Assessing the measurement properties of two commonly used measures of therapeutic relationship in physiotherapy.
Abstract: This study aimed to: (1) explore client and physiotherapist perspectives of the acceptability and usability of the Working Alliance Inventory-Short Form and Helping Alliance Questionnaire-Version Two; (2) assess the test-retest reliability of these measures; and (3) explore their convergent and face validity. Twenty-two physiotherapist/client dyads from three practices completed both measures twice, three to seven days apart; and answered a series of semi-structured questions regarding their acceptability and face validity. The measures were found to be generally acceptable and usable, and had good convergent validity (Spearman's rank correlation coefficient of 0.72 (p<0.001) and 0.87 (p<0.001) for client and therapist data respectively). While most items were reported to be relevant, some were considered irrelevant or inappropriate, and several were thought to be missing; indicating the measures may lack some conceptual clarity regarding the therapeutic relationship in a physiotherapy context. Both measures had reasonable testretest reliability with kappa values for individual items ranging from 0.26 to 0.73 indicating fair to substantial correlation between repeated measures. Further work is needed to develop a more conceptually sound measure of therapeutic relationship specific to physiotherapy. However, in the absence of such a measure both the measures tested here displayed adequate measurement properties in this population.

Besley J, Kayes NM, McPherson KM. Assessing the measurement properties of two commonly used measures of therapeutic relationship in physiotherapy New Zealand Journal of Physiotherapy 39(2) 75-80.

Key words: Working Alliance Inventory, Helping Alliance Questionnaire (Version 2), Therapeutic relationship, Physiotherapy, Physical Therapy
Authors: Besley, Jessica
Kayes, Nicola M.
McPherson, Kathryn M.
Pub Date: 07/01/2011
Publication: Name: New Zealand Journal of Physiotherapy Publisher: New Zealand Society of Physiotherapists Audience: Academic Format: Magazine/Journal Subject: Health Copyright: COPYRIGHT 2011 New Zealand Society of Physiotherapists ISSN: 0303-7193
Issue: Date: July, 2011 Source Volume: 39 Source Issue: 2
Accession Number: 288537989
Full Text: INTRODUCTION

In many areas of health care, especially psychotherapy, the relationship between therapist and client has been well established as an important determinant for successful treatment outcomes (Leach 2005, Roberts and Bucksey 2007, Schonberger et al 2006). In the literature, a range of terms are used interchangeably to refer to this concept including 'therapeutic relationship', 'therapeutic alliance' and 'working alliance.' However, although the therapeutic relationship is considered to be important, this concept has been underexplored in the field of physiotherapy to date (Bellner 1999, Leach 2005, Potter et al 2003a, Potter et al 2003b, Roberts and Bucksey 2007, Williams and Harrison 1999).

Two commonly used and widely validated measures of the working alliance are the Working Alliance Inventory (WAI) and the Helping Alliance Questionnaire Version Two (HAQ-II) (Cecero et al 2001, De Weert-Van Oene et al 1999, Fenton et al 2001, Horvath and Greenberg 1989, Kermarrec et al 2006, Le Bloc'h et al 2006, Luborsky et al 1996). While these measures were originally intended to be applicable to all kinds of health care professional relationships, their use has been largely limited to psychotherapy (Guedeney et al 2005, Hatcher and Gillaspy 2006). These measures are however, being increasingly adopted, and a recent review reported the WAI to be the most common measure used to evaluate the therapeutic relationship in physical rehabilitation (Hall et al 2010). Despite this, the measurement properties of the WAI and the HAQ-II have not been investigated in the physiotherapy context.

Furthermore, a recent literature review we carried out has critically explored the conceptual basis of the WAI and the HAQII by contrasting their item content against existing knowledge regarding the therapeutic relationship in physiotherapy and identified some clear limitations to both measures (Besley et al 2011). In the absence of other more conceptually sound measures, and in view of their expanding use, exploration of the psychometric properties of the WAI and the HAQ-II in physiotherapy seems warranted.

