Rapid Responses to:
|
|
Rapid Responses: Rapid Responses published:
-
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif)
Giving too little equal throwing away everything
- Antonio Lopez Pena (16 October 2000)
-
Response to 'Giving too little'
- Sue Roulstone, Margaret Glogowska (18 October 2000)
-
Methodological issues in SaLT research
- Leena Bhagwan (19 October 2000)
-
Speech Therapy for preschool children.
- Anis Mustafa (29 November 2000)
-
concerns about SLT provision
- Sarah Hulme (15 December 2000)
-
Re: Methodological issues in SaLT research
- Sue Roulstone (23 December 2000)
-
Re: Speech Therapy for preschool children.
- Sue Roulstone (23 December 2000)
-
Re: concerns about SLT provision
- Sue Roulstone (23 December 2000)
|
|
|||
|
Antonio Lopez Pena Santo Domingo
Send response to journal:
|
Giving 47 minutes service to these children is perhaps cheap...but not efficient!(so, not cheap at all). As a comparison, therapists in European School in Brussels were providing (long time ago) a mean of two-and-a-half 45 minute individual sessions ...per week! And, even if this school doesn't have accurate statistics (as far as I know), believe me, there are results and outcomes in it. And I can assure that therapy will not be discontinued until results are there. Expensive? Perhaps. But how expensive is a "cheap" system with no outcomes? Let's compare services...and not only approaches. |
|||
|
|
|||
|
Sue Roulstone, Research Speech & Language Therapists Speech & LAnguage Therapy Research Unit, Frenchay Hospital, Margaret Glogowska
Send response to journal:
|
We agree that giving too little therapy may mean that you achieve no outcomes - we love your title! But as we argued in the paper, this was essentially an evaluation of existing services and therefore therapy was unspecified in advance. (ie, the amount of therapy provided was one of the process outcomes) We believe that a next step would be to investigate the effects of different amounts of therapy to discover optimum amounts rather than assuming that an all round increase is the answer. Furthermore, the nature and timing of therapy for young children are also worthy of attention. |
|||
|
|
|||
|
Leena Bhagwan, 4th year medical student Department of epidemiology ang Public Health Medicine, University of Newcastle upon Tyne
Send response to journal:
|
Editor We are writing in response to Glogowska et al’s paper about speech and language therapy in pre-school children.(1) While we agree that the objectives of the study are worthwhile, we feel there is insufficient support for their conclusions. Firstly, the authors calculate that a total of 146-166 children are needed to detect a clinically significant 20% difference between the two trial arms. However the sample size falls below this figure due to missing data. For example, the outcome measure phonology error rate was based on 119 children. The only outcome measurement that was considered to be significantly different was auditory comprehension. Even this was based on less than the required sample size. Secondly, we agree with Law and Conti-Ramsden that a mean of six hours of therapy over an average of eight months is probably not enough to show the value of speech therapy.(2) We appreciate the pragmatic nature of the study reflects the services offered in the community. But how many hours of therapy have been shown to make a difference? The consistency of therapy delivered is another issue. Table 2 of the full text version shows that the number of hours of therapy, number of contacts with therapists and frequency and length of sessions vary considerably.(3) We note that sixteen clinics were involved and expect differences in the quality of services provided. In the presence of unsubstantiated results we feel that the conclusions drawn are not valid. In the future studies should use standardised researched protocol for speech and language therapy and larger groups. Yours sincerely, Leena Bhagwan, Ella Gupta, Kathryn Higgins, Joanna Schmidt, Liza Thoppil Fourth Year Medical Students Department of Epidemiology and Public Health Medical School, University of Newcastle Upon Tyne NE2 4HH Email leena.bhagwan@ncl.ac.uk 1 Glogowska M, Roulstone S, Enderby P, Peters TJ. Randomised controlled trial of community based speech and language therapy for preschool children. BMJ 2000; 321:923-6 2 Law J, Conti-Ramsden G. Treating children with speech and language impairments. BMJ 2000;321:908-9 3 Glogowska M, Roulstone S, Enderby P, Peters TJ. Randomised controlled trial of community based speech and language therapy for preschool children. BMJ 2000; 321:923-6 Full text version available at http:\\www.bmj.com/cgi/content/full/321/7266/923 |
|||
|
|
|||
|
Anis Mustafa
Send response to journal:
|
