[ research report ]

J. HAXBY ABBOTT, DPT, PhD, FNZCP1 CATHERINE M. CHAPPLE, PT, MManipPhty, PhD2 G. KELLEY FITZGERALD, PT, PhD, FAPTA3
JULIE M. FRITZ, PT, PhD, ATC4 JOHN D. CHILDS, PT, PhD5 HELEN HARCOMBE, BPhty, MPH, PhD1,6 KIRSTEN STOUT, RN1

The Incremental Effects of Manual

Therapy or Booster Sessions
in Addition to Exercise Therapy
Downloaded from www.jospt.org at University of Otago on October 1, 2015. For personal use only. No other uses without permission.

for Knee Osteoarthritis:

A Randomized Clinical Trial

O
steoarthritis (OA) is a
Copyright ${year} Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

TTSTUDY DESIGN: A factorial randomized
controlled trial.
TTOBJECTIVES: To investigate the addition of
manual therapy to exercise therapy for the reduc-
tion of pain and increase of physical function in
people with knee osteoarthritis (OA), and whether
booster sessions compared to consecutive ses-
sions may improve outcomes.
TTBACKGROUND: The benefits of providing
manual therapy in addition to exercise therapy, or
Journal of Orthopaedic & Sports Physical Therapy
TTRESULTS: Of 75 participants recruited, 66 common disorder that
(88%) were retained at 1-year follow-up. Factorial
affects up to 40% of
analysis of covariance of the main effects showed
significant benefit from booster sessions adults and, according to
(P = .009) and manual therapy (P = .023) over
exercise therapy alone. Group analysis showed that
some sources, may affect more
than 80% of those over 65 years
exercise therapy with booster sessions (WOMAC
score, 46.0 points; 95% confidence interval [CI]: of age.4,12 Knee and hip OA
80.0, 12.0) and exercise therapy plus manual
therapy (WOMAC score, 37.5 points; 95% CI: 69.7, are among the most common causes of
5.5) had superior effects compared with exercise pain and disability in older adults. Ex-

of distributing treatment sessions over time using
periodic booster sessions, in people with knee OA
are not well established.
TTMETHODS: All participants had knee OA and
were provided 12 sessions of multimodal exercise
therapy supervised by a physical therapist. Par-
ticipants were randomly allocated to 1 of 4 groups:
exercise therapy in consecutive sessions, exercise
therapy distributed over a year using booster
sessions, exercise therapy plus manual therapy
without booster sessions, and exercise therapy
plus manual therapy with booster sessions. The
primary outcome measure was the Western Ontar-
io and McMaster Universities Osteoarthritis Index
(WOMAC score; 0-240 scale) at 1-year follow-up.
Secondary outcome measures were the numeric
pain-rating scale and physical performance tests.
therapy alone. The combined strategy of exercise
therapy plus manual therapy with booster sessions
was not superior to exercise therapy alone.
TTCONCLUSION: Distributing 12 sessions of exer-
cise therapy over a year in the form of booster ses-
sions was more effective than providing 12 consec-
utive exercise therapy sessions. Providing manual
therapy in addition to exercise therapy improved
treatment effectiveness compared to providing 12
consecutive exercise therapy sessions alone. Trial
registered with the Australian New Zealand Clinical
Trials Registry (ACTRN12612000460808).
TTLEVEL OF EVIDENCE: Therapy, level 1b.
J Orthop Sports Phys Ther. In Press. Epub 28
Sep 2015. doi:10.2519/jospt.2015.6015
TTKEY WORDS: arthralgia, OA, physical therapy
techniques, randomized controlled trial
ercise therapy is known to be effective19
and is recommended as the first line of
treatment for reducing pain and disabil-
ity in individuals with knee OA.13 How-
ever, recent systematic reviews indicate
that these benefits are generally modest
for improvements in pain and function,19
and although similar to those of simple
analgesics and nonsteroidal anti-inflam-
matory drugs, effects are often short
lived.8,19,22,30 Strategies are needed to im-
prove treatment effect.
Manual therapy is an intervention
commonly combined with therapeutic
exercise in clinical practice,9 although the
research available to date cannot provide

Centre for Musculoskeletal Outcomes Research, Orthopaedic Surgery Section, Department of Surgical Sciences, Dunedin School of Medicine, University of Otago, Dunedin, New Zealand.
1

2
School of Physiotherapy, University of Otago, Dunedin, New Zealand. 3University of Pittsburgh, Pittsburgh, PA. 4University of Utah, Salt Lake City, UT. 5US Army-Baylor University, Schertz,
TX. 6Department of Preventive and Social Medicine, University of Otago, Dunedin, New Zealand. The study protocol was approved by the Lower South Regional Ethics Committee of the New
Zealand Ministry of Health (LSR/10/11/055), and registered with the Australian New Zealand Clinical Trials Registry (ACTRN12612000460808). This research was supported in part by the
New Zealand Lottery Grants Board, the New Zealand Society of Physiotherapists Scholarship Trust, the Health Research Council of New Zealand, and a University of Otago Research Grant.
Dr Abbott was supported in part by a Sir Charles Hercus Health Research Fellowship from the Health Research Council of New Zealand. The funders have had no influence on the content
of that work or the current article. The authors certify that they have no affiliations with or financial involvement in any organization or entity with a direct financial interest in the subject
matter or materials discussed in the article. Address correspondence to Dr J. Haxby Abbott, Centre for Musculoskeletal Outcomes Research, Department of Surgical Sciences, Dunedin
School of Medicine, University of Otago, PO Box 913, Dunedin, New Zealand. E-mail: haxby.abbott@otago.ac.nz t Copyright 2015 Journal of Orthopaedic & Sports Physical Therapy

