RESEARCH ARTICLE Choice of Treatment Modalities was not Influenced by Pain, Severity or Co-Morbidity in Patients with Knee Osteoarthritis Gro Jamtvedt1, Kristin Thuve Dahm2, Inger Holm3, Jan Odegaard-Jensen4 & Signe Flottorp5 1 Norwegian Knowledge Centre for the Health Services, Oslo, and Centre for Evidence-Based Practice, University College Bergen, Bergen, Norway 2 Norwegian Knowledge Centre for the Health Services, Oslo, Norway 3 Section of Health Science, Faculty of Medicine, University of Oslo, and Rikshospitalet University Hospital, Oslo, Norway 4 Norwegian Knowledge Centre for the Health Services, Oslo, Norway 5 Norwegian Knowledge Centre for the Health Services, Oslo, Norway Abstract Background and Purpose. Patients with knee osteoarthritis (OA) are commonly treated by physiotherapists in primary care. The physiotherapists use different treatment modalities. In a previous study, we identified variation in the use of transcutaneous electrical nerve stimulation (TENS), low level laser or acupuncture, massage and weight reduction advice for patients with knee OA. The purpose of this study was to examine factors that might explain variation in treatment modalities for patients with knee OA. Methods. Practising physiotherapists prospectively collected data for one patient with knee osteoarthritis each through 12 treatment sessions.We chose to examine factors that might explain variation in the choice of treatment modalities supported by high or moderate quality evidence, and modalities which were frequently used but which were not supported by evidence from systematic reviews. Experienced clinicians proposed factors that they thought might explain the variation in the choice of these specific treatments. We used these factors in explanatory analyses. Results. Using TENS, low level laser or acupuncture was significantly associated with having searched databases to help answer clinical questions in the last six months (odds ratio [OR] = 1.93, 95% confidence interval [CI] = 1.08–3.42). Not having Internet access at work and using more than four treatment modalities were significant determinants for giving massage (OR = 0.36, 95% CI = 0.19–0.68 and OR = 8.92, 95% CI = 4.37–18.21, respectively). Being a female therapist significantly increased the odds for providing weight reduction advice (OR = 3.60, 95% CI = 1.12–11.57). No patient characteristics, such as age, pain or co-morbidity, were significantly associated with variation in practice. Conclusions. Factors related to patient characteristics, such as pain severity and co-morbidity, did not seem to explain variation in treatment modalities for patients with knee OA. Variation was associated with the following factors: physiotherapists having Internet access at work, physiotherapists having searched databases for the last six months and the gender of the therapist. There is a need for more studies of determinants for physiotherapy practice. Copyright © 2009 John Wiley & Sons, Ltd. Received 18 November 2008; Accepted 4 July 2009 Keywords knee; osteoarthritis; physiotherapy practice *Correspondence Gro Jamtvedt, Researcher, Norwegian Knowledge Centre for the Health Services, PO Box 7004, St. Olavs plass, N-0130 Oslo, Norway. Email: grj@kunnskapssenteret.no Published online 23 December 2009 in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/pri.452 16 Physiother. Res. Int. 15 (2010) 16–23 © 2009 John Wiley & Sons, Ltd. G. Jamtvedt et al. Introduction Internationally, concerns about health-care quality have stimulated improvement efforts at all levels of health-care systems (Committee on the Quality of Health Care in America, 2001; National Strategy for Quality Improvement in Health and Social Services (2005–2015), 2005). Measuring performance is a necessary tool for judging quality, and essential before planning or evaluating any improvement initiative (Committee on Redesigning Health Insurance Performance Measures Payment and Performance Improvement Programs, 2006; Grol et al., 2004). Measuring performance is defined as the measurement of actual clinical practice and its comparison to desired clinical practice (Akl, 2007). Desired clinical practice or evidence-based practice should not only be based on high-quality clinical research, but should also be informed by patients’ preferences and knowledge, and practitioners’ clinical expertise