Effectiveness of enhanced communication therapy in the first four months after stroke for aphasia and dysarthria: a randomised controlled trial
BMJ 2012; 345 doi: https://doi.org/10.1136/bmj.e4407 (Published 13 July 2012) Cite this as: BMJ 2012;345:e4407
All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
Dear Editor,
The ACTNoW study is an enormously important addition to aphasia research. We concur with Rudd and Wolfe in their BMJ editorial (2012:345:e4870. Doi:10.1136/bmj) that the Bowen et al ACTNoW study (Assessing Communication Therapy in the North West) should stimulate important debate on the way that speech therapy is delivered in the early recovery phases after stroke. We also reiterate that there is currently insufficient evidence to make large-scale service delivery changes (increase or decrease) to speech therapy services for people with communication disorders in the early recovery phase based on the results of the ACTNoW study.
As Rudd and Wolf state, the recent Cochrane Review1 for speech and language therapy for aphasia following stroke continues to suggest that some form of intervention is better than none. The following letter reflects the views of 21 researchers in Australia from the Centre for Clinical Research Excellence (CCRE) in Aphasia Rehabilitation. In particular, we suggest future directions to assist in determining how best to help and provide services for those with aphasia and dysarthria after stroke.
Is there a selection bias in early intervention? Previous rehabilitation studies consistently exclude people with communication disorders2 and current literature shows that people with either mild or very severe communication disorders (aphasia and or dysarthria and or dyspraxia) are substantially under-represented.2,3 Rudd and Wolfe commented that recruitment for ACTNoW was “slow and selective, with only 21% of patients with suspected acute onset communication problems being recruited and only 44% of patients who were identified as eligible finally giving consent (pg 1).” The low recruitment rate in the ACTNoW trial is likely to have been influenced by the lack of clearly defined inclusion criteria. The pragmatic design of the ACTNoW trial would suggest that all those positively identified with either aphasia and or dysarthria should have been offered an invitation to participate in the trial. Instead up to 690 possible participants were excluded because the treating speech and language therapist did not consider them “likely to benefit from this intervention” thus introducing a serious selection bias.
We suggest there is a need for international agreement around the participant selection criteria for large scale clinical trials in early aphasia intervention. Additionally, an updated epidemiological study to determine the incidence and prevalence of communication disorders following stroke is warranted. The proposed epidemiological study should be in the context of stroke registers that routinely use a communication screening tool with high sensitivity for aphasia and dysarthria. Defining the population of people with communication difficulties as they move through the service delivery continuum will allow for appropriate health services to be targeted at the most beneficial periods in recovery.
Getting the outcome measure right. There has been little consensus about whether the primary aim of speech and language therapy is to improve the language or speech impairment, the disability, or quality of life following onset of aphasia and/or dysarthria. The ACTNoW study used a new method for measuring the primary outcome and opted to measure changes at the Activity level of the ICF. While the outcome measure was shown to be reliable, there has been no demonstration of its validity or sensitivity using the exact methods outlined in the ACTNoW study. There is an urgent need for consensus around the primary aim of aphasia and/or dysarthria rehabilitation. Following that, researchers and clinicians need to agree on an outcome measure with strong psychometric properties which can be used in a research setting while providing a meaningful clinical change to measure recovery. Preferably, this should be a universal outcome measure, used with all intervention studies specific to aphasia and/or dysarthria rehabilitation.
Is there an active ingredient of SLT? As the Rudd and Wolfe editorial suggests, we don’t know which therapy works. We also don’t know which of the specific therapies provided by SLTs may be the most potent active ingredient. This is due to the limited detail describing the exact therapeutic methods used in the majority of published aphasia intervention studies. It is generally considered that impairment based therapies are the specialty of SLTs, although SLTs would also consider they have specialist skills in reducing the overall communication disability. The ACTNoW study provides little insight into the aspects of the intervention that were (or were not) effective, thus providing little in the way of guidance for future intervention studies.
To enhance the quality of communication intervention trials, future studies should provide a highly detailed methodological description of the exact therapy types, techniques and targets used in all intervention. This will provide the much needed transparency required for communication therapy to be easily replicable in future trials.