The aims of this study were to: (1) explore client and physiotherapist perspectives of the acceptability and usability of the Working Alliance Inventory Short Form (WAI) and Helping Alliance Questionnaire Version Two (HAQ-II); (2) assess the test-retest reliability of the two measures; and (3) explore their convergent and face validity.

METHODS

Participants

Client participants. People accessing physiotherapy were invited to participate if they: (1) were over 18 years of age; (2) had attended at least four physiotherapy sessions with the same therapist for the current problem for which they were seeking treatment; (3) were able and willing to complete the questionnaire without having to refer to the physiotherapist for assistance (though a researcher was available for assistance); and (4) were having their next physiotherapy session within 3-7 days of when they first completed the questionnaires. People were excluded if they were unable to give informed consent, were unable to understand the questions sufficiently to complete the questionnaire, or if they could not make themselves available to participate in the study.

Physiotherapist participants. Physiotherapists were invited to participate if they: (1) were working for one of three local service providers; (2) were currently working with a client who had consented to participate; and (3) were able and willing to complete the study questionnaires at the same time as their client.

The aim was to recruit 30 participant pairs (physiotherapist and client) which would provide 95% power at p<0.05 to detect a correlation of 0.6 between the WAI and the HAQ.

Measures

Working Alliance Inventory. The WAI (short form) is designed to measure therapeutic alliance, and there are both patient and therapist versions available. It is a 12-item self-report measure; subdivided into three sub-scales (goals, bonds, and tasks). Each item is scored using a seven-point Likert scale (1 = never, 7 = always). The measure has two negatively and 10 positively worded items. An overall index of alliance is calculated by summing the 12 items (with negatively worded items reverse-scored), yielding a score range of 12 to 84.

Helping Alliance Questionnaire-Version Two. The HAQ-II is also a measure of therapeutic alliance, available in both therapist and patient versions. It is a 19-item self-report measure and each item is scored using a six-point Likert scale (1 = strongly disagree, 6 = strongly agree). The measure has five negatively and 14 positively worded items. An overall index of alliance is calculated by summing the 19 items (with negatively worded items reverse-scored), yielding a score range of 19 to 114.

Feedback questionnaires. The feedback questionnaires included a range of questions aiming to gain information regarding the acceptability, usability, and face validity of the WAI and the HAQ-II. Participants could respond 'yes' or 'no' to questions and were also given an opportunity to provide comments. The content of questions related to factors such as the relevance, appropriateness and completeness of the WAI/ HAQ-II, and whether the items were hard or uncomfortable to answer.

Demographic variables. A range of demographic variables were collected including age, gender, and ethnicity. For clients, information was also sought on the problem for which they were attending physiotherapy and the duration of treatment. In addition, the physiotherapists were asked how long they had been working as a physiotherapist, and to provide details regarding their previous experience.

Procedure

Ethical approval was gained from the Northern X Regional Ethics Committee. At the beginning of the study, participating organisations identified physiotherapists working in their organisation who they believed would be interested in participating and provided them with a study information sheet. Their written consent was then obtained if they were able and willing to participate. All physiotherapists were then asked to identify any of their clients who were eligible to participate. An independent person (e.g. manager or receptionist) approached eligible clients to ask if they would be interested in participating. If interested, the client was asked to meet with a researcher on completion of their next scheduled physiotherapy session. Clinicians notified the research team when the session was due to take place so that a researcher could be available at the locality to meet with the client.

On session completion, the researcher: a) gave the potential participants a participant information sheet and provided the opportunity to ask any questions; b) obtained written consent if they were willing to participate; and c) if consent was obtained, participants were asked to complete the first set of questionnaires. This contained a demographics questionnaire, the WAI and HAQ-II, and a feedback questionnaire. All consenting clients were then asked to arrive 15 minutes prior to their next scheduled appointment to complete the second set of questionnaires. This contained only the WAI and HAQ-II. The second appointment took place between 3 to 7 days after their first session in order to balance the competing effects of recall bias and potential for change in the therapeutic relationship. It was also a requirement that clients must have had no intervening sessions with their physiotherapist in the interim period (to ensure a stable period of time between repeated measures). Physiotherapists were asked to complete therapist versions of the measures at the same time points as their clients.