and practice knowledge (Herbert et al., 2005). Desired practice is often formulated as guideline recommendations or quality indicators. By measuring performance, we can identify variation in clinical practice, such as the manner in which physiotherapists treat patients with a similar condition, and we can estimate the amount of undesirable practice and identify determinants for variation (AKL, 2007; Committee on Redesigning Health Insurance Performance Measures Payment and Performance Improvement Programs, 2006; Lilford et al., 2007). Some variation in clinical practice can be expected, for example, between clinical experts and novices because of development of clinical reasoning by years of clinical experience (Jensen et al., 1992). Variation might also be because of characteristics of patients and to legitimate differences in preferences among patients and among physiotherapists. However, much of the variation in health-care delivery has been considered unwarranted because it cannot be explained by type or severity of illness, by patients’ preferences or by the experience of the practitioners (Grol, 2004; Resnik and Hart, 2003). It has been suggested that other factors such as a patient-centred approach, practicalities, incentives and social influence are important determinants of practice, and of the adoption of evidence-based physiotherapy (Metcalfe et al., 2001; Resnik and Hart, 2003; Bridges et al., 2007; Upton and Upton, 2006). We would also expect larger variation in practice in Physiother. Res. Int. 15 (2010) 16–23 © 2009 John Wiley & Sons, Ltd. Practice Variation circumstances where there is uncertainty because of lack of evidence (Wennberg, 2002). In recent years, several studies have measured physiotherapy performance by comparing practice to guidelines (Foster et al., 1999; Gracey et al., 2002; Mikhail et al., 2005; Swinkels et al., 2005), but only a few studies have examined determinants of practice variation (Swinkels et al., 2005; van der Wees et al., 2007; Freburger and Mielenz, 2008). Two Dutch studies examined factors related to number of treatment sessions for low back pain (Swinkels et al., 2005) and ankle injuries (van der Wees et al., 2007). Both studies concluded that some variations could be explained at patient level, typically by severity of the problem and co-morbidity. They found that very little variation was related to characteristics of the physiotherapists. Osteoarthritis is one of the most prevalent diseases among the elderly (Peat et al., 2001). Pain and disability are the main presenting problems in patients with knee osteoarthritis (OA) and the targets for seeking primary health care, including physiotherapy (Peat et al., 2001). In a previous study, we summarized the evidence from systematic reviews on the effects of physiotherapy interventions for patients with knee OA, and we graded the quality of the evidence as high, moderate or low, or as no evidence from systematic reviews (Jamtvedt et al., 2008a). The overview showed that exercise and weight reduction decreased pain and improved function, and that patient education, transcutaneous electrical nerve stimulation (TENS), low level laser and acupuncture could improve patient outcomes. However, for many interventions, there was a lack of evidence from systematic reviews. In a prospective study, we compared actual clinical physiotherapy practice with the evidence from this overview (Jamtvedt et al., 2008a). When the study was carried out, there was no national clinical guideline for treatment of knee OA in Norway, but one Health Technology Assessment report on physiotherapy for knee OA had been published (Nasjonalt kunnskapssenter for helsetjenesten, 2004). Our study showed that almost all therapists (98%) used exercise, which is supported by high-quality evidence. Less than 35% used TENS, low level laser or acupuncture, although these treatment modalities are supported by moderate quality evidence. Almost half of the therapists provided advice about weight reduction to patients that the therapists assessed as overweight. There is high-quality evidence that weight reduction improves outcomes in patients with 17 Practice Variation knee OA. The therapists also used many treatment modalities for which evidence from systematic reviews is lacking. Massage was the most frequently used treatment modality