Manual Therapy 20 (2015) 38e45

Contents lists available at ScienceDirect

Manual Therapy
journal homepage: www.elsevier.com/math

Original article

Physiotherapy triage assessment of patients referred for orthopaedic

consultation e Long-term follow-up of health-related quality of life,
pain-related disability and sick leave
Karin S. Samsson a, b, *, Maria E.H. Larsson a, c
a
b
c

Department of Clinical Neuroscience and Rehabilitation, The Sahlgrenska Academy, Gothenburg University, Box 430, 405 30 Gothenburg, Sweden
rhalsan Tjo
stra Go
g 1, 471 94 Klleka
rr, Sweden
rn Rehabilitation Clinic, Primary Healthcare, Region Va
taland, Syster Ebbas va
Na
rhalsan Research and Development, Primary Healthcare, Region Va
stra Go
taland, Kungsgatan 12, level 6 411 18 Gothenburg, Sweden
Na

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 20 December 2013
Received in revised form
9 May 2014
Accepted 24 June 2014

Introduction: The literature indicates that physiotherapy triage assessment can be efcient for patients
referred for orthopaedic consultation, however long-term follow up of patient reported outcome measures are not available.
Aim: To report a long-term evaluation of patient-reported health-related quality of life, pain-related
disability, and sick leave after a physiotherapy triage assessment of patients referred for orthopaedic
consultation compared with standard practice.
Methods: Patients referred for orthopaedic consultation (n 208) were randomised to physiotherapy
triage assessment or standard practice. The randomised cohort was analysed on an intention-to-treat
(ITT) basis. The patient reported outcome measures EuroQol VAS (self-reported health-state), EuroQol
5D-3L (EQ-5D) and Pain Disability Index (PDI) were assessed at baseline and after 3, 6 and 12 months. EQ
VAS was analysed using a repeated measure ANOVA. PDI and EQ-5D were analysed using a marginal
logistic regression model. Sick leave was analysed for the 12 months following consultation using a Mann
eWhitney U-test.
Results: The patients rated a signicantly better health-state at 3 after physiotherapy triage assessment
[mean difference 5.7 (95% CI 11.1; 0.2); p 0.04]. There were no other statistically signicant differences in perceived health-related quality of life or pain related disability between the groups at any of
the follow-ups, or sick leave.
Conclusion: This study reports that the long-term follow up of the patient related outcome measures
health-related quality of life, pain-related disability and sick leave after physiotherapy triage assessment did
not differ from standard practice, indicating the possible benets of implementation of this model of care.
2014 Elsevier Ltd. All rights reserved.

Keywords:
Randomised controlled trial
Primary care
EQ-5D
PDI

1. Introduction
Musculoskeletal pain often results in functional limitations in
rnsdo
 ttir et al., 2013) and
daily life (Bingefors and Isacson, 2004; Bjo
many patients with persistent pain experience a high level of
disability, affecting work capacity (Gureje et al., 2001; Gerdle et al.,
2004; Landmark et al., 2013). Persistent pain has also been found to
negatively inuence quality of life (Kroenke et al., 2013; Landmark,
et al., 2013). Studies show that between 45% and 74% of the

rhalsan Tjo
rn Rehabilitation Clinic, Syster Ebbas v. 1,
* Corresponding author. Na
47194 Kllek
arr, Sweden. Tel.: 46 707 5262604; fax: 46 304 660342.
E-mail addresses: karin.samsson@vgregion.se, karin.samsson@gmail.com
(K.S. Samsson).
http://dx.doi.org/10.1016/j.math.2014.06.009
1356-689X/ 2014 Elsevier Ltd. All rights reserved.

population reports musculoskeletal pain over the course of a year

(Picavet and Schouten, 2003; Bingefors and Isacson, 2004) and that
patients with musculoskeletal pain represent up to half of the
consultations in primary care (Picavet and Schouten, 2003; Jordan
et al., 2010; Mnsson et al., 2011). About 20% of patients with
musculoskeletal pain are referred for consultation with an orthopaedic surgeon (Canizares et al., 2009; MacKay et al., 2010) but the
proportion of patients found appropriate for orthopaedic surgery
varies between studies, ranging from 30% to 68% (McHugh et al.,
2011; Menzies and Young, 2012; Samsson and Larsson, 2015). To
provide optimal care for patients with musculoskeletal disorder,
the existing roles of health professionals, such as physiotherapists,
has been extended (Department of Health, 2006) and internationally these specially trained physiotherapists are referred to as