journal of orthopaedic & sports physical therapy | ahead of print | 1

45-12 Abbott-AoP.indd 1 9/28/2015 3:09:00 PM

 [ research report
evidence as to whether it may improve
the overall effectiveness of rehabilita-
tion for reducing pain and disability in
patients with knee OA.1,14,15 A recent trial
that included patients with knee OA who
were assigned exercise, manual therapy,
or a combination of exercise therapy plus
manual therapy found that these inter-
ventions were superior to usual care with
Downloaded from www.jospt.org at University of Otago on October 1, 2015. For personal use only. No other uses without permission.
no additional therapy; however, no sig-
nificant differences were found between
the 3 intervention groups.1 In that trial,
the overall treatment contact time was
equivalent across treatments; however,
this meant that the participants in the
combined-therapy group did not receive
a comparable dose of supervised exercise
therapy to that received by the exercise
Copyright ${year} Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.
therapy group within the same allocated
time.1 There is a need to investigate the
incremental benefit of providing manual
therapy in addition to exercise therapy
for patients with knee OA compared with
exercise therapy alone.
A weakness of most studies of exer-
cise therapy or manual therapy interven-
tions has been the duration of follow-up,
such that most have only provided evi-
Journal of Orthopaedic & Sports Physical Therapy
dence of short- to medium-term ef-
fectiveness. Few trials have followed
participants beyond 3 to 6 months, and
those that did have generally shown sig-
nificant diminishment of effectiveness in
the longer term.30 Strategies are needed
to facilitate longer-term maintenance of
beneficial effects. Previous investigators
and expert opinion14,17,18,30 recommend
that patients receive regular follow-ups
or booster sessions in the delivery of
physical therapy. A recent systematic re-
view heralded the benefits of booster ses-
sions in the delivery of exercise therapy
for people with OA of the hip or knee30;
however, critical examination of the
studies from which these recommen-
dations were made reveals that most of
those primary studies did not, in fact, in-
vestigate the effectiveness of booster ses-
sions compared to no booster sessions.
Only 1 study testing this proposition has
appeared in the literature, which found
no benefit from booster sessions.10 There
2 | ahead of print | journal of orthopaedic & sports physical therapy
45-12 Abbott-AoP.indd 2
is a pressing need to evaluate the effec-
tiveness of regular follow-up via booster
sessions28,30 for maintaining the benefi-
cial effect of interventions over longer-
term follow-up.
In this randomized clinical trial, we
investigated, in people with knee OA,
the effectiveness of providing manual
therapy in addition to exercise therapy
as a strategy for increasing the ben-
eficial effects of exercise therapy, and
of regularly scheduled booster sessions
as a strategy for optimizing the ben-
eficial effects of interventions to 1-year
follow-up.
The specific aims of the trial were
(1) to investigate the effects of manual
therapy combined with exercise therapy,
compared with exercise therapy alone,
in improving pain, disability, and physi-
cal function; and (2) to compare the ef-
fects of delivering the physical therapy
intervention using periodic booster ses-
sions versus not using booster sessions in
improving pain, disability, and physical
function at 1-year follow-up.
METHODS
Design
T
he present study was a random-
ized controlled trial (RCT) with a
parallel-group, factorial design and
a 1-year follow-up period. Data were col-
lected at the Outpatient Physiotherapy
and Orthopaedics Departments, Duned-
in Hospital, New Zealand. The study was
approved by the Lower South Regional
Ethics Committee of the New Zealand
Ministry of Health (LSR/10/11/055)
and registered with the Australian New
Zealand Clinical Trials Registry (AC-
TRN12612000460808). Participants
were recruited in Dunedin, New Zea-
land from 3 sources: patients present-
ing to physical therapy with knee pain,
patients referred for orthopaedic consul-
tation for knee OA but not eligible for
joint replacement surgery, and people
with knee OA on our clinical trials mail-
ing list. All participants provided written
informed consent.
]
A research nurse screened each po-
tential participant against the inclusion
and exclusion criteria by chart review
and telephone interview. Eligible partici-
pants had to be 40 years of age or older
and meet the American College of Rheu-
matology clinical criteria for a diagnosis
of knee OA.5,6 Exclusion criteria were
rheumatoid arthritis; previous knee or
hip joint replacement surgery of the af-
fected joint; any other surgical proce-
dure on the lower limbs in the previous 6
months; surgical procedure on the lower
limbs planned in the next 6 months; ini-
tiation of opioid analgesia or corticoste-
roid or analgesic injection intervention
for hip or knee pain within the previous
30 days; physical impairments unrelated
to the hip or knee that would prevent safe
participation in exercise, manual therapy,
walking, or stationary cycling; inability to
comprehend and complete study assess-
ments or comply with study instructions;
or stated inability to attend or complete
the proposed course of intervention and
follow-up schedule.
Potential participants attended an
appointment, at which an assessor con-
firmed their eligibility and obtained
written informed consent and baseline
measures. Baseline and follow-up testing
was conducted by research staff blinded
to group allocation. Eligible participants
were randomly allocated to each group
by a researcher who was not involved in
participant assessment or treatment. The
random allocation sequence was gener-
ated (by J.H.A.) with an online service
(http://www.randomization.com), in-
cluded randomly permuted blocks of 8
and 12 participants per block, and was
concealed from recruitment staff, asses-
sors, and treatment providers. The allo-
cation ratio was 1:1:1:1.
Interventions
Following baseline testing, participants
with knee OA were randomized to 1 of
4 groups: (1) exercise therapy without
booster sessions (Ex), (2) exercise therapy
with booster sessions (ExB), (3) exercise
therapy plus manual therapy with no
9/28/2015 3:09:01 PM
 TABLE 1 Brief Description of the Exercise Therapy and Manual Therapy Interventions*

Therapy Type/Intervention Description

Exercise
Mandatory interventions 1. Aerobic exercise: up to 10 minutes, cycle or walk
2. Strengthening: 3 sets of 10 repetitions of knee extension, hip extension, knee flexion. Resistance adjusted
as appropriate
Downloaded from www.jospt.org at University of Otago on October 1, 2015. For personal use only. No other uses without permission.

