Physiotherapy 95 (2009) 110119

A treatment schedule of conventional physical therapy provided to enhance upper limb sensorimotor recovery after stroke: Expert criterion validity and intra-rater reliability
Catherine Donaldson, Raymond C. Tallis, Valerie M. Pomeroy
St Georges University of London, Geriatric Medicine, Cramer Terrace, London SW17 0RE, UK

Abstract Background Inadequate description of treatment hampers progress in stroke rehabilitation. Objective To develop a valid, reliable, standardised treatment schedule of conventional physical therapy provided for the paretic upper limb after stroke. Design, setting and participants Eleven neurophysiotherapists participated in the established methodology: semi-structured interviews, focus groups and piloting a draft treatment schedule in clinical practice. Different physiotherapists (n = 13) used the treatment schedule to record treatment given to stroke patients with mild, moderate and severe upper limb paresis. Rating of adequacy of the treatment schedule was made using a visual analogue scale (0 to 100 mm). Mean (95% condence interval) visual analogue scores were calculated (expert criterion validity). For intra-rater reliability, each physiotherapist observed a video tape of their treatment and immediately completed a treatment schedule recording form on two separate occasions, 4 to 6 weeks apart. The Kappa statistic was calculated for intra-rater reliability. Results The treatment schedule consists of a one-page A4 recording form and a user booklet, detailing 50 treatment activities. Expert criterion validity was 79 (95% condence interval 74 to 84). Intra-rater Kappa was 0.81 (P < 0.001). Conclusion This treatment schedule can be used to document conventional physical therapy in subsequent clinical trials in the geographical area of its development. Further work is needed to investigate generalisability beyond this geographical area. 2008 Chartered Society of Physiotherapy. Published by Elsevier Ltd. All rights reserved.
Keywords: Rehabilitation; Stroke; Physical therapy

Introduction Upper limb weakness is a frequent occurrence after stroke [1] and recovery is often limited. At 6 months post stroke, only approximately 12% of stroke survivors who initially experienced severe motor impairment were found to regain full use of their arm and hand [2]. Conventional physical therapies, those used in routine clinical practice, are known to contribute to the benecial effects of stroke units [3] but remain under-evaluated. Further evidence is needed to decide which conventional physical therapy (CPT) interventions are most effective for which patients, at which dose, and at what time in the recovery process [4]. Progress in stroke rehabilitation is
Corresponding author at: Faculty of Health, University of East Anglia, Norwich NR4 7TJ, UK. Tel.: +44 1603 591724; fax: +44 1603 593166. E-mail address: v.pomeroy@uea.ac.uk (V.M. Pomeroy).

impeded by insufcient description of CPT treatment activities [5,6] and lack of use of packages of treatment activities [7], which limits replication of research studies and incorporation of research ndings into clinical practice [4]. A central difculty is that CPT is complex. Indeed, it has been described as both a black box and a Russian doll [8,9]. Such terms imply that CPT cannot be described precisely [10]. Published reports of evaluative research often describe CPT treatment activities using labels such as routine care, even though this gives little information about the specic treatment activities that are used [5,11]. However, it is known that the content of CPT differs between clinical settings and over time as it is inuenced by what physiotherapists learned during undergraduate training [12]. The use of treatment labels has been questioned further by the nding that even though physiotherapists around a major UK city claimed to use an eclectic approach, they actually used a traditional

0031-9406/$ see front matter 2008 Chartered Society of Physiotherapy. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.physio.2008.11.005