not supported by evidence, provided by 54% of therapists (Jamtvedt et al., 2008b). Our previous study showed that there was variation between physiotherapists in the use of treatment modalities for patients with knee OA. The purpose of this study was to examine factors that might explain variation in treatment modalities for patients with knee OA. Methods The present study was conducted among private physiotherapy practitioners working in outpatient clinics in Norway. These clinics are integrated into primary health care and patients with knee OA are treated in outpatient clinics in Norway. The Regional Committee for Medical Research Ethics in Norway approved the protocol for the study. For the purpose of the study, we developed a paperbased data collection form. The physiotherapists used the data collection form to prospectively register, by self-report, actual clinical practice for one patient with knee OA. We worked closely with a group of clinicians to develop and test the data collection form (Jamtvedt et al., 2008b). The final form had three parts, part one collected data on the patient, part two on the treatments and part three on the physiotherapist. A designer contributed to the layout to create a user-friendly form. In February 2006, we invited all private practitioners in Norway, identified by membership of The Norwegian Physiotherapy Association, to participate in the study (n = 2798). For various reasons, mainly related to not treating patients with osteoarthritis, 744 physiotherapists were not eligible (Jamtvedt et al., 2008b). In August 2006, we distributed the anonymous data collection form with a pre-paid return envelope to the remaining therapists (n = 2054). We asked the physiotherapists to report the management of the first patient with knee OA referred to their practice (one patient), and to complete the form at every treatment session through 12 treatment sessions. We chose 12 treatment sessions because this is the number commonly referred by general practitioners in Norway, and because we considered this number to be sufficient for measuring performance and examining variation in physiotherapy. The diagnosis was to be confirmed by x-ray or 18 G. Jamtvedt et al. magnetic resonance imaging. Patients who had undergone knee arthrosplasty or post-operative treatments were excluded. We tried to increase the response rate using followup reminders, telephone contacts and information in journals and newsletters. The data collection period was nine months, from August 2006 to May 2007. Treatment modalities supported by high or moderate quality evidence (TENS, low level laser or acupuncture, weight reduction advice), or modalities that were frequently used but not supported by evidence from systematic reviews (massage), were the focus of this study. We invited the clinicians who participated in the development of the data collection form and researchers involved in studies of osteoarthritis to a one-day meeting in September 2007. We presented results from the study of performance to the participants and, by interactive discussions and small group work, they discussed and identified factors that they thought might explain variation in the treatment modalities that we had chosen to focus upon. The factors they suggested, such as characteristics of the patient or the physiotherapist, were used in the analyses. Analysis We used SPSS 15.0 (Chicago, IL, USA) in the analyses, and to describe physiotherapy practice and the characteristics of patients and physiotherapists. We merged TENS, low level laser and acupuncture into one variable in the analyses. The reasons for merging these modalities were that each modality was used by less than 20% of the physiotherapists, they can all be classified as physical treatments and they were all supported by moderate quality evidence for pain reduction. By merging these variables, we also restricted the number of dependent variables in the analysis. The outcome variables (massage, weight reduction advice and TENS, low level laser or acupuncture) were either: 1) the treatment modality classified as a dichotomous outcome of ‘used’ or ‘not used’, or 2) the treatment modality classified as a continuous variable: the total number of times an intervention was used during the treatment period (12 sessions). We carried out the explanatory analyses in two