total hip replacement; total
research. knee replacement;
Results: Thirty individual studies and seven systematic reviews were included, with most research examin- rehabilitation; exercise-
ing the effectiveness of physiotherapy-based exercise rehabilitation after total knee replacement using based physiotherapy;
randomized control trial methods. Rehabilitation after hip and knee replacement whether carried out at home-based exercise
the clinic or monitored at home, appears beneficial but type, intensity and duration of interventions were programs; therapeutic
not consistently associated with outcomes. The burden of comorbidities rather than specific rehabilitation attention; therapy alliance
approach may better predict rehabilitation outcome. Monitoring of recovery and therapeutic attention
appear important but little is known about optimal levels and methods required to maximize outcomes.
Conclusions: More work exploring the role of comorbidities and key components of therapeutic attention
and the therapy relationship, using a wider range of study methods may help to advance the field.

ä IMPLICATIONS FOR REHABILITATION

Physiotherapy-based exercise rehabilitation after total hip replacement and total knee replacement,
whether carried out at the clinic or monitored at home, appears beneficial.

Type, intensity, and duration of interventions do not appear consistently associated with outcomes.

Monitoring a patient’s recovery appears to be an important component. The available research pro-
vides limited guidance regarding optimal levels of monitoring needed to achieve gains following hip
and knee replacement and more work is required to clarify these aspects.

The burden of comorbidities appears to better predict outcomes regardless of rehabilitation
approach.

Introduction However 15–30% of knee replacement patients and a smaller

number of hip replacement patients report little or no improve-
Joint replacement is the most common and cost-effective
ment usually because of ongoing pain, restricted range of motion,
elective surgical intervention for end-stage degenerative joint dis-
reduced function and low quality of life [1,2,6,7]. A number of
ease such as osteoarthritis [1,2] and surgery rates are escalating demographic and clinical factors have been shown to influence
worldwide [3]. In New Zealand, the rates of primary hip and knee these outcomes including age, gender, general health and comor-
replacement have doubled and tripled respectively over the past bidities, pain, weight, post-operative complications, surgical wait
15 years and these rates continue to rise as the population ages time, and rehabilitation [6,8–10].
[4]. Osteoarthritis is the primary diagnosis leading to hip and knee Rehabilitation, with its emphasis on physiotherapy and exer-
replacement and accounts for up to 81% of hip and 94% of knee cise, is widely encouraged after hip and knee replacement surgery.
replacements [5,6]. Rehabilitation seeks to help individuals regain and maintain phys-
The long-term outcomes for people with osteoarthritis follow- ical, sensory, intellectual, psychological, and social function [11].
ing primary joint replacement are generally favorable with most However, approaches to rehabilitation vary and rehabilitation is
experiencing positive functional outcomes [6,7]. Prospective, complex and multi-dimensional, including but not limited to vari-
observational studies have shown the greatest functional gains ous physiotherapy approaches, occupational therapy, psycho-
take place within the first 6 months following hip replacement logical support, education, and assistive technology. Influencing
and the first 12 months following knee replacement [6]. factors include choice of outcomes and outcome measures, the

CONTACT Deborah Snell debbie.snell@otago.ac.nz University of Otago Christchurch, New Zealand

ß 2017 Informa UK Limited, trading as Taylor & Francis Group
2 D. L. SNELL ET AL.

timing, intensity, and type of intervention as well as the nature of the Boolean operator ‘AND.’ Searching was an iterative process
the setting (e.g., inpatient versus outpatient, funding differences). moving back and forth between search terms to capture relevant
There are a number of recently published systematic reviews research. Search limits were applied including studies published in
evaluating effectiveness of rehabilitation after hip and knee joint English, full text availability and human participants. We recognize
replacement [12–16]. These reviews summarize the literature pub- that including studies published only in English may result in
lished to 2013 and involved over 6400 patients. Invariably some relevant literature being overlooked.
although, review authors conclude that the quality of the research
Inclusion criteria: Publications were included if (1) a full report
is poor overall and comparison across studies is difficult because
was available; (2) the study was written in English; (3) participants
of significant methodological variability. Evaluating the effective-
had undergone THR or TKR as a result of osteoarthritis; (4) all
ness of rehabilitation after joint replacement is further compli-
patients were adults; (5) post-operative outcome measures were
cated by inconsistency across studies in relation to the various
used; (6) a relevant post-operative intervention was trialed; and
rehabilitation dimensions described above. For example, the type,
(7) pain, physical function and/or quality of life were assessed as
timing and duration of rehabilitation, choice of outcomes, and
outcomes. All study designs were included. A wide view of
outcome measures vary across studies, meaning study to study
rehabilitation was adopted for the scoping study and an interven-
comparisons are difficult.
tion was considered relevant for this review if this was described
It may be time to step back and reconsider the way rehabilitation
by study authors as a rehabilitation intervention regardless of set-
is evaluated in this population, using a wider range of methods and
ting (e.g., clinic, home, community setting); did not solely involve
tools. To better understand the breadth of the topic and assess
surgical procedures or enhancements; or pharmacological
inconsistencies and gaps in the research, a scoping study may be a
interventions.
helpful starting point. The scoping study is one method amongst
Exclusion criteria: Articles were excluded if rehabilitation inter-
many, to review a body of research. A scoping study examines the
ventions were medical interventions (e.g., pharmacological), or
extent, range, and nature of research on a topic and specifically
conducted on children or animals, revision patients, and with pri-
seeks to identify gaps in the research evidence [17,18]. This method
marily non-osteoarthritic patients. Papers were excluded along
entails several key iterative steps: specifying the research question,
with protocol reports and conference proceedings if insufficient
identifying relevant studies, selecting studies for closer examination,
information was available to evaluate the study.
charting and collating data, summarizing results, and reporting con-
clusions. A scoping study differs from a systematic review in that
the latter generally focuses on a narrowly defined question, specific Selection of relevant studies
research methods such as randomized controlled trials, and in par-
Two researchers (J. H. and D. S.) applied the inclusion and exclu-
ticular rigorous assessment of study quality. In comparison, a scop-
sion criteria to all titles and abstracts. Copies of full articles were
ing review seeks to summarize the literature within a defined set,
obtained for those studies that appeared to represent a “best fit”
regardless of quality, in order to examine the range of studies that
with the research question. If the relevance of a study was unclear
exist and consistency of the findings [17].
from the title and abstract, then the full article was retrieved.
Our objective was to conduct a scoping study to map the
breadth and depth of research examining the effectiveness of
Charting and collating data, summarizing results, and reporting
rehabilitation after primary hip and knee replacement. We aimed
conclusions
to build on findings of previously published systematic reviews,
We used a data-charting form based on that described by Arskey
update, and highlight gaps in the evidence, and identify new
and O’Malley [17] to collate the data. Data charted included
research avenues with potential to advance the field.
author(s), year of publication and study location; description of
study sample; intervention type; study aims; methods; outcome
Methods measures; important results; and study limitations. Consistent with
our choice of method (scoping study), we made no systematic
Study design
attempt to evaluate the weight or quality of the evidence. Having
A scoping study was used to map the key issues for research eval- charted information from included studies, we conducted a narra-
uating the effectiveness of rehabilitation following total hip tive analysis of included studies, highlighting key themes and
replacement (THR) and total knee replacement (TKR). This involved issues. First, basic numerical analysis was undertaken of the
the following steps: extent, nature and distribution of the studies included in the
review. Second tables and charts mapped the distribution of stud-
The research question ies geographically; the range of interventions; the research meth-
The aim of this scoping study was to identify what is known from ods adopted and outcomes. Systematic reviews were included
the existing scientific literature about the effectiveness of rehabili- and review findings were integrated with results from individual
tation after THR and TKR for people with osteoarthritis. studies in order to provide a more complete overview of the evi-
dence. Third, a narrative summary of included studies was com-
Identification of relevant studies and study selection pleted where we looked for similarities and inconsistencies across
Five electronic databases (EMBASE, MEDLINE, Pubmed, Cinhahl, study interventions and outcomes, to present a descriptive over-
and Web of Science) were searched between January 2013 and view of the review findings. Results have been organized and
December 2016 for relevant research. We initiated searching from summarized by joint replaced because research consistently shows
2013 because studies evaluating the effectiveness of rehabilitation differing recovery timelines following THR and TKR [6].
after joint replacement published up to 2013 have been thor-
oughly summarized by many systematic reviews. Search terms
Results
included knee replacement, hip replacement, arthroplasty,
joint replacement, combined with rehabilitation, exercise, physio- Initial searching for relevant articles yielded over 2000 hits. This
therapy, physical therapy, occupational therapy, education using was reduced to 1033 studies once additional search terms were
 REHABILITATION AFTER JOINT REPLACEMENT 3