C. Donaldson et al. / Physiotherapy 95 (2009) 110119

111

Bobath approach [13]. Treatment labels mean different things to different people. Using labels as descriptors of physical therapy in evaluative research could result in CPT treatment activities being discarded inappropriately [4]. For example, some effective CPT treatment activities could be discarded because a different CPT intervention with the same descriptive label was found to be ineffective. Continuing to use labels is therefore unsatisfactory. Before CPT can be evaluated, it needs to be described in sufcient detail so that a physiotherapist with appropriate training will know exactly what treatment activity was tested and be able to replicate it in clinical practice. Treatment schedules are a means of providing a standardised treatment protocol for evaluation [10,14,15]. A systematic review [10] identied 15 reports of development of treatment schedules, 10 of which are related to stroke care and three to the upper limb. The alleviation/remediation of hemiplegic shoulder pain forms the focus for two studies [14,16], and Hunter et al. [10] concentrated on a small part of CPT called mobilisation and tactile stimulation which is designed to provide therapist hands-on sensory stimulation for stroke survivors with paralysis or severe paresis. Although these studies provide sufcient description of CPT and enable recording of both content and dose, they are focused on relatively small areas for CPT treatment activities for the upper limb after stroke. Furthermore, neither the validity nor reliability of these treatment schedules have been tested. The aims of the present study were: to develop a treatment schedule of CPT treatment activities aimed at enhancing sensorimotor recovery of the upper limb after stroke, which has sufcient precision to enable treatment activities to be recorded in a standardised way so that they can be communicated to clinicians and researchers; to determine whether senior neurophysiotherapists agree that the record of the treatment generated using the treatment schedule accurately describes the treatment activities actually given (expert-criterion-related validity) [17]; and to determine whether individual senior neurophysiotherapists provide an identical list of treatment activities when, on two separate occasions, they view video tapes of the same treatment sessions given on an earlier occasion (intrarater reliability).

Participants and clinical centres All physiotherapists who participated in this study were experienced in stroke rehabilitation. Twelve physiotherapists were recruited for participation in development of the treatment schedule. The interviews were conducted at six different National Health Service (NHS) trusts, in or around the London area, including teaching hospitals, district general hospitals and clinical settings in the community. Different physiotherapists (band 6 or 7, n = 14) from different London NHS trusts participated as raters in investigation of validity and reliability of the treatment schedule. For investigation of validity and reliability, three stroke patients who had suffered a stroke at least 7 days prior to recruitment were recruited for each of the raters (i.e. 30 stroke patients). All stroke patients exhibited loss of functional ability in the paretic upper limb, as assessed by a score of 56 or less (maximum score 57) on the Action Research Arm Test (ARAT) [18,19]. To maximise the likelihood that all treatment activities listed in the treatment schedule were used by the raters, each rater (physiotherapist) treated three stroke patients, one with mild (ARAT score 0 to 19), one with moderate (ARAT score 20 to 38) and one with severe (ARAT score 39 to 56) loss of upper limb functional ability. Stroke patients were identied from each raters NHS caseload so that specic assessment activity required to make decisions about what treatment to provide did not take up time during the recording session. Procedure The treatment schedule was developed using the ve stages identied in an earlier systematic review [10] (Fig. 1). Generation of a treatment list The themes for the semi-structured interview were provided as questions: What physiotherapy treatment activities would you give to patients who t these criteria when they present for rehabilitation? How would you progress the treatment? The clinical focus for the semi-structured thematic interviews was provided by the study criteria to be used in a subsequent Phase I trial to begin evaluation of the resultant treatment schedule: infarction of the anterior cerebral circulation (diagnosed through neuro-imaging) between 1 week and 3 months after stroke; some voluntary muscle activity in the paretic upper limb; no obvious unilateral visuospatial neglect on clinical observation of subjects ability to orientate towards objects and people in their environment;

Methods Design Semi-structured thematic interviews, content analysis, focus groups and piloting of a draft schedule in clinical practice were undertaken [10]. Validity and reliability were tested using a prospective agreement design [10].

112

C. Donaldson et al. / Physiotherapy 95 (2009) 110119

Fig. 1. Flow chart to illustrate study procedure. VAS, visual analogue scale; CI, condence intervals.

able, prior to their stroke, to use the paretic upper limb to lift a cup to drink from; and able to follow a one-stage command (touch your nose with your stronger hand). Renement of list into treatment schedule Audio tapes of the interviews were transcribed verbatim. Content analysis was undertaken independently by two researchers to produce a list of treatments used. Differences between the lists were discussed and resolved by the two researchers with reference to the original transcripts. The resultant treatment list was discussed at a focus group of participating physiotherapists to check for accuracy and omissions. Those unable to attend were invited to give written/verbal feedback. The treatment list was modied following the feedback and a draft treatment schedule was produced.