Consent forms and questionnaires were stored in separate locked cabinets to ensure confidentiality. Questionnaire instructions clearly stated that information would not be divulged to clients, therapists, or other staff. Therefore clients and physiotherapists were blind to each other's appraisal of the working alliance.

Data analysis

Microsoft Excel was used to calculate summary statistics for all participant characteristics and for each measure (including median, range and inter-quartile range). Client and physiotherapist perspectives of the acceptability and usability of the two measures was determined by: collating qualitative data from the open-ended questions into recurring categories/themes regarding the perceived strengths and limitations of the questionnaires; and by calculating the proportion of yes/no responses on the feedback questionnaire. For test-retest reliability and convergent validity, data were analysed using Stata version 10.0. Cohen's kappa coefficient was calculated to determine the test-retest reliability of each item in the WAI and the HAQ-II. Prior to importing data into Stata for analysis, reverse-score items were entered into Excel and the polarity reversed. Landis and Koch (1977) suggested the following standards for strength of agreement: <0=poor; 0.01-0.2=slight; 0.21-0.40=fair; 0.41-0.60=moderate; 0.61-0.80=substantial; and 0. 81-1.0=almost perfect. Therefore the Kappa coefficients were interpreted accordingly. For convergent validity, Shapiro Wilks tests of normality revealed that the client participant data were significantly different from 'normal' (WAI p<0.001; HAQ-II p=0.04), confirming nonparametric testing was most appropriate. Although the normality tests revealed the therapist data to approximate a normal distribution (WAI p=0.74; HAQ-II p=0.09), the relatively small sample size and ordinal level of measurement supported the use of nonparametric testing. Therefore, the Spearman's rank correlation coefficient was calculated to determine the strength of association between the two questionnaires (convergent validity). Face validity was determined using the same methods as those described above for acceptability and usability

RESULTS

Participant characteristics

Client and physiotherapist demographics are displayed in Table

1. The client sample was diverse on all participant characteristics representing age, ethnicity and treatment duration, although there were no Maori participants. Conversely, the physiotherapist sample was less diverse: all were female and the majority NZ European in ethnicity. There was however diversity in number of years of experience as a physiotherapist (0 to 40 years).

Summary statistics

Summary statistics for both measures are provided in Table 2.

Acceptability and usability

Please note, where a direct participant quote has been used, a pseudonym has been applied to maintain confidentiality. Also, findings, where relevant to more than one aim, may be repeated in brief.

The majority of participants, 73% and 63% for the WAI and 82% and 57% of respondents for the HAQ-II (clients and physiotherapists respectively), indicated the questionnaires did not make them feel uncomfortable. Similarly, 76% and 86% of respondents for the WAI and 73% and 100% of respondents for the HAQ-II (clients and physiotherapists respectively) found that the questions were not difficult to answer in a physiotherapy context.

However, a number of participants reported contrasting views. In particular, the questions relating to 'liking the therapist/ patient' were often commented on as difficult to answer. The main reason cited regarding difficulty answering these questions, was that participants found the meaning of the word 'like' ambiguous, as illustrated by the following quotes:

"[The] word 'like' contains lots of meaning for example trust, in this question 'like' is hard to define-it could be 'likes working with me' or 'likes spending time with me'" (Louise, aged 22)

"[I] had not thought about personally likes or dislikes, [it's a] bit too subjective in my view" (Annette, aged 47)

One participant noted discomfort relating to the subjectivity of the question 'the therapist and I sometimes have unprofitable exchanges':