K.S. Samsson, M.E.H. Larsson / Manual Therapy 20 (2015) 38e45

advanced physiotherapy practitioners (APP), or extended scope

physiotherapists (ESP) (Desmeules et al., 2012; Stanhope et al.,
2012). In order to reduce the load on orthopaedic surgeons, to
minimise the number of patients inappropriate for surgery, as well
as shorten waiting times, ESPs or APPs have been used to triage,
diagnose, and determine management plans; to refer for investigations, orthopaedic surgery or conservative management
(Maddison et al., 2004; Aiken and McColl, 2008; Bath and Janzen,
2012). Studies of this model of care have reported a high agreement on diagnoses [good to excellent (kappa (k) 0.69e1.00) and
treatment approach (fair to very good (kappa (k) 0.52e0.70)]
between physiotherapists and orthopaedic surgeons (Desmeules
et al., 2012), as well as a decrease in referrals for orthopaedic
consultation (Rabey et al., 2009). Furthermore, there are indications
of decreased average waiting times for consultation (Blackburn
et al., 2009; Morris et al., 2011) and for surgery (Aiken et al.,
2007). Daker-White et al. (1999) reported in their randomised
controlled trial that they found no differences in patient-centred
measures of pain, functional disability or perceived handicap, and
that this model resulted in lower direct hospital costs. Due to
methodological quality however, scientic evidence is still scarce
(Desmeules et al., 2012). Considering international differences in
healthcare systems, studies of such a model need to be conducted
in each respective country (Stanhope et al., 2012). Research evaluating APPs and ESPs has predominately been carried out in the UK,
Australia and Canada. A formal recognition such as APP or ESP does
not exist in Sweden, nor does the described model of care with
physiotherapy triage. Moreover, to our knowledge, only one study
has been conducted in a Swedish primary healthcare setting. Our
previous paper (Samsson and Larsson, 2015) reports the main
outcomes of this study; physiotherapy triage assessment of patients referred for orthopaedic consultation in primary healthcare
resulted in signicantly higher selection accuracy for appropriateness for orthopaedic surgery [30% (95% CI 11; 49), p 0.002], as
well as a signicantly smaller proportion of referrals back to the
referring general practitioners (GP) [19% (95% CI 29; 9),
p < 0.001] and a larger proportion to physiotherapy [26% (95% CI 13;
39), p < 0.001] when compared with standard practice. Also waiting time was signicantly shorter in the triage group [mean score
19 days (SD 12) versus 28 days (SD 14)] (p < 0.001).
There are indications that this model of care could be effective
and provide a more efcient use of resources; however, Patient
Reported Outcome Measures (PROMs) for this model have not been
evaluated. The aim of this paper was therefore to report a long-term
follow-up of patient-reported health-related quality of life, painrelated disability, and sick leave after a physiotherapy triage
assessment of patients referred for orthopaedic consultation
compared with standard practice.
2. Methods
2.1. Study design
This paper is part of a larger clinical trial, and the full study
design and method has been reported previously (Samsson and
Larsson, 2015). The study design used was a randomised
controlled trial.
2.2. Participants
The study took place at a primary healthcare centre in a Swedish
municipality. Consecutive recruitment was performed between
August 2009 and January 2011, including patients referred for orthopaedic consultation at the healthcare centre with the following
inclusion criteria; working age (between 18 and 67 years of age),

39

subacute (four weeks to three months) or persistent (>three

months) musculoskeletal pain, and the ability to understand
written and spoken Swedish. The exclusion criteria were chosen
together with the orthopaedic surgeon in the study, and patients
were excluded if the stated diagnosis on the referral was hallux
valgus, ganglion or trigger nger, since the GPs were entrusted to
have high specicity of managing these specic diagnoses.
2.3. Procedure
Upon inclusion, patients (n 208) gave verbal consent to
participate, and were then randomised by the administrator
drawing an opaque envelope containing notes marked physiotherapy triage assessment or standard practice, from a box, were
the envelopes were put in bundles of 20 (ten of each) in order to
ensure an even distribution. Prior to the consultation, participants
completed a written informed consent form.
2.3.1. Physiotherapy triage assessment
The physiotherapist, also the rst author of this paper (KS), did
not receive any training specic for this trial. She had specialist
training in form of postgraduate qualications that included a
master's degree in Manipulative Therapy, one year of mentored
clinical practice within the scope of orthopaedic manual therapy
(OMT) and eight years of clinical experience in primary care, four of
which were within the scope of OMT. The triage assessment was
based on a 60-minute screening, with the main aims of diagnosis
and the most appropriate management pathway. In conjunction
with the triage, the patients also received brief treatment
comprised of advice on ergonomics and/or exercises when appropriate, however only during the one visit. Management pathways
consisted of one or more of the following; referral for further
investigation, for orthopaedic surgeon consultation (i.e. appropriate candidate for surgery), back to the patient's GP, or if conservative management with ongoing support was needed, referral
to physiotherapy or occupational therapy. If referral for orthopaedic
surgeon consultation was found appropriate, the physiotherapist
had the authority to make an appointment without consideration
of the waiting list. Referrals for further investigations were
requested and sent via the patient's GP and the images could be
assessed together with the orthopaedic surgeon, if needed. One or
two optional follow-up visits were offered when needed, for
example follow-up after treatment or investigations.
2.3.2. Standard practice
The orthopaedic surgeon had 26 years of experience in orthopaedic medicine, 21 of which were as an orthopaedic specialist. The
duration of the appointment was 15 min, with the same main aims
of diagnosis and the most appropriate management pathway. The
patients received advice, prescriptions or injections, when appropriate. Management pathways consisted of one or more of the
following; further investigation, orthopaedic intervention (i.e. minor surgery at the present healthcare centre), referral to orthopaedic clinics for orthopaedic intervention (i.e. appropriate
candidates for surgery), back to the patient's GP, or if conservative
management with ongoing support was needed, referral to physiotherapy or occupational therapy. One or two optional follow-up
visits were offered when needed, for example follow-up after
investigations.
2.4. Patient related outcome measures
Demographic data were collected to describe the study population and to determine any differences between the groups at
inclusion. PROMs were distributed to the patients at the healthcare