Piloting the draft treatment schedule in clinical practice Participating physiotherapists were invited to pilot the draft treatment schedule with their teams in clinical practice. Each therapist received training in how to use the treatment schedule and was asked to use it over a 2-week period with patients who met the study criteria. Appropriate changes were made and the nal treatment schedule was produced. Validation of the treatment schedule Each rater was familiarised with the recording form, the denitions of terms used, the associated manual and the method for completing the recording form. Raters had a 2-week period to practice using the treatment schedule in routine clinical practice. Following training, each rater treated three stroke patients, one from each category; i.e. mild, moderate and severe loss of functional ability of the paretic upper limb. Raters were asked

C. Donaldson et al. / Physiotherapy 95 (2009) 110119

113

to provide a treatment session focused entirely on improving functional ability in the paretic upper limb. Within this broad guidance, the specic content of the treatment sessions was decided by the therapist. The rst author videotaped the treatment session which lasted for a maximum of 20 minutes and a minimum of 10 minutes, and told raters when the recording started and stopped. Immediately after the treatment session, raters completed the treatment schedule recording form independently by ticking the relevant boxes describing the treatment provided. They immediately rated the adequacy of the description on the recording form of the treatment they had given using a 100 mm vertical visual analogue scale (VAS) with the lowest anchor point corresponding to did not provide a description of the treatment provided (score 0), and the highest anchor point corresponding to completely described the treatment provided (score 100). Raters used a ballpoint pen to mark the VAS with a cross (x). If the centre of the cross was off-line, a horizontal line was projected from the centre of the cross to the VAS line [19]. Specic criteria for the rating of adequacy of description were not provided to raters, who were simply asked to use their expert clinical judgement. Raters did not keep copies of completed forms. Testing reliability of the treatment schedule Intra-rater reliability was tested. The same stroke patient treatment session was rated on two separate occasions to avoid change in clinical presentation between treatment sessions on different occasions. The video tapes of the treatment sessions were therefore used [20,21]. The video tapes of the three treatment sessions for each rater were placed into a computer generated random order and copied on to Tape 1 (10 tapes, one for each rater), and then placed into a different computer generated random order and copied on to Tape 2 (10 tapes, one for each rater) [20]. Each rater was sent their copy of Tape 1, viewed each subject session and then completed a recording form to describe the treatment given (i.e. they completed recording forms for their own treatment sessions). After completing the three recording forms for Tape 1, each rater returned them and the tape to the rst author (no copies kept by raters). Four to six weeks later, each rater was sent their copy of Tape 2 and completed the recording forms using the same procedure. Data analysis: validity and reliability The VAS score was measured using an acetate sheet with a 100 mm line divided into 1 mm markings placed over the completed VAS and read by two independent assessors. The mean of the two readings was used for analysis. To estimate expert criterion validity, the mean and 95% condence intervals (95% CI) of VAS scores were calculated from the 30 separate ratings (three stroke patients per rater). Adequate description of treatment was set at a mean score of 75% or above based on recommendations that an acceptable level for

reliability is 0.75 [21] and in the absence of specic published statistical guidance. To estimate intra-rater reliability for each pair of recording forms, all tick boxes for the activities section of the recording form were allocated a code of 1 if ticked or not ticked on both occasions, i.e. in agreement. A code of 0 was allocated if a box was ticked on one occasion but not on the other, i.e. not in agreement. The unweighted Kappa statistic and its Pvalue were then calculated, using Statistical Package for the Social Sciences Version 10 (SPSS Inc., Chicago, IL, USA), to estimate the level of agreement. The acceptable level for reliability was set at 0.75 [22]. A comparison of the record of treatment activities made immediately after a session and from a videotaped recording on a later occasion was not the focus of this study. Therefore, data were not analysed to investigate this issue.