"[The question] was not worded very well and had a huge interpretation meter which could be taken in either a bad or a good way-therefore I did not answer that question" (Joyce, aged 24)

Similarly, a number of participants identified the question 'the therapist and I have meaningful exchanges' as making them feel uncomfortable and being difficult to answer. Again, the reason cited was ambiguity:

"It feels like you are asking if me and my therapist have some sort of [romantic] relationship" (Zac, aged 39)

"What is meaningful? [It has] many different interpretations" (Brent, aged 63)

One physiotherapist indicated the question 'I have doubts about what we are trying to accomplish in therapy' made her feel uncomfortable. She said:

"[you] shouldn't be having doubts about therapy itself" (Sally, aged 51)

Test-retest reliability

Fifty percent of the WAI kappa values and 63% of the HAQII kappa values were found to be below 0.60 (Table 3). This indicates that, at best, these questionnaire items were only moderately correlated between repeated measures. Interestingly, all the items in which the question or statement was phrased 'negatively' (i.e. the reverse-scored items), had particularly low kappa scores (ranging from 0.32 to 0.58). On the other hand, 50% of the WAI kappa values, and 37% of the HAQ-II kappa values were between 0.61 and 0.80 indicating 'substantial' agreement at different time points for those items. There were no kappa scores in the 'almost perfect' category of between 0.81 and 1.0.

Convergent validity

The Spearman's rank correlation coefficient was: 0.72 (p<0.001) for client data, indicating a moderate to strong positive relationship between measures; and 0.87 (p<0.001) for therapist data, indicating a strong positive correlation (Zou et al 2003).

Face validity

Ninety-one percent of clients and 88% of physiotherapists agreed that the questions in both the WAI and the HAQ-II were relevant, given their individual experiences of physiotherapy. Similarly 90% of clients and 75% of physiotherapists felt that the questions asked in the WAI were appropriate to find out how the client feels about the therapeutic relationship with their physiotherapist and 86% of both clients and physiotherapists felt the questions in the HAQ-II were appropriate.

"[There was a] good broad range of questions [in the WAI], [and the HAQ] covered all aspects of relationship" (Gary, aged 45)

However, as noted above, a number of participants indicated specific questions which lacked relevance and/or appropriateness. In particular, the questions relating to 'liking the therapist/patient' were frequently commented on as being irrelevant. Some physiotherapists said:

"[Whether my client likes me] is not relevant as long as the client engages effectively in therapy...As long as the client is prepared to engage fully in therapy, it doesn't matter whether they like you or not-[it] just makes it easier to chat and pass time if the therapy is repetitive" (Sally, aged 51)

"[I felt] some questions lacked relevance. Particularly questions about whether or not patient/therapist should like each other" (Elizabeth, aged 46)

Several clients made similar comments:

"Whether [my therapist] likes me or not, should not affect what we are trying to achieve" (Ian, aged 50)

"The questionnaire is fine, but you do not have to like the physio to have a successful outcome. You may hate them but it is still very beneficial to you" (Brent, aged 63)

Similarly, irrelevance and lack of clarity prompted two participants to comment on the WAI question 'I feel that my therapist appreciates me' as hard to answer:

"[I'm] not sure I understand why I need to be appreciated" (Annette, aged 47)

"[My] initial impression was "why" does the therapist need to appreciate me" (Philippa, aged 31)

Further, although most participants felt the questionnaires were comprehensive, some participants identified certain factors they felt important, were absent. Eighteen percent of clients and 14% of physiotherapists for the WAI, and 29% and 43% for the HAQ-II (respectively), felt something was missing that might help to ascertain how the client feels about the therapeutic relationship with their physiotherapist. These participants made suggestions regarding additional factors they felt important to the therapeutic relationship. For example, two clients made comments related to the physiotherapist's role as an educator; and in maintaining professionalism:

"Perhaps a question about talking through the treatment methods" (Brian, aged 59)

