‘Physiotherapy in Parkinson’s disease. Towards evidence-based practice’
Leiden University, 29 April 2010
Awarded the PhD Price 2011 by Dutch Scientific Board for Physiotherapy (part of the Royal Dutch Society for Physical Therapy)
Pdf available for free at hdl.handle.net/1887/15341
Chapter 1. General introduction
Parkinson’s disease (PD) is a complex neurodegenerative disease. Despite optimal medical
management, mounting limitations in the performance of activities develop. To counteract
these limitations, many PD patients resort to physiotherapy. However, information about the current use and efficacy of physiotherapy in PD is sparse. The aim of this thesis was to gain insight into the everyday use and added value of physiotherapy in PD and, in addition, to make a step forward in the development of evidence-based physiotherapy for patients with PD.
Chapter 2. The efficacy of physiotherapy in PD
Until 2000, insights into the quality and quantity of physiotherapy in PD were very limited. To estimate the efficacy of physiotherapy in PD, a systematic review was carried out. Twelve controlled clinical trials that evaluated the effects of physiotherapy for patients with PD were identified. Summary effect sizes were calculated. The results showed that PD patients benefit from physiotherapy when this is added to their medication. Significant improvements were seen with regard to activities of daily living, stride length and walking speed. However, the findings also highlighted that many studies contained serious methodological flaws and enrolled an very limited number of patients. Moreover, the specific physiotherapy interventions that were evaluated in the different studies varied widely. Therefore, these results were promising, but far from conclusive.
Chapter 3. Threats for optimal care
The next step was to gain more insights into the current care of physiotherapy for PD. For this purpose, we mailed questionnaires mailed to 300 PD patients. In addition, we also contacted 129 physiotherapists who had treated any of these patients during the previous year. The results showed that high numbers of patients (up to 70% annually) received physiotherapy. Most of these patients were treated for prolonged periods of time, with a mean of nearly one and a half years. Several specific threats that jeopardised the delivery of optimal physiotherapy care were identified. Our first striking observation was the lack of PD-specific expertise among the physiotherapists. Second, most patients were treated by physiotherapists who only incidentally treated PD patients. The referred patients had been ‘diluted’ across the many physiotherapists that are active in the Netherlands. It was rare for therapists to treat more than three patients per year and this low patient volume (or case load) appears insufficient to gain and maintain an acceptable level of expertise. Third, referral seemed, to some extent, arbitrary. This could be due to the absence of an accepted set of criteria for referral to a physiotherapist. We concluded that it was at least questionable whether patients received adequate and evidence-based care
Chapter 4. Evidence-based guideline
To counteract these threats to optimal care, we took several measures. First, an evidence-based guideline was developed to facilitate the uniformity and efficacy of care. Scientific evidence was systematically selected and its quality appraised. The results were supplemented with clinical expertise and patient values and translated into recommendations for clinical practice. Four of these recommendations were based on firm scientific evidence: cueing strategies to improve gait, cognitive movement strategies to improve transfers, specific exercises to improve balance and training of joint mobility and muscle power to improve physical capacity.
Evidence-based clinical decisions can only be facilitated when the recommendations are combined with the patient’s individual preferences and limitations. To achieve this, patients need to be empowered in managing their disease and to participate in making treatment decisions. Therefore, the guideline also provides specific information about the process of goal setting by the patient, supported by the therapist and about the correct use of outcome measures.