40

K.S. Samsson, M.E.H. Larsson / Manual Therapy 20 (2015) 38e45

centre before the consultation, and by mail at 3, 6 and 12 months

after consultation.
2.4.1. Health-related quality of life
The European Quality of Life-5 Dimensions Questionnaire (EQ5D-3L) was used to assess perceived health-related quality of life
(EuroQol Group, 1990; Brooks, 1996). Subjects are asked to rate
their health state on three severity levels ranging from 1 (no
problems), to 3 (extreme problems) in the areas mobility, self-care,
usual activities, pain and anxiety/depression. Also, the subjects are
asked to rate their current health state on a vertical visual analogue
scale (EQ VAS) with the endpoints 0 (worst imaginable health state)
and 100 (best imaginable health state). The EQ-5D has evidence of
validity and reliability in different musculoskeletal disorders (Soer
et al., 2012a) and orthopaedics (Beard et al., 2010; Devine et al.,
2011). The Swedish version of the EQ-5D was used for this study.
This version has been translated using forward-backward translation (EuroQol Group, 2013).

to the analysis, the ve items of the EQ-5D were dichotomised

according to the User guide (EuroQol Group, 2013); level 1 into no
problems, level 2 and 3 into problems. The PDI was dichotomised
intomedian or below, and above median, using the median score
at baseline for each item (reported in brackets below); family/home
responsibility (3), recreational (5), social activity (0), occupation (2),
sexual behaviour (0), self care (1) as well as life-support activity (0).
Between-group comparison regarding sick leave (days) was performed with a ManneWhitney U-test using the IBM SPSS. Due to
the level of missing data being higher than 20% for various items, at
various times, a between-group drop-out analysis was performed.
We chose the self-reported health state at baseline (EQ VAS) as the
main outcome variable for this analysis. However, since baseline
data for EQ VAS was missing for some participants for whom data
were missing, analysis had to be made using the actual respondents
and was performed using a t-test. Level of signicance for all analyses was set to p < 0.05.
3. Results

2.4.2. Pain-related disability

The Pain Disability Index (PDI) was used to evaluate pain-related
disability. This questionnaire is a seven-item inventory, designed to
measure self-reported disability on a participation level, due to
pain in all body regions (Tait et al., 1987, 1990). Subjects are asked to
rate disability on a numeric rating scale ranging from 0 (no
disability) to 10 (maximum disability) in the areas family and home
responsibilities, recreation, social activity, occupation, sexual
behaviour, self-care, and life-support activity. The instrument has
been tested and is considered valid and reliable for pain-related
disability (Tait, Chibnall, 1990; Soer et al., 2012b, 2013). The
Swedish version of the PDI was used for this study.
2.4.3. Sick leave
Sick leave was measured as the number of days off work
registered at the Swedish Social Insurance Agency for the 12
months following the assessment, due to the disorder or diagnosis
that originated the referral for orthopaedic consultation. Data was
collected in collaboration with the Swedish Social Insurance
Agency.
2.5. Statistical analysis
Sample size was originally calculated for the study's main
outcome variable, management outcome (management pathway),
as reported previously (Samsson and Larsson, 2015). A retrospective sample size calculation was made for this study using the EQ
VAS, which have the suggested minimal clinically important
change (MCIC) of 10.5 units (Soer et al., 2012a). Calculation showed
that 63 patients in each group would have been sufcient to detect
differences in MCIC at the p < 0.05 level, with a power of 80%.
Demographic data was analysed using descriptive statistics with
IBM SPSS, version 18.0 (IBM Corp, Armonk, NY, USA). The randomised cohort was analysed on an intention-to-treat (ITT) basis.
Data from the EQ VAS was treated as a continuous quantitative
variable and analysed with a linear longitudinal model by using
PROC MIXED, which yielded beta (regression coefcients) estimates of least square means, which subsequently were analysed for
between-group differences with a repeated measure ANOVA.
Outcomes of the EQ-5D and PDI were treated as ordinal variables,
and were analysed using a marginal logistic regression model (i.e.
the generalised estimating equations, GEE, model) and estimated
population average odds ratio in SAS for windows, version 9.3 (SAS
Institute Inc. Cary, NC, USA). Odds ratios for binary outcomes were
estimated using repeated logistic regression with PROC GENMOD
(Hosmer and Lemeshow, 2005; Grimby-Ekman et al., 2009). Prior

3.1. Participants
The inclusion process and follow-up is presented in Fig. 1. There
were no statistically signicant differences between the two groups
regarding patient demographics at baseline (Table 1).
3.2. Health-related quality of life
The patients in the physiotherapy triage assessment group rated
a signicantly better health state at 3 months following consultation compared with the standard practice group [mean difference
5.7 (95% CI 11.1; 0.2); p 0.04]. No statistically signicant
differences were found between the groups at baseline (p 0.11) or
after 6 (p 0.06) or 12 months (p 0.14) (Fig. 2). There were no
statistically signicant differences between the groups in odds ratios for reporting no problemsin the EQ-5D at baseline or after 3, 6
and 12 months (Table 2).
3.3. Pain-related disability
We found no statistically signicant differences between the
groups in odds ratios for reporting baseline or below in the PDI at
baseline or after 3, 6 and 12 months (Table 3).
3.4. Sick leave
A small number of patients were on sick leave during the 12
months following consultation; seven in the physiotherapy triage
assessment group [mean days 146, (SD 128)] and 15 in the standard
practice group [mean days 72, (SD 81)]. No statistically signicant
difference between the groups was found (p 0.113).
3.5. Missing data analysis
The patients in the physiotherapy triage assessment group who
responded to EQ VAS at 3 months reported a statistically signicantly better health state at baseline, compared with the patients in
the standard practice group (difference in mean score 7.2, 95%
CI 13.6; 0.8). This means that the statistically signicant ndings
in the main analysis for EQ VAS must be interpreted with caution.
There were similar ndings of signicant differences for self care
(difference in mean score 6.2, 95% CI 12.2; 0.1), anxiety
(difference in mean score 6.2, 95% CI 12.2; 0.1) and mobility
(difference in mean score 6.5, 95% CI 12.6; 0.5) at 3 months.

