Professional Documents
Culture Documents
DESCRIPTIVE REPORT
ABSTRACT
The purpose of present study was to investigate important measurement properties of the Body Awareness
Rating Questionnaire (BARQ) subscales: Function; Mood; Feelings; and Awareness. Samples of 50 patients
with prolonged musculoskeletal pain and psychosomatic disorders, and 50 gender and age-matched healthy per-
sons participated in the study. Relative (ICC 2,1) and absolute reliability (Sw) were calculated. Construct validity
was examined by testing hypothesis using Pearson (r) or Spearman rank (rs) correlation. Discriminate ability was
examined using a receiver operating characteristic (ROC) curve, the area under the curve (aROC) being the
measure of discriminative validity. All patients reported more or less improvement after treatment, using the
Patient Global Impression of Change (PGIC) as an external indicator of important change. Responsiveness to
For personal use only.
important change was therefore examined by one-way repeated measures analysis of variance (ANOVA), relat-
ing change scores of BARQ subscales to the PGIC categories. Relative and absolute reliability values were within
recommended limits for all four subscales. Construct and discriminate validity was indicated for Function,
Feelings and Awareness, but not for Mood. Responsiveness was indicated for the subscales Function and
Awareness, but not for Feelings and Mood. Further research is needed to complement the subscales of
BARQ, found with appropriate measurement properties.
515
516 Tove et al.
defined, but Mølstad et al. (1989) points to the impor- different aspects of the phenomenon of body aware-
tance of contact with one’s own body, that a positive ness (Dragesund, Ljunggren, Kvåle, and Strand,
feeling for your own body includes both accepting and 2010) and should thus be examined separately in
being confidential with its reactions. They emphasize further clinometric analysis. The purpose of the
that the patients’ ability to sense body movements and present study was to examine test-retest reliability,
their emotional aspects, reflect how they experience construct and discriminative validity and responsive-
and relate to their own body. Working on body aware- ness to important change of the BARQ subscales.
ness is considered essential in musculoskeletal and psy- Internal consistency was also re-examined in the new
chosomatic conditions in order to relieve pain and sample of patients.
tension and to improve respiration and bodily function
in general (Dragesund and Råheim, 2008; Mølstad
et al, 1989; Thornquist and Bunkan, 1991; Øien, MATERIALS AND METHODS
Physiother Theory Pract Downloaded from informahealthcare.com by University of Connecticut on 10/28/14
(Dragesund, Ljunggren, Kvåle, and Strand, 2010). musculoskeletal pain (ICPC-2-2004, L-codes) and/
Initially, a pool of 66 test items was collected. Then or psychosomatic disorders (ICD-10-2005, codes:
300 patients scored each of the items on a Likert F45.0-F45.4); a condition that had lasted for more
scale. The underlying factor structure and internal than 3 months; and not earlier being treated with
consistency was examined, excluding items that did NPMP. Exclusion criterion was severe psychiatric dis-
not contribute to increased internal consistency. orders. In Norway, physicians and manual therapists
Four dimensions (factors) were identified and diagnose and refer patients with the above conditions
the six items with highest loadings within each sub- to NPMP. Fifty patients who were asked to participate
dimension were retained, resulting in a 24-item in the present study were recruited consecutively from
questionnaire, called the Body Awareness Rating the waiting- lists of six physiotherapists during a period
Questionnaire (BARQ). of 15 months (October 2008 to December 2009). The
When developing a new assessment tool, evidence physiotherapists had been specialists in NPMP for
must be provided that it has satisfactory measurement more than 6 years, and all worked in primary health
properties. Terwee et al. (2007) has presented quality care.