applied and articles were screened for relevance based on their outcomes such as increased activity levels; and pain reduction. A
title and abstracts. Another 873 studies were excluded when inclu- small number of studies measured impact of interventions on
sion and exclusion criteria were applied to the abstracts. After full quality of life [22,26,40]. Three studies considered psychosocial
text or abstract review of the remaining 160 studies, 30 studies factors such as fear and anxiety [31,40] and self-efficacy [27].
and seven systematic reviews were retained and included in the Choice of outcome measures varied considerably across studies
review (see Figure 1). (see Tables 1–5). The Western Ontario and McMaster Osteoarthritis
Index was the most consistently used across studies with 17 stud-
ies including this patient reported outcome measure[19,22,24,
Overview of extent, nature, and distribution of included studies
28–30,33–37,39,41,42,44,49,50]. Other measures included either
The majority of the 30 included individual studies (n ¼ 21) were the 12-item or 36-item versions of the Medical Outcomes Short
randomized controlled trials (RCTs), with two of these being pilot Form Questionnaire (10 studies) [19,22,26,28,33,35,38–40,44];
studies (see Tables 1–5). Two studies used non-RCT methods. The Oxford Hip or Knee Scores (four studies) [24,32,33,38]; Hip
remainder of the included studies were prospective observational Disability and Osteoarthritis Outcome Score/Knee Injury and
studies (n ¼ 4), a retrospective cohort study (n ¼ 1), and a case ser- Osteoarthritis Outcome Score (six studies) [21,22,27,32,40,45]; Knee
ies (n ¼ 1). The tables also show the geographical distribution of Society Score (two studies) [34,35]; Visual Analogue Scale for pain
included studies. The seven systematic reviews only included RCTs or anxiety (five studies) [23,31,32,35,46]; range of movement
as expected. measures (13 studies) [27–31,34,37,39,41,43,45,46,50], and various
Eight individual studies [19–26] and three systematic reviews additional measures of walking speed and muscle strength.
[12–14] evaluated rehabilitation following THR; 20 studies [27–46]
and four systematic reviews [15,16,47,48] focused on TKR and two
Intervention type and outcomes following total hip replacement
studies [49,50] included both THR and TKR participants (see Tables
(Tables 1 and 2)
1–5). The RCTs compared an intervention with usual or standard
care, or two different physiotherapy approaches or settings for Three RCTs evaluating interventions following THR compared
example home-based exercises versus clinic-based outpatient ther- active clinic-based physiotherapy interventions with usual care
apy. Most physiotherapy interventions included components such [19,21,23]. Interventions involved 10–12-week gait retaining,
as walking skills and gait training, range of movement and muscle strengthening, and progressive resistance training programs. Usual
strengthening, cycle ergometer training, continuous passive care involved home-based exercise programs supplemented by
motion, and progressive resistance training. Two studies [26,35] community sessions and varying levels of follow up. There were
evaluated educational interventions and one study [46] examined no between group differences reported in two of these studies
the effectiveness of adding a traditional Chinese medicine [19,21], with all patients benefitting from rehabilitation. The third
approach to standard rehabilitation. Usual care control interven- study [23] compared a home-based walking training program with
tions were often vaguely described for example as standard out- fortnightly clinic visits, with a control condition where participants
patient physiotherapy or standard rehabilitation. were not given a program but attended fortnightly clinic visits
Outcomes across studies were typically physical outcomes such over 12 weeks, in 28 women after THR. In this study, the interven-
as improved range of movement, functional capacity (e.g., walking tion group showed greater improvement in pain, function and
speed, distance, and endurance), muscle strength; functional energy expenditure compared with controls. Very small samples

Figure 1. Selection of studies for inclusion in scoping study.

 4
Table 1. Studies included in the review by location, methods, sample size, type of rehabilitation, and outcomes after total hip replacement (n ¼ 8).
Study and location Sample characteristics Study methods Rehabilitation type(s) Outcome Measures Outcomes
Beaupre et al. (2014) [19] N ¼ 21 Single blind randomized Outpatient strengthening, gait retraining, þ WOMAC; Rand 36 item Health Both groups improved – no
Canada Mean age: 53.4 pilot study daily home exercises (intervention) for 3- Survey at end point of 12- between group differences
Female: 47.6% months after THR compared with usual months post-operation.
care (home exercise program for 3
months þ4–6 community based
rehabilitation sessions). (Pilot)
D. L. SNELL ET AL.

Judd et al. (2016) [20] N ¼ 10 Case series Five participants in an 8-week, 2 weekly Stair climb test, 4-meter walk All participants showed
United States Mean age: 60.3 exercise program following THR, empha- test, the 6-minute walk test, changes and improvements
Female: 50% sizing targeted neuromuscular reeduca- and the Fullerton Advanced but more participants in the
tion techniques hallmarked by specific, Balance Scale to assess in the neuromuscular reedu-
weight-bearing exercise to improve hip physical function cation program showed
abductor performance and pelvic stabil- changes across measures
ity. Five additional participants were
supervised and followed for comparison
with a home program

Mikkelsen et al. (2014) [21] N ¼ 62 RCT Rehabilitation intervention compared with Primary outcome ¼ change in Both groups improved with no
Denmark Mean age: 65.0 usual care after THR. Intervention group leg extension at 10-week significant between group
Female: 41.9% received 1:1 physio supervision at a follow up; secondary outco- differences
community center twice a week for 10 mes ¼ isometric hip muscle
weeks focusing on progressive resistance strength; sit-to-stand test;
training þ home-based exercises. Usual stair climb test; 20 m walk-
care involved 10 weeks of home-based ing speed; HOOS
exercises. All patients were seen at 4
and 10 weeks by a physiotherapist
Monticone et al. (2014) [22] N ¼ 100 RCT Inpatient-based task-oriented exercises with WOMAC; Pain Numerical Intervention group showed
Italy Mean age: 69.2 early full weight bearing (intervention) Rating Scale; Functional greater improvements in
Female: 60% compared with exercise and partial Independence Measure; pain, activity levels and
weight bearing (control) following THR SF-36 at 12-months quality of life and gains sus-
post-operation tained at 12 months post-
operation (p < 0.01)
Morishima et al. (2014) [23] N ¼ 28 Randomised pilot study Interval walking training with close moni- Isometric knee extension and Intervention group showed
Japan Mean age: 60.1 toring compared with usual care in flexion forces, VO2peak, and greater improvement in
Female: 100% women after THR (pilot). Usual care anaerobic threshold; VAS for pain, function and energy
included advice to remain as active as pain. expenditure (p < 0.05).
usual with follow up sessions every two
weeks during 12 week training period.
Okoro et al. (2013) [24] N ¼ 50 POCS Monitored home based exercise program WOMAC (function scale); OHS All patients improved. Pre-op
United Kingdom Mean age: 64.9 (intervention) for 6 weeks after THR at 12-months post- status best predictor of
Female: 57.1% compared with standard rehabilitation operation post-op outcome
(standard outpatient physiotherapy)
Rapp et al. (2015) [25] N ¼ 59 POCS Patients following THR compared with age Gait symmetry and walking Both groups showed improve-
Germany Mean age: 65.7 matched healthy controls after patients speed at an average of 27 ments in walking speed
Female: 48.3% had 4 weeks of inpatient rehabilitation. days post-operation (p < 0.01) but THR patients
Healthy controls participated in a com- slower to improve than con-
munity based exercise program for 4 trols (p < 0.01)
weeks
Umpierres et al. (2014) [26] N ¼ 106 RCT Verbal instruction and education regarding Goniometry and muscle Intervention group improved
Brazil Mean age: 61.4 exercise program (controls) compared strength assessment; SF-36; more than controls regard-
Female: 53.8% with verbal instruction and educa- Merle d’Aubign
e and Postel ing functional capacity,
tion þ daily physiotherapy supervised Scores at 15 d post- mobility, muscle strength,
exercise practice (intervention) for 15 d operation pain and quality of life
following THR (p < 0.05)
THR: total hip replacement; POCS: prospective observational cohort study; RCT: randomized control trial; WOMAC: Western Ontario and McMaster Universities Arthritis Index; OHS: Oxford Hip Score; HOOS: Hip Disability
and Osteoarthritis Outcome Score.
 REHABILITATION AFTER JOINT REPLACEMENT 5