3. Stretching: 60-second passive stretch of knee flexors, knee extensors, ankle plantar flexors
4. Neuromuscular coordination control exercises: 3 sets of 2 minutes of (choose from) standing weight shifting,
standing balance on uneven surfaces, sidestepping, forward/backward and shuttle walking drills, stair walking
Secondary (nonmandatory) interventions, prescribed when 1. Ankle plantar flexor strengthening, hip abductor strengthening, hip lateral rotator strengthening, hip flexor
indicated by assessment findings and knee extensor stretching, trunk muscle strengthening
Home exercise program 1. Prescribe up to 6 of the above activities to reinforce clinic interventions
Manual
Mandatory interventions 1. Knee flexion: nonthrust physiologic motion
Copyright ${year} Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

2. Anteroposterior-directed force to the tibia, tibiofemoral joint: nonthrust

Secondary (nonmandatory) interventions prescribed when
indicated by assessment findings
Journal of Orthopaedic & Sports Physical Therapy
3. Knee extension: nonthrust physiologic motion
4. Posteroanterior-directed force to the tibia, tibiofemoral joint: nonthrust
5. Patellar gliding force: nonthrust
6. Manual stretch to quadriceps, hamstrings, triceps surae muscles
7. Soft tissue manipulation: quadriceps and peripatellar connective tissue, hamstrings, hip adductors, and triceps
surae muscles
1. Long-axis hip distraction with thrust
2. Lateral hip distraction: nonthrust

3. Anteroposterior-directed force to proximal femur: nonthrust

4. Posteroanterior-directed force to proximal femur: nonthrust
5. Medial hip rotation: nonthrust
6. Soft tissue manipulation to hip and thigh musculature and fascia
7. Manual stretches to connective tissue of hip and thigh
8. Ankle and talocalcaneal joint distraction: thrust or nonthrust
9. Ankle talocrural anteroposterior-directed force: nonthrust
10. Anteroposterior-directed force to distal fibula, tibiofibular joint: nonthrust
11. Soft tissue manipulation: ankle plantar flexor muscle group
12. Lumbopelvic rotation: thrust manipulation
Home program of reinforcing activities 1. Prescribe up to 6 range-of-motion activities to reinforce clinic interventions
*A full description of the intervention protocol and procedures is available from the corresponding author.

Supervised exercise physical therapy. The exercise therapy protocol did not allow therapist-applied manual forces.

Individualized manual physical therapy. Manual therapy was defined as skilled therapistapplied manual procedures intended to modify the quality
and range of motion of the target joint and associated soft tissue structures. The manual therapy protocol did not provide or prescribe aerobic, strengthening,
or neuromuscular control exercises. The home program of reinforcing activities did not include any exercise therapy exercises.

booster sessions (Ex+MT), or (4) exer-
cise therapy plus manual therapy with
booster sessions (ExB+MT). Factor 1 was
therefore booster versus no booster; fac-
tor 2 was exercise alone versus exercise
plus manual therapy.
The exercise therapy protocol consist-
ed of a multimodal, supervised program
of warm-up/aerobic, muscle strengthen-
ing, muscle stretching, and neuromuscu-
lar control exercises (TABLE 1). Additional
exercise therapy interventions were pre-
scribed individually for each participant,
based on the physical examination find-
ings, from a limited list of interventions.
In the case of bilateral symptoms, the
limb with the greatest numeric pain-rat-
ing scale (NPRS) score at baseline assess-

journal of orthopaedic & sports physical therapy | ahead of print | 3

45-12 Abbott-AoP.indd 3 9/28/2015 3:09:01 PM

 [ research report ]
ment was considered the affected limb,
and treatment was targeted to that limb. Characteristics of Participants
TABLE 2
All participants were provided twelve at Entry to the Trial*
45-minute sessions of exercise therapy,
supervised and progressed by a physical ExB+MT
therapist. The dose of exercise therapy Ex (n = 19) ExB (n = 19) Ex+MT (n = 18) (n = 19)
was therefore balanced across all of the Demographic
groups. Men, n (%) 8 (42) 8 (42) 6 (33) 7 (37)
Participants randomized to the no- Women, n (%) 11 (58) 11 (58) 12 (67) 12 (63)
Downloaded from www.jospt.org at University of Otago on October 1, 2015. For personal use only. No other uses without permission.

booster condition received 12 consecutive

Age, y 64 10 65 10 61 12 64 10.2
sessions of their assigned intervention in
Body mass index, kg/m2 29.2 6.1 30.2 5.6 27.6 4.7 29.8 6.6
the first 9 weeks of the study. We defined
booster sessions as sessions of supervised Clinical
therapy provided at time intervals sepa- WOMAC score (0-240) 70.9 45.1 108.4 54.8 71.1 42.8 93.5 50.1
rated from the consecutive sessions of the Pain-intensity score (0-10) 2.1 1.2 3.4 2.1 2.8 1.9 2.5 1.5
initial episode of care, with intervening Timed up-and-go test, s 7.8 1.8 7.8 2.3 7.2 2.1 9.2 3.1
periods of no supervised therapy provi-
40-meter self-paced walk time, s 31.7 5.8 32.2 7.3 30.1 7.3 35.7 10.1
sion. Participants randomized to receive
30-second sit-to-stand test, n 10.8 5.1 10.8 3.1 12.2 4.7 8.3 5.0
Copyright ${year} Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

booster sessions had their assigned inter-

vention distributed in the following man- Bilateral symptoms, n (%) 10 (53) 12 (63) 6 (33) 13 (68)
ner: 8 consecutive sessions in the first 9 Duration of symptoms, n
weeks, 2 booster sessions at 5 months, 1 Less than 1 y 3 4 4 5
booster session at 8 months, and 1 boost- 1-2 y 2 3 4 4
er session at 11 months, also for a total
3-5 y 3 3 1 3
of 12 sessions. We selected 3-month in-
5-10 y 9 4 2 3
tervals between booster sessions because
it appears likely that benefits from exer- More than 10 y 2 5 7 4

cise programs may diminish within this Abbreviations: Ex, exercise therapy without booster sessions; Ex+MT, exercise therapy plus manual
Journal of Orthopaedic & Sports Physical Therapy