K.S. Samsson, M.E.H. Larsson / Manual Therapy 20 (2015) 38e45

41

Fig. 1. CONSORT ow chart of the patients through the study. PT physiotherapy, m months, EQ VAS EuroQol VAS (Self-reported health-state), EQ-5D EuroQol-5D,
PDI Pain Disability Index.

However, the ndings in the main analysis for these items were not
statistically signicant.
4. Discussion
The ndings in this study show that patients in the physiotherapy triage assessment group perceived better self-rated health
state (EQ VAS) after 3 months compared with patients in the
standard practice group. However, this nding must be interpreted
with caution since the missing data analysis showed a betweengroup difference for the respondents self-reported health state at
baseline. The study also shows that health-related quality of life,
pain-related disability and sick leave after physiotherapy triage
assessment did not differ from standard practice at the long-term
follow-up. We have earlier demonstrated that physiotherapy
triage assessment of patients referred for an orthopaedic consultation in primary healthcare can reduce referrals for orthopaedic
consultation, result in more appropriate referrals, as well as shorter
waiting times (Samsson and Larsson, 2015). By analysing long term
patient-related outcomes, yet another dimension is added, and our
ndings, in combination with the existing scientic evidence,
suggest that a model using physiotherapy triage assessment of
patients referred for orthopaedic consultation could be suitable for

implementation in primary healthcare in Sweden. This model of

care is suggested to be a more efcient use of resources (Desmeules
et al., 2012) since it aims for the most appropriate healthcare provider to attend to the patient's needs; to improve access to care
with equal or better effectiveness, improve health outcomes, and
maintain or decrease costs, without compromising patient satisfaction (Kersten et al., 2007; Desmeules et al., 2012; Stanhope,
2012).
The scope of this study was to perform a long-term follow up of
patient-related variables after a physiotherapy triage assessment of
patients referred for orthopaedic consultation compared with
standard practice. Due to the rather limited intervention of this
study, we did not expect it to have any major long-term effect on
outcomes. The interest lied in establishing whether there were any
negative effects in patient-related outcome variables with this new
model. As reported previously (Samsson and Larsson, 2015), a
majority of the patients in the study were referred for different
interventions and management, which could have affected the
outcome in both groups. Moreover, since healthcare utilisation for
these patients was outside the scope of this study, it is unknown
whether the patients have sought additional orthopaedic consultation or other healthcare treatment, which could have inuenced
the results.

42

K.S. Samsson, M.E.H. Larsson / Manual Therapy 20 (2015) 38e45

Table 1
Demographic characteristics of the participants at baseline.

Age
Min
Max
Mean
SD
Sex (%)
Male
Female
Civil status (%)
Married/living
together
Single/living alone
Missing
Birth country (%)
Sweden
Other
Missing
Education (%)
Elementary School
Upper secondary
school
University
Missing
Occupation (%)
Working
Student
Other
Missing
Anatomic region (%)
Spinal
Cervical
Thoracic
Lumbar
Upper extremity
Shoulder
Arm/Wrist/Hand
Lower extremity
Hip
Knee
Leg/Ankle/Foot
Other
Duration of symptoms
Subachute
Chronic
Missing

PT triage assessment
(n 102)

Standard practice
(n 101)

18
67
51
13

21
67
53
12

45 (44)
57 (56)

45 (45)
56 (55)

82 (80)

86 (85)

16 (16)
4 (4)

14 (14)
1 (1)

93 (91)
5 (5)
4 (4)

92 (91)
8 (8)
1 (1)

11 (11)
48 (47)

23 (22)
47 (47)

39 (38)
4 (4)

30 (30)
1 (1)

71 (70)
2 (2)
25 (24)
4 (4)

70 (69)
3 (3)
27 (27)
1 (1)

7 (7)
2 (2)
18 (17)

8 (8)
0
15 (15)

13 (13)
12 (12)

14 (14)
16 (16)

12 (12)
21 (20)
15 (15)
2 (2)
(%)
18 (18)
75 (73)
9 (9)

9 (9)
30 (30)
6 (6)
3 (3)
14 (14)
70 (69)
17 (17)

Parametric variables were compared using an Independent t-test, non parametric

variables using the chi-squared test.
PT Physiotherapy, SD Standard Deviation.