steps. Firstly, all variables specified by the clinicians were examined in a univariate manner using the χ2 test (categorical independent variables), t-tests (continuous Physiother. Res. Int. 15 (2010) 16–23 © 2009 John Wiley & Sons, Ltd. G. Jamtvedt et al. Practice Variation independent variables for dichotomous outcome) and univariate linear regression analysis (continuous independent variable for continuous outcome). Secondly, all variables identified in the first step with a p value less than 0.3 were entered as independent variables in a multiple regression analysis or logistic regression analysis. Results The results are based on information from 297 data collection forms. Table 1 outlines the characteristics of the physiotherapists. The mean age of the therapists was 47 years (SD = 11) and 47% were women. The mean age of the patients was 65 years (SD = 11), and 67% were women. Pain intensity during the last week was 5.9 (SD = 2.1) on a 10-point visual analog scale. Almost half of the patients (46%) suffered from pain during the night or at rest. More than half of the patients had bilateral knee OA, and 32% were diagnosed more than five years ago. Thirty-three per cent were considered overweight, and 31% had important co-morbidity, most frequently cardiovascular diseases or low back pain. Fifty per cent of the patients were referred to physiotherapy for knee OA for the first time. The most important goal of treatment, as reported by the therapists, was to reduce pain (92%), followed by increasing muscle strength (85%). The factors that might explain variation in practice, as suggested by the clinical experts, overlapped highly across the treatment modalities (Table 2). Univariate analyses showed that using TENS, low level laser or acupuncture was associated with patient pain and overweight, therapist gender, practice setting, having Internet access at work, search of databases and number of articles read in the last six months, using a p value of Table 1. Characteristics of physiotherapists (n = 297) Variable Age, mean (SD) Years since graduation, mean (SD) Female Single practice Postgraduate education Specialist Exercise facility Internet at work Used databases in last six months Read more than five articles in the last six months 47 (11) 21 (12) 47% 15% 31% 11% 95% 58% 46% 82% Physiother. Res. Int. 15 (2010) 16–23 © 2009 John Wiley & Sons, Ltd. <0.3. These variables were entered as independent variables into the multivariate analysis. Only one of these variables, having searched databases to help answer clinical questions in the last six months, was significant in the multiple regression analysis, p = 0.03 (Table 3). The odds of receiving TENS, low level laser or acupuncture were nearly doubled if the therapist had searched databases in the last six months (odds ratio [OR] = 1.93, 95% confidence interval [CI] = 1.08–3.42). Many factors, both related to the patients and the physiotherapists, were associated with provision of massage in the univariate analyses, given a p value < 0.3 (Table 2). In multivariate analyses, two factors were significantly associated with massage: not having Internet at work and using more than four treatment modalities through the 12 sessions (OR = 0.36, 95%CI = 0.19–0.68 for Internet and OR = 8.92, 95% CI = 4.37– 18.21 for more than four modalities) (Table 3). Giving advice about weight reduction was also associated with several variables in the univariate analyses, given a p value < 0.3 (Table 2). However, multiple regression analyses showed that physiotherapists who provided information about physical activity also gave advice about weight reduction (OR = 11.46, 95% CI = 1.170–112.36), and female therapists were more likely to give advice than their male colleagues (OR = 3.60, 95% CI = 1.12–11.57) (Table 3). The alternative analyses using the total number of each modality as outcome variables confirmed these results. Having searched databases in the last six months was a predictor for using more TENS, low level laser or acupuncture (p = 0.01). Not having Internet access at work and having used more than four treatment modalities were predictors for more use of massage (p = 0.03 for Internet and p < 0.001 for more than four modalities). Giving information about physical activity was associated with giving advice about weight reduction (p = 0.01). Discussion