criteria for health measurement questionnaires pro- To evaluate the BARQ’s discriminative validity, a
posed by the Scientific Advisory Committee of the convenience sample of 50 persons who considered
Medical Outcomes Trust (2002), and these criteria themselves as healthy were asked to participate. They
were taken into consideration in the present study. were recruited among part and full time workers in
The content validity and internal consistency of six different private companies during the same
BARQ have recently been demonstrated (Dragesund, period of time as the patients (October 2008 to
Ljunggren, Kvåle, and Strand, 2010). It is realized December 2009). They should not have been on sick
that each measurement property may not be equally leave due to musculoskeletal pain or psychosomatic
important. The intended use of the questionnaire disorders during the preceding year and not be familiar
decides which measurement properties that should with BARQ prior to the study. To match the patient
be examined. The four subscales of BARQ were group, the healthy persons were stratified according
aimed to be both discriminative and evaluative to age groups (20-35, >35-50, >50) and gender.
(i.e. being able to discriminate between healthy indi- Both samples were given written and oral infor-
viduals and patients with long-lasting musculoskeletal mation about the study, and the participants signed a
pain and/or psychosomatic disorders, and be able to consent form. The study was approved by the Regional
capture change in body awareness). The four Committee for Medical Research Ethics in Western
subscales of BARQ were shown to reflect somewhat Norway and the Norwegian Data Inspectorate.
scored on a seven-point Likert scale. Subscale scores (Bagby, Parker, and Taylor, 1994; Loas et al, 1996).
range from 6 to 42, and high scores reflect better A higher score indicates alexithymia. The Norwegian
body awareness (Dragesund, Ljunggren, Kvåle, and version has demonstrated good psychometric proper-
Strand, 2010). The BARQ is presented with subscales ties (Graeme, Taylor, Bagby, and Parker, 2002;
and single items in the Appendix. Taylor, Bagby, and Parker, 2003).
To examine construct validity, two other question-
naires were filled in by the patients. These question- Patient Global Impression of Change (PGIC)
naires were the Short Form-36 Health Survey PGIC is a 7-point ordinal scale, used as an external
(SF-36) and the Toronto Alexithymia Scale (TAS- criterion to measure clinical important change after
20). The SF-36 is designed to measure perceived treatment (Farrar et al, 2001). Scores <4 denote
health related quality of life, and body awareness as improvement (1 = very much improved, 2 = much
captured by BARQ was hypothesized to be related to improved, 3 = slightly improved), and scores >4 a
For personal use only.
this construct. However, the constructs assessed by worsening (5= minimal worse, 6= much worse, 7 =
the different subscales were expected to be more or very much worse). Studies indicate that the PGIC is
less related, as reflected in the purpose statement. valid as a measure of important change in patients
Since patients with long-lasting musculoskeletal pain with low back pain (Grotle, Brox, and Vollestad,
and/or psychosomatic disorders often have problems 2004; Ostelo, de Vet, Knol, and van den Brandt,
describing their feelings and bodily function 2004). In the present study the questionnaire was
(Braatøy, 1947), the TAS-20 was also chosen for con- used to report change in body awareness.
struct validation.
After treatment the patients filled in the Patients
Global Impression of Change as an external anchor Treatment
of important change.
Norwegian Psychomotor Physiotherapy (NPMP) is a
body-oriented physiotherapeutic approach, taking
Short Form-36 Health Survey (SF-36)
into consideration the association between body and
SF-36 is a 36-item self-report questionnaire used to mind. The whole body is examined and treated as a
measure health-related quality of life (Ware, 2000). functional unit. Respiration, muscle tension and
The questionnaire includes eight subscales: 1) Phys- emotions are considered to be intertwined, and the
ical Functioning (PF); 2) Role function - Physical patients’ respiration pattern is particularly a focus of
aspects (RP),;3) Bodily Pain (BP); 4) General attention in examination as well as in treatment. The
Health (GH); 5) Vitality (V); 6) Social Functioning aim of NPMP is to address dysfunction by facilitating
(SF); 7) Role function - Emotional aspects (RE); change through massage and movements, closely
and 8) Mental Health (MH). The score for each of linked to respiration. Movements include grounding,
the eight subscales ranges from 0 to 100. A higher balancing, stretching and relaxation. Verbal reflection
score indicates better health (Ware, 2000). SF-36 on embodied experiences is part of the therapy. Body
has been used in several studies on musculoskeletal awareness is considered essential to improve function,
disorders and has been found to be reliable, valid and is stimulated by encouraging the patient to sense
and responsive to change (Kvien, Kaasa, and Smed- bodily reactions during treatment (Thornquist and
stad, 1998; Schlenk et al, 1998; Tuttleman et al, Bunkan, 1991; Øien, Råheim, Iversen, and Steihaug,
1997). Good psychometric properties have also been 2009; Øvreberg and Andersen, 1986). Each treatment
demonstrated in the Norwegian version (Loge, session lasts one hour and usually the patients receive
Kaasa, Hjermstad, and Kvien, 1998). treatment once a week or once every second week.