Results Participants One physiotherapist who consented to participate in the development of the treatment schedule did not adhere to the interview themes despite repeated facilitation, and therefore this interview could not be used. Consequently, 11 physiotherapists participated in the interviews. Four physiotherapists attended the focus group, two gave written feedback, and two could not be contacted due to change of workplace. The remaining physiotherapists did not respond. Thirteen physiotherapists were involved in piloting the draft treatment schedule in inpatient, outpatient and community clinical settings. To test validity and reliability, 14 physiotherapists and 35 stroke patients were recruited from ve different sites (two London teaching hospitals, two community centres and one rehabilitation centre). Full sets of data were obtained from 10 physiotherapists. Four physiotherapists were unable to complete data collection because of rotating to another post or being placed on long-term sick leave. All collected data were included in the analysis. Content and format of the treatment schedule The treatment schedule consists of a treatment recording form (Fig. 2) and explanatory booklet (available from corresponding author). The recording form is on one side of A4 paper and is divided into ve categories: aims of treatment, gross position of patient during activities used, setting, equipment used and treatment activities (divided into 11 subcategories). Information regarding the date, the duration of the upper limb treatment intervention, the number of therapists and assistants involved, and their identication can also be recorded. The reverse of the page gives instructions for completion and a condensed glossary of any abbreviations and terms used (Box 1). The explanatory booklet is

114

C. Donaldson et al. / Physiotherapy 95 (2009) 110119

Fig. 2. The treatment schedule recording form.

C. Donaldson et al. / Physiotherapy 95 (2009) 110119

115

Box 1 Instructions for completing the treatment schedule recording form. Please note that this treatment schedule is not suitable for patients demonstrating neglect. 1. One form for each treatment session Please complete one form for each treatment session given to patients included as subjects in the Functional Strength Training (FST) Upper Limb Clinical Trial. 2. To complete the aims section Please place a tick in the box that best describes the aims relevant to the particular treatment session being recorded. Unless stated otherwise, it is assumed that the aim is to improve/optimise in each case. 3. To complete the gross position section Place a tick in the box for every gross position used to deliver physiotherapy treatment during treatment sessions being recorded. 4. To complete the equipment section Please write the name of any equipment used during the particular treatment session being recorded. Please refer to the booklet for further details of equipment. 5. To complete the treatment activities section Please place a tick in the boxes which best describe the treatment that was given to the patient during the particular treatment session being recorded. 6. For further description of items on recording form, see overleaf Please refer to the accompanying document Description of Upper Limb Treatment for Patients in FST Trial. 7. Completed forms given to research team Completed forms will be collected weekly by Catherine Donaldson (researcher conducting the FST Trial).

A gliding manipulation performed with light centripetal pressure that deforms subcutaneous tissue down to the investing layer of the deep fasciaa Est. duration of session Estimated duration of treatment session Estimated percentage of time on upper limb Estimated percentage of treatment session spent working directly on the upper limb Facilitation The application of an appropriate mode and dose (frequency, duration and intensity) of sensory stimulus provided by the therapist to access a desired active response from the patientb Friction A repetitive, specic, non-gliding technique that produces movement between the bres of connective tissue, increasing tissue extensibility and promoting ordered alignment of collagen within the tissuesa Lymph drainage techniques A non-gliding technique performed in the direction of lymphatic ow, using short, rhythmical strokes with minimal to light pressure, which deforms subcutaneous tissue without engaging musclea Mental imagery Mental rehearsal of a motor act that occurs in the absence of overt motor output Myofascial release A technique that combines a non-gliding fascial traction with varying amounts of orthopaedic stretch to produce a moderate, sustained tensional force on the muscle and its associated fascia, which results in palpable visco-elastic lengthening and plastic deformation of the fasciaa Petrissage A group of related techniques that repetitively compress, shear and release muscle tissue with varying amounts of drag, lift and glidea Physiotherapist Person with professional physiotherapy qualication Specic compression A non-gliding technique that is applied with a specic contact surface to muscle, tendon or connective tissue; the compression and release is applied in a direction that is perpendicular to the target tissue, and the compression is often sustaineda Stroking Gliding over the patients skin (unidirectionally) with minimal deformation of subcutaneous tissuesa Rehabilitation assistant Person assisting the physiotherapist but who is not a qualied physiotherapist Andrade C-K, Clifford P. Outcome-based massage. London: Lippincott Williams and Wilkins; 2001. Hunter SM, Crome P, Sim J, Donaldson C, Pomeroy VM. Development of treatment schedules for research: a structured review to identify methodologies used and a worked example of mobilisation and tactile stimulation for stroke patients. Physiotherapy 2006;92:195207.
b a