Dear Sir - With reference to your paper, Randomised Controlled Trial With Community Based Speech and Language Therapy in Pre-School Children by Margaret Glogowska et al1 and the editorial I was not surprised with the effectiveness of 6 hours of speech therapy over a period of 12 months with pre-school children who were delayed with speech and language. Due to the severe shortage of speech therapists in some of the places where I have worked the waiting time for assessment by a speech therapist is more than 6-8 months and they are so overloaded with work that in the majority of cases they will only assess the child and the actual therapy sessions will be handed over to be provided by a special needs teacher at nursery or school, or sometimes instructions are given to the parents to carry out speech therapy at home. In one case I remember the speech therapist had sent instructions to the parents by post. In my experience working in several child development centres as consultant community paediatrician I have found only a few children received consistent regular speech therapy until they improved. So finding the patchy inconsistent service my individual practice is to involve the parents as much as possible with speech therapy which I practice as follows: 1. I warn the parents about the state of the service, what they can expect and how soon the speech therapist will assess. 2. I ask them to get involved and obtain instruction and training as much as possible from the speech therapist. I also ask them to obtain information from other sources and try to provide speech therapy to their children with the guidance of a speech therapist in their free time at home as much as possible. I have found that most parents are more than willing to cooperate and do as much as they can themselves. This practice is based for obvious reasons, which are: 1. The parents are readily available, free of charge, day and night,
365 days a year.
It has been proven with evidence from several studies that parents trained with speech therapy could provide effective and beneficial speech therapy. I refer here to two papers published, the first by Gibbard, D.2 , 1994, where they did a randomised controlled trial. In the experimental group the parents took part in speech therapy who had parental language training sessions and the control group without the parents involvement only the speech therapists provided the speech therapy. The results showed significant greater gain in the expressive language skills to the experimental group than the controlled group. The same author did a second experiment and in this study three groups were selected for speech therapy. The experimental group was formed of parents who had attended parent language training sessions and the controlled group was formed of parents who had training sessions with emphasis on general learning skills rather than language and speech therapy. A third group of children received individual direct speech and language therapy from a speech therapist. The results showed significantly greater language gain in the parental language training group. I refer to a second study by Pamplona, M.C. et al3 where children with cleft palates had the involvement of the parents with their speech therapy. The first group received speech therapy by only a speech therapist. The second group received speech therapy accompanied by their mothers who were allowed to provide linguistic interaction with the children facilitating their communication. The result was significant higher linguistic advance with the mothers' involvement. At this stage I think we should utilise the involvement of the parents and giving them even basic training and instruction enables them to give much more to the children than what can be given from the speech therapists. References 1 Margaret Glogowska - Randomised controlled trial of community based speech and language therapy in pre-school children. BMJ 2000, 321, page 923-926 (14th October). 2 Gibbard, D - Parental based interaction with pre-school language delayed children. Eur. J. Disorder Commun. 1994; 29(2); 131-50. 3Pamplona M.C. Ysuza Uriostegues C. International J of Paediatrics otorhinolaryngeal 1996 Sept, 37(1); 17-27. Dr Anis Mustafa
|
|||
|
|
|||
|
Sarah Hulme, Principal Speech & Language Therapist for Early Years Services Camden & Islington Community Health Services NHS Trust
Send response to journal:
|
Glogowsak's study of the effectiveness of community clinic based speech and language therapy raises questions about the provision of speech and language therapy around the country. The study has attempted to compare the effectiveness of speech and language therapy with 'watchful waiting', using random control methodology. Whilst we agree wholeheartedly that research is important and necessary for the profession, and that the therapy we provide needs to be evidence-based, we would question the appropriateness of randomised control trials (RCTs) for an area of healthcare such as speech and language therapy. We are aware that RCTs can be a powerful tool for investigating the efficacy of drug treatments for defined populations. However this study was analysing a heterogeneous population who received individualised therapy. There was no control over the type of therapy provided, and the data indicates that some of the children in the therapy group actually received no therapy at all! (mean 6.2 hours; range 0-15 hours). RCT design does not provide us with qualitative information about which therapies were effective for which disorders. In recent years speech and language therapy services such as our own have used qualitative outcome measures to define which are the most effective types of speech and language therapy for which disorders, and to define the most appropriate age for intervention. Services such as our own offer a continuum of packages of care, prioritising those with severe disorders for intensive therapy. In addition, mindful of the holistic nature of language development, therapy involves each of the child's learning environments - nursery/school as well as home. The advantage of using RCTs is that their rults can be generalised. However, it is clear that the therapy and provision described is not representative of all speech and language therapy services around the country. It is a shame that this potential advantage has been lost. In agreement with the suthors we would wwelcome detailed studies comparing the effectiveness of different types of therapy for specific disorders. In the meantime we hope that service managers will use the already available research to put into practice the most effective continuum of packages of care for children with speech and language disorders. Sharon Millard & Sarah Hulme Gila Falkus, |