Defining best practice rehabilitation processes are essential. There are no high quality aphasia specific clinical guidelines which comply with rigorous scientific requirements.3 SLTs in the North East of England developed their own best practice guidelines for the ACTNoW study but these have yet to be published. It seems vital that aphasia and dysarthria rehabilitation have published clinical guidelines which have achieved requirements of consensus statements such as AGREE II5 and GLIA.6
Does conversation practice work? Participants in the Attention Control (AC) arm of the study primarily received conversation practice and some SLT (for those who violated the AC protocol). The AC intervention was provided at a similar dosage (a mean total of 18 hours) of contact as the intervention group (SLT). It is therefore not surprising that the AC group made gains equivalent to those in the intervention arm of the trial, given that neuroplasticity principles suggest that people with aphasia would benefit from talking more in everyday life. There is a need to examine whether a communicatively enriched environment (as provided by the visitors) enhances communication recovery across the ICF domains.
Dosage matters. The ACTNoW study provided an average of 1.3 hours per week (a mean of 18 hours total (range: 4-18) over 16 weeks) of speech and language therapy to those in the intervention group. This is less than the amount of therapy per week noted in research that has reported positive results7-9 for aphasia therapy and is less than 5 hours per week which is recommended in some stroke guidelines.10 Research is needed to examine the most effective and appropriate dose of aphasia and dysarthria intervention at each phase (very early, subacute and chronic) of recovery. Godecke et al7 and Bakheit et al9 have shown positive effects of SLT in very early and early post stroke recovery when therapy was provided for 2-3 hours per week. Bakheit et al9 also demonstrated that a higher dose (>5 hours per week) of aphasia intervention in the subacute phase was less well tolerated and provided no additional benefit than therapy given at the lesser dose. The dearth of research relating to dysarthria trials indicates this is an area in dire need of attention.
The most cost effective methods (computer based therapy, group therapy, non-SLT delivered communication practice) of delivering the optimum dosage of the right therapy to the right people, at the right time, remains to be determined. It is unlikely that one study will provide the answers to these questions. Rather we support an integrated, international collaborative approach to answering these important research questions.
As is often the case, this research leads to more questions than it does answers. The CCRE in Aphasia Rehabilitation has developed an Australian research and clinical network and increased the capacity to conduct some of the research suggested above; however an international effort is required to give people with post-stroke communication disorders the services they deserve. Additionally, health service providers need to be convinced that services are effective and delivered in a cost effective way. The ACTNoW study is just the catalyst for such international collaboration.
Erin Godecke (e.godecke@ecu.edu.au) and Linda Worrall (l.worrall@uq.edu.au)
on behalf of the CCRE in Aphasia Rehabilitation; Australia.
References
1. Brady MC, Kelly H, Godwin J, Enderby P. Speech and Language therapy for aphasia following stroke. Cochrane Database of Systematic Reviews 2012, Issue 5. Art. No.: CD000425. DOI: 10.1002/14651858.CD000425.pub.3.
2. Townsend E, Brady M, McLaughlan M. Exclusion and inclusion criteria for people with aphasia in studies of depression after stroke: a systematic review and future recommendations. Neuroepidemiology 2007; 29: 1-17.
3. Dalemans R, Wade DT, van den Heuvel WJA, de Witte LP. Facilitating the participation of people with aphasia in research: a description of strategies. Clinical Rehabilitation 2009; 23:948-959. DOI: 10.117/0269215509337197
4. Rohde A, Worrall L, Le Dorze G. (Submitted) Systematic review of the quality of clinical guidelines for aphasia in stroke management. Journal of Evaluation in Clinical Practice.
5. Brouwers MC, Kho ME, Browman GP, Burgers JS et al. AGREE II: advancing guideline development, reporting and evaluation in health care. CMJA 2010. DOI 10.1503/cmaj.090440
6. Shiffman R, Dixon J, Brandt C, Essaihi A, Hsiao A, Michel G, O’Connell. The guideline implementability appraisal (GLIA): development of an instrument to identify obstacles to guideline implementation. BMC Medical Informatics and decision making. 2005; 5:23. DOI: 10.1186/1472-6947-5-23.
7. Bhogal SK, Teasell R, Foley NC, Speechley MR. Rehabilitation of aphasia: more is better. Topics in Stroke Rehabilitation. 2003;10(2):66-76.
8. Godecke E, Hird K, Lalor EE, Rai T, Phillips MR. Very early aphasia therapy: A pilot randomized controlled efficacy trial. International Journal of Stroke 2011; DOI: 10.1111/j.1747-4949.2011.00631.x
9. Bakheit AMO, Shaw S, Barrett L, Wood J, Carrington S, Griffiths S, et al. A prospective randomised, parallel group, controlled study of the effect of intensity of speech and language therapy on early intervention from poststroke aphasia. Clinical Rehabilitation 2007;21,885-894.