Table 2. Systematic reviews included in the review by methods and sample size, type of rehabilitation, and outcomes after total hip replacement (n ¼ 3).
Number of studies Type of rehabilitation interventions
Study and timeline for searching and participants Study methods compared Findings/conclusions
Coulter et al. (2013) [12] 5 studies Systematic review Post-operative physiotherapy- Physiotherapy-directed rehabilita-
Searched to March 2012 N ¼ 234 and meta- directed exercise programs fol- tion was similarly effective
analysis lowing THR delivered either whether exercises were per-
supervised in an outpatient cen- formed unsupervised at home or
ter or unsupervised at home supervised at an outpatient
center
Small sample sizes and unclear/vari-
able levels of supervision in the
home and clinic settings across
studies limited confidence in
results
Di Monaco et al. (2013) [13] Nine studies Systematic review Early rehabilitation following THR Insufficient evidence to build an
Searched Jan 2008-Dec 2012 N ¼ 835 to determine which type and evidence-based protocol for
timing of exercise was most rehabilitation following THR
effective because of the low quality of
included studies. Some support
for ergometer cycling and resist-
ance strength training in the
early postoperative phase, and
weight bearing exercises
Minns Lowe et al. (2015) [14] 11 studies Systematic review Compared outcomes for those who Physiotherapy interventions target-
Searched January N ¼ 506 received post-THR rehabilitation ing walking and muscle
2007–November 2013 with those who did not strengthening may confer
benefit
Quality of trials, while improving,
remains low and guidelines and
clear recommendations about
specific physiotherapy
approaches following THR
remain elusive
THR: total hip replacement.

(range n ¼ 21–62) and inconsistent choice of outcome measures establish whether post discharge physiotherapy exercise was
across these studies make study to study comparisons difficult. effective after THR because of the low quality of included studies.
Umpierres et al. [26], using an RCT design, compared education The systematic review and meta-analysis by Coulter and col-
coupled with active physiotherapy-led exercise (the intervention) leagues [12] examined the efficacy of home based, unsupervised
with education alone for 106 participants and reported the inter- exercise programs versus physiotherapy-directed programs. Five
vention group demonstrated significantly better activity levels, studies were included and findings showed similar gains whether
pain control and quality of life than the controls 15 days post- exercises were performed unsupervised at home or supervised at
operation. In the only other trial to demonstrate greater improve- the clinic. Finally, the review by Di Monaco et al. [13] examined
ments in the intervention group, Monticone et al. [22] compared timing of exercise-based rehabilitation following THR, including
early full weight bearing coupled with task-oriented exercises, nine studies. Like the review by Minns Lowe et al., while there
with exercises and partial weight bearing, in a sample of THR was some support for early interventions targeting weight bearing
patients (n ¼ 100). Intervention participants demonstrated signifi- and strength training, no clear conclusions regarding the efficacy
cantly better and sustained improvements in pain, activity levels of early exercise after THR could be made because of the poor
and quality of life 12 months’ post-operation. quality of the evidence.
Okoro et al. [24] conducted a prospective observational study
following 50 patients who participated in a 6-week monitored
Intervention type and outcomes following total knee
home exercise program. The authors reported all patients
replacement (Tables 3 and 4)
improved but pre-operative status was the best predictor of out-
come. Judd et al. [20] described a series of 10 cases, five under- Twenty individual trials and four systematic reviews involving 63
going an 8-week, twice weekly exercise program involving RCTs evaluating rehabilitation interventions after TKR were
neuromuscular re-education and weight-bearing exercise, and five included. Of the 63 studies included in the systematic reviews,
were supervised with a general exercise-based home program. three studies identified by our search [27,40,41] overlapped with
They reported that all participants made improvements but the review by Artz et al. [47] and two studies [34,36], overlapped
gains were more marked for the neuromuscular re-education with the review by Pozzi et al. [16]
participants. Ten of the individual studies evaluated a home-based exercise
Three systematic reviews [12–14] evaluated effectiveness of program compared with outpatient physiotherapy or no interven-
rehabilitation after THR (see Table 2). All 25 studies included tion [28,29,33–35,37,38,40,44,45]. Where a trial involved a compari-
across these reviews were published before 2013 and were RCTs. son between a home-based exercise program and outpatient
The systematic review by Minns Lowe et al. [14] evaluated effect- clinic-based physiotherapy [28,29,33,34,37,38,45], there were no
iveness of exercise-based physiotherapy rehabilitation following between group differences and all groups benefitted from inter-
hospital discharge. Eleven studies were included and while there ventions provided. Where a home-based exercise program was
was some support for interventions targeting walking and muscle compared with no intervention or advice [35,40,44], intervention
strengthening, the authors concluded it was not possible to groups demonstrated greater gains than controls with respect to
 6
Table 3. Studies included in the review by location, methods, sample size, rehabilitation type and outcomes after total knee replacement (n ¼ 20).
Study and location Sample characteristics Study methods Rehabilitation type(s) Outcome measures Outcomes
B€
uker et al. (2014) [28] N ¼ 34 RCT Physiotherapy exercise sessions 5 WOMAC; ROM; SF-36; BDI at Both groups improved with no
Turkey Mean age: 66.2 week for 4 weeks compared with 10 weeks post-operation between group differences
Female: 91.2% monitored home exercise program
for 4 weeks (1 week ses-
sion þ home program). All patients
had same inpatient intervention.
D. L. SNELL ET AL.