Abbreviations for glossary of terms used in recording form Efeurage

a 10-page manual with the following sections: explanatory background/introduction, the recording form and instructions, examples of equipment used, examples of specic treatment activities, examples of materials used for splinting, glossary of terms and references.

Validity and reliability Expert validity Ten raters provided three VAS scores (one mild, one moderate and one severely impaired stroke patient). Four raters

116

C. Donaldson et al. / Physiotherapy 95 (2009) 110119

Table 1 Visual analogue readings per rater by loss of functional activity category. Rater Visual analogue score (0 to 100), loss of functional ability Mild 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Mean (95% condence interval) 90 100 87 93 77 53 89 88 85 94 No data No data No data No data 86 (76 to 95) Moderate 86 98 79 73 57 75 69 67 50 93 No data 84 60 91 76 (67 to 84) Severe 85 99 90 51 76 64 77 85 54 82 92 No data 65 No data 77 (67 to 86) Mean (95% condence interval) of all treatments for all raters

79 (74 to 84)

did not provide a full set of three VAS scores. The mean score for all treatment sessions was 79 (95% CI 74 to 84) which is above the 75% level set for acceptability, although the lower 95% CI was just below acceptability (Table 1). Intra-rater reliability Of the 50 treatment activities, 41 (82%) were used by physiotherapists in this study. Those not used were: soft tissue mobilisation: lymph drainage techniques (treatment activity 1.5, Fig. 2) and frictions (treatment activity 1.7, Fig. 2); positioning: half lying (treatment activity 4.4, Fig. 2);

specic sensory input: electrical stimulation (treatment activity 5.3, Fig. 2); splinting techniques: shoulder support (treatment activity 6.1, Fig. 2) and elbow support (treatment activity 6.2, Fig. 2); balance and mobility incorporating upper limb activity: in walking (treatment activity 8.5, Fig. 2); and other treatment activities/techniques: acupuncture (treatment activity 11.1, Fig. 2) and ultrasound (treatment activity 11.2, Fig. 2). Data for agreement for the ratings of the same treatment session on two different occasions are shown in Table 2. The Kappa statistic for the 95% agreement was 0.81 (P < 0.001).

Table 2 Intra-rater reliability (agreement between treatment recording forms for the same treatment on two different occasions for each rater). Rater Percentage agreement by impairment level Mild 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Mean % agreement Kappa statistic P-value 96 98 94 92 94 100 96 98 94 96 No data No data No data No data 96 0.85 <0.001 Moderate 94 96 86 96 94 98 96 94 92 92 No data 96 100 No data 95 0.79 <0.001 Severe 90 94 94 96 90 100 98 92 96 92 86 No data 94 No data 94 0.78 <0.001 Mean, all categories for all raters Kappa statistic and P-value, all categories for all raters

94.5

0.81, P < 0.001

C. Donaldson et al. / Physiotherapy 95 (2009) 110119

117

Thus, intra-rater reliability was found to be above the acceptable level of 0.75. Intra-rater reliability was also acceptable for each of the separate stroke patient categories (Table 2): 0.85 (P < 0.001 for mild, 0.79 (P < 0.001) for moderate and 0.78 (P < 0.001) for severe.