Chapter 5. The feasibility of a large RCT
The new guideline served two purposes. Pending further evidence, it intends to provide a decision supporting tool for physiotherapists who treat PD patients in everyday clinical practice. In addition, the guideline can serve to inform the active arm in a trial aiming to evaluate the efficacy of physiotherapy in PD. For this purpose, we designed a randomised controlled trial (RCT). Following recommendations for complex interventions, we first performed a pilot study to assess the feasibility of the actual RCT. In the feasibility study, 173 patients were asked to participate. An important inclusion criterion was having an indication for referral to a physiotherapist, as described in the guideline. Subjects were randomly allocated to either ‘best practice physiotherapy’ (i.e. according to the new evidence-based guideline) or to ‘no physiotherapy’. The primary outcome measures were the Parkinson’s Disease Questionnaire-39 (PDQ-39), the Parkinson Activity Scale (PAS) and a generic patient preference outcome scale, the Patient Specific Index (PSI). Patients were treated by two physiotherapists who had been trained in the correct use of the guideline. Only 14% of the eligible patients could be included. This finding concurred with the experience of others. Moreover, the results showed that the guideline was a feasible tool to guide evidence-based physiotherapy. The PAS, designed for functional assessment in PD, showed a low and non-significant effect size. This could be explained by a possible ceiling effect in the PAS. Activity limitations are mainly experienced in the patient’s own home environment, but the current PAS tests patients in a (presumably easier) hospital setting. Therefore, we advised that in future, researchers should deliver the PAS in the patient’s home. We also considered that the PAS was simply not sensitive enough for mildly affected patients. Therefore, we also advised to modify the PAS, among others by making several tests more difficult. The effect size for the PDQ-39 could be translated into a minimal clinically important difference. The PSI showed the largest effect size and a significant group effect. This was conforming our expectations, as it evaluates the patient specific treatment goal.
Chapters 6 & 7. Improvement of outcome measures
The findings of the feasibility study inspired us to optimise the PAS and the PSI.
First, we developed a Modified PAS, taken into account the findings of the feasibility study (Chapter 6). Our aims were: (a) to introduce a Modified PAS with unambiguous scoring options and without ceiling effect; (b) to evaluate its inter-rater agreement, using physiotherapists with and without PD-specific expertise; and (c) to examine the concurrent validity with the visual analogue scale for Global Functioning and the motor section of the Unified Parkinson’s Disease Rating Sclae (UPDRS-III). The Modified PAS was scored by a large number of physiotherapists, with or without PD-specific expertise, in 15 patients. The results, based on 195 observations, showed absence of the ceiling effect, good concurrent validity, good inter-rater agreement and no differences between physiotherapists with and without PD specific expertise. Second, the generic PSI that we used in the feasibility study was now made PD-specific (Chapter 7). The improvements were based on the evidence-based guideline and the results of a questionnaire identifying patients reported problems. The scale was relabeled as ‘Patients Specific Index for PD’ (PSI-PD). With this new PSI-PD, patients are asked: (a) to select their limited activities out of a predefined list, including one self-report item; (b) to rank the (maximum) five most important of the selected activities in order of importance; and (c) to rate the severity of the limitations in the performance of these activities on a visual analogue scale. De list contains activities related to the five domains of physiotherapy in PD (gait, transfers, posture, dexterity and physical capacity).
To examine the test-retest reliability, a cohort of patients (n=122) was asked to complete the PSI-PD twice within a time span of two weeks. Afterwards, validity was evaluated using a telephone interview. Test-retest agreement on domains was 74-80%, whereas the agreement of the PSI-PD with information collected during the telephone interview was 96%. Only 4 patients used self-report item. Thereby, the results showed that the PSI-PD is a relevant, reliable and valid instrument to identify limitations in everyday activities that are important for both PD patients and physiotherapists. However, the results also showed that the ranking of problems by patients was not reliable. Therefore, assistance by an interviewer (e.g. a physiotherapist) may be necessary to facilitate the ranking process within the PSI-PD. We concluded that the Modified PAS and the PSI-PD offer two promising outcome measures to evaluate the efficacy of physiotherapy in PD in future trials.