10. National Stroke Foundation. Clinical Guidelines for Stroke Management 2010. Melbourne Australia.
Competing interests: No competing interests
http://www.bmj.com/content/345/bmj.e4407/rr/602841
Borthwick’s main concern appears to be our study’s recruitment rate. It would be naive to expect any research to achieve a 100% recruitment rate (or “reflect service demands in the City hospitals” where Borthwick works) for two reasons: eligibility and consent rates. The latter reflects the important ethical issue that research participation is a voluntary altruistic act. Randomised trials are particularly strong at making a fair comparison within recruited participants (‘internal validity’). For ‘external validity’, or generalisability, the relevant question to ask is: are there any differences between the 44% who consented and those who were eligible but declined? The answer, as we reported, was that there were not on any of our measured variables. Of course you cannot (and should not) measure everything, especially in the first few days post stroke when some of Borthwick’s suggestions (e.g. level of acceptance, personal goals) are inappropriate.
Regarding the eligibility rate, Borthwick has misunderstood our data. Her 8% is a red herring as it is based on the 2074 people seen for screening as potential research participants rather than the correct proportion who failed the NHS speech and language therapists’ communication screen and were deemed eligible for intervention. The main reason for exclusion (690/2074) was that the therapist found no evidence of a communication problem (e.g. resolved, too mild). Borthwick’s suggestion that the exclusion of these non-aphasic/non-dysarthric people weakens the generalisability of our findings is absurd, as is the implication that inclusion of others (e.g. the 340 deceased, too severely ill or too cognitively impaired for rehabilitation) would have improved the study.
A real strength of this study is that NHS speech and language therapists (and not researchers) confirmed eligibility and excluded people (quite appropriately). This was because these highly experienced clinicians were familiar with the traumatic impact of stroke and decision making about the limitations of intervention. Even if they were overly cautious in their exclusions (understandable given the unpopularity of RCTs in aphasia research) the logical conclusion is that since the intervention had no added benefit with the included sample it is even less likely to have proven worthwhile with those deemed unlikely to benefit by the recruiting therapists.
On the matter of outcome measures we reiterate:
1. The lack of difference between the two study arms on the primary outcome measure (modified TOM activity subscale) was echoed on every secondary outcome measure from every stakeholders’ perspective (including those she suggests e.g. writing skills, naturalness of speech) as well as in our qualitative study of service users’ views1,2.
2. Our previous paper on the TOM did show good inter-rater agreement3 (77% of ratings were within a half point) and 93% were within a half point for intra-rater agreement. In the current trial the TOM does appear sensitive to change as we observed a 0.8 difference for the overall sample (therapy and control group) from baseline to six months. The point is that the therapy did not result in a half point (or even a 0.3) difference when compared with the control.
1. Young A et al. Trial participants’ experiences...: a qualitative study nested within a RCT. Clinical Rehabilitation (Online First 26/7/12).
2. Bowen A et al. Clinical effectiveness, cost effectiveness and service users’ perceptions...(the ACT NoW Study). Health Technol Assess 2012;16(26):1-160.
3. Hesketh A et al. The reliability of rating conversation... Aphasiology, 2008, 22 (9) 970-984.
Competing interests: None beyond those in our original paper.
Apologies to Chris Code for not replying sooner, we have since covered some of his queries in replies to others.
1. Spontaneous recovery – The study was designed with two predictions in mind (a) the likely spontaneous recovery during the first few months of stroke and (b) the likely attention (placebo+) effect of receiving regular contact with a therapist or visitor. Randomising participants means that spontaneous recovery is likely to occur in both the therapy and the control group. Therefore we tested for between group differences at six months i.e. whether there was any added benefit of seeing a qualified speech and language therapist over and above spontaneous recovery and any general psychosocial support from regular contact. So spontaneous recovery is not a difficulty to be avoided in future trials and we encourage further research in the early phase of the stroke pathway.
2. We entirely agree with Code’s suggestion that our employed visitors provided social interaction much like a good aphasia therapist would. However this was without years of degree level training to qualify as a therapist (and on a much lower salary) so it is reasonable to expect greater benefit from seeing a qualified therapist who offers not just interaction but theoretically driven therapy too. We also agree that we selected part-time visitors with excellent (not just averagely good) social skills but feel that this is misunderstood. Our visitors were individuals recruited through job adverts in the local newspaper for a rather low paid (~£15,000 pro rata) and short term salary. It was not difficult to recruit such excellent people and this could easily be replicated by NHS services. However they were deliberately not trained in stroke and were not trained ‘conversation partners’ as this term has come to be understood in aphasia. Their ‘training’ and monitoring related more to research governance and might be far less in clinical practice.