period.30
The manual therapy protocol consist-
ed of procedures intended to modify the
quality and range of motion of the target
joint and associated soft tissue structures.
Additional manual therapy interventions
were prescribed individually for each
participant randomized to this interven-
tion, based on the physical examination
findings, from a limited list of interven-
tions defined in our protocol (TABLE 1).
This manual therapy protocol has been
shown to be effective in earlier research.1
Participants allocated to manual therapy
were provided twelve 30- to 45-minute
sessions of manual therapy in addition to
the exercise therapy sessions.
All interventions were provided by
physical therapists at Dunedin Hospi-
tal, following training in delivery of the
intervention protocols. Audits of treat-
ment records were conducted through-
out the trial to assess provider adherence
and treatment progression. Due to the
therapy with no booster sessions; ExB, exercise therapy with booster sessions; ExB+MT, exercise
therapy plus manual therapy with booster sessions; MT, manual therapy; WOMAC, Western Ontario
and McMaster Universities Osteoarthritis Index.
*Values are mean SD unless otherwise indicated.
Lower scores represent less pain, stiffness, and disability.
Higher scores represent more pain.
nature of the interventions, it was not We also assessed treatment success,
possible to blind treatment providers to defined according to the Outcome Mea-
group allocation. sures in Rheumatoid Arthritis Clinical
Trials-Osteoarthritis Research Society
Outcome Measures International (OMERACT-OARSI) re-
The primary outcome variable was sponder criteria29 as either (1) greater
change in the Western Ontario and Mc- than or equal to 50% improvement in the
Master Universities Osteoarthritis Index WOMAC pain or function subscales and
(WOMAC) total score. We used WOMAC an absolute improvement of greater than
Version NRS 3.1, in which each of the or equal to 20 points when transformed
24 items is rated on a 0-to-10 numeric to a 0-to-100 scale, or (2) at least 2 of
rating scale, for a total scale range of 0 the following: (a) pain reduction greater
to 240 points. The WOMAC is a well- than or equal to 20% from baseline and
established, disease-specific measure of absolute change greater than or equal to
pain, stiffness, and physical function for 10 (WOMAC pain score, 0-100 scale), (b)
individuals with knee OA.25 Minimum functional improvement greater than or
important change has been shown to be equal to 20% from baseline and abso-
approximately 20% of baseline score.7 lute change greater than or equal to 10

4 | ahead of print | journal of orthopaedic & sports physical therapy

45-12 Abbott-AoP.indd 4 9/28/2015 3:09:01 PM

 Assessed for eligibility, n = 134 Declined to participate, n = 3
Did not meet inclusion criteria, n = 4
Less than 40 years of age, n = 3
Did not meet ACR criteria, n = 1
Excluded, n = 52
Unable to attend, n = 16
Inflammatory arthritis, n = 10
Prior TKR or TKR planned, n = 6
Other prior surgery, n = 2
Downloaded from www.jospt.org at University of Otago on October 1, 2015. For personal use only. No other uses without permission.

Insufficient communication ability, n = 2

Randomized, n = 75 Other reasons, n = 16
Allocation

Ex, n = 19 ExB, n = 19 Ex+MT, n = 18 ExB+MT, n = 19

Received allocated interven- Received allocated interven- Received allocated interven- Received allocated interven-
tion, n = 19 tion, n = 19 tion, n = 18 tion, n = 19
Copyright ${year} Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

Lost to follow-up, n = 1 Lost to follow-up, n = 3 Lost to follow-up, n = 1 Lost to follow-up, n = 4

12-month
follow-up

Unable to contact, n = 1 Illness, n = 1 Time commitments, n = 1 Unable to contact, n = 1

Personal reasons, n = 1 Declined, n = 2
Unknown, n = 1 Declined due to pain, n = 1
Journal of Orthopaedic & Sports Physical Therapy

Analysis

Included in complete case Included in complete case Included in complete case Included in complete case
analysis, n = 18 analysis, n = 16 analysis, n = 17 analysis, n = 15

FIGURE. Flow diagram. Abbreviations: ACR, American College of Rheumatology; Ex, exercise therapy without booster sessions; Ex+MT, exercise therapy plus manual therapy
with no booster sessions; ExB, exercise therapy with booster sessions; ExB+MT, exercise therapy plus manual therapy with booster sessions; TKR, total knee replacement.

(WOMAC function score, 0-100 scale), or
(c) patient global assessment of change of
4 or greater (scale, 7 to 7) assessed using
the global rating of change instrument.21
Secondary outcome measures in-
cluded the NPRS,2 the timed up-and-go
test, the 30-second sit-to-stand test, and
the 40-meter fast-paced walk test.16 As-
sessors were trained in the assessment
methods and blinded to group allocation.
Statistical Analysis
An analysis of covariance (ANCOVA) is
the preferred method of analyzing RCTs
with paired baseline and follow-up val-
ues.32 We therefore used ANCOVA linear
regression as the primary analysis to as-
sess main effects and the interaction ef-
fect between groups, adjusting for age,
sex, and bilateral symptoms at base-
line.24,27 To determine whether the data
violated statistical assumptions of linear
regression, we used Shapiro-Wilk and
skewness/kurtosis tests of the regression
residuals and residuals plots. An inten-
tion-to-treat analysis, with 20 multiple
imputations for each missing value,33
was conducted with the mi suite of
commands in the Stata Version 13.1 sta-
tistical package (StataCorp LP, College
Station, TX), using the same explanatory
variables as those of the primary analysis.
The appropriateness of imputed values
was assessed. Multiple imputed (pri-
mary) and complete case analyses were
performed for the primary outcome,
as recommended by CONSORT, with
the imputed analysis performed for the
intention-to-treat analysis and the com-
plete case analysis to reveal any sensitiv-
ity to the imputed values.26
We also reported the number needed
to treat (NNT) by the trial intervention
groups (groups 2-4) compared with the
reference group (group 1, Ex) to achieve
a gain of 1 additional OMERACT-OAR-
SI responder, as well as mean change in
the secondary outcome measures from
baseline to 1-year follow-up. Differences
in means were analyzed using an AN-
COVA adjusted for age, sex, and bilateral

journal of orthopaedic & sports physical therapy | ahead of print | 5

45-12 Abbott-AoP.indd 5 9/28/2015 3:09:01 PM

 [ research report ]
symptoms at baseline. Only the results of
complete case analyses were reported for Results of the Adjusted Analysis-of-
TABLE 3
secondary outcomes. Covariance Regression Models*

RESULTS Multiple Imputation, m = 20 (n = 75) Complete Case Analysis (n = 66)

O
Coefficient P Value Coefficient P Value
f 75 participants recruited
Treatment groups
from April 2011 to June 2012, 66
(88%) were retained at 1-year fol- Ex Reference ... Reference ...
Downloaded from www.jospt.org at University of Otago on October 1, 2015. For personal use only. No other uses without permission.

low-up. The participants are described ExB 46.0 (80.0, 12.0) .009
46.9 (79.4, 14.4) .005
in TABLE 2. The FIGURE shows the flow of Ex+MT 37.5 (69.7, 5.3) .023 37.3 (68.8, 5.8) .021
participants through the trial. ExB+MT 1.5 (35.3, 32.3) .928 2.0 (34.9, 30.9) .905
In the intention-to-treat ANCOVA fac-
Age 0.2 (1.4, 1.0) .747 0.2 (1.3, 0.9) .727
torial model with multiple imputation of
Sex
missing outcome data, the main effects
for both manual therapy (WOMAC score, Male Reference ... Reference ...
37.49 points; 95% confidence interval Female 5.2 (29.2, 18.9) .669 4.4 (28.6, 19.7) .715
[CI]: 69.72, 5.26; P = .023) and booster Bilateral symptoms
Copyright ${year} Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

sessions (WOMAC score, 46.02 points; No Reference ... Reference ...