"Nothing was asked about their [your physio]'s professionalism at all times, or [their] ability to teach you new exercises" (Joyce; aged 24)

Other topics participants' felt were either missing were related to communication and aspects of 'relationship':

"There needs to be a focus on communication, [for example] 'I communicate easily with my therapist/patient'" (Sally; aged 51)

"[There could be] something about how the patient can aid in the building of trust and empathy" (Brian; aged 59)

"Developing understanding and respect for the process through professional relationship seems most on topic" (Elizabeth; aged 46)

DISCUSSION

This project investigated the measurement properties of two commonly used measures of the therapeutic relationship, the WAI and the HAQ-II, in a physiotherapy population. Findings of this study suggest both measures to be generally acceptable, usable, valid, and reasonably reliable in the physiotherapy setting. However there are several points worthy of discussion.

First, while most participants agreed the contents of the WAI and the HAQ-II were relevant and appropriate given their individual experiences of physiotherapy, a number of participants indicated specific questions which lacked relevance and/or appropriateness, therefore limiting their face validity. Our recent review identified several aspects of the WAI and the HAQ-II which did not seem evident in the physiotherapy literature such as whether or not the client believes the therapist likes them and vice versa (Besley et al 2011). The lack of reference to this factor in the physiotherapy literature would indicate that it may not be a priority in the physiotherapeutic relationship. Certainly, the findings of the current study would support this premise with several participants suggesting the questions relating to 'like' were irrelevant, difficult, and/or uncomfortable to answer within the physiotherapy context. This finding is consistent with research in other populations. For example, Kermarrec et al (2006) explored acceptability of the HAQ within a psychiatric population and while they found acceptability to be adequate, a number of participants did not respond to the item 'I believe my therapist likes me'. One participant stated "the doctor is not there to like me that is not his job" (Kermarrec et al 2006, p. 916). Furthermore, several participants in this study identified the questions regarding meaningful and unprofitable exchanges as ambiguous and confusing. This indicates that, should future therapeutic relationship questionnaires include such questions, it may be beneficial to use monosemous words and clearly phrased questions.

Second, while most participants felt the questionnaires were comprehensive, some identified certain factors they felt important missing from the questionnaires further limiting their face validity. For example, a few participants made comments related to the physiotherapist's role as an educator; and in maintaining professionalism as important. This also confirms findings from our recent review which suggested there to be some items considered core to the physiotherapeutic relationship not considered by these existing measures (Besley et al 2011). These findings further challenge the conceptual basis of these commonly used measures in the physiotherapy context and highlight the need for the development of a measure of therapeutic relationship specific to physiotherapy (or rehabilitation more broadly).

In terms of the reliability of both measures, Kappa values for individual items ranged from having 'slight' to 'substantial' agreement between repeated measures. As mentioned previously, the reverse-score items had particularly low Kappa scores. A possible explanation for this may be that the use of double negatives confuses people, especially if they are reading the questions hurriedly. This possibility is supported in existing literature, which suggests reverse-score items may reduce score reliability (Weems and Onwuegbuzie 2001). While the purpose of using positively and negatively worded items is to minimise acquiescent behaviors and the tendency for participants to generally agree rather than disagree (Barnette 2000), one study found that score reliability was 0.1 lower when 'mixed stems' were used (Barnette 2000).

To compare current findings with previous work, Luborsky et al (1996) reported on the test-retest reliability of the HAQ-II. While they did not report on the reliability of individual questionnaire items (as in the current study), their overall score was 0.78 for the patient version, and 0.56 for the therapist version. This indicates a moderate to strong agreement between repeated measures, a slightly stronger level of agreement than in this study, in which 63% of HAQ-II questionnaire items scored, at best, moderate levels of agreement. Further, the moderate to strong correlation between the WAI and the HAQ-II found in the current study is comparable with past work which reported a Spearman's rank correlation coefficient of 0.76 between the two alternative versions of these measures (Bale et al 2006).