In the present study, we have identified factors that are associated with and might explain variation in treatment modalities for knee OA. We found that use of TENS, low level laser or acupuncture, which is supported by moderate-quality evidence, was associated with having searched databases in the last six months. Not having Internet at work and using more than four treatment modalities were determinants of the use of 19 Practice Variation G. Jamtvedt et al. Table 2. Results from univariate analyses of variables that might explain practice variation, as suggested by clinicians Explanatory variables Dependent variables Weight reduction advice OR (95% CI) Patient characteristics Pain intensity (VAS) Pain type (pain at night/rest) Female Age >5 years since diagnosis Employed First-time treatment Overweight Need for more activity Co-morbidity Physiotherapist characteristics Female Age Specialist Solo practice Exercise room Used database in the last six months Internet at work Read >5 articles in the last six months Read article about knee OA Course/continuing education Other variables Provided advice about physical activity >4 treatment modalities 0.82 (0.44–1.52) 1.29 (0.59–2.84) 1.00 (0.40–2.53) 0.98 (0.94–1.02) 0.40 (0.13–1.23) 1.92 (0.77–4.82) 2.08 (0.90–4.80) TENS, laser or acupuncture p value 0.53 0.52 0.99 0.27* 0.11* 0.16* OR (95% CI) p value Massage OR (95% CI) p value 1.10 (0.98–1.24) 1.74 (1.07–2.82) 0.82 (0.49–1.40) 1.01 (0.99–1.03) 0.10* 0.03* 0.47 0.38 0.98 (0.88–1.09) 1.01 (0.68–1.07) 1.37 (0.82–2.27) 0.10 (0.99–1,03) 0.72 0.76 0.23* 0.39 1.05 (0.65–1.69) 1.42 (0.86–2.36) 0.84 0.18* 0.77 (0.49–1.22) 0.81 (0.50–1.30) 0.27* 0.39 0.77 (0.46–1.33) 0.38 1.43 (0.86–2.36) 0.17* 0.09* 1.57 (0.69–3.59) 1.02 (0.98–1.06) 0.28* 0.25* 0.61 (0.37–1.01) 0.98 (0.97–1.01) 0.67 (0.30v1.51) 0.53 (0.24–1.19) 0.05* 0.31 0.34 0.13* 1.81 (1.12–2.92) 1.00 (0.98–1.02) 0.45 (0.21–0.95) 1.49 (0.77–2.85) 0.02* 0.84 0.04* 0.23* 0.92 (0.41–2.04) 0.81 (0.37–1.83) 1.35 (0.47–3.85) 1.74 (0.64–4.74) 1.32 (0.59–2.99) 0.83 0.62 0.58 0.28* 0.50 2.39 (1.46–3.91) 1.39 (0.85–2.27) 1.87 (0.93–3.76) 0.98 (0.58–1.75) 0.83 (0.51–1.36) 0.01* 0.19* 0.08* 0.98 0.46 0.70 (0.44–1.11) 0.34 (0.27–0.55) 0.49 (0.26–0.93) 0.87 (0.51–1.45) 0.66 (0.42–1.07) 0.13* <0.001* 0.03* 0.62 0.10* 1.10 (1.01–1.20) 0.023* 7.20 (3.92–13.19) <0.001* * Entered as independent variables in the multivariate analyses. Empty cells indicate that the factors were not suggested for these treatment modalities. CI = confidence interval; OA = osteoarthritis; OR = odds ratio; TENS = transcutaneous electrical nerve stimulation; VAS = visual analog scale. massage. There is no evidence from systematic reviews of the effect of massage for patients with knee OA. Finally, we found that advice about weight reduction among patients assessed as overweight by the physiotherapist was provided more frequently by female physiotherapists and by therapists who also gave advice about physical activity. The findings suggest that characteristics of physiotherapists may be more important determinants of practice than characteristics of patients across all treatment modalities. In the present study, practice does not seem to vary because of severity or duration of disease or other patient-specific factors. The present study has several limitations. Firstly, because of the cross-sectional design, we cannot claim to have identified any causal relationships between treatment modalities and explanatory factors, only associations. Secondly, the study has limited explana20 tory power. We selected a limited number of factors that might explain variation, based on suggestions from experienced clinicians and researchers and the data that we had collected. There are probably other factors, such as individual treatment preferences, and access to or contraindications to certain treatment modalities that might be important determinants of practice. On the other hand, adding too many variables into the analysis would have increased the risk of identifying spurious associations. In the present study, we collected data by self-report from physiotherapists. They might have reported what they thought was a socially desirable practice. However, instead of using clinical vignettes, we collected information from real patients through every treatment session to try to capture actual