3. It is not accurate to describe the intervention designed and delivered by the speech and language therapists as “generalized”. Instead therapy was individualised within an agreed approach.
4. Aphasia is not more complex than other impairments and plenty of good stroke rehabilitation trials have demonstrated effectiveness where that is present. Advances in service delivery could potentially be made if theoretical expertise in aphasia were coupled with methodological developments in the evaluation of complex interventions.
Competing interests: Nothing additional to those stated in our original paper.
There are several interesting points to debate in Nouwens’ response. However we did not claim, as Nouwens writes, that “speech and language therapy might be discarded in the early stage of recovery after stroke”. We wrote “There is no suggestion that speech and language therapists should be removed from the early stage of the stroke pathway”. Instead our conclusions were about specific therapy approaches and when to use them. We called for an evaluation of service reorganisation to increase functional communication practice in the early stroke pathway through greater use of supervised assistants under the direction of speech and language therapists.
1. We agree that the current evidence base for intensive therapy (for aphasia or dysarthria) is unclear, especially in the first few months of stroke. An Australian pilot study[1] suggested some promise for intensive early (2.5 hours per week for 3 weeks) aphasia therapy for the subgroup who could tolerate interacting for up to 30 minutes.
However a larger Swedish trial[2] reported no evidence for the effectiveness of an intensive aphasia programme of impairment focused therapy “very early” after a stroke (21 days of 45 minutes per day). A previous UK aphasia study[3] found no evidence of difference for five versus two hours per week of speech and language therapy for 12 weeks (and lower user acceptability for the higher intensity). The Cochrane review of aphasia warns of the higher dropout from intensive interventions across the stroke pathway[4]. It seems there is no clear evidence for the oft touted ‘two hours per week’ whilst there are warning bells about poor acceptability amongst service users early in the stroke pathway.
Nouwens and others have queried whether our trial was not intensive enough to show an effect. We must emphasise that our study was not a trial of intensity but a comparison to an attention control and the studies above2,3 address the intensity question. We enhanced services (in terms of quality and not just quantity) and secured resources to offer up to three times a week for up to 16 weeks where therapists and users felt this was appropriate. These were upper limits not targets. We found that on average people accepted less than we offered, on average 22 contacts over 13 weeks. This provides useful data on acceptability to service users, their autonomous right and preferences - important components of clinical effectiveness. Our nested qualitative study[5] showed that users valued flexibility and a patient-centred approach to the frequency of contact. More contact equated with more benefit but only if they retained the possibility of indicating when contact/therapy was too much or too intense at a given time. There is no point offering something that service users do not want or do not feel is worth their time.
As to whether the amount of therapy offered and used was sufficient to test whether therapy added benefit to attention control, in practice the control group received a similar amount of contact (from an employed visitor) but no/negligible communication therapy or contact with an NHS therapist. We argue that 22 versus 0 contacts with a therapist is a considerable NHS resource use, should be expected to show a difference and therefore we do not feel any justification to repeat the study with a higher ‘dose’. Instead our hypothesis is that it is the quality of the everyday communicative interaction with a therapist/visitor/assistant, and not the impairment-based therapy approach, that may be the active ingredient in early communication intervention and warrants further exploration.
2. It is potentially misleading to say we ‘combined aphasia and dysarthria to one population’. There were certainly common features to the overall therapeutic model such as the requirement for further assessment, the involvement of carers and multidisciplinary team working. However the initial screening (pre randomisation) by the speech and language therapists clearly distinguished between these two impairments and their impact, and further assessment and treatment pathways were specific to the communication problem rather than a one size fits all approach. Furthermore separate planned subgroup analyses are provided for aphasia and dysarthria. The bottom line for service delivery is that these problems frequently co-occur and are not likely to be treated by different therapists.
3. Nouwens makes an interesting methodological point, that we measured outcome once at six months and therefore do not have data on whether the likely recovery rate was faster in one group compared with the other. Actually, we did not choose six months because of predictions about spontaneous recovery. The randomisation process itself dealt with that. Our trial was designed to identify between group differences beyond the end of therapy (maintenance of benefit) over and above the predicted occurrence of spontaneous recovery. Time to achieve recovery is an interesting outcome measure (economically and potentially from service users’ perspectives) however if this is so short-lived that a possible difference no longer exists at six months it is of questionable value.