Limitations

There were several limitations to the study which impact on interpretation of findings. One of these was the way in which the feedback questionnaires for the WAI and the HAQ-II were worded. While certain questions were designed to investigate acceptability and others face validity, when analysing participant responses, it was difficult to identify which comments related to acceptability and which related to face validity given that a number of participant comments relating to face validity also clearly offered some insight into acceptability. In future studies, it would be worthwhile clearly distinguishing between these two measurement properties when designing feedback questionnaires, to assist the data analysis process.

A further limitation is due to practical constraints (e.g. consecutive appointments with other patients), is that it was not always possible for physiotherapists to complete their version of the questionnaires at exactly the same time points as their clients. This was managed by ensuring that physiotherapists completed the first set of questionnaires later on the same day as seeing the client, and the second set at some stage prior to their scheduled appointment with the client on the same day.

Another limitation to this research was the small sample size. Due to the study's short duration, it was not possible to recruit the target of 30 participant dyads, thus decreasing the statistical strength of the findings. The small sample size also limited the variety of participant characteristics. For example, the lack of Maori participants, given the New Zealand context of the study, is a limitation. Maori perceptions of the therapeutic relationship and what factors are perceived to be important or missing may well be different from other cultural groups, given their differing cultural perspective where face to face engagement, whanau involvement and spirituality are important (Mauri Ora Associates 2006). Further, the fact that all the therapist participants were females means findings cannot be generalised to the male physiotherapist population.

CONCLUSION

The current study suggested both the WAI and the HAQII were acceptable, usable, valid, and reasonably reliable in a physiotherapy setting. However, there were a number of items included in the measures reported to be irrelevant or inappropriate by participants; as well as a number of items thought to be important missing. These findings suggest the WAI and the HAQ-II may lack some conceptual clarity regarding the therapeutic relationship in a physiotherapy context. Additional research is needed to further develop our understanding of the therapeutic relationship in physiotherapy, and this in turn will inform the development of a more conceptually sound measure of the therapeutic relationship specific to that population. That said, these findings suggest the WAI and HAQ-II have adequate measurement properties in the physiotherapeutic population, and therefore both could be adopted in research and practice in the absence of a better available measure.

KEY POINTS

* The WAI and the HAQ-II were generally acceptable, usable, valid, and reasonably reliable in a physiotherapy setting.

* There were a number of items included in the measures reported to be irrelevant or inappropriate by participants; as well as a number of items thought to be important missing. This suggests the measures may lack some conceptual clarity regarding the therapeutic relationship in a physiotherapy context.

* Additional research is required to further develop our understanding of the physiotherapeutic relationship, and to develop a more conceptually sound measure of the therapeutic relationship specific to physiotherapy.

* The WAI and the HAQ-II could be adopted in research and practice in the absence of a better available measure.

ACKNOWLEDGEMENTS

Thanks to AUT University Summer Studentship Scholarship for providing funding for this research. Thanks also to the Person Centred Research Team (PCRT) at AUT University for providing support and assistance throughout the research process. Finally, thanks are given to those localities that facilitated recruitment for the study and to the physiotherapists and their clients for their participation.

ADDRESS FOR CORRESPONDENCE

Nicola M Kayes, AUT University, Northcote, Private Bag 92006, Auckland 1142, New Zealand; nkayes@aut.ac.nz Phone (09) 921 9999 extn 7309 Fax (09) 921 9620

SOURCE OF FUNDING

Funded by AUT University Summer Studentship Scholarship

REFERENCES

Bale R, Catty J, Watt H, Greenwood N and Burns T (2006): Measures of the Therapeutic Relationship in Severe Psychotic Illness: A Comparison of Two Scales. International Journal of Social Psychiatry 52: 256-266.

Barnette JJ (2000): Effects of stem and likert response option reversals on survey internal consistency: If you feel the need, there is a better alternative to using those negatively worded stems. Educational and Psychological Measurement 60: 361-370.