Discussion The authors have developed a treatment schedule to describe and record CPT provided by physiotherapists to enhance sensorimotor recovery of the upper limb after stroke. This treatment schedule is probably the rst to codify a comprehensive list of treatments for the upper limb after stroke to enable research evaluation of the content and quantity of CPT. Other CPT treatment schedules, developed using similar methodology, have focused on other aspects of stroke rehabilitation [15,23] and precise clinical presentations [14], and have included activities usually undertaken by other professional groups [24] or specic CPT treatment activities [10]. The present treatment schedule therefore adds to our knowledge of the content of CPT by describing treatment activities in sufcient detail to allow for replication of evaluative studies. In addition, it is probably one of the rst to have been tested formally for aspects of validity and reliability. Intra-rater reliability was found to be acceptable but expert criterion validity was borderline. Although the mean VAS score for accuracy of treatment description exceeded the acceptable level of 75, the lower 95% CI was just below at 74. However, guidelines for interpretation would place 74 in the category of substantial agreement and the upper 95% CI of 84 in the category of almost perfect [25]. Methodological strengths of the study reported in this paper include: the different randomised orders of the two video tapes used for the intra-rater reliability investigation, the broad spectrum of clinical areas in which the raters worked, and ensuring that the treatment schedule was used across different clinical presentations of stroke. These strengths enhance internal validity of the results. There are, however, some potential limitations to use of these results. Firstly, this treatment schedule was developed as a precursor to a trial involving stroke survivors with the specic characteristics which formed the focus for the semi-structured interviews. Caution is needed, therefore, in generalising the results of this study to stroke survivors with different clinical characteristics. Secondly, a time limit was imposed on the length of a treatment session for pragmatic reasons. If a longer time period had been given, the physiotherapists might have provided different treatments which could have affected the estimates of intra-rater reliability obtained. Thirdly, not all the physiotherapists interviewed attended the subsequent focus groups, and this could have inuenced the content of the treatment schedule. However, the draft of the treatment schedule was piloted in clinical practice with physiotherapists who were not interviewed, but whose input resulted in

subsequent changes before the nal version was produced. It would be advisable, however, to test the generalisability of the treatment schedule beyond the geographical area in which it was developed. Although acceptable intra-rater reliability has been found for this treatment schedule, it has not yet been tested for inter-rater reliability. Investigation of inter-rater reliability will need to consider that different physiotherapists may use different treatment activities for the same clinical presentation [9,11,13,2628]. In addition, it cannot be assumed that subject presentation will remain the same between treatment sessions on different occasions (e.g. natural recovery, treatment effectiveness). Two physiotherapists treating the same patient on two separate occasions might therefore provide different treatment activities. Testing of inter-rater reliability could be confounded when wanting to investigate whether two different physiotherapists would give the same treatment to the same stroke survivor, although published ndings indicate acceptable inter-rater reliability when independent observers decide treatment activities from video tapes of treatment sessions [24]. A way forward, therefore, could be for different physiotherapists to watch the same video tape of a treatment session, but the present treatment schedule has more specic items than those tested earlier [24]. It is possible that physiotherapists could make different interpretations of the same activity (e.g. distinguishing tactile and proprioceptive stimulation during facilitated movement could be difcult). It is also possible that there might be differences between treatment activities recorded immediately after provision of a treatment session and those recorded from watching a video tape. This possibility needs to be tested for both intra- and inter-rater reliability. Clearly further work is required to test reliability. In the meantime, it is recommended that stroke survivors should be treated mainly by one physiotherapist in subsequent clinical trials evaluating this treatment schedule (this recommendation reects routine clinical practice in the UK). It is of interest that not all treatment activities were recorded from watching the video tapes of treatment sessions (nine of 50 treatment activities were not used by raters). It is important to note that the treatment activities not ticked by raters were nevertheless identied as CPT by senior physiotherapists in the interviews. The absence of records of the use of some treatment activities could indicate that these treatment activities are not used routinely in clinical practice at all, that these treatment activities were not used in practice by the raters, that they may be used less frequently, that restrictions on the treatment activities used were imposed by the need to video tape sessions, and/or that the relatively small number of stroke patients included meant that not all treatment activities were appropriate. It is therefore important not to withdraw these treatment activities from the schedule until further investigation of its content has been undertaken. Such investigation also needs to consider the possibility that treatment activities are missing from this treatment sched-