4. We agree with Nouwens, people with aphasia and dysarthria need more trials to improve service delivery as there are still many uncertainties, and services are vulnerable in the absence of evidence. We await the results of Nouwens latest trial (RATS-3) to see how they compare to Laska’s2.
5. We will respond to the repeated query about adequate contrast between therapist and visitor directly in a response to Chris Code.
1. Godecke E, Hird K, Lalor E, Rai T, Phillips MR. Very early poststroke aphasia therapy: a pilot randomized controlled efficacy trial. International Journal of Stroke 2011;1-10.
2. Laska AC, Kahan T, Hellblom A, Murray V, von Arbin M. A randomized controlled trial on very early speech and language therapy in acute stroke patients with aphasia. Cerebrovascular Diseases Extra, 2011;1:66-74.
3. Bakheit A, Shaw S, Barrett L, Wood J, Carrington S, Griffiths S, Searle K, Koutsi F. A prospective randomised parallel group controlled study of the effect of intensity of speech and language therapy on early recovery from poststroke aphasia. Clinical Rehabilitation, 2007; 21:885-94.
4. Brady MC, Kelly H, Godwin J, Enderby P. Speech and language therapy for aphasia following stroke. Cochrane Database of Systematic Reviews 2012, Issue 5. Art. No.: CD000425.
5. Young A, Gomersall T, Bowen A, et al. Trial participants’ experiences of early, enhanced speech and language therapy after stroke compared with employed visitor support: a qualitative study nested within a RCT. Clinical Rehabilitation (Online First 26/7/12). DOI: 10.1177/0269215512450042
Competing interests: Nothing additional to those stated in our original paper.
http://www.bmj.com/content/345/bmj.e4407/rr/603217
It is most helpful to hear the views of an NHS therapist supporting the use of assistants. Your other comments on diagnosis and psychological support very much reflect those we expressed in the paper. We hope you will find this paper worth reading in detail and may we suggest you quote it when developing future services for people with stroke.
As we wrote “There is no suggestion that speech and language therapists should be removed from the early stage of the stroke pathway. The approaches requiring replacement are one to one, impairment focused therapy and detailed cognitive neuropsychological assessment in the first few months after stroke...Extreme caution is required to avoid over-extrapolation of the evidence leading to damage to vulnerable services for people with high unmet need but potential to benefit. People with communication problems after stroke will continue to need support. Replacement approaches are required, and some services will welcome the justification to develop their activity level and social participatory approaches.”
Competing interests: Nothing additional to declarations in original paper.
As a practicing speech and language therapist I can offer no real academic review of this paper and I have only read the paper briefly. However I felt strongly that I needed to give my view. I am particularly concerned about the comment made by Jim Thornton that the " NHS stops funding this ineffective treatment". If Speech and Language Therapists where not funded at all to work with stroke survivors where would this leave us ? Do we not have a role in the assessment and diagnosis of communication disorders follow stroke ? Do we not need to work with other health professionals so they are aware of patient’s communication problems and adapt their therapy as required. Is there not a role for speech and language therapists to offer psychological support to those patients who have severe aphasia ?
I believe that it does not need to be a speech and language therapist delivering the therapy and significant cost savings could be made if therapy assistants delivered the intensive therapy that is required. But to suggest that the NHS stops funding speech and language therapy completely ? Surely this statement needs to be reconsidered?
Competing interests: No competing interests
Whilst this research may be described as a robust RCT, I want to set out the contextual concerns that are immediately obvious to any speech and language therapist working in acute stroke care.
The writers claim that concerns about the feasibility of this study proved unfounded, yet they do not explore the issue of slow recruitment. Recruitment was not just slow but appears completely bizarre. Examining the recruitment totals for each centre on the trial website reveals that Central Manchester only recruited 16 people throughout the 3 years and 4 month, i.e. less than 5 patients per year. If we apply the stated consent rate (44%), this suggests that there are approximately 12 patients per year candidates for communication therapy. I can confidently state that this does not reflect service demands in the City hospitals where I work. The researchers claim that they scrutinised the recruitment process. However, when the number of patients recruited equals the number excluded, for no known reason, confidence in this claim is severely diminished.