Chapter 8. The ParkinsonNet trial
The findings of the feasibility study (Chapter 5) inspired us to develop the ParkinsonNet trial, a multi-centre trial with a cluster randomised design. The aim of this trial is to evaluate the merits of the ParkinsonNet concept, a new approach to improve the delivery of allied health care to patients with PD. Briefly, ParkinsonNet is a professional regional network within the catchment area of hospitals. ParkinsonNet aims to: (a) improve PD-specific expertise among allied health personnel, by training a selected number of therapists according to evidence-based guidelines; (b) enhance the accuracy of referrals by neurologists; (c) boost patient volumes per therapist, by stimulating preferred referral to ParkinsonNet therapists; and (d) stimulate collaboration between therapists, neurologists and patients. The design, conduct and analysis of the ParkinsonNet trial are carefully selected, in order to enable us to report about the trial according to the CONSORT Statement for cluster RCTs. Possible risks of bias associated with the design and conduct of cluster RCTs (e.g. foreknowledge of allocation, differential recruitment between experimental and control clusters and differential application of inclusion and exclusion criteria) have been avoided. In the trial, sixteen regions in the Netherlands (clusters) were randomly divided (Chapter 8). In the eight experimental clusters, a ParkinsonNet was implemented. In the eight control clusters, the organization of care was left unchanged. Participating patients were followed for six months to evaluate the implementation process, health benefits and costs of the intervention. In the ParkinsonNet clusters, 46 therapists were trained and 358 patients were included. In the control clusters, 341 patients were included. The data show that the baseline characteristics of the eight ParkinsonNet clusters and the eight control clusters are comparable. Although we encountered comparable recruiting problems as in previous RCTs on physiotherapy in PD, the multi-centre design and cluster randomization enabled the inclusion of a sufficient number of patients. With 699 participating patients, this is the largest allied health study in PD to date. Results of the ParkinsonNet trial will be available in the second half of 2009.
Chapter 9. Evolution of physiotherapy in Parkinson’s disease
A systematic literature search and analysis of the accumulating evidence of physiotherapy in PD analysis showed that since the publication of the first RCT in 1981, the quantity and quality of clinical trials evaluating the efficacy of physiotherapy in PD has evolved rapidly. A total of 38 controlled clinical trials, 11 systematic reviews and two best-evidence summaries of systematic reviews were identified. Of these, more than half were published after the publication of the guideline in 2004. The recommendations of the guideline remained valid, albeit that the supporting evidence for specific recommendations increased. In addition to the literature search, we personally contacted 35 experts on physiotherapy in PD worldwide. The response to our query for research information was 74% and 19 ongoing, or completed but yet still unpublished, trials were reported. Except for dexterity, all core areas for physiotherapy in PD were involved in the evaluations. Unfortunately, still a large number of these trials were not registered in a database for clinical trials before the start of patient inclusion. The results of several large and well-designed ongoing trials are expected to fill the current knowledge gaps of evidence-based physiotherapy in PD. Consequently, current levels of evidence that support the use of physiotherapy strategies, e.g. to prevent falls or to improve physical capacity, may be increased.
Chapter 10. Future challenges
Significant milestones in the quality of studies have een reached, but many more goals remain.
Four important issues should be taken into account. First, study designs should be improved (e.g. proper randomisation) support sufficient recruitment. Second, insights in the optimal delivery of physiotherapy interventions over time should be gained. Third, appropriate outcome measures, tailored to the patients and intervention studied should be chosen. Fourth, all new trials should be registered in a database for clinical trials before the start of patient inclusion and should be published according to the CONSORT statement. At the same time, sustained attention needs to be given to implement any new acquired evidence into clinical care. One option for this is the development of community-based networks of dedicated health professionals specialised in PD (e.g. ParkinsonNet in the Netherlands, Chapter 8). Given the complex nature of PD, these networks should be multidisciplinary. The recently developed, Dutch national multidisciplinary guideline for PD provides recommendations for the optimal organisation of this care. When indeed referrals are structured to this selected group of health professionals, who are trained to work according to evidence-based guidelines supplemented with the latest evidence and communication between participating health professionals and patients is supported, evidence-based care may be achieved.