95% CI: 80.04, 11.99; P = .009) were
clinically and statistically significant. How-
ever, the factorial analysis also identified a
statistically significant interaction effect (P
= .001) between interventions of manual
therapy and booster sessions, which there-
fore required an analysis of all results by
group, with group 1 (Ex) as the reference
Yes 12.2 (13.4, 37. 8) .342 12.7 (11.3, 36.8) .293
Abbreviation: Ex, exercise therapy without booster sessions; Ex+MT, exercise therapy plus manual
therapy with no booster sessions; ExB, exercise therapy with booster sessions; ExB+MT, exercise
therapy plus manual therapy with booster sessions; MT, manual therapy.
*Values are Western Ontario and McMaster Universities Osteoarthritis Index points (0-240 scale),
with lower scores (negative difference) indicating improvement.
Adjusted for age, sex, and bilateral symptoms at baseline.
Values in parentheses are 95% confidence interval.
Statistically significant at 2-sided P<.05.

group.24,27 The results, reported in TABLE

Journal of Orthopaedic & Sports Physical Therapy

3, indicate that groups 2 and 3 (ExB and DISCUSSION builds on a recent report that found both
Ex+MT) showed outcomes superior to exercise therapy and manual therapy in

those of the reference group (Ex), while the
outcomes of group 4 (ExB+MT) were not
superior to those of Ex. The model using
only complete cases was consistent with
the imputed intention-to-treat model.
Treatment success, as defined by the
OMERACT-OARSI responder criteria, was
observed among 36 of 66 (54.5%) complete
cases. The NNT, reported in TABLE 4, was
statistically significant for groups 2 and 3
(ExB and Ex+MT) compared with the ref-
erence group (Ex). Group 4 (ExB+MT) was
not superior to Ex. These results were con-
sistent with the ANCOVA model.
Change in pain intensity (NPRS)
significantly favored the Ex+MT group
(TABLE 4). While change in physical per-
formance test scores generally favored
group 2 (ExB), the only statistically
significant difference was seen for the
30-second sit-to-stand test, with ExB su-
perior to Ex (TABLE 4).
O
ur results indicated that pro-
viding either manual therapy or
booster sessions, in addition to
exercise therapy, conferred incremental
benefits over providing exercise therapy
alone. However, our results did not sup-
port the hypothesis that providing both
manual therapy and booster sessions in
addition to exercise results in incremental
benefit. In fact, we detected a strong in-
teraction effect between manual therapy
and booster sessions that resulted in a di-
minished effect in that combined group.
These results indicate that benefits,
in terms of pain and self-reported dis-
ability at 1-year follow-up, result from
providing 12 sessions of manual therapy
in addition to exercise therapy. Earlier
studies showed conflicting results re-
garding the effectiveness of combined
manual therapy plus exercise therapy
for knee OA.14,15,20,23 The present finding
addition to usual medical care to provide
benefits superior to those of usual medi-
cal care alone.1 That trial also included
combined exercise therapy plus manual
therapy, which did not provide signifi-
cant benefits over either intervention
alone. However, in that trial, all condi-
tions received approximately the same
treatment contact time. Consequently,
in the combined exercise therapy plus
manual therapy condition the delivered
dose was reduced compared with that in
the separately delivered exercise therapy
and manual therapy conditions.1 This is
an important distinction between that
trial and the present trial, in which all
exercise conditions were equivalent and
the manual therapy sessions were addi-
tional. This trial indicates that, keeping
content and dose of exercise equivalent,
the addition of manual therapy confers
additional benefits. However, that in-

6 | ahead of print | journal of orthopaedic & sports physical therapy

45-12 Abbott-AoP.indd 6 9/28/2015 3:09:02 PM

 TABLE 4 Changes in Secondary Outcome Measures From Baseline to 1-Year Follow-up*

Secondary Outcome Ex ExB Ex+MT ExB+MT

Mean SD (range) total treatment time, min 440 137 (60-570) 346 167 (0-555) 741 192 (227-960) 622 237 (180-970)
Total WOMAC score (0-240) 5.0 (14.2, 24.3) 51.1 (82.2, 20.0) 34.2 (57.5, 11.0) 3.3 (30.9, 24.2)
Pain-intensity score (0-10) 1.0 (0.3, 2.3) 1.0 (2.3, 0.3) 1.3 (2.5, 0.1) 1.2 (0.4, 2.8)
Downloaded from www.jospt.org at University of Otago on October 1, 2015. For personal use only. No other uses without permission.

Timed up-and-go test, s

0.4 (0.4, 1.3) 0.6 (1.9, 0.7) 0.4 (0.8, 1.5) 0.3 (1.4, 2.0)
40-meter self-paced walk time, s 0.5 (3.7, 2.7) 2.4 (6.1, 1.5) 1.4 (4.3, 1.6) 1.1 (1.5, 3.6)
30-second sit-to-stand test, n 0.2 (1.8, 2.1) 2.7 (1.4, 3.9) 2.1 (0.7, 3.4) 1.2 (0.9, 3.3)
Adverse events, n 1# 0 0 1**
NNT to gain 1 additional OMERACT-OARSI responder

Reference 2.8 (1.7, 50.5) 2.7 (1.7, 23.4) NA
Abbreviations: Ex, exercise therapy without booster sessions; Ex+MT, exercise therapy plus manual therapy with no booster sessions; ExB, exercise therapy
with booster sessions; ExB+MT, exercise therapy plus manual therapy with booster sessions; MT, manual therapy; NA, not available; NNT, number needed
to treat; OMERACT-OARSI, Outcome Measures in Rheumatoid Arthritis Clinical Trials-Osteoarthritis Research Society International responder criteria;
WOMAC, Western Ontario and McMaster Universities Osteoarthritis Index.
*Complete case analysis (n = 66). Values are mean change (95% confidence interval) in test score (or time) from baseline to 1 year unless otherwise indicated.
Copyright ${year} Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

A lower score (negative difference) is better.