Exercise programs focused on

range of movement, strengthening
following TKR
Brunn-Olsen et al. (2013) N ¼ 57 RCT Walking skills group based training Primary outcome ¼6 min walk Improved walking outcomes
[27] Mean age: 69.0 compared with usual care (individ- test at 9-months post-oper- for intervention group sus-
Norway Female: 56.1% ual physiotherapy care) after TKR ation; secondary outco- tained at 9 months post-op
mes ¼ timed stair climbing, (p ¼ 0.02). No other signifi-
timed stands, Figure-of- cant between group
eight test, Index of muscle differences
function, active knee ROM,
KOOS
Han et al. (2014) [29] N ¼ 390 Multi-center RCT Monitored home-based exercise pro- WOMAC at 6-weeks post-oper- Both groups improved with no
Australia Mean age: 64.8 gram (strength and range of ation; ROM significant between group
Female: 54.4% movement) compared with usual differences
care after TKR. Intervention group
had two sessions with a physio
and were phoned weekly for 6
weeks. Usual care involved referral
for clinic based outpatient physio-
therapy as recommended by dis-
charging hospital
Herbold et al. (2015) [30] N ¼ 141 RCT Continuous passive motion compared Primary outcome measur- Both groups improved with no
United States Mean age: 72.0 with conventional physiotherapy e ¼ active knee flexion ROM. significant between group
Female: 70.2% during inpatient rehabilitation Secondary outcome measur- differences.
after TKR es ¼ active knee extension
ROM length of stay, FIM
and Timed Up and Go test,
girth measurement, WOMAC
Hsu et al. (2016) [31] Experimental group Non randomized Patients in the experimental group Anxiety measured with visual Experimental group exhibited
Taiwan (n ¼ 49): 2-group experi- listened to music from 10 min analog scale (VAS), heart significantly lower anxiety
Mean age 73.9 mental design before continuous passive motion rate, ROM levels (p < 0.05) and
Female: 69.4% (CPM) until the end of the session increased CPM angles
Control group (25 min in total) on the first and (p < 0.05) during treatment
(n ¼ 42): second days after surgery. and increased active flexion
Mean age: 71.3 Controls did not listen to music ROM (p < 0.05) upon
Female: 78.6% and rested in bed 10 min before discharge
CPM. The patients routinely
received CPM rehabilitation twice
daily (10 a.m. and 4 p.m.) begin-
ning the first day after surgery
Jakobsen et al. (2014) [32] N ¼ 82 RCT Physiotherapy with progressive resist- Primary outcome ¼6-minute No between group differences.
Denmark Mean age: 64.5 ance training was compared with walk test 26 weeks post- Progressive resistance train-
Female: 58.3% physiotherapy without progressive operations; secondary ing did not confer add-
resistance training after TKR. outcomes ¼ VAS for pain; itional benefit
Interventions were for 7 weeks KOOS; OKS; EQ-5D
Ko et al. (2013) [33] N ¼ 249 RCT 1:1 exercise-based therapy compared OKS; WOMAC; SF-12 10-weeks All groups improved – no sig-
Australia Mean age: NR with either group-based exercise post-operation nificant between group dif-
Female: NR therapy or home exercise program ferences. Improvements
for 6 weeks after TKR sustained at 12 months
post-op
(continued)
Table 3. Continued
Study and location Sample characteristics Study methods Rehabilitation type(s) Outcome measures Outcomes
Levine et al. (2013) [34] N ¼ 70 RCT NMES with unsupervised at-home Flexion (degrees); Extension No differences between groups
United States NMES group: ROM exercises compared with (degrees); KSS, WOMAC; Get across any measures
Mean age: 68.1 therapist-managed physical ther- up and Go test; ROM
Female: 76% apy after TKA. Physical therapy
Physical therapy was the standard of care
group: Mean age: Participants were followed at 6 weeks
65.1 Female: 62% and 6 months post-operation
Li et al. (2015) [35] N ¼ 50 RCT Education and monthly phone follow Accelerometer (Fitbit Inc., New
China Mean age: NR ups compared with no education York., NY); WOMAC; SF-12;
Female: NR and no phone monitoring after KSS score; VAS for pain 6-
TKR months post-operation
Liao et al. (2013) [36] N ¼ 113 Prospective inter- The control group received conven- Functional reach test; WOMAC; Experimental group better
Taiwan Experimental group: vention study tional function training for eight single leg stand test; timed than controls (all p < 0.001)
Mean age 71.4 and RCT weeks. The experimental group 10m walk test; timed up for distance of functional
Female: 79.3% received the same conventional and go test; stair climbing forward reach; single leg
Control group: Mean training as the control group- stance (eyes closed and
age 72.9 þ additional balance exercises at open); timed sit-to-stand
Female: 67.3% each study follow up over an 8- test; stair climbing test;
week time period timed 10-m walk; timed up-
and-go test and the
WOMAC
L
opez-Liria et al. N ¼ 78 Non-randomized Home-based rehabilitation program WOMAC; ROM; muscle strength Both groups improved with no
(2015) [37] Age:71.3 controlled trial with physio sessions at commu- – timing of outcome meas- significant between group
Spain Female: 70.4% nity center compared with hos- urement unclear differences
pital-based outpatient
rehabilitation (physio led) follow-
ing TKR. Patients referred to
groups based on recommendation
of their physician
Madsen et al. (2013) [38] N ¼ 80 RCT Group-based exercise program com- Primary outcome ¼ OKS; sec- Both groups improved with no
Denmark Mean age: 66.5 pared with supervised home- ondary outcomes ¼ SF 36; significant between group
Female: 48.5% based training for 6 weeks after EQ-5D final end point 6- differences
TKR months post-operation
Mau-Moeller et al. (2014) N ¼ 125 RCT Continuous passive motion compared Primary outcome ¼ passive Sling training showed superior
[39] Mean age: 68.0 with sling training after TKR knee flexion range of results at 3 months post-op
Germany Female: 42.1% motion at 3-months post- (p ¼ 0.02) in ROM but group
operation; secondary differences disappeared by
measures ¼ ROM, SF-36, HSS 9 months
and WOMAC
Monticone et al. (2013) N ¼ 110 RCT For the intervention group, before KOOS; SF-36; Tampa Scale for Significant time by group
[40] Mean age: 67.0 returning home after TKR, patients Kinesiophobia at 12 months interaction in all the varia-
Italy Female: 63.6% were asked to continue the func- following hospital discharge bles in favor of the experi-
tional exercises learned during mental group and sustained
hospitalization and were given a for 6 months for quality of
book containing theoretical infor- life and functional activity
mation about the management of levels (p < 0.01)
kinesiophobia. They received
monthly phone follow up calls for
support for 6 months. In the con-
trol group, the patients were
advised to stay active and grad-
ually resume their usual activities
Piqueras et al. (2013) [41] N ¼ 142 RCT Physiotherapist supervised virtual Main outcome ¼ ROM Both groups improved with no
Spain Mean age: 73.3 reality tele-rehabilitation (exercise 3-months post-operation; significant between group
REHABILITATION AFTER JOINT REPLACEMENT

Female: 72.4% program) for 2 weeks compared Secondary outco- differences

with "conventional" outpatient mes ¼ muscle strength, walk
physiotherapy for 2 weeks after speed, pain, WOMAC
7

TKR
(continued)
 8
Table 3. Continued
Study and location Sample characteristics Study methods Rehabilitation type(s) Outcome measures Outcomes
Shanb and Youseff N ¼ 45 RCT Active exercise based training with Isometric peak torque of the Both groups improved but bio-
(2014) [42] Mean age: 63.3 biofeedback compared with active quadriceps and WOMAC feedback group improved
Saudi Arabia Female: 49.7% exercise training without biofeed- after 4-months more with respect to knee
back for 4 months after TKR rehabilitation functional activity (p < 0.05)
Unver et al. (2016) [43] N ¼ 60 RCT An 8-week exercise program was Primary outcome: isometric Significant differences
Turkey Mean age: 69.6 designed for bilateral TKA patients muscle strength of QF and (p < 0.01) for all parameters
Female: 72.5% who were 4 or more years post- hamstring muscles. (except rest pain and ROM)
D. L. SNELL ET AL.

surgery. Patients in one group Secondary outcomes: pain in favor of the weighted
were assigned to weighted exer- intensity, 30 s sit-to-stand group. QF muscle strength
cise group. Patients in the other test, 10min walk test, active changes (kg); 30 s sit-to-
group were assigned to non- ROM, Hospital for Special stand test changes (repeti-
weighted exercise group Surgery (HSS) knee score tions); 10-min walk test
Subjects in both groups were changes (seconds)
instructed to perform 10 repeti-
tions of each exercise at home
one time a day for 8 weeks
Vuorenmaa et al. (2014) N ¼ 108 RCT Home-based exercise program þ phy- WOMAC; SF-36; maximal walk- Both groups improved but
[44] Mean age: 69.0 siotherapy sessions for 4 months ing speed; isometric knee trend to greater improve-
Finland Female: 61.1% after TKR compared with no inter- muscle strength; Timed Up ment in walking speed and
vention. All patients had the same and Go (TUG) test at 12 knee flexion in intervention
rehabilitation for first two months. months post-operation group. Controls had physio
Intervention group then had 4 from other providers dilut-
months supervised exercise train- ing impact
ing and controls had no further
input as per standard usual care
Warren et al. (2016) [45] N ¼ 109 Retrospective The purpose of this retrospective 6-minute walk test (6MWT), No significant differences for
United States OP group: cohort study cohort study was to describe out- KOOS, and ROM KOOS, 6MWT, knee ROM
Mean age: 69.5 patient physical therapy (PT) after between groups after post-
Female: 56.3% TKR and compare short-term (2 operative PT. Time in OP PT
HH group: months) functional and clinical (p ¼ 0.55) and the number
Mean age: 74.4 outcomes of patients following of outpatient PT visits
Female: 86.4 TKR who were discharged from (p ¼ 0.68) were similar
the hospital to home and received between groups
(a) outpatient PT immediately (OP)
or (b) home health PT before out-
patient PT (HH)