with less improvement of body awareness reported Function; 3.2 for Mood; 2.7 for Feelings; and 3.1
on PGIC. This trend was examined by a one-way for Awareness. The percentage of Sw related to the
repeated measures analysis of variance (ANOVA). total score ranged between 7.5% and 8.9%
(Table 2). The smallest detectable change (SDC)
TABLE 1 Background characteristics of the participants values are also presented in Table 2. A tendency
(n = 100) towards higher scores from test to retest was seen,
and when drafting the Bland-Altman plot (Figure 1),
Patients Healthy this drift must be taken into consideration to arrive
Variables N = 50 N = 50 at the 95% limits of agreement, which indicate the
limit an individual has to exceed to demonstrate a
Age, years: mean, (SD) 42.2 (13.2) 40.3 (12.8)
change above measurement error.
17–35 years, men, women 2/15 2/15
36–50 years, men, women 7/13 7/13
>51 years men, women 4/10 4/10 Construct Validity
Gender: men, women, % 12/38 12/38
The Kolmogorov-Smirnov test for normality of
(23.5/75.5 %) (23.5/75.5 %)
Duration of pain problems, 6.6 (6.6) 0
subscale data showed p-values ranging between
years: mean (SD) p < 0.001 and 0.200, thus deciding the use of
Work: blue collar, white 10/35/5 5/37/8 Pearson or Spearman rank correlation. All 11 hypoth-
collar, students eses regarding a correlation between the Function
Sick listed: yes, no, % 22/17 0/50 subscale and the subscales of the other measures
(43.1/33.3 %) (0/100%) (Table 3) were confirmed, with medium correlations
TABLE 2 Cronbach’s α values and reliability of the BARQ subscales by ICC(2,1), Sw, also in per cent of the maximum score and the
smallest detectable change (SDC).
Subscales (6-42) Cronbach’s α values ICC (95% CI) Sw Sw in per cent (%) of max. scale score SDC
FIGURE 1 Intra-individual differences between test and retest plotted against the mean of the two tests for the four subscales, Func-
tion, Mood, Feelings and Awareness. The central horizontal line represents the mean of the intra-individual differences, and the flank-
ing lines the mean difference + /- 1.96 × SDdifference, which are the limits of agreement (see inserted values).
TABLE 3 Correlation between the subscale scores of BARQ and the subscale scores of SF-36 and of TAS-20.
BARQ subscales
SF-36
Physical Functioning (PF) 0.468∗ 0.285∗ 0.029 −0.068
Role function-Physical aspects (RP) 0.538∗ 0.268 0.101 0.154
Bodily Pain (BP) 0.437∗ 0.085 −0.026 −0.089
General Health (GH) 0.119 0.277 0.042 −0.130
Vitality 0.279 −0.074 0.253 −0.079
Social Functioning (SF) 0.205 −0.261 0.369∗ 0.154
Role function-Emotional aspects (RE) 0.089 0.422∗ 0.222 −0.105
Physiother Theory Pract Downloaded from informahealthcare.com by University of Connecticut on 10/28/14
p ≤ 0.05∗ . Predefined hypotheses of at least moderate correlations (r > 0.30) should be seen in relation to the bold face values, most
confirming the hypothesis, others rejecting the hypothesis.