P<.05 compared with exercise therapy only (reference), adjusted for baseline value, age, sex, and bilateral symptoms at baseline.

Negative scores indicate reduced pain.

Negative times represent shorter time to complete, indicating improvement.

Positive values represent more repetitions, indicating improvement.
#
Possibly trial-related hip pain associated with exercise.
**Possibly trial-related fall onto knee associated with exercise.

As defined in Pham et al.29

Unable to calculate (nonsignificantly less than reference).

cremental benefit carries the additional was observed performing the home ex- An important difference between this
cost of providing the manual therapy in- ercises, whereas in the present trial the trial and the previous trial investigating
Journal of Orthopaedic & Sports Physical Therapy

tervention, which required an additional
300 minutes per participant of therapist
time (TABLE 4).
The results indicating benefit from
booster sessions are contrary to the re-
cent findings of Bennell et al,10 who found
that 2 booster sessions did not influence
pain or physical function outcomes in
patients with knee OA who completed
a 12-week course of physical therapist
supervised exercise. The age of the par-
ticipants, the duration and severity of
their symptoms, the content of the inter-
vention, its intensity, and the number of
visits were similar to those of the present
study. However, the booster sessions were
shorter in the trial by Bennell et al10 (30
minutes) compared to the present trial
(approximately 45 minutes in the ExB
group and 75 minutes in the ExB+MT
group). The content of the booster ses-
sions might have differed slightly. In the
trial by Bennell et al,10 the home exercise
program was reviewed and the patient
full supervised intervention protocol was
completed and the home exercises were
reviewed. Also, the duration of follow-up
was shorter in that trial (36 weeks) com-
pared to that of the current trial (1 year),
while within-group treatment standard-
ized effect size (calculated from reported
results) was greater in the trial by Bennell
et al10 (approximately 0.78 at 13 weeks11)
than it was in the present trial (0.1 for ex-
ercise alone at 1 year). This indicates that
there was considerably more room for
improvement in treatment effect in the
current trial than there was in the study
by Bennell et al.10,11 It also reveals an
unexpectedly low treatment effect from
the Ex intervention group in the current
trial, compared with the previous trial on
which its treatment protocol was based.1
However, the current trial had a lower
number of participants per group, and
therefore greater uncertainty of treat-
ment effect and higher risk of chance
findings compared with larger trials.
the effectiveness of booster sessions10 was
that, in the present trial, the nonbooster
and booster conditions had equal treat-
ment time, whereas in the study by Ben-
nell et al,10 the booster condition had
2 additional sessions. In this trial, all
groups had 12 exercise therapy sessions,
only distributed differently over the
12-month duration of the trial. In fact,
the ExB group participants consumed
93 minutes less therapist contact time,
on average, over the 12 months. Yet, we
found superior benefits in terms of self-
reported disability and the 30-second
sit-to-stand test from distributing ses-
sions over 12 months, to include booster
sessions, rather than providing all 12 ses-
sions in the first 9 weeks.
The finding of an adverse interac-
tion effect between manual therapy and
booster sessions, that is, lower WOMAC
change in the combination condition
compared with each intervention indi-
vidually, was perplexing. The ExB+MT