Yang et al. (2013) [46] N ¼ 41 RCT Chinese traditional medici- VAS for pain and ROM day 5 Chinese traditional medicine
Taiwan Mean age: 69.5 ne þ rehabilitation compared with post-operation approach effective in reduc-
rehabilitation alone following TKR. ing pain compared with
Chinese traditional medicine rehabilitation alone
involved light massage, range of (p < 0.05)
movement and slow mobilization
with pain reduction as primary
outcome
TKR: total knee replacement; RCT: randomized control trial; WOMAC: Western Ontario and McMaster Universities Arthritis Index; SF-36: Medical Outcomes Short Form 36; SF-12: Medical Outcomes Short Form-12; BDI:
Beck Depression Inventory; OKS: Oxford Knee Score; EQ-5 D: EuroQoL-5 Dimensions Scale; HSS: Hospital for Special Surgery Score; KSS: Knee Society Score; VAS: Visual Analogue Scale; KOOS: Knee Injury and
Osteoarthritis Outcome Score; FIM: Functional Independence Measure; ROM: range of motion; CPM: continuous passive motion; NR: not reported; NMES: neuromuscular electrical stimulation.
 REHABILITATION AFTER JOINT REPLACEMENT 9

function and quality of life. These studies taken together, suggest group but no differences in pain, function or perceived self-
the setting (home versus clinic) did not significantly contribute to efficacy 9 months’ post-operation. Hsu et al. [31] investigated the
outcomes. impact of music coupled with continuous passive motion, early
Nine trials evaluated a particular physiotherapy approach com- after THR for 91 patients who either received 25 min of music
pared with usual care or with another physiotherapy approach leading up to and during continuous passive motion or who
after TKR [27,30–32,36,39,41–43]. One trial [32] compared physio- rested in bed ahead of continuous passive motion treatment.
therapy plus progressive resistance training with physiotherapy They reported a significant reduction in anxiety during treatments
without progressive resistance training with 82 participants and and improved range of movement at discharge from hospital, for
did not find between group differences. Two trials [30,39] eval- patients in the music condition. Finally, one trial evaluated
uated continuous passive motion with either physiotherapy with- rehabilitation with a traditional Chinese medicine approach (light
out continuous passive motion or sling training, respectively. massage) with rehabilitation alone [46] with 41 patients after TKR
While sling training conferred short-term benefit, over time there and reported improvement favoring the intervention group in
were no between group differences across these two studies. One pain (primary outcome).
trial evaluated a 3-week tele-rehabilitation exercise program with Four systematic reviews evaluated physiotherapy rehabilitation
usual care defined as two weeks of clinic based physiotherapy after TKR [15,16,47,48] (see Table 4). Three of these reviews
[41] with 142 patients after TKR. No significant between group dif- [15,16,47] evaluated physiotherapy interventions at home or in
ferences were reported. One trial added biofeedback to exercise the outpatient clinic compared with no physiotherapy. There was
based physiotherapy with 45 participants after TKR [42] and some overlap across these three reviews with regard to included
reported that adding biofeedback conferred benefit with respect studies. All three reported that physiotherapy supervision whether
to knee function but there were no other group differences in in the outpatient clinic or at home, appeared better than no
outcomes. The trial by Unver et al. [43] compared an 8 week physiotherapy for pain reduction and functional improvement, but
home-based exercise program with or without weight training felt the low quality of studies limited confidence in the findings.
with 60 participants. They reported the weighted exercise group Finally, the systematic review by Moutzouri et al. [48] included six
demonstrated greater gains with respect to strength and walking, studies evaluating the effects of sensorimotor training on balance
but not angle of movement or pain. Similarly, Liao et al. [36] com- and function. While their results favored sensorimotor training
pared an 8 week exercise training program with and without bal- groups, once again they felt the low quality and underpowered
ance exercises with 113 participants and reported the nature of included studies was limiting.
experimental group (exercise plus balance training) made more
gains compared with controls, across measures of function and
Studies including both hip and knee replacement participants
pain.
(Table 5)
One trial compared walking focused physiotherapy with usual
care defined as individual physiotherapy [27] with 57 participants Two studies, both prospective observational studies, evaluated
and reported improved walking outcomes for the intervention rehabilitation following both THR and TKR [49,50]. In the study by

Table 4. Systematic reviews included in the review by methods and sample size, type of rehabilitation, and outcomes after total knee replacement (n ¼ 4).
Number of studies and Type of rehabilitation interven-
Study and timeline for searching participants Study methods tions compared Findings/conclusions
Artz et al. (2014) [47] 18 studies Systematic review Physiotherapy versus no Physiotherapy appeared better
Searched to Oct 2013 N ¼ 1739 physiotherapy; home versus than no physiotherapy for
outpatient clinic setting pain reduction and function
but type of approach and the
setting (home versus clinic)
seemed less important. Low
quality of trials overall limited
confidence in these findings
Moutzouri et al. (2016) [48] Six studies Systematic review Effect of sensori-motor training Clinical performance-based tests
Searched to Sep 2014. N ¼ 409 on functional and balance (more than relevant patient-
performance reported measures) showed
An exercise-based intervention that functional ability and
incorporating sensori-motor balance were improved com-
components was pared to controls. The robust-
compared with another thera- ness of evidence was
peutic intervention, placebo compromised because most
or control of the studies were
underpowered
Papalia et al. (2013) [15] 18 studies Systematic review Home-based versus outpatient Supervision and location/setting
Searched to March 2013 N ¼ 1489 supervised rehabilitation did not seem to directly
determine the final outcomes.
However studies were too
heterogeneous to draw any
firm conclusions
Pozzi et al. (2013) [16] 19 studies Systematic review Review of the effectiveness of Support was found for super-
Searched Jan N ¼ 1613 postoperative outpatient care vised outpatient and home-
2003 – Jun 2013 on short- and long-term func- based programs but low
tional recovery quality of the studies such as
poorly described usual care
interventions, limited the con-
clusions able to be drawn
TKR: total knee replacement.
10 D. L. SNELL ET AL.

Table 5. Studies included in the review by methods, location, type of rehabilitation and outcomes after total hip replacement where both total hip and total knee
replacement participants were included (n ¼ 2).
Study and location Sample characteristics Study methods Rehabilitation type(s) Outcome measures Outcomes
Benz et al. (2015) [49] N ¼ 201 POCS Patients followed up by WOMAC after 4 weeks of All patients improved.
Switzerland Mean age: 68.7 outpatient physiother- rehabilitation post- Pre-op status and
Female: 56.7% apy or admitted to a operation comorbidities best pre-
convalescence center or dictors of outcome.
an inpatient program Older patients more
following THR and TKR. likely to receive
Type of rehabilitation inpatient rehabilitation
referral determined by
patient’s surgeon
Zech et al. (2015) [50] THR: POCS Intervention for THR and WOMAC; hip and knee All patients improved. No
Germany N ¼ 58 TKR patients involving ROM after rehabilitation associations found
Mean age:62.5 hands on physiotherapy, (up to 25 d post- between duration or
Female: 51.7% group exercise program, operation) intensity of rehabilita-
TKR: strength and gait train- tion and outcomes
N ¼ 65 ing as per standard
Mean age: 66.6 rehabilitation protocols
Female: 64.6% for the hospital (no con-
trol group)
THR: total hip replacement; TKR: total knee replacement; POCS: prospective observational cohort study; RCT: randomized controlled trial; WOMAC: Western Ontario
and McMaster Universities Arthritis Index; ROM: range of movement.