practice. Finally, the response rate was low and the respondents might be selected. The physiotherapists who participated were comparable Physiother. Res. Int. 15 (2010) 16–23 © 2009 John Wiley & Sons, Ltd. G. Jamtvedt et al. Practice Variation Table 3. Factors explaining variation in practice for patients with knee osteoarthritis, based on logistic regression analyses* Explanatory variables Dependent variables Weight reduction advice OR (95% CI) Patient characteristics Pain intensity (VAS) Pain type (pain at night/rest) Female Age >5 years since diagnosis Employed First-time treatment Overweight Need for more activity Co-morbidity Physiotherapist characteristics Female Age Specialist Solo practice Exercise room Used database in the last six months Internet at work Read >5 articles in the last six months Read article about knee OA Course/continuing education Other variables Provided advice about physical activity >>4 treatment modalities TENS, laser or acupuncture p value OR (95% CI) p value Massage OR (95% CI) p value 0.20 0.29 0.94 0.07 0.15 0.16 0.41 0.40 0.70 0.66 3.60 (1.12–11.57) 0.03 0.22 0.07 0.10 0.31 0.24 0.97 1.93 (1.08–3.42) 0.03 0.91 0.46 0.36 (0.19–0.68) 0.77 11.46 (1.17–112.37) 0.66 <0.001 0.59 0.42 0.03 8.92 (4.37–18.21) <0.001 * p Values are reported for all variables included in the analysis; OR (95% CI) reported only for variables that were significant in the analyses. CI = confidence interval; OA = osteoarthritis; OR = odds ratio; TENS = transcutaneous electrical nerve stimulation; VAS = visual analog scale. with private practitioners in Norway regarding age. The mean age among members in the Norwegian Physiotherapists Association is 48 years, compared with 47 years in our study. A higher proportion of men responded to our study. We have no additional i nformation about the non-responding physiotherapists. Our findings do not support results from studies that explained variation in the number of physiotherapy treatment sessions for low back pain and ankle injuries (Swinkels et al., 2005; van der Wees et al., 2007). They found that variation mainly depended on characteristics of the patients, such as duration of complaints, prior therapy, recurrence of injury and patients’ age and gender. On the other hand, our results are comparable with findings from a recently published study on the quality of physiotherapy care for patients with low back pain from the United States (Resnik et al., 2008). Resnik and colleagues (2008) found that clinics with Physiother. Res. Int. 15 (2010) 16–23 © 2009 John Wiley & Sons, Ltd. fewer physical therapist assistants were more likely to provide better care, indicating that factors related to practice circumstances might explain variation in the quality. To our knowledge, the present study is the first to examine factors that might explain variation in physiotherapy performance for patients with a chronic rheumatic condition. We found that variation seemed to be more associated with characteristics of the physiotherapists than of the patients. It has been suggested that practice variation explained by determinants at the level of the therapist is not desirable, based on the assumption that variation should primarily be explained by characteristics of the patient and heterogeneity among patients (Swinkels et al., 2005; Leemrijse et al., 2006). We agree that treatment should be tailored to the individual patient, and that patients should be involved in decisions about treatments. However, not all variation 21 Practice Variation explained by factors related to the therapist is undesirable. Factors associated with education, specialization and information behaviour may be important facilitators of ‘best practice’(Bridges et al., 2007). Quality improvement strategies have often included educational activities, for example, to implement clinical guidelines and improve skills in evidence-based practice (Bekkering et al., 2005; Forsetlund et al., 2009). Bridges and colleagues (2007) found that personal characteristics, especially a desire for learning and selfdirected learning, were associated with the propensity to adopt evidence-based physiotherapy, whereas characteristics of the social system made only a minimal contribution to the observed variation in the propensity to adopt evidence-based physiotherapy (Bridges et al., 2007). They suggested that the