journal of orthopaedic & sports physical therapy | ahead of print | 7

45-12 Abbott-AoP.indd 7 9/28/2015 3:09:02 PM

 [ research report
group did receive less therapist contact
time than the Ex+MT group (622 min-
utes compared with 740 minutes, on
average), but more than the Ex and ExB
groups (TABLE 4). This trial was not de-
signed to reveal mechanisms, so potential
explanations are conjecture. The simplest
and therefore most likely explanation is
that the interaction effect may be sim-
Downloaded from www.jospt.org at University of Otago on October 1, 2015. For personal use only. No other uses without permission.
ply a chance finding due to the relatively
small individual group sizes (n = 18 or
19 per group) in this trial, which was
intended to test the main effects within
the factorial design (ie, n = 37 or 38 per
factor). This might have introduced in-
stability in the resulting per-group find-
ings, leading to type II error in finding no
significant effect for the ExB+MT group.
Copyright ${year} Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.
This explanation may indicate that in-
terpretation should rest with the pri-
mary factorial analysis, which indicated
that providing either manual therapy or
booster sessions, in addition to exercise
therapy, conferred incremental benefits
over providing exercise therapy alone. If
the interaction results represent a real
phenomenon, the mechanism cannot be
determined and would require further
Journal of Orthopaedic & Sports Physical Therapy
investigation.
A strength of the present trial was
the use of exercise therapy and manual
therapy protocols previously shown to be
effective.1 Also, all participants were allo-
cated a standard dose of twelve 45-min-
ute sessions of standardized exercise
therapy, so all comparisons can be attrib-
uted to the allocated group conditions.
We consider the results to be generaliz-
able to clinical practice, as we excluded
few patients with knee OA, the included
patients represented a wide spectrum of
symptom severity, the trial was set in a
busy hospital physical therapy clinic, and
the clinicians providing the interventions
were not selected on the basis of special-
ist skills or qualifications and received
only in-service training. A potential
threat to the analysis was the continu-
ous outcome (WOMAC) being skewed
at follow-up assessment, which risks
violating the assumptions of common
parametric statistical methods, such as
8 | ahead of print | journal of orthopaedic & sports physical therapy
45-12 Abbott-AoP.indd 8
linear regression. Given that supervised
exercise is a known effective therapy,13,19
it was not surprising that improvement
in all groups at follow-up led to skewed
distribution of the WOMAC scores. This
is common following treatment in mus-
culoskeletal conditions.3 However, AN-
COVA has been established as robust to
nonnormally distributed data31 and is the
preferred statistical analysis for RCTs,32
and our statistical analysis of the regres-
sion residuals indicated acceptable con-
formance to a normal distribution. The
NNT analysis of difference in propor-
tions of responders between the groups
(TABLE 4) is not subject to distributional
assumptions and was consistent with
the ANCOVA. Another potential weak-
ness was the chance inequivalence in the
primary outcome (WOMAC) at baseline;
however, the ANCOVA controls for base-
line, and all of our analyses resulted in
the same interpretations for both self-
reported disability and pain. Therefore,
we believe that our results are statisti-
cally robust. Both the factorial and the
per-group analyses showed statistically
significant results in favor of the booster
and manual therapy interventions. How-
ever, because we were forced to conduct
a per-group analysis following the facto-
rial analysis, due to the adverse inter-
action effect between the main factors,
this resulted in a relatively small sample
size per group. Therefore, the data may
be variable and susceptible to chance
findings, so we recommend further
investigation.
While the results of this trial indicate
that booster sessions or manual therapy
provide incremental benefits in addition
to exercise therapy, these results are not
definitive: the booster session results in
this trial do not concur with the results
of Bennell et al,10 the Ex group in the
current trial showed lower-than-expect-
ed treatment effect compared with the
protocol on which it was based,1 and the
interaction effect seen in the ExB+MT
group was contradictory. The dose of in-
terventions delivered has differed among
trials in the literature and is very likely to
]
be a factor determining treatment effect.
The incremental effectiveness of these
interventions therefore remains unre-
solved, and further research will be nec-
essary to resolve the aims of this trial. We
are aware of a larger, recently completed
study (ClinicalTrials.gov; NCT01314183),
the results of which provide further
evidence toward establishing whether
booster sessions and/or manual therapy
provide incremental benefits in addition
to exercise therapy.
CONCLUSION
D
istributing supervised exercise
therapy sessions over the course of
1 year, in the form of 8 initial ses-
sions in the first 2 months, then 4 booster
sessions at 5, 8, and 11 months, improved
outcomes at 1 year compared with deliv-
ery of 12 consecutive sessions of super-
vised exercise therapy within 2 months,
without additional therapist contact
time. The addition of 12 sessions of in-
dividually tailored manual therapy to 12
sessions of supervised exercise therapy,
delivered over 2 months, also improved
outcomes at 1 year, while also requiring
additional therapist time. However, pro-
viding manual therapy distributed over
12 months using booster sessions in ad-
dition to supervised exercise therapy did
not provide incremental benefit at 1 year
compared with delivery of 12 consecutive
sessions of supervised exercise therapy
alone, and required additional thera-
pist time. Further research is required
to establish the incremental benefits of
booster sessions and/or manual therapy
in addition to exercise therapy. t
KEY POINTS
FINDINGS: Providing manual therapy in
addition to exercise therapy improved
treatment effectiveness compared with
providing 12 consecutive exercise thera-
py sessions only. Distributing 12 exercise
therapy sessions over a year in the form
of booster sessions was more effective
than providing consecutive exercise
therapy sessions.
9/28/2015 3:09:02 PM
 IMPLICATIONS: The addition of manual
therapy, or distribution of exercise ther-
apy sessions over a year, enhances the
treatment effectiveness of 12 sessions of
supervised exercise therapy.
CAUTION: The findings, including that of
an adverse interaction effect between
manual therapy and booster sessions,
may be due to chance, as a result of
Downloaded from www.jospt.org at University of Otago on October 1, 2015. For personal use only. No other uses without permission.
small per-group sample size.
REFERENCES
1. Abbott JH, Robertson MC, Chapple C, et al.
Manual therapy, exercise therapy, or both, in
addition to usual care, for osteoarthritis of the
hip or knee: a randomized controlled trial. 1:
clinical effectiveness. Osteoarthritis Cartilage.