The patients in the physiotherapy triage assessment group

perceived better self-rated health state (EQ VAS) at 3 months.
Although statistically signicant, the mean difference was 4.9 for
the physiotherapy group and 3.8 for the standard practice group
(Fig. 2) and might not be clinically relevant, considering previous
reported MCIC being 10.5 (Soer et al., 2012a). As previously reported, waiting times were signicantly shorter for the patients in
the physiotherapy triage assessment group, which could have
affected the outcome (Samsson and Larsson, 2015). It has previously been reported that a waiting time of less than 14 days was the
most important factor for patient outcome, regardless of intervention (Fritz et al., 2012). Although this indicates the importance
of a short waiting time, it does not explain our ndings since the
mean waiting time in our study was 19 and 28 days respectively.
Another aspect could be the outcome of management as reported
previously (Samsson and Larsson, 2015). A larger proportion of the
patients in the physiotherapy triage assessment group, 62%,
compared with 36% in the standard practice group, were referred to

Fig. 2. Long-term follow up for self-rated health state (EQ VAS) (predicted means,
standard errors); physiotherapy triage assessment (PT) versus standard practice (SP).
Analysis was made using an ANOVA. * A statistically signicant difference between
the groups [mean difference 5.7 (95% CI 11.1; 0.2); p 0.04].

physiotherapy due to a need for conservative management with

ongoing support. This could have inuenced the outcome
positively.
The ndings of no differences in self-reported health state at 6
and 12 months, as well as in health-related quality of life and painrelated disability at all time points, are consistent with previous
studies of similar models of care in various settings, such as orthopaedic outpatient departments (Daker-White et al., 1999) and
emergency rooms (Richardson et al., 2005). However, an observational study of patients with musculoskeletal disorders referred for
primary care assessment, found a signicant improvement in
health-related quality of life after 3 months, which was maintained
at 12 months (Sephton et al., 2010).
In this study, the majority of patients had persistent pain, but a
very low number of patients were on sick leave during the 12
months following the consultation. Additionally, they reported a
relatively high level of health-related quality of life and low level of
disability. These ndings differ from those previously reported;
that many patients with persistent pain experience a high level of
disability, affecting both work capacity (Gureje, Simon, 2001;
Gerdle et al., 2004; Landmark et al., 2013) and quality of life
(Kroenke et al., 2013; Landmark et al., 2013). However, Mewes et al.
(2009) report that approximately 30% of the general population
with one or more somatic complaints indicated no disability on the
PDI. Furthermore, Denison et al. (2007) presented three subgroups
of patients with musculoskeletal pain in primary care, where the
largest group reported low levels of pain and mean score on PDI, as
well as low levels of fear of movement, and a high level of selfefcacy. It is possible that these bio-psychosocial factors could
have inuenced the outcome for the participants in our study and it
could be of interest to include outcome measures for such factors in
future studies.
Participants in this study were randomised to either physiotherapy triage assessment or standard practice. In a clinical setting
there are different examples of how patients are included in this
model, such as; all patients referred for orthopaedic consultation
being re-routed to the physiotherapists for triage (Canada) (Bath
and Janzen, 2012); a physiotherapy-initiated telephone triage of
all orthopaedic waitlist patients (Australia) (Morris, et al., 2011),
physiotherapy triage of all incoming referrals (Canada, UK) (Robarts
et al., 2008; Sephton et al., 2010), or screening of incoming referrals
by an orthopaedic surgeon followed by physiotherapy triage

K.S. Samsson, M.E.H. Larsson / Manual Therapy 20 (2015) 38e45

43

Table 2
Health-related quality of life; Odds ratios (OR) and condence intervals (95% CI) for rating no problems in the physiotherapy triage assessment group.
EQ-5D

Activity
Anxiety
Mobility
Pain
Self care

Group

PT
SP
PT
SP
PT
SP
PT
SP
PT
SP

Baseline

3 months

6 months

12 months

OR

95% CI

OR

95% CI

OR

95% CI

OR

95% CI

101
99
101
97
99
96
101
97
101
96

1.4
Ref
1.4
Ref
1.2
Ref
0.7
Ref
4.1
Ref

0.8; 2.4

80
75
80
76
80
75
80
75
80
76

1.1
ref
0.9
Ref
1.8
Ref
0.8
Ref
1.5
Ref

0.3; 4.0

78
68
79
67
79
68
77
65
78
68

1.1
Ref
1.9
Ref
1.2
Ref
0.9
Ref
4.6
Ref

0.3; 3.8

82
79
83
79
83
78
84
78
83
79

1.8
Ref
1.6
Ref
1.0
Ref
0.7
Ref
1.6
Ref

0.5; 6.3

0.8; 2.4
0.7; 2.0
0.2; 3.3
0.8; 19.9

0.3; 3.1
0.5; 5.8
0.0; 21.1
0.1; 36.4

0.5; 8.1
0.3; 4.2
0.0; 2.1
0.2; 1215.3

0.5; 5.2
0.3; 3.9
0.0; 16.5
0.0; 59.0

The number of respondents in each group varied (65e101) because of incomplete data.
Analysis was made using a marginal logistic regression model.
EQ-5D EuroQol 5D, PT Physiotherapy triage assessment group, SP Standard practice group, Ref reference group for analysis.