For personal use only.
FIGURE 2 Area under the Receiver Operating Characteristic (ROC) curves demonstrating discriminate ability of the BARQ sub-
scales Function, Mood, Feelings and Awareness, contrasting test scores of patients with prolonged musculoskeletal pain and psycho-
somatic disorders and gender and age-matched healthy participants.
TABLE 4 Discriminate ability of the four subscales of BARQ, contrasting test scores of healthy persons and patients with prolonged
musculoskeletal pain and psychosomatic disorders using ROC curve analysis, reporting the area under the curve (aROC) with a 95%
confidence interval (CI).
Function 36.74 (4.71) 17.76 (7.13) 0.97 0.95,1.00 21.5 1.00 0.76
Mood 30.22 (8.14) 31.40 (7.58) 0.46 0.35,0.58 30.5 0.50 0.42
Feelings 37.04 (5.48) 31.82 (7.60) 0.73 0.63,0.83 33.5 0.84 0.50
Awareness 30.06 (6.32) 21.68 (7.19) 0.80 0.71,0.88 21.5 0.94 0.52
Physiother Theory Pract Downloaded from informahealthcare.com by University of Connecticut on 10/28/14
TABLE 5 Scores on the four subscales of BARQ at baseline and after 6 months NPMP treatment and change in scores, with p-values,
after 6 months treatment.
Items Baseline Scores Mean (SD) After Treatment Mean (SD) Change Mean (SD), p-values
TABLE 6 Change in scores on the four subscales of BARQ according to categories of self-reported change on the Patient Global
Impression of Change (PGIC) after rehabilitation, examined for linear trend by One way ANOVA. N = 50.
For personal use only.
BARQ Very much Improved Mean Much Improved Mean Slightly Improved Mean Linear trend, Contrast and
subscales (SD) n = 11 (SD) n= 33 (SD) n= 6 p-values
improvement on the PGIC (Table 6), therefore ROC not for Mood. Positive improvement after therapy
curve analysis could not be performed discriminating was indicated by all the four subscales, but responsive-
improved from not improved. There was a significant ness to important change was only demonstrated by
linear trend of decreasing change scores according to Function and Awareness.
decreasing improvement on the PGIC, for the sub- The internal consistency of all BARQ subscales was
scales Function and Awareness, but not for Mood. highly satisfactory, and even better than found in a pre-
vious study (Dragesund, Ljunggren, Kvåle, and
Strand, 2010) of a similar patient group (n = 300).
DISCUSSION Relative reliability was satisfactory for each subscale,
with ICC (2,1) values between 0.79 (Awareness) and
The aim of this study was to examine important 0.87 (Feelings). One prerequisite when examining the
measurement properties of BARQ which includes test-retest reliability of a measure is that the patients’
four subscales intended to assess different aspects of condition does not change between tests. Because
body awareness. Test-retest reliability was found to phenomena and conditions tend to change over time,
be satisfactory for all four subscales, since both relative the time period between test and retest should be
and absolute reliability were within recommended short enough to ensure that a clinical change has not oc-
limits. Construct validity was acceptable for Function, curred, but long enough to prevent recall of the pre-
Feelings and Awareness, with moderate associations vious scoring (Terwee et al, 2007). In the present
with other related scales, while Mood demonstrated study, three to six days elapsed between test and
inconsistent results. The Function subscale demon- retest. The first test was performed in the waiting
strated very high discriminate ability, and this property room shortly after examination by an NPMP specialist.