Zech et al. [50], 123 patients referred for physiotherapy were fol- Rehabilitation and TKR
lowed after either THR or TKR for up to 3 weeks. Therapy included
Studies evaluating interventions after TKR were of varying quality
hands on physiotherapy, a group exercise program, and strength
involving a mix of intervention types, different outcome measures
and gait training. All patients made gains between pre- and post-
intervention assessments and no associations were found between and outcomes making across study comparisons difficult.
outcomes and intensity or duration of rehabilitation. In the study However, and consistent with suggested findings for rehabilitation
by Benz et al. [49], 201 patients referred by their surgeon for after THR, any follow up regardless of the setting (home versus
either inpatient rehabilitation, outpatient physiotherapy, or admis- hospital), type, intensity, or method of delivering rehabilitation,
sion to a convalescence center after either THR or TKR, were appears to confer benefit. These findings appear consistent across
monitored for a 6-month period. Older patients were more likely many studies including the systematic reviews shown in Table 4.
to have been referred for inpatient rehabilitation and while all There was some specific support for the benefits of biofeedback,
patients made gains, pre-operative status and comorbidity were sensorimotor training and focus on walking and pain. The two
the best predictors of outcome. prospective observational studies examining rehabilitation after
TKR [49,50] highlighted the importance of premorbid status and
comorbidities regarding responsiveness to intervention. However,
Discussion
this is not a consistent finding across the research. For example, a
We conducted a scoping study to review research studies pub- large multi-site study conducted in the United States involving
lished from January 2013 to December 2016, evaluating effective- more than 2000 joint replacement rehabilitation patients failed to
ness of rehabilitation following THR and/or TKR. We sought to find any associations between comorbidities and outcomes
review this work in the context of outcomes described in previ- [51,52].
ously published systematic reviews. In this way, we aimed to map
the key issues for rehabilitation following THR and TKR and high-
light potential avenues for new research. We found 30 studies Is monitoring progress key?
meeting criteria for inclusion in the review, adding to seven sys-
There appears to be converging evidence that rehabilitation
tematic reviews. This vast body of work including more than 100
regardless of intensity, type, method of delivery, or setting
studies examined the effectiveness of rehabilitation after total
appears helpful following both THR and TKR. Many published
joint replacement, mostly using RCT methods. True to the scoping
study method we made no attempt to weight this research based studies now, albeit of generally poor and inconsistent quality have
on quality and risk for bias. evaluated many different combinations of rehabilitation type, set-
ting, and intensity. When patients are either not provided with
any advice or intervention, or they are provided with education
Rehabilitation after THR and/or an exercise program at discharge but progress is not moni-
Collectively, studies evaluating effectiveness of rehabilitation after tored, outcomes appear poor. This suggests monitoring progress
THR suggested physiotherapy-based exercise programs were and recovery after joint replacement even remotely, has value but
beneficial whether these were clinic-based, home-based or moni- research examining this issue appears limited.
tored by phone. There was some specific support for early full It is not surprising that monitoring of progress in itself might
weight bearing approaches. This raises a question regarding the be important. The value of attention and the therapeutic alliance
value of follow up irrespective of setting, intensity or type of inter- have been widely discussed in the psychotherapy literature for
vention. Consistent with these findings, prospective observational decades [53] and are becoming an increasing focus for physiother-
studies did not show associations between duration and intensity apy [53,54]. Two recently published systematic reviews of research
of rehabilitation, and outcomes after THR [24,25,49,50]. Rather examining the role of therapeutic relationships in physiotherapy
age, comorbidity and functional disability appeared more relevant. [53,54] report that effective patient-practitioner interactions result
 REHABILITATION AFTER JOINT REPLACEMENT 11

in improved outcomes, reduced symptoms, increased motivation, need for more patient-centric outcome measures. Most of the
and better adherence to treatment. work reviewed here focuses on physical and functional gains,
We also found three qualitative studies examining stakeholder while essential, they may not be the sole priorities for the person
perspectives on rehabilitation following joint replacement where at the center of the intervention. For example, many of the pub-
relationships between the patient and their health practitioner lished trials included in our review and the systematic reviews
were explored [55–57]. First, a longitudinal qualitative study [56] described above did not include outcome measures related to
focused on understanding patient use of support after TKR. The quality-of-life, psychosocial outcomes, and patient expectations.
researchers reported that their participants (n ¼ 10) felt more con- Second, there is a need for agreed-upon outcomes and out-
fident and reassured by contact with health practitioners after come measures. This need is underscored by the inconsistency of
their surgery and that missing support increased anxiety and dis- outcome measures across the research reviewed here making it
tress. Their participants highlighted the value of having a point of difficult to compare findings and synthesize the evidence. Work
contact and knowing who to go to for advice for example, with developing guidelines and recommendations for choice of out-
regard to return to activities like driving, when to stop using walk- comes and outcome measures after joint replacement is underway
ing aids, how to progress exercise intensity, and what to do when [59,60]. The use of standardized outcome measures has potential
pain management is inadequate. to greatly improve understandings about the benefits of rehabili-
These experiences were echoed by participants in the qualita- tation after joint replacement and to advance rehabilitation
tive study by Westby and Blackman [57] who interviewed health science.
practitioners (n ¼ 44) and patients after THR and TKR (n ¼ 32). Research regarding best practice in joint replacement rehabili-
Patient participants in this study also reported increased anxiety tation is needed more than ever. The United States Medicare pro-
and vulnerability, and felt forgotten by their health practitioners gram, aimed mainly at older Americans, for example, is testing
when no follow up was provided. Finally, Crepeau [55] used new payment models for joint replacement that entail a fixed
vignettes to highlight the importance of therapeutic attention price for 90-d episodes of care commencing with a hospitalization
after TKR and indicated attention is a precursor to establishing a and includes all acute interventions and post-acute rehabilitation
therapeutic relationship and has a positive psychosocial impact in in the ensuing days. Under this bundled payment scheme, pro-
its own right. viders face both upside and downside risk for both costs and out-
This collection of work offers promising insights regarding the comes and thus the scheme provides powerful incentives to
value and positive impact of therapeutic attention for recovery identify best practice and better care coordination across acute
after joint replacement particularly given the nature of this patient and post-acute providers. In short, it intensifies the need for
population (generally older, higher rates of comorbidities). More research as providers scramble to identify the optimal mix of serv-
work is needed, however, to better understand key aspects such ices that will provide best value for both patients and payers.
as optimal mode, duration, and frequency of post-surgical con- The results of this review suggest that some form of home
tacts, as well as which patients might benefit from remote moni- monitoring, for example, should be an essential component in an
toring and which patients need more intensive rehabilitation. episode-based plan of care for patients with joint replacement in
a bundled payment scheme – or any payment scheme for that
matter. Beyond this finding, this review offers mixed guidance.
Clinical, research, and policy implications
Given the incentives inherent in a bundled payment system, pro-
There is converging evidence from many studies and reviews of viders now also have an incentive to deploy their own data ana-
studies that exercise-based rehabilitation following THR and TKR is lytic tools as they try through trial and error to find high-value,
beneficial, whether this is clinic or home based, low or high inten- best practice. This review should give such providers a head start
sity. However, there is less evidentiary consensus about the best about where to look and what previous research has uncovered,
mix of pre- and post-operative rehabilitation services in terms of inconclusive as it maybe.
setting, mode, frequency, intensity, and duration of post-operative
therapy and patient monitoring. Multi-phase RCTs that examine
Limitations
just one or two of these variables at a time do not always provide
the granularity of information needed to address clinical and pol- Limitations of the scoping method include absence of any weight-
icy issues that are inherently multi-dimensional. ing or ranking of quality of included research. Recently published
More work using a wider range of research methods is needed systematic reviews have consistently highlighted concerns with
to help clarify key questions such as which patients are best the quality of trials evaluating rehabilitation after joint replace-
suited to low-intensity home- or clinic-based care versus more ment indicating these have typically involved small samples sizes,
intensive interventions in these settings. Prospective observational inadequate randomization procedures, absence of blind outcome
study methods offer alternatives to RCTs that may allow for closer measurement, and high risk for bias. However, as we have shown
examination of associations between a larger array of variables at the increasing consistency of findings across this growing body of
any one time such as pre-, peri-, and post-surgical factors and out- work, involving both THR and TKR participants, does allow cau-
comes. Using these methods, the characteristics of people at risk tious confidence in reported findings. We acknowledge some
for poor outcomes after THR and TKR, for example, can be better research may have been overlooked because we limited searching
understood. Identification of at-risk patients will enable better pre- to studies published in English. Finally, our search strategy
and post-surgical care decisions regarding rehabilitation needs fol- involved updating previously published systematic reviews of
lowing surgery. In addition, mixed methods studies involving both RCTs. This means that intervention studies published before 2013
quantitative and qualitative methods may help to triangulate the that did not involve RCT methods may have been overlooked.
evidence, provide more patient-centered research foci, and Both a strength and limitation of our scoping study is that it
explore seemingly puzzling and inconsistent findings across quan- entailed studies from many different countries. It is a strength
titative studies [58]. because it helps to generalize findings across national boundaries.
More work is needed with respect to the development and It is also a limitation because it obscures differences in practice
selection of joint replacement outcome measures. First, there is a settings, practice patterns, and payment systems across countries
12 D. L. SNELL ET AL.