(Ireland) (Ashmore et al., 2014). The different models vary

depending on the clinical setting.
The strengths of this study are that it is randomised, the relative
large number of participants, and the long term follow-up of
patient-related outcome. However, the study has some potential
limitations. The patients were assessed by either one physiotherapist or one orthopaedic surgeon. Preferably, the protocol should
consist of several physiotherapists and orthopaedic surgeons and
also, by both healthcare professionals assessing the same patients.
Due to the clinical reality at the present healthcare centre such a
protocol was not feasible. Also, on national level, many primary
healthcare centres are small, and very few have orthopaedic surgeons or physiotherapists with appropriate level of education. The
patients and the healthcare personnel involved in the study knew
which group they were randomised to; therefore there was a potential risk of performance bias in this study. However, all healthcare personnel involved in the study were briefed on the study
protocol and it is therefore likely that everyone performed at their
best. Furthermore, KS was the main author and the physiotherapist
performing the triage assessment, as well as responsible for the
analysis, which causes a risk of bias in the analysis and the interpretation. However, measures were taken to minimise this risk; KS
was not involved in the eligibility assessment, randomisation,
coding, or data collection and all data were coded before the
analysis. The choice of randomisation was made to enable an even
clinical distribution of patients to the respective group. It is possible
that the administrator could keep track on the respective envelopes, which causes a risk of selection bias.

The PROMs were self-administered at the patients' home to

minimise observer bias. Due to it being 3 months between follow
ups, the risk for detection bias is considered to be low. Attrition bias
may have been present since those patients in the physiotherapy
triage assessment group who responded to EQ VAS at the 3 month
follow-up had a higher self-reported health state at baseline. The
total response rates for the questionnaires for the various followups could be considered acceptable (ranging from 71% to 80%); a
response rate of 70% has been suggested as acceptable for questionnaires (Edwards et al., 1998). However, since the analysis was
made on an item-level, the level of missing data varied up to as
much as 38%, which is why we did a missing-data analysis. The
choice to treat the EQ-5D and the PDI as ordinal data and not use
the indexes could have affected the reliability and validity of the
instruments. This was decided in order to extract as much information as possible on item level from the repetitive measures.
Moreover, the dichotomisation of the response categories could
have further inuenced the results. The analysis of odds ratios
resulted in wide condence intervals, suggesting considerable
imprecision in these estimates, thus lowering the condence in our
results.
The patients in this study generally perceived high healthrelated quality of life as well as low levels of pain-related
disability at baseline and there were oor effects for a large number of patients for both EQ-5D and PDI (more than 20% of respondents reported lowest possible score) (Fitzpatrick et al., 1998).
Evidence suggests that the more specic and sensitive the outcome
measure is, the more sensitive the response becomes (Devine et al.,

Table 3
Pain-related Disability; Odds ratios (OR) and condence intervals (95% CI) for rating median or below in the physiotherapy triage assessment group.
PDI

Family/home responsibility
Recreational
Social activity
Occupation
Sexual behaviour
Self Care
Life-support activity

Group

PT
SP
PT
SP
PT
SP
PT
SP
PT
SP
PT
SP
PT
SP

Baseline

3 months

6 months

12 months

OR

95% CI

OR

95% CI

OR

95% CI

OR

95% CI

98
94
99
96
100
97
99
94
96
94
100
98
100
96

1.5
Ref
1.2
Ref
1.3
Ref
1.4
Ref
1.6
Ref
1.1
Ref
1.1
Ref

0.8; 2.6

80
75
80
74
80
75
78
75
75
73
79
74
79
75

1.0
Ref
1.4
Ref
1.4
Ref
1.8
Ref
1.4
Ref
1.4
Ref
1.0
Ref

0.3; 3.8

77
66
77
64
78
65
78
65
76
63
78
66
79
65

1.5
Ref
1.5
Ref
1.7
Ref
1.8
Ref
2.0
Ref
3.1
Ref
1.4
Ref

0.4; 5.8

82
76
82
75
82
78
83
75
80
73
81
75
81
76

1.3
Ref
1.2
Ref
1.4
Ref
1.1
Ref
1.3
Ref
1.5
Ref
1.2
Ref

0.3; 5.0

0.7; 2.1
0.7; 2.3
0.8; 2.5
0.9; 2.8
0.6; 1.9
0.6; 1.8

0.4; 5.1
0.4; 4.4
0.5; 6.5
0.5; 3.9
0.4; 4.7
0.3; 3.5

0.4; 5.3
0.5; 4.7
0.5; 5.2
0.6; 5.5
0.8; 14.6
0.4; 5.2

The number of respondents in each group varied (63e100) because of incomplete data.
Analysis was made using a marginal logistic regression model.
PDI Pain Disability Index, PT Physiotherapy triage assesment group, SP Standard practice group, Ref reference group for analysis.