|||
|
|
|||
|
Sue Roulstone, Clinical Research Director SLT Research Unit, Frenchay Hospital
Send response to journal:
|
We found this letter interesting and encouraging, to know that medical students are keenly scrutinising research from the wider health care setting such as speech and language therapy. On the question of power and the impact of missing data, it is quite correct that all other things being equal a reduction in sample size reduces the sensitivity of a trial to differences that may exist. However, once a trial has been completed and analysed, the crucial issue (in this respect) is the achieved precision of the results as portrayed in the confidence intervals comparing the trial groups. Specifically, interest should move from the power (planned or achieved) as a concept rooted in a hypothesis-testing paradigm, to the likely clinical importance of the estimates and the confidence limits. This latter issue is fully addressed in the paper, including how the CIs compare with pre-specified thresholds for clinical significance, and the limitations in conclusions that accrue from the width of some of them. More minor points are first that the extent of missing data varied across the outcome measures (and in some cases was trivial), even though the conclusions drawn were relatively consistent. Secondly, missing data did not preclude one outcome reaching statistical significance. We point this out not in order to sidestep the issue but so as to reinforce the point that sample size estimates are based on imperfect information - at times observed standard deviations are smaller than we predict and/or observed differences are relatively large. This then reinforces the importance of considering the (attained) comparative confidence intervals when judging the nature and strength of evidence that a study provides. Finally, this letter does highlight an inconsistency in the conventional approach to designing and reporting trials. That is, most sample size calculations are performed using a power approach and hence focus on hypothesis testing; the current emphasis for presenting results, however, is quite rightly on estimation, confidence intervals and hence clinical as well as statistical significance. There may be little alternative to this inconsistency in a given situation due to limited information, but it's worth noting here that this emphasises the need for a clear statement and justification of the target difference, and also that precision-based sample size calculations should perhaps be performed more often than they are at present. On the question of how many hours therapy are needed to make a difference, the literature is far from conclusive on this issue and there have been few investigations of differential ‘dosage rates’. From the 10 high quality RCTs of therapy identified by a recent systematic review (Law et al, 1998)1, nine of which were found to reveal statistically significant treatment outcomes, there was a median of 9 direct contact hours over period of up to 4 months. Regarding the consistency of therapy delivered, as this was a service evaluation rather than the evaluation of any particular therapy approach, the number of therapists involved was a necessary part of the study and the quality of therapy delivered by any one therapist could indeed be an issue. The numbers of therapists involved for this overall sample level makes the investigation of therapist variation through inferential statistics unviable. This does not diminish the conclusion about the overall impact of the services for those children who received them. 1 Law J, Boyle J, Harris F, Harkness A, Nye C. (1998) Screening for Speech and Language Delay: A Systematic Review of the Literature. Health Technology Assessment, 2 (9), 1-184. |
|||
|
|
|||
|
Sue Roulstone, Clinical Research Director SLT Research Unit, Frenchay Hospital
Send response to journal:
|