Besley J, Kayes NM, McPherson KM (2011): Assessing therapeutic

relationships in physiotherapy: literature review. New Zealand Journal of Physiotherapy 39(2): 81-91.

Bellner AL (1999): Senses of responsibility: A challenge for occupational and physical therapists in the context of ongoing professionalization. Scandinavian Journal of Caring Sciences 13: 55-62.

Cecero JJ, Fenton LR, Frankforter TL, Nich C and Carroll KM (2001): Focus on therapeutic alliance: The psychometric properties of six measures across three treatments. Psychotherapy 38: 1-11.

De Weert-Van Oene GH, De Jong CAJ, Jorg F and Schrijvers GJP (1999): The Helping Alliance Questionnaire: Psychometric properties in patients with substance dependence. Substance Use & Misuse 34: 1549-1569.

Fenton LR, Cecero JJ, Nich C, Frankforter TL and Carroll KM (2001): Perspective is everything: the predictive validity of six working alliance instruments. The Journal Of Psychotherapy Practice And Research 10: 262-268.

Guedeney N, Fermanian J, Curt F and Bifulco A (2005): Testing the Working Alliance Inventory (WAI) in a French primary care setting. Social Psychiatry and Psychiatric Epidemiology 40: 844-852.

Hall AM, Ferreira PH, Maher CG, Latimer J and Ferreira ML (2010): The influence of the therapist-patient relationship on treatment outcome in physical rehabilitation: A systematic review. Physical Therapy 90: 1099-1110.

Hatcher RL and Gillaspy JA (2006): Development and validation of a revised short version of the working alliance inventory. Psychotherapy Research 16: 12-25.

Horvath AO and Greenberg LS (1989): Development and validation of the Working Alliance Inventory. Journal of Counseling Psychology 36: 223-233.

Kermarrec S, Kabuth B, Bursztejn C and Guillemin F (2006): French adaptation and validation of the Helping Alliance Questionnaires for child, parents, and therapist. Canadian Journal of Psychiatry 51: 913-922.

Landis JR and Koch GG (1977): The measurement of observer agreement for categorical data. Biometrics 33: 159-174.

Le Bloc'h Y, De Roten Y, Drapeau M and Despland JN (2006): New, but improved? Comparison between first and revised version of the Helping Alliance Questionnaire. Schweizer Archiv fur Neurologie undPsychiatrie 157: 23-28.

Leach MJ (2005): Rapport: A key to treatment success. Complementary Therapies in Clinical Practice 11: 262-265.

Luborsky L, Barber JP, Siqueland L, Johnson S, Najavits LM, Frank A and Daley D (1996): The revised Helping Alliance Questionnaire (HAQ-II): Psychometric properties. Journal of Psychotherapy Practice and Research 5: 260-271.

Mauri Ora Associates (2006). Best health outcomes for Maori: Practice implications.

Potter M, Gordon S and Hamer P (2003a): The difficult patient in private practice physiotherapy: A qualitative study. Australian Journal of Physiotherapy 49: 53-61.

Potter M, Gordon S and Hamer P (2003b): The physiotherapy experience in private practice: the patients' perspective. Australian Journal of Physiotherapy 49: 195-202.

Roberts L and Bucksey SJ (2007): Communicating with patients: What happens in practice? Physical Therapy 87: 586-594.

Schonberger M, Humle F, Zeeman P and Teasdale TW (2006): Working alliance and patient compliance in brain injury rehabilitation and their relation to psychosocial outcome. Neuropsychological Rehabilitation 16: 298-314.

Weems GH and Onwuegbuzie AJ (2001): The impact of midpoint responses and reverse coding on survey data. Measurement & Evaluation in Counseling & Development 34: 166-176.

Williams S and Harrison K (1999): Physiotherapeutic interactions: A review of the power dynamic. Physical Therapy Reviews 4: 37-50.

Zou KH, Tuncali K and Silverman SG (2003): Correlation and simple linear regression. Radiology 227: 617-628.