118

C. Donaldson et al. / Physiotherapy 95 (2009) 110119 [2] Kwakkel G, Kollen B, van der Grond J, Prevo AJ. Probability of regaining dexterity in the accid upper limb. Stroke 2003;34:21816. [3] Stroke Unit Triallists Collaboration. Organised inpatient (stroke unit) care for stroke. Cochrane Database of Systematic Reviews 2007; 4, art no: CD000197, doi:10.1002/14651858 CD000197, pub 2. [4] Pomeroy VM, Tallis RC. Neurological rehabilitation: a science struggling to come of age. Physiother Res Int 2002;7:7689. [5] Pomeroy VM, Tallis RC. Need to focus research in stroke rehabilitation. Lancet 2000;355:8367. [6] Marsden J, Greenwood R. Physiotherapy after stroke: dene, divide and conquer. J Neurol Neurosurg Psychiatry 2005;76:4656. [7] Grimmer K, Bialocerkowski A, Kumar S, Milanese S. Implementing evidence in clinical practice: the therapies dilemma. Physiotherapy 2004;90:18994. [8] Bode R, Heinemann A, Semik P, Mallinson T. Patterns of therapy: activities across length of stay and impairment levels: peering inside the black box of inpatient stroke rehabilitation. Arch Phys Med Rehabil 2004;85:19018. [9] DeJong G, Horn SD, Gassaway JA, Slavin MD, Dijkers MP. Toward a taxonomy of rehabilitation interventions: using an inductive approach to examine the black box of rehabilitation. Arch Phys Med Rehabil 2004;85:67886. [10] Hunter SM, Crome P, Sim J, Donaldson C, Pomeroy VM. Development of treatment schedules for research: a structured review to identify methodologies used and a worked example of mobilisation and tactile stimulation for stroke patients. Physiotherapy 2006;92:195207. [11] Pomeroy VM, Niven DS, Barrow S, Faragher EB, Tallis RC. Unpacking the black box of nursing and therapy practice for post-stroke shoulder pain: a necessary precursor to evaluation. Clin Rehabil 2001;15:67 83. [12] Turner PA, Whiteld TA. Physiotherapists reasons for selection of treatment techniques: a cross-national survey. Physiother Theory Pract 1999;15:23546. [13] Tyson SF, Selley AB. The effect of perceived adherence to the Bobath concept on physiotherapists choice of intervention used to treat postural control after stroke. Disabil Rehabil 2007;29:395401. [14] Edwards S, Partridge C, Mee R. Treatment schedules for research, a model for physiotherapy. Physiotherapy 1990;76:6057. [15] Pomeroy VM, Cooke E, Hamilton S, Whittet A, Tallis RC. Development of a schedule of current physiotherapy treatment used to improve movement control and functional use of the lower limb after stroke: a precursor to a clinical trial. Neurorehabil Neural Repair 2005;19: 3509. [16] Chatterton HJ, Pomeroy VM. Positioning for stroke patients: how and why? A survey of physiotherapists aims and practices. Disabil Rehabil 2001;23:41321. [17] Wade DT. Measurement in neurological rehabilitation. Oxford: Oxford University Press; 1992. p. 38. [18] Lyle RC. A performance test for assessment of upper limb function in physical rehabilitation treatment and research. Int J Rehabil Res 1981;4:48392. [19] Hsieh CL, Hsueh IP, Chiang FM, Lin PH. Inter-rater reliability and validity of the Action Research Arm Test in stroke patients. Age Ageing 1998;27:10713. [20] Pomeroy VM, Pramanik A, Sykes L, Richards J, Hill E. Agreement between physiotherapists on quality of movement rated via videotape. Clin Rehabil 2003;17:26472. [21] Slagle J, Weinger B, Dinh M-T, Brumer V, Williams K. Assessment of the intrarater and interrater reliability of an established clinical task analysis methodology. Anesthesiology 2002;96:112939. [22] Fleiss JL. Statistical methods for rates and proportions. 2nd ed. New York: John Wiley & Sons; 1981. [23] Tyson SF, Selley A. The development of the Stroke Physiotherapy Intervention Recording Tool (SPIRIT). Disabil Rehabil 2004;26:11848. [24] De Wit L, Kamsteegt H, Yadav B, Verheyden G, Feys H, De Weerdt W. Dening the content of individual physiotherapy and occupational therapy sessions for stroke patients in an inpatient rehabilitation setting.