We do not disagree with the basic premise underlying and concluding this response: that the involvement of and suitable support and training for parents is a fundamental part of the process of delivering speech and language therapy, particularly for preschool children. The trial did not preclude this and indeed the availability of a parent or carer to attend and be present within all therapy sessions was a criterion for inclusion. However, the results did not directly document the level of involvement of carers. Data was collected on the therapy goals: approximately half of the therapists indicated that they directed therapy towards changing parents strategies although therapists varied in how explicitly these were set out in their aims statements. The study included a qualitative investigation of parents’ views (Glogowska and Campbell, 2000)1. The results support the comments made in this response: that parents do indeed appreciate their role in intervention. But parents also highlighted the need for regular support suggesting that using parents as a means of compensating for limited amounts of therapy is not necessarily possible. Indeed, in the Gibbard study (Gibbard, 1994)2, the amount of input received in the parent groups (one hour a fortnight for six months) is more than that received by some of the children in our study. Finally, the involvement of parents must give due respect to their role as parents (and not as educators) and recognise the cost to a family of high levels of focus on a particular child: to see parents as available ‘free of charge’ as this response suggests, fails to recognise the very real costs to families. 1 Glogowska, M. & Campbell, R. (2000) Investigating parental views of involvement in pre-school speech and language therapy, International Journal of language and Communication Disorders, 35 (3), 391 - 405. 2 Gibbard, D. 1994 Parental based interaction with pre-school language delayed children. European Journal of Disorders of Communication, 29 (2); 131-50 |
|||
|
|
|||
|
Sue Roulstone, Clinical Research Director SLT Unit, Frenchay Hospital
Send response to journal:
|
Within the profession of speech and language therapy, there continues to be huge anxiety and scepticism about the value of randomised controlled trials as a suitable means of evaluating therapy and services. However, they remain one of best means of evaluating the effects of treatment (from drugs to acupuncture and homeopathy) because it is the methodology least prone to bias. As such, we do not believe the profession can afford to abandon its potential as a tool in our research repertoire. Any treatment, as it moves from the theoretical, idealised mode of delivery into a service context undergoes change. Drug treatments are just as susceptible – people forget their tablet one week and take a double dose for several days to catch up or throw them away to avoid detection! Knowing what happens in practice – the effectiveness of treatment – is just as important as knowing the efficacy or idealised effect of a particular therapy. So documenting the process of service delivery was an important part of the study. If one attempts to deliver a service to children under 3;06years, it is important to know what proportion will not attend and what impact that has on the overall effectiveness of the service; if one attempts to deliver a 12 month monitoring service we need to know what proportion of families will adhere to that policy and the impact on the overall results. The study was analysed on an ‘intention to treat’ basis in order to take account of such departures from protocol without introducing bias into the two groups. The children were indeed heterogeneous in terms of their severity, but those with identified learning disabilities, autism, oromotor dysfunction were excluded from the outset. The children’s difficulties can therefore be considered primary speech and language difficulties and typical of the range commonly dealt with in community clinics. As a clinical sample, recruited from successive referrals, this study improves on many studies which select on the basis of advertisement or reflect a narrowly based sample which is not representative of a clinical caseload. Given that our intention was to investigate the impact of therapy within the clinics, and that we have accepted the typical variation of a clinical caseload, then the individualisation of therapy becomes a requirement rather than a hazard – otherwise the therapy regimes would have been deemed inappropriate by their therapists. We accept that the RCT design does not provide qualitative data about the effect of therapy. To this end, a qualitative study was conducted alongside the trial to investigate the acceptability of the two trial conditions to parents. (Glogowska and Campbell, 2000). Interestingly, as well as providing much valuable data about the parents’ views, the emerging story does not conflict with the quantitative results. We were interested to hear of the work being carried out to define the most effective types of speech and language therapy for particular disorders and the outcome measures on which such evaluations were based. The production of ever more sensitive outcome measures is indeed an issue for us as a profession. Sensitive outcome measures are one of the ways in which trial methodologies can be improved. We hope that the authors are considering peer reviewed publication as a means of disseminating their work. Finally, we continue to believe that the children included in this trial are typical of many community clinics around the country. The structure of the service and the amount of therapy delivered have been documented clearly so that other services can judge how far this can be compared with the service they offer. The potential for generalisation to other similar services should therefore not be problematic. Glogowska, M. & Campbell, R. (2000) Investigating parental views of involvement in pre-school speech and language therapy, International Journal of language and Communication Disorders, 35 (3), 391- 405. | |||