Jessica Besley, BHSc (Physiotherapy), BSc, NZRP

Physiotherapist (Counties Manukau District Health Board)

Nicola M Kayes, BSc, MSc(Hons), PhD

Senior Research Officer, Person Centred Research Centre, Health and Rehabilitation Research Institute, AUT University

Kathryn M McPherson, rn, BA(Hons), PhD

Professor of Rehabilitation, Person Centred Research Centre, Health and Rehabilitation Research Institute, AUT

University
Table 1: Client and therapist participant characteristics

                                   Clients             Physiotherapists

Number of participants               22                        8

Gender
 Females                           7 (32%)                 8 (100%)
 Males                            15 (68%)                  0 (0%)

Age (years)
 Mean                                44                       39
 Range                            18 to 78                 22 to 61

Ethnicity
 NZ European                      13 (59%)                  7 (88%)
 Pacifica                          3 (14%)                  0 (0%)
 Asian                             1 (5%)                   0 (0%)
 European                          1 (5%)                   0 (0%)
 North American                    1 (5%)                   0 (0%)
 Other                             2 (9%)                   1 (13%)
 Did not answer                    1 (5%)                   0 (0%)

Duration of treatment
for the current problem
 Mean (n=17)                      14 weeks                    n/a
 Range (n=17)                      1 to 43                    n/a
                                   weeks
Years working as a
physiotherapist
 Mean                                n/a                   16 years
 Range                               n/a               0 (new graduate)
                                                          to 40 ye

Note. (n/a) not applicable. Unless otherwise specified
data are n (% proportion of sample).

Table 2: Descriptive statistics including median, range, and
interquartile range for the WAI (short form) and the HAQ-II
questionnaires

                                  Median       Range        IQR

WAI total (Client time 1)         79.0         56-84        74-84
WAI total (Client time 2)         81.5         56-84        72-84
WAI total (Physiotherapist        71.5         52-84        65-75
 time 1)
WAI total (Physiotherapist        69.0         57-84        63-77
 time 2)
HAQ-II total (Client time 1)      107.0        83-114       101-110
HAQ-II total (Client time 2)      103.5        93-114       97-112
HAQ-II total (Physiotherapist     97.0         87-114       94-105
 time 1)
HAQ-II total (Physiotherapist     101.0        88-114       94-109
 time 2)

Note. (IQR) Interquartile range.

Table 3: Kappa coefficient and standard error for each
individual item of the WAI (short form) and the HAQ-II

Question         Kappa               Question         Kappa
                 (standard                          (standard
                  error)                             error)

W 1              0.69 (0.10)          H 1            0.52 (0.14)
W 2              0.26 (0.09)          H 2            0.73 (0.12)
W 3              0.50 (0.09)          H 3            0.54 (0.15)
W 4 *            0.44 (0.10)          H 4 *          0.39 (0.11)
W 5              0.66 (0.10)          H 5            0.62 (0.13)
W 6              0.57 (0.10)          H 6            0.54 (0.12)
W 7              0.66 (0.11)          H 7            0.58 (0.15)
W 8              0.53 (0.11)          H 8 *          0.52 (0.13)
W 9              0.67 (0.11)          H 9            0.58 (0.12)
W 10 *           0.54 (0.09)          H 10           0.64 (0.13)
W 11             0.72 (0.10)          H 11 *         0.41 (0.12)
W 12             0.63 (0.10)          H 12           0.63 (0.13)
                                      H 13           0.60 (0.12)
                                      H 14           0.39 (0.13)
                                      H 15           0.63 (0.11)
                                      H 16 *         0.32 (0.12)
                                      H 17           0.53 (0.09)
                                      H 18           0.63 (0.13)
                                      H 19 *         0.58 (0.11)

Note. (W) WAI short form. (H) HAQ-II. * Reverse-score items
Gale Copyright: Copyright 2011 Gale, Cengage Learning. All rights reserved.