informationseeking behaviour common to both self-directed learning and evidence-based physiotherapy may account for the association. One important question is whether informationseeking behaviour, such as having Internet access or having searched databases, can explain practice variation in itself. Physiotherapists who are interested in information seeking might differ from physiotherapists without this attitude (behaviour), and our findings might therefore be confounded by other factors related to these personalities. Information-seeking behaviour might be confounded with other characteristics of therapists, such as being an innovator who wants to use new methods and new technology. Providing massage and TENS, low level laser or acupuncture was associated with information-seeking behaviour, although in opposite directions. Not having Internet at work predicted use of massage whereas having searched databases predicted use of TENS, low level laser or acupuncture. There is no evidence for the effect of massage, while there is moderate-quality evidence that TENS, low level laser and acupuncture reduce pain. For interventions supported by highor moderate-quality evidence, information-seeking behaviour might be a determinant of being up to date. Therapists who read articles regularly, and are up to date, might choose not to use massage based on the current lack of evidence. This is supported by our findings suggesting that having read more than five articles in the last six months was associated with using TENS, low level laser or acupuncture, and not using massage. On the other hand, we found no association between practice variation and having read specific articles 22 G. Jamtvedt et al. about knee OA or having attended a course about knee OA. The factors that might explain variation in giving advice about weight reduction were associated with different types of factors than the other modalities examined. The clinicians strongly suggested that providing weight reduction advice was commonly applied by giving advice about physical activity. The results confirm that giving these two types of advice were associated. Patients who were given advice about physical activity were also more often provided with advice about weight reduction. Thus, the findings do not support the hypothesis introduced by the clinicians. Rather, it suggests that these types of advice were linked. Because weight reduction can improve function and pain in patients with knee OA (Christensen et al., 2007; Jamtvedt et al., 2008a) physiotherapists should focus more on such advice. Educational activities, both at graduate and postgraduate levels, should address this issue. Implications Knowledge about factors associated with the choice of physiotherapy modalities in patients with knee OA may inform strategies to reduce unwarranted variation in practice. We need more research on the characteristics of patients and physiotherapists, and attitudes, values and preferences related to practice patterns in physiotherapy. Acknowledgements We thank Saga Høgheim for assistance with data entering, Julia Tavridou for layout and the Norwegian physiotherapists for completing the data collection form. The study was funded by The Norwegian Fund for Post-Graduate Training in Physiotherapy. Dr Lucy Hyatt provided editorial assistance for the final manuscript. REFERENCES Akl EA, Treweek S, Foy R, Francis J, Oxman AD. NorthStar, a support tool for the design and evaluation of quality improvement interventions in healthcare. Implementation Science 2008; 2: 19. Bekkering GE, Hendriks HJM, van Tulder MW, Knol DL, Hoeijenbos M, Oostendorp RAB, Bouter LM. Effect on the process of care of an active strategy to implement Physiother. Res. Int. 15 (2010) 16–23 © 2009 John Wiley & Sons, Ltd. G. Jamtvedt et al. clinical guidelines on physiotherapy for low back pain: a cluster randomised controlled trial. Quality & Safety in Health Care 2005; 14: 107–112. Bridges PH, Bierema LL, Valentine T. The propensity to adopt evidence-based practice among physical therapists. BMC Health Services Research 2007; 7: 103. Christensen R, Bartels EM, Astrup A, Bliddal H. The effect of weight reduction in obese patients diagnosed with knee osteoarthritis (OA): a systematic review and metaanalysis. Annals of the Rheumatic Diseases 2007; 66: 433–439. Committee on Redesigning Health