Considering the study design, it would be reasonable to expect poor recruitment. Selecting an option that raises a 50% chance of being denied a standard treatment for 6 months can hardly be considered an incentive. The researchers claim that patients were offered the possibility of enhanced therapy; however, they were also warned against signing up if they were not prepared to take the risk of receiving no therapy. Approximately half of the study group received fewer than 20 sessions during the 6 month period within the trial protocol. This would equate to less than 1 therapy session per week and therefore cannot be described as enhanced and is well below recommendations based on available evidence (1). It is perhaps too easy to assume that the subjects were those less motivated or concerned about their communication, but the relatively low uptake when up to 48 sessions were available fully supports this assumption.
They claim that the external validity of participants was good and unsurprisingly there was no difference on characteristics such as gender, age, impairment level, which are easily quantifiable. However, they made no attempt to measure motivation, level of acceptance, personal goals, readiness for therapy or family and self coping strategies. They state that 60% of those who declined entry into the trial gave their reason for refusal as a strong preference to see a SLT. In addition to this, 25% of the control group did appear to regret their decisions and accessed SLT out with the trial.
My concern regarding the validity of the sample goes beyond the low acceptance rate for patients. The study excluded those with severe cognitive issues, global communication problems or other health issues but there is nothing to define these criteria. Many patients in the first 2 weeks post stroke could be described within these categories. Any clinician familiar with the distress, confusion and bewilderment demonstrated by the newly aphasia patient will understand the ethical challenge to recruit to such a study in the first 12 days post stroke (2). The study only recruited 8% of the patients referred for communication screening and elimination of more typical patients makes it impossible to generalise the results.
The choice of the activity scale, which is only one of four sections within the Therapy Outcome Measure (TOM) (3) as the primary outcome measure also causes concern. Although the extremes are easy to assign, the mid point scoring is challenging. Relevant goals such as improving writing skills, improving the naturalness of speech quality would all be impossible to score within this outcome measure. In a paper including the same authors, published in 2008, they concluded, “further research is needed to explore whether a half-point difference for the TOM is clinically meaningful i.e. whether the degree of unreliability allows the measure to be sensitive to change” after finding the inter rater reliability was only 40%. (4)
If we are prepared to be convinced by the confident claims made by this research group we risk committing a severe disservice to the overwhelming number of patients who refused to take part and the many patients who value and campaign for SLT services. Communication therapy is not merely administered to an individual. Passive acceptance and participation can of course happen but all therapists will confirm that it is when patients are actively engaged, seeking solutions and in the driving seat that real measurable improvements are likely to happen. For patients after stroke the timing of when therapy feels right will vary. Some immediately, some after 1 month, some after 6 months but this study could never tell us when communication therapy is most effective for each sub-group. The study suggests we re-arrange SLT services due to their results. Well, I would argue that services are already very aware that there is no one-size fits all approach and we continue to try to provide a flexible, patient focused service that promotes self management, patient-centred goal setting and offers a range of intervention approaches.
The authors’ acknowledgement that this study does not suggest that there is no value from speech and language therapy services in the acute phase after stroke is very welcome. However with such a clearly flawed recruitment the authors should be sharing lessons learned from attempting such a complex trial rather than defending and constructing unjustifiable conclusions that risk damaging service provision to this patient group.
1. Scottish Intercollegiate Guidelines Network (SIGN) (2010). Management of patients with stroke: Rehabilitation, prevention and management of complication and discharge planning. SIGN 118
2. Holland A., Fridriksson J. (2001) Aphasia management during the early phases of recovery following stroke. Jrnl of speech-language pathology 10.
3. Enderby P., John A., & Petheram B. Therapy outcome measures for rehabilitation professionals 2nd edition. Wiley. (2006)
4. Hesketh A., Long A., Patchick E., Lee J & Bowen A. (2008): The reliability of rating conversation as a measure of functional communication following stroke. Aphasiology, 22:9
Competing interests: No competing interests
Reply to Rapid Responses from Enderby, Meteyard and Thornton:
It is encouraging to see the Royal College of Speech and Language Therapists (Enderby) supporting RCTs. Meteyard worries that RCTs will not cope with the complexity inherent post-stroke. However, many RCTs have demonstrated the effectiveness of a range of complex interventions for heterogeneous populations (e.g. Stroke Unit care, Occupational Therapy).
As Enderby notes the Cochrane review [1] finds benefit of therapy compared with nothing. However, like us it also finds no benefit over attention control. So “some is better than none” [2] but we must be open-minded about what is done and by whom. Despite Meteyard’s concerns we can rule out those activities only provided to the intervention group (e.g. one-to-one impairment based therapy). In the first four months of stroke they added nothing to the outcome for participants from any measured perspective.