was also satisfactory for Feelings and Awareness, but Since the patients filled in a form with demographic
data and two other questionnaires on the same be at least moderately associated with the three SF-
occasion, they probably did not recall their previous 36 subscales: 1) Social Functioning; 2) Role func-
scoring on the retest, which was performed at home. tion-Emotional aspects; and 3) Mental Health and
There was a tendency, however, of drift in data, with the two TAS-20 items: 1) Difficulty identifying feel-
patients scoring somewhat better the second time. ings; and 2) Difficulty describing feelings, but not
This drift could have been caused by the different test with the other (six) subscales. All the hypotheses
situations and/or by a possible treatment effect. Some- were confirmed except the hypothesis of a moderate
what better scores on the second test may reflect a more association between Feelings and Role function-
relaxed state of mind and less dysfunction due to the Emotional aspects. There is, accordingly, evidence
initiation of NPMP after several months on a waiting for construct validity of Feelings.
list. Another explanation may be familiarity with the The Awareness subscale reflects awareness of body
questionnaire from the first test. movements and actions in items such as: ‘I am not
Physiother Theory Pract Downloaded from informahealthcare.com by University of Connecticut on 10/28/14
Absolute reliability indicates how much error we can aware of how I breathe’, and ‘I always push myself to
expect (Domholdt, 2005). Absolute agreement for the my limits’. It appears to assess a different construct
subscales was satisfactory, since the Sw values related than the SF-36 and TAS-20 subscales. We therefore
to the total score were less than 10% for all subscales did not expect a correlation with these subscales, as
(Ostelo, de Vet, Knol, and van den Brandt, 2004). was in fact demonstrated. As we lack a validated
The Bland-Altman plots show the limits of agreement, measure of body awareness, it is difficult to find
including the systematic drift that should be taken evidence for construct validity by showing moderate
into consideration when judging changed scores for or high associations with other measures.
individual patients. To be 95% confident that the Regarding discriminative validity, the ability of the
change is not simply a measurement error, a change BARQ subscales to discriminate between patients
should be above the limits of agreement. with prolonged musculoskeletal pain and/or psychoso-
The construct validity of a measure should be matic disorders and healthy participants was satisfac-
For personal use only.
examined if the construct in question cannot be tory (aROC > 0.70) in all subscales except Mood.
observed objectively or directly and no gold standard The Function subscale demonstrated an almost
is available (Streiner and Norman, 2008). perfect ability to discriminate between patients and
The Function subscale mainly captures awareness healthy participants (aROC = 0.97). The statements
of bodily pain and tension during daily functioning. included in this scale seem to characterise very well
It seems to assess a somewhat similar phenomenon functional problems of patients referred to NPMP,
as captured by the SF-36 subscales Physical Function- for instance in the items ‘My muscles are often
ing, Role function-Physical aspects and Bodily pain. tense’ and ‘I never sit comfortably’. Work on body
We hypothesised at least a medium correlation with awareness in NPMP aims to harmonise muscle
these subscales, but a low or no correlation with the tension, breathing, movement and daily functioning.
other subscales of SF-36 and TAS-20, which appar- These aspects are addressed in treatment, for instance
ently assess different phenomena. All these hypotheses by repeatedly asking the patient what he/she senses
were confirmed. and/or feels in different positions and movements
The Mood subscale measures awareness of the (Dragesund and Råheim, 2008; Øien, Råheim,
mood’s influence on the body in items such as: ‘My Iversen, and Steihaug, 2009).
body feels different when I am happy and when I am The discriminative ability of the Awareness subscale
sad’, and ‘I breathe more easily when I am in a good was very satisfactory (aROC of 0.80). A lack of aware-
mood’. We expected at least a moderate correlation ness of how they move, breathe and expend effort on
between this scale and the three SF-36 subscales: 1) daily activities seems common in patients referred to
Social Functioning; 2) Role function-Emotional NPMP, and the subscale appears to capture a differ-
aspects; and 3) Mental Health, and the two TAS-2- ence in this phenomenon between patients and
subscales: 1) Difficulty identifying feelings; and 2) healthy participants. The Feelings subscale had also
Difficulty describing feelings. However, Mood corre- sufficient discriminative ability (aROC = 0.73), but
lated only moderately with Role function-Emotional negative perceptions of the body (dislike and shame)
aspects and negatively and moderately with Mental as captured by this scale may not just characterise
Health of SF-36. There is, accordingly, no evidence NPMP patients, but also healthy persons, as suggested
for construct validity of this scale. when comparing scores on this scale (Table 4). The
The Feelings subscale largely consists of items Mood subscale lacks evidence for construct validity.