ule. Although physiotherapists involved in this study were experienced in stroke rehabilitation and worked in different clinical sites, it is possible that they may not be representative of physiotherapists outside the geographical area in which this study was conducted. Uncertainty about external validity indicates that the present treatment schedule should be used with caution outside the geographical area in which it was developed before a generalisability study has been conducted to investigate: whether the list of treatment activities incorporates all those used by physiotherapists throughout the UK to enhance recovery of the upper limb in the group of stroke survivors for whom this treatment schedule was designed; and whether the treatment activities already identied are described appropriately and adequately. With these caveats, this treatment schedule can be used to document the CPT provided by physiotherapists experienced in stroke rehabilitation in the geographical area in which it was developed. It is, however, important to note that it does not provide guidelines on how to treat the upper limb after stroke and is not prescriptive. Rather, within the context of a clinical trial, it is expected to enable the use of clinical judgement to choose appropriate activities for individuals, and a way to record CPT in sufcient detail for research replication and translation of ndings into clinical practice.

Potential wider implications The methods used to develop the treatment schedule reported in this paper are expected to contribute to minimising recognised aws in rehabilitation research [4,5,29,30]. Specically, valid and reliable treatment schedules could be used in studies to establish exact differences between CPT in different geographical locations, and which form/s of CPT produce best motor and/or functional outcome for stroke survivors. Such knowledge should enhance the evidence base for clinical practice. The methodology is also expected to be transferable to other aspects of rehabilitation, e.g. occupational therapy treatment activities. Ethical approval: Local Ethics Committee. Participants were recruited by written invitation and gave written informed consent. LREG: 03.0190 and 04/Q0803/171. Funding: The Wellcome Trust and the St Georges Charitable Foundation. Conict of interests: None declared.

References
[1] Nakayama H, Jorgensen S, Raaschou HO, Olsen TS. Recovery of upper extremity function in stroke patients: the Copenhagen stroke study. Arch Phys Med Rehabil 1994;75:3948.

C. Donaldson et al. / Physiotherapy 95 (2009) 110119 Development, validation and inter-rater reliability of a scoring list. Clin Rehabil 2007;21:4509. [25] Pereira-Maxwell F. AZ of medical statistics: a companion for critical appraisal. London: Arnold Publishers; 1998, ISBN 0340719400. [26] Nilsson L, Nordholm L. Physical therapy in stroke rehabilitation: bases for Swedish physiotherapists; choice of treatment. Physiother Theory Pract 1992;8:4955. [27] Carr J, Mungovan S, Shepherd RB, Dean CM, Nordholm LA. Physiotherapy in stroke rehabilitation: bases for Australian physio-

119

therapists choice of treatment. Physiother Theory Pract 1994;10:201 9. [28] Sackley C, Lincoln N. Physiotherapy treatment for stroke patients: a survey of current practice. Physiother Theory Pract 1996;12:8796. [29] Whyte J. Clinical trials in rehabilitation: what are the obstacles? Am J Phys Med Rehabil 2003;82:S1621. [30] Dobkin BH. Confounders in rehabilitation trials of task-orientated training: lessons from the designs of the EXCITE and SCILT multicenter trials. Neurorehabil Neural Repair 2007;21:313.

Available online at www.sciencedirect.com