Meteyard is incorrect to say treatment was unconstrained and that we examined variation in current practice. Each site altered their previous practice by adopting manualised assessment and treatment pathways, tools and techniques as agreed by consensus. As Enderby recommends, our therapists targeted therapy to those most likely to benefit and selected appropriately tailored interventions.
We are grateful to Enderby for quoting our cautionary warnings regarding misinterpretation of the findings, especially given Thornton’s reaction. Our nested qualitative study [3] showed people with stroke valued increased early support (regardless of whether therapy or control). Interaction with a good communicator, as much as formal therapy, may be beneficial. We recommend evaluating reorganised early services that retain therapists to supervise increased time with less qualified staff, stepping up later for therapists’ direct involvement for persisting problems.
In response to Thornton, the £1.5m NIHR HTA funding supported a series of studies recruiting over 700 participants, including developing patient-centred outcome measures [4,5] that have had good international uptake.
1. Brady MC, Kelly H, Godwin J, Enderby P. Speech and language therapy for aphasia following stroke. Cochrane Database Syst Rev 2012;5:CD000425.
2. Rudd A and Wolfe C. Is early speech and language therapy after stroke a waste? Perhaps, but some intervention to promote communication is better than none. BMJ 2012;345:e4870
3. Young A, Gomersall T, Bowen A. Trial participants’ experiences of early, enhanced speech and language therapy after stroke compared with employed visitor support: a qualitative study nested within a RCT. Clin Rehabil (Online First 26/7/12). http://cre.sagepub.com/content/early/2012/07/25/0269215512450042.full
4. Long AF, Hesketh A, Paszek G, Booth M, Bowen A, on behalf of The ACT NoW study team. Development of a reliable, self-report outcome measure for pragmatic trials of communication therapy following stroke: the Communication Outcome After Stroke (COAST) scale. Clin Rehabil 2008:22:1083-1094.
5. Long AF, Hesketh A, Bowen A, on behalf of The ACT NoW study team. Communication outcome after stroke: a new measure of the carer’s perspective. Clin Rehabil 2009:23:846-856.
Competing interests: Declared on original paper BMJ 2012;345:e4407
With great interest we have read the report of the scrupulously designed study by Bowen and colleagues. We laud their complex trial design with which they have studied not only the effectiveness of speech and language therapy for patients with various communication deficits after stroke, but also the cost effectiveness of speech and language therapy and conducted a qualitative study of experiences in patients and their proxies with the provided intervention. We feel however that the authors jump to a conclusion by claiming that speech and language therapy might be discarded in the early stage of recovery after stroke.
The recently updated Cochrane review concludes that language therapy is more effective if it is provided intensively (2). Hów intensive however is still unclear. Bowen et al report that the intervention group on average received 22 contacts (18 hours) over 13 weeks, which comes down to a little over 1,5 hours of therapy per week. Only 53% of this time was spent on direct one-to-one contacts (including assessments) and half of this limited time consisted of impairment based therapy. Several studies suggest a minimally required treatment intensity of two hours per week to add an effect of therapy to spontaneous recovery (3-6). Hence, the impairment-based treatment delivered in the ACT NoW study was probably not sufficiently intensive to generate an additional effect.
Prof. Code already questioned whether the contrast between the two interventions; SLT with a professional versus AC with a trained visitor, was substantial enough. From this point of view it is interesting to note that even with such a small treatment contrast, the 95% confidence interval of the primary effect measure included the clinically worthwhile difference of 0.5 points on the TOM, and cannot be used to rule out a treatment effect.
We entirely agree with dr. Meteyard that it is peculiar to combine aphasia and dysarthria to one population of patients with communication deficits. Prof. Code emphasizes that a population of aphasic patients alone is highly heterogeneous and challenges researchers to design proper studies. Adding another impairment increases heterogeneity even more, hampering correct interpretation of Bowens’ findings.
Bowen et al timed their primary outcome at six months after stroke. Agreed, this is an important time, because spontaneous recovery should have occurred already and patients now enter the chronic stage. However, in the absence of earlier follow-up measurements, we are not informed about recovery processes in both groups during the first six months. We wonder if recovery is equal in both groups or if one group reached better communication skills earlier, which is not a negligible benefit of treatment.