relating to respondents’ feelings about their own The scale seems to reflect a general awareness of how
bodies, such as: ‘I am ashamed of my body’ and ‘I being happy, sad etc. influences the body, and may
dislike my body’. We therefore expected this scale to explain why it does not discriminate between healthy
persons and patients with prolonged musculoskeletal Our results are based on a rather small sample, and
pain and/or psychosomatic disorders. we cannot be sure that they apply to the whole popu-
A recommended method of examining responsive- lation of patients with long-lasting musculoskeletal
ness is the ROC curve analysis, distinguishing patients pain and/or psychosomatic disorders although demo-
who have and have not improved according to an graphic characteristics were similar to a broader
external criterion. In the present study 44 patients sample including 300 patients in a previous study
reported much or very much improvement of body (Dragesund, Ljunggren, Kvåle, and Strand, 2010).
awareness on the PGIC after treatment and six The sampling procedure used in the study secured a
patients reported to be slightly improved making it high compliance, preventing missing data, but the in-
impossible to contrast patients who had improved volvement of the therapists might have influenced the
versus those that had not improved. The subscales scoring of the questionnaires.
Function and Awareness demonstrated, however, a
Physiother Theory Pract Downloaded from informahealthcare.com by University of Connecticut on 10/28/14
Braatøy T 1947 Nervous Minds [De nervøse sinn][in Norwegian] Ostelo RWJG, de Vet HCW, Knol DL, van den Brandt PA 2004 24-
Oslo, Cappelen item Roland-Morris Disability Questionnaire was preferred out
Cohen JW 1998 Statistical Power Analysis for Behavioral Sciences of six functional status questionnaires for post-lumbar disc
(2nd edn), pp 79–81. Hillsdale, NJ, Lawrence Erlbaum surgery. Journal of Clinical Epidemiology 57: 268–276
Associates Rothschild B 2000 The Body Remembers: The Psychophysiology
Domholdt E 2005 Rehabilitation Research. Principles and Appli- of Trauma and Trauma Treatment, p 101. New York,
cations. St. Louis, MO, Elsevier Saunders W. W. Norton & Company
Dragesund T, Ljunggren AE, Kvåle A, Strand LI 2010 Body Aware- Roxendal G 1985 Body Awareness Therapy and the Body Aware-
ness Rating Questionnaire. Development of a self-administered ness Scale, Treatment and Evaluation in Psychiatric Physiother-
questionnaire for patients with long-lasting musculoskeletal and apy. Thesis, Gothenburg: University of Gothenburg
psychosomatic disorders. Advances in Physiotherapy 12: 87–94 Schlenk EA, Erlen JA, Dunbar-Jacob J, McDowell J, Engberg S,
Dragesund T, Råheim M 2008 Norwegian Psychomotor Phy- Sereika SM, Rohay Jm, Bernier MJ 1998 Health-related
siotherapy and patients with chronic pain. Patients’ perspective quality of life in chronic disorders: A comparison across
on body awareness. Physiotherapy Theory and Practice 24: studies using the MOS SF-36. Quality of Life Research 7: 57–65
Physiother Theory Pract Downloaded from informahealthcare.com by University of Connecticut on 10/28/14
Guyatt GH, Deyo RA, Charlson M, Levine MN, Mitchell A 1989 Intelligence, pp 40–67. Jossey-Bass, San Francisco, CA