and the terms used to describe them – making it more difficult to [7] Lavernia C, Alcerro J, Brooks L, et al. Mental health and out-
compare results across countries. For example, a home-based comes in primary total joint arthroplasty. J Arthroplasty
post-op TKR therapy program in one country may consist of a dif- 2012;27:1276–1282.
ferent constellation of interventions than a corresponding pro- [8] Berges I-M, Kuo Y-F, Ostir G, et al. Gender and ethnic differ-
gram in another country. Moreover, in some countries, patients ences in rehabilitation outcomes after hip-replacement sur-
may receive rehabilitation in multiple settings within an episode gery. Am J Phys Med Rehabil. 2008;87:567–572.
of care, making it difficult to isolate the contribution of any one [9] Bischoff-Ferrari H, Lingard E, Losina E, et al. Psychosocial
setting or practice pattern. and geriatric correlates of functional status after total hip
replacement. Arthritis Rheum—Arthritis Care Res. 2004;
51:829–835.
Conclusions [10] Hooper G, Rothwell A, Hooper N, et al. The relationship
Rehabilitation after THR or TKR, whether carried out at the clinic between the American Society of Anesthesiologists physical
or monitored at home, appears beneficial but type, intensity, and rating and outcome following total hip and knee arthro-
duration of interventions do not appear consistently associated plasty: an analysis of the New Zealand Joint Registry.
with outcomes. Burden of comorbidities may better predict J Bone Joint Surg Am. 2012;94:1065–1070.
rehabilitation outcomes. Monitoring of progress and therapeutic [11] World Health Organization (WHO). Available from: http://
www.who/.int/topics/rehabilitation/en/-
attention appear important however the available research pro-
[12] Coulter CL, Scarvell JM, Neeman TM, et al. Physiotherapist-
vides limited guidance regarding who would best benefit from
directed rehabilitation exercises in the outpatient or home
remote monitoring versus more intensive intervention, or the opti-
setting improve strength, gait speed and cadence after
mal intensity of monitoring needed to achieve gains following
elective total hip replacement: a systematic review.
THR and TKR. More work exploring the key components (e.g.,
J Physiother. 2013;59:219–226.
intensity, frequency, and duration) of monitoring and therapeutic [13] Di Monaco M, Castiglioni C. Which type of exercise therapy
attention using a wider range of study methods and more consist- is effective after hip arthroplasty? A systematic review of
ent choice of outcomes and outcome measures is needed. randomized controlled trials. Eur J Phys Rehabil Med.
2013;49:893–907.
[14] Minns Lowe CJ, Davies L, Sackley CM, et al. Effectiveness of
Acknowledgements
land-based physiotherapy exercise following hospital dis-
We wish to acknowledge and thank the Editor and Peer Reviewers charge following hip arthroplasty for osteoarthritis: an
for their constructive feedback during the peer review process. updated systematic review. Physiotherapy 2015;101:
252–265.
[15] Papalia R, Vasta S, Tecame A, et al. Home-based vs super-
Disclosure statement vised rehabilitation programs following knee surgery: a sys-
The authors report no declarations of interest. tematic review. Br Med Bull. 2013;108:55–72.
[16] Pozzi F, Snyder-Mackler L, Zeni J. Physical exercise after
knee arthroplasty: a systematic review of controlled trials.
Funding Eur J Phys Rehabil Med. 2013;49:877–892.
This research was support by a Project Grant from the Canterbury [17] Arksey H, O'Malley L. Scoping studies: towards a methodo-
Medical Research Foundation (Grant no. 14/07), and a 2015 logical framework. Int J Soc Res Methodol. 2005;8:19–32.
University of Otago summer student scholarship from Aged [18] Levac D, Colquhoun H, O'Brien K. Scoping studies: advanc-
Concern Trust. ing the methodology. Implement Sci. 2010;5:69.
[19] Beaupre LA, Masson EC, Luckhurst BJ, et al. A randomized
pilot study of a comprehensive postoperative exercise pro-
References gram compared with usual care following primary total hip
arthroplasty in subjects less than 65 years of age: feasibility,
[1] Collins N, Roos E. Patient reported outcomes for total hip
selection of outcome measures and timing of assessment.
and knee arthroplasty: commonly used instruments and
BMC Musculoskelet Disord. 2014;15:192.
attributes of a “good” measure. Clin Geriatr Med. [20] Judd DL, Winters JD, Stevens-Lapsley JE, et al. Effects of
2012;28:367–394. neuromuscular reeducation on hip mechanics and func-
[2] Jeffery A, Wylde V, Blom A, et al. “It's there and I'm stuck tional performance in patients after total hip arthroplasty: a
with it”: patients' experiences of chronic pain following case series. Clin Biomech (Bristol, Avon). 2016;32:49–55.
total knee replacement surgery. Arthritis Care Res. [21] Mikkelsen L, Mechlenburg I, Soballe K, et al. Effect of early
2011;286–292. supervised progressive resistance training compared to
[3] Jordan RW, Smith NA, Chahal GS, et al. Enhanced education unsupervised home-based exercise after fast track total hip
and physiotherapy before knee replacement; is it worth it? replacement applied to patients with preoperative func-
A systematic review. Physiotherapy 2014;100:305–312. tional limitations: a single-blinded randomised controlled
[4] New Zealand Orthopaedic Association (NZOA). New trial. Osteoarthritis Cartilage 2014;22:2051–2058.
Zealand Joint Registry: Sixteen Year Report (January 1999 [22] Monticone M, Ambrosini E, Rocca B, et al. Task-oriented
to December 2014). Christchurch New Zealand; 2015. exercises and early full weight-bearing contribute to
[5] Turner A, Barlow J, Buszewicz M, et al. Beliefs about the improving disability after total hip replacement: a rando-
causes of osteoarthritis among primary care patients. mised controlled trial. Clin Rehabil. 2014;28:658–668.
Arthritis Rheum. 2007;57:267–271. [23] Morishima Y, Mizushima T, Yamauchi K, et al. Effects of
[6] Westby MD. Rehabilitation and total joint arthroplasty. Clin home-based interval walking training on thigh muscle
Geriatr Med. 2012;28:489–508. strength and aerobic capacity in female total hip
 REHABILITATION AFTER JOINT REPLACEMENT 13