In our opinion there is a need for further trials of SLT with a strong intervention contrast and proper evaluation. We are conducting a multi centre RCT in the Netherlands on early intensive cognitive-linguistic language treatment (CLT) in aphasia after stroke, called “Rotterdam Aphasia Therapy Study (RATS) – 3; The efficacy of cognitive-linguistic therapy in the acute stage of aphasia: a randomised controlled trial”. We will randomise 150 patients with first ever aphasia after stroke within two weeks post stroke for either the therapy group (four weeks of intensive, one hour a day CLT) or the control group (in which regular language therapy is deferred for four weeks). Both groups continue with regular, less intensive, language therapy after these four weeks.
We chose three evaluation moments; directly after four weeks of either intensive therapy or no therapy and three and six months after inclusion. This design allows us to evaluate if there is a direct treatment effect of early intensive CLT (results after four weeks), and whether this supposed effect lasts beyond the acute phase (results after three and six months). The trial design is summarised in the Dutch Trial Register (www.trialregister.nl, NTR3271) and will be published later this year.
In conclusion, we agree with Prof. Enderby, Dr. Meteyard and Prof. Code that it is not possible to conclude from the ACT NoW study that speech and language therapy is equally effective as no therapy. In our opinion, further research is needed to establish the optimal timing and treatment intensity of speech and language therapy for patients with aphasia due to stroke.
On behalf of the Department of Neurology, Erasmus Medical Centre:
F. Nouwens MA, speech and language therapist and clinical linguist
Dr. L.M.L. de Lau, neurologist
Dr. E.G. Visch-Brink, clinical linguist
Prof. Dr. D.W.J. Dippel, neurologist
Correspondence to:
F. Nouwens, MA
Erasmus Medical Centre, Department of Neurology
Office Ee 2291
P.O. Box 2040
3000 CA Rotterdam
The Netherlands
f.nouwens@erasmusmc.nl
References
1. Bowen A, Hesketh A, Patchick E, Young A, Davies L, Vail A, et al. Effectiveness of enhanced communication therapy in the first four months after stroke for aphasia and dysarthria: a randomised controlled trial. BMJ. 2012;345:e4407.
2. Brady MC, Kelly H, Godwin J, Enderby P. Speech and language therapy for aphasia following stroke. Cochrane Database Syst Rev. 2012;5:CD000425.
3. Robey RR. A meta-analysis of clinical outcomes in the treatment of aphasia. J Speech Lang Hear Res. 1998 Feb;41(1):172-87.
4. Bhogal SK, Teasell R, Speechley M. Intensity of aphasia therapy, impact on recovery. Stroke. 2003 Apr;34(4):987-93.
5. Salter K, Teasell R, Bhogall S, al e. Evidence-based review of stroke rehabilitation: Aphasia: V.11 2007: Available from: http://www.ebrsr.com.
6. Bakheit AM, Shaw S, Barrett L, Wood J, Carrington S, Griffiths S, et al. A prospective, randomized, parallel group, controlled study of the effect of intensity of speech and language therapy on early recovery from poststroke aphasia. Clin Rehabil. 2007 Oct;21(10):885-94.
Competing interests: No competing interests
Re: Effectiveness of enhanced communication therapy in the first four months after stroke for aphasia and dysarthria: a randomised controlled trial
Authors’ response to Godecke, Worrall and the Australian CCRE http://www.bmj.com/content/345/bmj.e4407?tab=responses
We warmly welcome the comments from our Australian colleagues at the Centre for Clinical Research Excellence (CCRE) in Aphasia Research.
Their suggestion of an international collaboration for rehabilitation research into aphasia and dysarthria is one we wholly endorse. Readers may be interested to note that Marian Brady from the UK recently led a bid to the EU for infrastructure and networking funding for such a venture ‘Collaboration of Aphasia Trialists (CATs)’, which extended beyond EU countries to include Australia.
We support the CCRE researchers’ suggestions for prioritisation of research into: epidemiology, universal outcome measures, group therapy, computer based therapy, and evaluating the provision of a communicatively enriched environment.
Godecke and Worrall’s response seems to have crossed in the mail with our replies to previous authors correcting misunderstandings about recruitment, dosage and outcome measurement in ACT NoW. We respectfully ask all interested parties to read our replies e.g. 690 people were not excluded because they were not likely to benefit but because they only ‘potentially’ had a communication problem on admission and this was no longer present (false positive/recovered/too mild to require intervention) at screening.
Competing interests: None beyond those in our original paper