0.3; 4.4
0.4; 4.6
0.3; 3.8
0.4; 4.3
0.4; 5.3
0.3; 4.0

44

K.S. Samsson, M.E.H. Larsson / Manual Therapy 20 (2015) 38e45

2011). The outcome measures were chosen due to them being

generic, short, reliable, and valid for a wide range of musculoskeletal disorders and due to the large spectrum of diagnoses in this
study, choosing more specic outcome measures was difcult.
Nevertheless, we were able to detect a statistically signicant difference between the groups for EQ VAS at 3 months following
consultation. The study was also powered sufciently for EQ VAS at
all follow-ups. Future studies might benet from adding a minimum score on PROMs as an inclusion criterion, or to choose more
specic PROMs.
4.1. Future studies
For this model of care to be implemented, nationally as well as
internationally, further studies, preferably multicenter studies,
including several physiotherapists and orthopaedic surgeons, are
needed. Also, studies investigating cost efciency of the model are
necessary.
5. Conclusion
This study shows that health-related quality of life, pain-related
disability and sick leave after physiotherapy triage assessment did
not differ from standard practice at the long-term follow-up. The
ndings indicate the possible benets of implementing this model
of care; however, more research on the model is still needed.
Acknowledgement
Ethical approval for the study was obtained from the Regional
Ethical Review Board in Gothenburg. Financial support was obstra
tained from the Department of Healthcare services, Region Va
taland. The authors gratefully acknowledge Maria Dottori for
Go
helping with the data collection, Susanne Bernhardsson for support during the writing process and prof. em. Jane Carlsson for
general support. The authors declare that they have no competing
interests.
References
Aiken AB, McColl MA. Diagnostic and treatment concordance between a physiotherapist and an orthopedic surgeon e a pilot study. J Interprof Care 2008;22:
253e61.
Aiken AB, Atkinson M, Harrison MM, Hope J. Reducing hip and knee replacement
wait times: an expanded role for physiotherapists in orthopedic surgical clinics.
Healthc Q 2007;10:88e91. 6.
Ashmore K, Smart K, O'Toole G, Doody C. Triage of knee pain by an Extended Scope
Physiotherapist (ESP) in an orthopaedic clinic: a clinical audit. Physiotherapy
Pract Res 2014;35:25e32.
Bath B, Janzen B. Patient and referring health care provider satisfaction with a
physiotherapy spinal triage assessment service. J Multidiscip Healthc 2012;5:
1e15.
Beard DJ, Knezevic K, Al-Ali S, Dawson J, Price AJ. The use of outcome measures
relating to the knee. Orthop Trauma 2010;24:309e16.
Bingefors K, Isacson D. Epidemiology, co-morbidity, and impact on health-related
quality of life of self-reported headache and musculoskeletal pain e a gender
perspective. Eur J Pain 2004;8:435e50.
rnsdo
ttir SV, Jo
nsson SH, Valdimarsdo
 ttir UA. Functional limitations and physical
Bjo
symptoms of individuals with chronic pain. Scand J Rheumatol 2013;42:59e70.
Blackburn MS, Cowan SM, Cary B, Nall C. Physiotherapy-led triage clinic for low
back pain. Aust Health Rev 2009;33:663e70.
Brooks R. EuroQol: the current state of play. Health Policy 1996;37:53e72.
Canizares M, MacKay C, Davis AM, Mahomed N, Badley EM. A population-based
study of ambulatory and surgical services provided by orthopaedic surgeons for
musculoskeletal conditions. BMC Health Serv Res 2009;9.
Daker-White G, Carr AJ, Harvey I, Woolhead G, Bannister G, Nelson I, et al.
A randomised controlled trial. Shifting boundaries of doctors and physiotherapists in orthopaedic outpatient departments. J Epidemiol Community Health
1999;53:643e50.
f P, Sandborgh M, Lindberg P. Musculoskeletal pain in primary
Denison E, senlo
health care: subgroups based on pain intensity, disability, self-efcacy, and fearavoidance variables. J Pain 2007;8:67e74.