arthroplasty patients: a randomized, controlled pilot study. home-exercises after total knee arthroplasty. Dan Med J.
PLoS One 2014;9:e108690. 2013;60:A4607.
[24] Okoro T, Morrison V, Maddison P, et al. An assessment of [39] Mau-Moeller A, Behrens M, Finze S, et al. The effect of con-
the impact of behavioural cognitions on function in tinuous passive motion and sling exercise training on clin-
patients partaking in a trial of early home-based progres- ical and functional outcomes following total knee
sive resistance training after total hip replacement surgery. arthroplasty: a randomised active-controlled clinical study.
Disabil Rehabil. 2013;35:2000–2007. Health Qual Life Outcomes 2014;12:68–78.
[25] Rapp W, Brauner T, Weber L, et al. Improvement of walking [40] Monticone M, Ferrante S, Rocca B, et al. Home-based func-
speed and gait symmetry in older patients after hip arthro- tional exercises aimed at Managing Kinesiophobia contrib-
plasty: a prospective cohort study. BMC Musculoskelet ute to improving disability and quality of life of patients
Disord. 2015;16:291. undergoing total knee arthroplasty: a randomized con-
[26] Umpierres CS, Ribeiro TA, Marchisio AE, et al. Rehabilitation trolled trial. Arch Phys Med Rehabil. 2013;94:231–239.
following total hip arthroplasty evaluation over short fol- [41] Piqueras M, Marco E, Coll M, et al. Effectiveness of an inter-
low-up time: randomized clinical trial. J Rehabil Res Dev.
active virtual telerehabilitation system in patients after total
2014;51:1567–1578.
knee arthoplasty: a randomized controlled trial. J Rehabil
[27] Brunn-Olsen V, Heiberg KE, Wahl AK, et al. The immediate
Med. 2013;45:392–396.
and long-term effects of a walking-skill program compared
[42] Shanb A-SA, Youssef EF. Effects of adding biofeedback
to usual physiotherapy care in patients who have under-
training to active exercises after total knee arthroplasty.
gone total knee arthroplasty (TKA): a randomized controlled
trial. Disabil Rehabil. 2013;35:2008–2015. J Musculoskelet Res. 2014;17:1450001.
[28] Buker N, Akkaya S, Akkaya N, et al. Comparison of effects [43] Unver B, Bakirhan S, Karatosun V. Does a weight-training
of supervised physiotherapy and a standardised home pro- exercise programme given to patients four or more years
gram on functional status in patients with total knee after total knee arthroplasty improve mobility: a random-
arthroplasty: a prospective study. J Phys Ther Sci. 2014;26: ized controlled trial. Arch Gerontol Geriatr. 2016;64:45–50.
1531–1536. [44] Vuorenmaa M, Ylinen J, Piitulainen K, et al. Efficacy of a 12-
[29] Han AS, Nairn L, Harmer AR, et al. Early rehabilitation after month, monitored home exercise programme compared
total knee replacement surgery: a multicenter, noninferior- with normal care commencing 2 months after total knee
ity, randomized clinical trial comparing a home exercise arthroplasty: a randomized controlled trial. J Rehabil Med.
program with usual outpatient care. Arthritis Care Res 2014;46:166–172.
(Hoboken) 2015;67:196–202. [45] Warren M, Kozik J, Cook J, et al. A comparative study to
[30] Herbold JA, Bonistall K, Blackburn M, et al. Randomized determine functional and clinical outcome differences
controlled trial of the effectiveness of continuous passive between patients receiving outpatient direct physical ther-
motion after total knee replacement. Arch Phys Med apy versus home physical therapy followed by outpatient
Rehabil. 2014;95:1240–1245. physical therapy after total knee arthroplasty. Orthop Nurs.
[31] Hsu C-C, Chen W-M, Chen S-R, et al. Effectiveness of music 2016;35:382–390.
listening in patients with total knee replacement during [46] Yang TH, Yeh WL, Chen HY, et al. Compare the Traditional
CPM rehabilitation. Biol Res Nurs. 2016;18:68–75. Chinese Medicine manipulation with rehabilitation on in-
[32] Jakobsen TL, Kehlet H, Husted H, et al. Early progressive patients after total knee arthroplasty. J Arthroplasty 2013;
strength training to enhance recovery after fast-track total 28:954–959.
knee arthroplasty: a randomized controlled trial. Arthritis [47] Artz N, Elvers KT, Lowe CM, et al. Effectiveness of physio-
Care Res (Hoboken). 2014;66:1856–1866. therapy exercise following total knee replacement: system-
[33] Ko V, Naylor J, Harris I, et al. One-to-one therapy is not atic review and meta-analysis. BMC Musculoskelet Disord.
superior to group or home-based therapy after total knee 2015;16:15.
arthroplasty: a randomized, superiority trial. J Bone Joint [48] Moutzouri M, Gleeson N, Billis E, et al. What is the effect of
Surg Am. 2013;95:1942–1949.
sensori-motor training on functional outcome and balance
[34] Levine M, McElroy K, Stakich V, et al. Comparing conven-
performance of patients' undergoing TKR? A systematic
tional physical therapy rehabilitation with neuromuscular
review. Physiotherapy 2016;102:136–144.
electrical stimulation after TKA. Orthopedics 2013;36:
[49] Benz T, Angst F, Oesch P, et al. Comparison of patients in
e319–e324.
three different rehabilitation settings after knee or hip
[35] Li Z, Jiang L, Lin J. The effect of education for daily physical
arthroplasty: a natural observational, prospective study.
activity level recovery of osteoarthritis patients after total
knee arthroplasty. A prospective randomized controlled BMC Musculoskelet Disord. 2015;16:317.
clinical trial using accelerometry. Osteoarthritis Cartilage [50] Zech A, Hendrich S, Pfeifer K. Association between exercise
2015;23:A373. therapy dose and functional improvements in the early
[36] Liao CD, Liou TH, Huang YY, et al. Effects of balance train- postoperative phase after hip and knee arthroplasty: an
ing on functional outcome after total knee replacement in observational study. Pm R. 2015;7:1064–1072.
patients with knee osteoarthritis: a randomized controlled [51] De Jong G, Tian W, Smout R, et al. Long-term outcomes of
trial. Clin Rehabil. 2013;27:697–709. joint replacement rehabilitation patients discharged from
[37] Lopez-Liria R, Padilla-Gongora D, Catalan-Matamoros D, skilled nursing and inpatient facilities. Arch Phys Med
et al. Home-based versus hospital-based rehabilitation pro- Rehabil. 2009;90:1306–1316.
gram after total knee replacement. Biomed Res Int. [52] DeJong G, Horn S, Smout R, et al. Joint replacement
2015;2015:1 rehabilitation outcomes on discharge from skilled nursing
[38] Madsen M, Larsen K, Madsen I, et al. Late group-based facilities and inpatient rehabilitation facilities. Arch Phys
rehabilitation has no advantages compared with supervised Med Rehabil. 2009;90:1284–1296.
14 D. L. SNELL ET AL.

[53] Besley J, Kayes N, McPherson K. Assessing therapeutic rela- [58] O'Cathain A, Murphy E, Nicholl J. Why, and how, mixed
tionships in physiotherapy: literature review. N Z J methods research is undertaken in health services research
Physiother. 2011;39:81–91. in England: a mixed methods study. BMC Health Serv Res.
[54] Pinto R, Ferreira M, Oliveira V, et al. Patient-centred com- 2007;7:85–96.
munication is associated with positive therapeutic alliance: [59] Rolfson O, Bohm E, Franklin P, et al. Patient-
a systematic review. J Physiother. 2012;58:77–87. reported outcome measures in arthroplasty registries:
[55] Crepeau E. “I need someone to keep an eye on me:” the Report of the Patient-Reported Outcome Measures
power of attention in patient-practitioner interactions. Working Group of the International Society of
Disabil Rehabil. 2016;12:1–9. Arthroplasty Registries. Part II. Recommendations for
[56] Johnson E, Horwood J, Gooberman-Hill R. Trajectories of selection, administration, and analysis. Acta Orthop.
need: understanding patients' use of support during the 2016;87:9–23.
journey through knee replacement. Disabil Rehabil. [60] Rolfson O, Wissig S, van Maasakkers L, et al. Defining an
2016;10;1–14. international standard set of outcome measures for
[57] Westby M, Backman C. Patient and health professional patients with hip or knee osteoarthritis: consensus of the
views on rehabilitation practices and outcomes following international consortium for health outcomes measurement
total hip and knee arthroplasty for osteoarthritis: a focus hip and knee osteoarthritis working group. Arthritis Care
group study. BMC Health Serv Res. 2010;10:119–134. Res. 2016;60:1631–1639.