Insurance Performance Measures Payment and Performance Improvement Programs. Performance measurement: accelerating improvement. Washington; D.C., 2006. Committee on the Quality of Health Care in America. Crossing the Quality Chasm: A New Health System for the 21st Century. Washington, D.C.: National Academy Press, 2001. Forsetlund L, Bjørndal A, Rashidian A, Jamtvedt G, O’Brien MA, Wolf F, Davis D, Odgaard-Jensen J, Oxman AD. Continuing education meetings and workshops: effects on professional practice and health care outcomes. Cochrane Database of Systematic Reviews 2009; Issue 2. Art. No.: CD003030. DOI: 10.1002/14651858.CD003030.pub2. Foster NE, Thompson KA, Baxter GD, Allen JM. Management of nonspecific low back pain by physiotherapists in Britain and Ireland — A descriptive questionnaire of current clinical practice. Spine 1999; 24: 1332–1342. Freburger JK, Mielenz TJ. Invited commentary. Physical Therapy 2008; 88: 1007–1009. Gracey JH, McDonough SM, Baxter GD. Physiotherapy management of low back pain: a survey of current practice in Northern Ireland. Spine 2002; 27: 406–411. Grol R, Wensing M, Eccles M. Improving patient care; the implementation of change in clinical practice. London: Elsevier, 2004. Herbert R, Jamtvedt G, Mead J, Hagen KB. Practical Evidence-Based Physiotherapy. London: Elsevier, 2005. Jamtvedt G, Dahm KT, Christie A, Moe R, Haavardsholm E, Holm I, Hagen KB. Physical therapy for patients with osteoarthritis of the knee, an overview of systematic reviews. Physical Therapy 2008a; 88: 123–136. Jamtvedt G, Dahm KT, Holm I, Flottorp S. Measuring physiotherapy performance in patients with osteoarthritis of the knee: a prospective study. BMC Health Services Research 2008b; 8: 145. Jensen GM, Shepard KF, Gwyer J, Hack LM. Attribute dimensions that distinguish master and novice physical therapy clinicians in orthopedic settings. Physical Therapy 1992; 72: 711–722. Physiother. Res. Int. 15 (2010) 16–23 © 2009 John Wiley & Sons, Ltd. Practice Variation Leemrijse CJ, Plas GM, Hofhuis H, van den Ende CHM. Compliance with the guidelines for acute ankle sprain for physiotherapists is moderate in the Netherlands: an observational study. Australian Journal of Physiotherapy 2006; 52: 293–299. Lilford RJ, Brown CA, Nicholl J. Use of process measures to monitor the quality of clinical practice. BMJ 2007; 335: 648–650. Metcalfe C, Lewin R, Wisher S. Barriers to implementing the evidence base in four NHS therapies. Physiotherapy 2001; 87: 441–443. Mikhail C, Korner-Bitensky N, Rossignol M, Dumas JP. Physical therapists’ use of interventions with high evidence of effectiveness in the management of a hypothetical typical patient with acute low back pain. Physical Therapy 2005; 85: 1151–1167. Nasjonalt kunnskapssenter for helsetjenesten. Effekt av fysioterapi ved kneleddsartrose, begrenset til elektroterapi og øvelsesbehandling. Oslo, 2004. National Strategy for Quality Improvement in Health and Social Services (2005–2015), and it’s going to get better National Directorate for Health and Social Affairs. Oslo, 2005. Peat G, McCarney R, Croft P. Knee pain and osteoarthritis in older adults: a review of community burden and current use of primary health care. Annals of the Rheumatic Diseases 2001: 60; 91–97. Resnik L, Hart DL. Using clinical outcomes to identify expert physical therapists. Physical Therapy 2003: 83; 990–1002. Resnik L, Liu D, Mor V, Kart DL. Predictors of physical therapy clinic performance in the treatment of patients with low back pain syndromes. Physical Therapy 2008; 88: 989–1004. Swinkels IC, Wimmers RH, Groenewegen PP, van den Bosch WJ, Dekker J, van den Ende CH. What factors explain the number of physical therapy treatment sessions in patients referred with low back pain; a multilevel analysis. BMC Health Services Research 2005; 5: 74. Upton D, Upton P. Knowledge and use of evidence-based practice by allied health and health science professionals in the United Kingdom. Journal of Allied Health 2006; 35: 127–133. van der Wees PJ, Hendriks EJM, Jansen MJ, van Beers H, de Bie RA, Dekker J. Adherence to physiotherapy clinical guideline acute ankle injury and determinants of adherence: a cohort study. BMC Musculoskeletal Disorders 2007; 8: 45. Wennberg JE. Unwarranted variations in healthcare delivery: implications for academic medical centres. BMJ 2002; 325: 961–964. 23