2013;21:525-534. http://dx.doi.org/10.1016/j.
Copyright ${year} Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.
joca.2012.12.014
2. Abbott JH, Schmitt J. Minimum important
differences for the Patient-Specific Functional
Scale, 4 region-specific outcome measures, and
the numeric pain rating scale. J Orthop Sports
Phys Ther. 2014;44:560-564. http://dx.doi.
org/10.2519/jospt.2014.5248
3. Abbott JH, Schmitt JS. The Patient-Specific
Functional Scale was valid for group-level
change comparisons and between-group dis-
crimination. J Clin Epidemiol. 2014;67:681-688.
http://dx.doi.org/10.1016/j.jclinepi.2013.11.002
Journal of Orthopaedic & Sports Physical Therapy
4. Access Economics. The Economic Cost of
Arthritis in New Zealand in 2010. Kingston, Aus-
tralia: Access Economics; 2010.
5. Altman R, Asch E, Bloch D, et al. Development
of criteria for the classification and reporting of
osteoarthritis: classification of osteoarthritis of
the knee. Arthritis Rheum. 1986;29:1039-1049.
http://dx.doi.org/10.1002/art.1780290816
6. Altman RD. Criteria for classification of clinical
osteoarthritis. J Rheumatol Suppl. 1991;27:10-12.
7. Angst F, Aeschlimann A, Michel BA, Stucki G.
Minimal clinically important rehabilitation ef-
fects in patients with osteoarthritis of the lower
extremities. J Rheumatol. 2002;29:131-138.
8. Bennell KL, Dobson F, Hinman RS. Exercise
in osteoarthritis: moving from prescription to
adherence. Best Pract Res Clin Rheumatol.
2014;28:93-117. http://dx.doi.org/10.1016/j.
berh.2014.01.009
9. Bennell KL, Egerton T, Pua YH, Abbott JH, Sims
K, Buchbinder R. Building the rationale and
structure for a complex physical therapy inter-
vention within the context of a clinical trial: a
multimodal individualized treatment for patients
with hip osteoarthritis. Phys Ther. 2011;91:1525-
1541. http://dx.doi.org/10.2522/ptj.20100430
10. Bennell KL, Kyriakides M, Hodges PW, Hinman
45-12 Abbott-AoP.indd 9
RS. Effects of two physiotherapy booster sessions
on outcomes with home exercise in people with
knee osteoarthritis: a randomized controlled
trial. Arthritis Care Res (Hoboken). 2014;66:1680-
1687. http://dx.doi.org/10.1002/acr.22350
11. Bennell KL, Kyriakides M, Metcalf B, et al. Neu-
romuscular versus quadriceps strengthening ex-
ercise in patients with medial knee osteoarthritis
and varus malalignment: a randomized controlled
trial. Arthritis Rheumatol. 2014;66:950-959.
http://dx.doi.org/10.1002/art.38317
12. Buckwalter JA, Saltzman C, Brown T. The impact
of osteoarthritis: implications for research. Clin
Orthop Relat Res. 2004;427:S6-S15.
13. Conaghan PG, Dickson J, Grant RL. Care and
management of osteoarthritis in adults: summa-
ry of NICE guidance. BMJ. 2008;336:502-503.
http://dx.doi.org/10.1136/bmj.39490.608009.AD
14. Deyle GD, Allison SC, Matekel RL, et al. Physical
therapy treatment effectiveness for osteoarthri-
tis of the knee: a randomized comparison of
supervised clinical exercise and manual therapy
procedures versus a home exercise program.
Phys Ther. 2005;85:1301-1317.
15. Deyle GD, Henderson NE, Matekel RL, Ryder MG,
Garber MB, Allison SC. Effectiveness of manual
physical therapy and exercise in osteoarthritis
of the knee: a randomized, controlled trial. Ann
Intern Med. 2000;132:173-181. http://dx.doi.
org/10.7326/0003-4819-132-3-200002010-00002
16. Dobson F, Hinman RS, Roos EM, et al. OARSI
recommended performance-based tests to
assess physical function in people diagnosed
with hip or knee osteoarthritis. Osteoarthritis
Cartilage. 2013;21:1042-1052. http://dx.doi.
org/10.1016/j.joca.2013.05.002
17. Flanagan T, Coburn P, Harcourt P, Zylinski M,
Jull G. Justifying the on-going physiotherapy
management of long-term patients. Man Ther.
2003;8:254-256. http://dx.doi.org/10.1016/
S1356-689X(03)00055-9
18. Flanagan T, Green S. The concept of main-
tenance physiotherapy. Aust J Physiother.
2000;46:271-278.
19. Fransen M, McConnell S. Exercise for os-
teoarthritis of the knee. Cochrane Database
Syst Rev. 2008:CD004376. http://dx.doi.
org/10.1002/14651858.CD004376.pub2
20. French HP, Brennan A, White B, Cusack T. Manu-
al therapy for osteoarthritis of the hip or knee
a systematic review. Man Ther. 2011;16:109-117.
http://dx.doi.org/10.1016/j.math.2010.10.011
21. Jaeschke R, Singer J, Guyatt GH. Measurement
of health status. Ascertaining the minimal clini-
cally important difference. Control Clin Trials.
1989;10:407-415.
22. Jamtvedt G, Dahm KT, Christie A, et al. Physical
therapy interventions for patients with osteo-
arthritis of the knee: an overview of systematic
reviews. Phys Ther. 2008;88:123-136. http://
dx.doi.org/10.2522/ptj.20070043
23. Jansen MJ, Viechtbauer W, Lenssen AF, Hen-
journal of orthopaedic & sports physical therapy | ahead of print | 9
driks EJ, de Bie RA. Strength training alone,
exercise therapy alone, and exercise therapy
with passive manual mobilisation each re-
duce pain and disability in people with knee
osteoarthritis: a systematic review. J Physio-
ther. 2011;57:11-20. http://dx.doi.org/10.1016/
S1836-9553(11)70002-9
24. McAlister FA, Straus SE, Sackett DL, Altman
DG. Analysis and reporting of factorial trials: a
systematic review. JAMA. 2003;289:2545-2553.
http://dx.doi.org/10.1001/jama.289.19.2545
25. McConnell S, Kolopack P, Davis AM. The
Western Ontario and McMaster Universities
Osteoarthritis Index (WOMAC): a review of its
utility and measurement properties. Arthri-
tis Rheum. 2001;45:453-461. http://dx.doi.
org/10.1002/1529-0131(200110)45:5<453::AID-
ART365>3.0.CO;2-W
26. Moher D, Hopewell S, Schulz KF, et al. CONSORT
2010 explanation and elaboration: updated
guidelines for reporting parallel group ran-
domised trials. BMJ. 2010;340:c869. http://
dx.doi.org/10.1136/bmj.c869
27. Montgomery AA, Peters TJ, Little P. De-
sign, analysis and presentation of facto-
rial randomised controlled trials. BMC Med
Res Methodol. 2003;3:26. http://dx.doi.
org/10.1186/1471-2288-3-26
28. Ottawa Panel. Ottawa Panel evidence-based
clinical practice guidelines for therapeutic exer-
cises and manual therapy in the management of
osteoarthritis. Phys Ther. 2005;85:907-971.
29. Pham T, van der Heijde D, Altman RD, et al.
OMERACT-OARSI Initiative: Osteoarthritis Re-
search Society International set of responder
criteria for osteoarthritis clinical trials revisited.
Osteoarthritis Cartilage. 2004;12:389-399.
http://dx.doi.org/10.1016/j.joca.2004.02.001
30. Pisters MF, Veenhof C, van Meeteren NL, et al.
Long-term effectiveness of exercise therapy
in patients with osteoarthritis of the hip or
knee: a systematic review. Arthritis Rheum.
2007;57:1245-1253. http://dx.doi.org/10.1002/
art.23009
31. Vickers AJ. Parametric versus non-parametric
statistics in the analysis of randomized tri-
als with non-normally distributed data. BMC
Med Res Methodol. 2005;5:35. http://dx.doi.
org/10.1186/1471-2288-5-35
32. Vickers AJ, Altman DG. Analysing controlled
trials with baseline and follow up measure-
ments. BMJ. 2001;323:1123-1124. http://dx.doi.
org/10.1136/bmj.323.7321.1123
33. White IR, Royston P, Wood AM. Multiple imputa-
tion using chained equations: issues and guid-
ance for practice. Stat Med. 2011;30:377-399.
http://dx.doi.org/10.1002/sim.4067
@ MORE INFORMATION
WWW.JOSPT.ORG
9/28/2015 3:09:02 PM