Responsiveness and validity in health status measurements: A Taylor GJ, Bagby RM, Parker JDA 2003 The 20-item Toronto
clarification. Journal of Clinical Epidemiology 42: 403–408 Alexithymia Scale - IV: Reliability and factorial validity in differ-
Kvåle A, Ljunggren AE 2007 Body awareness therapies. In: Schmidt ent languages and cultures. Journal of Psychosomatic Research
RF, Willis WD (eds) Encyclopedia of Pain, pp 167–169. Berlin, 55(3): 277–283
Springer Verlag Terwee CB, Bot SD, Boer MR, van Der Windt DAWM, Knol DL,
Kvien TK, Kaasa S, Smedstad LM 1998 Performance of the Nor- Dekker J, Bouter LM, de Vet HCW 2007 Quality criteria were
wegian SF-36 Health Survey in patients with rheumatoid arthri- proposed for measurement properties of health status question-
tis. II. A comparison of the SF-36 with disease specific measures. naires. Journal of Clinical Epidemiology 60: 34–42
Journal of Clinical Epidemiology 51: 1077–1086 Terwee CB, Dekker FW, Wiersinga WM, Prummel MF, Bossuyt
Loas G, Otmani O, Verrier A, Fermaux D, Marchand MP 1996 PM 2003 On assessing responsiveness of health-related quality
Factor analysis of the French version of the 20-item Toronto of life instruments: Guidelines for instrument evaluation.
Alexithymia Scale (TAS-20). Psychopathology 29: 139–144 Quality of Life Research 12: 349–362
Loge JH, Kaasa S, Hjermstad MJ, Kvien TK 1998 Translation and Thornquist E, Bunkan BH 1991 What is Psychomotor Therapy?,
performance of the Norwegian SF-36 Health Survey in patients p 10–150. Oslo, Norwegian University Press
with rheumatoid arthritis. I: Data quality, scaling assumptions, Tuttleman M, Pillemer SR, Tilley BC, Fowler Se, Buckley LM,
reliability and construct validity. Journal of Clinical Epidemiol- Alarcon GS, Trentham De, Neuner R, Clegg Do, Leisen JC,
ogy 51: 1069–1076 Heyse SP 1997 A cross sectional assessment of health status in-
McGraw KO, Wong SP 1996 Performing inferences about some in- struments in patients with rheumatoid arthritis participating in a
terclass correlation coefficients. Psychological Methods 1: 30–46 clinical trial. Minocycline in Rheumatoid Arthritis Trial Group.
Mehling WE, Daubenmier GJ, Price CJ, Hecht FM, Stewart A 2009 Journal of Rheumatology 24: 1910–1915
Body awareness: Construct and self-report measures. Plos One 4 Ware JE 2000 SF-36 health survey update. Spine 25: 3130–3139
(5): e5614 Øien AM, Råheim M, Iversen S, Steihaug S 2009 Self perception as
Mølstad E, Nielsen G, Barth K, Haver B, Havik OE, Rogge H, embodied knowledge- changing processes for patients with
Skåtun M 1989 Body relationship in patients in psychotherapy. chronic pain. Advances in Physiotherapy 11: 121–129
Contribution from physiotherapists in the Bergen project to Øvreberg G, Andersen T 1986 Aadel Bülow-Hansen’s physiother-
short-term dynamic psychotherapy [Kroppsforhold hos pasien- apy. A method for change of tense muscles and inhibited respir-
ter i psykoterapi. Fysioterapeutens bidrag til Bergens-prosjektet ation [Aadel Bülow-Hansen’s fysioterapi. En metode til
om korttids dynamisk psykoterapi] [in Norwegian] Tidsskrift omstilling av anspent muskulatur og hemmet respirasjon][in
for Norsk Psykologforening 26: 73–79 Norwegian]. Tromsø, Tromsprodukt A/S
526
Tove et al.
528
Tove et al.