Department of Health. The musculoskeletal services Framework e a Joint responsibility: Doing it differently; 2006. Available at, http://webarchive.
nationalarchives.gov.uk/20130107105354/http://www.dh.gov.uk/prod_
consum_dh/groups/dh_digitalassets/@dh/@en/documents/digitalasset/dh_
4138412.pdf. accessed 21.11.13.
Desmeules F, Roy JS, Macdermid JC, Champagne F, Hinse O, Woodhouse LJ.
Advanced practice physiotherapy in patients with musculoskeletal disorders: a
systematic review. BMC Musculoskelet Disord 2012;13:107.
Devine J, Norvell DC, Ecker E, Fourney DR, Vaccaro A, Wang J, et al. Evaluating the
correlation and responsiveness of patient-reported pain with function and
quality-of-life outcomes after spine surgery. Spine 2011;36:S69e74.
Edwards SJL, Lilford RJ, Braunholtz DA, Jackson JC, Hewison J, Thornton J. Ethical
issues in the design and conduct of randomised controlled trials. Health
Technol Assess 1998;2. v-76.
EuroQol. Available at http://www.euroqol.org/home.html. accessed 18.11.13.
EuroQol Group. EuroQolea new facility for the measurement of health-related
quality of life. Health Policy 1990;16:199e208.
Fitzpatrick R, Davey C, Buxton MJ, Jones DR. Evaluating patient-based outcome
measures for use in clinical trials. Health Technol Assess 1998;2. i-74.
Fritz JM, Childs JD, Wainner RS, Flynn TW. Primary care referral of patients with low
back pain to physical therapy: Impact on future health care utilization and
costs. Spine 2012;37:2114e21.
Gerdle B, Bjork J, Henriksson C, Bengtsson A. Prevalence of current and chronic pain
and their inuences upon work and healthcare-seeking: population study.
J Rheumatol 2004;31:1399e406.
Grimby-Ekman A, Andersson EM, Hagberg M. Analyzing musculoskeletal neck pain,
measured as present pain and periods of pain, with three different regression
models: a cohort study. BMC Musculoskelet Disord 2009;10.
Gureje O, Simon G, Von Korff M. A cross-national study of the course of persistent
pain in primary care. Pain 2001;92:195e200.
Hosmer DW, Lemeshow S. Applied logistic regression. 2nd ed. Hoboken, NJ, USA:
John Wiley & Sons, Inc; 2005.
Jordan KP, Kadam UT, Hayward R, Porcheret M, Young C, Croft P. Annual consultation prevalence of regional musculoskeletal problems in primary care: an
observational study. BMC Musculoskelet Disord 2010;11.
Kersten P, McPherson K, Lattimer V, George S, Breton A, Ellis B. Physiotherapy
extended scope of practice e who is doing what and why? Physiotherapy
2007;93:235e42.
Kroenke K, Outcalt S, Krebs E, Bair MJ, Wu J, Chumbler N, et al. Association between
anxiety, health-related quality of life and functional impairment in primary care
patients with chronic pain. Gen Hosp Psychiatry 2013;35:359e65.
Landmark T, Romundstad P, Dale O, Borchgrevink PC, Vatten L, Kaasa S. Chronic
pain: one year prevalence and associated characteristics (the HUNT pain study).
Scand J Pain 2013;4:182e7.
MacKay C, Canizares M, Davis AM, Badley EM. Health care utilization for musculoskeletal disorders. Arthritis Care Res 2010;62:161e9.
Maddison P, Jones J, Breslin A, Barton C, Fleur J, Lewis R, et al. Improved access and
targeting of musculoskeletal services in northwest Wales: targeted early access to musculoskeletal services (TEAMS) programme. Br Med J 2004;329:
1325e7.
rkelund C. Reasons for encounters, inMnsson J, Nilsson G, Strender LE, Bjo
vestigations, referrals, diagnoses and treatments in general practice in Swedena multicentre pilot study using electronic patient records. Eur J Gen Pract
2011;17:87e94.
McHugh GA, Campbell M, Luker KA. GP referral of patients with osteoarthritis for
consideration of total joint replacement: a longitudinal study. Br J Gen Pract
2011;61:459e68.
Menzies RD, Young RA. Referrals from a primary care-based sports medicine
department to an orthopaedic department: a retrospective cohort study. Br J
Sports Med 2012;45:1064e7.
hler E. What is normal
Mewes R, Rief W, Stenzel N, Glaesmer H, Martin A, Bra
disability? An investigation of disability in the general population. Pain
2009;142:36e41.
Morris J, Grimmer-Somers K, Kumar S, Murphy K, Gilmore L, Ashman B, et al.
Effectiveness of a physiotherapy-initiated telephone triage of orthopedic
waitlist patients. Patient Relat Outcome Meas 2011;2:151e9.
Picavet HSJ, Schouten JSAG. Musculoskeletal pain in the Netherlands: prevalences,
consequences and risk groups, the DMC(3)-study. Pain 2003;102:167e78.
Rabey M, Morgans S, Barrett C. Orthopaedic physiotherapy practitioners: surgical
and radiological referral rates. Clin Gov Int J 2009;14:15e9.
Richardson B, Shepstone L, Poland F, Mugford M, Finlayson B, Clemence N. Randomised controlled trial and cost consequences study comparing initial physiotherapy assessment and management with routine practice for selected
patients in an accident and emergency department of an acute hospital. Emerg
Med J 2005;22:87e92.
Robarts S, Kennedy D, MacLeod AM, Findlay H, Gollish J. A framework for the
development and implementation of an advanced practice role for physiotherapists that improves access and quality of care for patients. Healthc Q
2008;11:67e75.
Samsson K, Larsson MEH. Physiotherapy screening of patients referred for orthopaedic consultation in primary healthcare e a randomised controlled trial. Man
Ther 2015;19(5):386e91.
Sephton R, Hough E, Roberts SA, Oldham J. Evaluation of a primary care musculoskeletal clinical assessment service: a preliminary study. Physiotherapy
2010;96:296e302.

K.S. Samsson, M.E.H. Larsson / Manual Therapy 20 (2015) 38e45

Soer R, Reneman MF, Speijer BLGN, Coppes MH, Vroomen PCAJ. Clinimetric properties of the EuroQol-5D in patients with chronic low back pain. Spine J
2012a;12:1035e9.
Soer R, Reneman MF, Vroomen PC, Stegeman P, Coppes MH. Responsiveness and
minimal clinically important change of the pain disability index in patients with
chronic back pain. Spine 2012b;37:711e5.
Soer R, Koke AJA, Vroomen P, Stegeman P, Smeets R, Coppes MH, et al. Extensive
validation of the pain disability index in three groups of patients with
musculoskeletal pain. Spine 2013;38:562e8.

45

Stanhope J, Grimmer-Somers K, Milanese S, Kumar S, Morris J. Extended scope

physiotherapy roles for orthopedic outpatients: an update systematic review of
the literature. J Multidiscip Healthc 2012;5:37e45.
Tait R, Pollard C, Margolis R, Duckro P, Krause S. The pain disability index: psychometric and validity data. Arch Phys Med Rehabil 1987;68:438e41.
Tait R, Chibnall J, Krause S. The pain disability Index: psychometric properties. Pain
1990;40:171e82.