Rapid Responses to:

PAPERS: Gabrielle J Laing, James Law, Abigail Levin, and Stuart Logan Evaluation of a structured test and a parent led method for screening for speech and language problems: prospective population based study BMJ 2002; 325: 1152 [Abstract] [Full text]

Rapid Responses: Submit a response to this article

Rapid Responses published:

Early Identification of Speech and Language Disorders

Melanie J. Merricks, Carol M. Stott, Patrick F. Bolton   (3 December 2002)

Early Identification of Speech and Language Disorders 3 December 2002
Melanie J. Merricks, NHS Fellow and Honorary Consultant in Child and Adolescent Psychiatry Developmental Psychiatry Section, University of Cambridge, 18b Trumpington Road, Cambridge, CB2 2AH, Carol M. Stott, Patrick F. Bolton Send response to journal: Re: Early Identification of Speech and Language Disorders	Editor – in assessing the performance of potential screening methods, Laing et al address an important issue regarding the early identification of children with significant speech and language disorders (1). However we feel that their conclusion that, ‘Screening is likely to be an ineffective approach to the management of speech and language problems…’ is unduly pessimistic given the study reported. This study was undertaken in the context of the child health surveillance programme, ‘a programme of care initiated and provided by professionals, with the aim of preventing illness and promoting good health’(2). The purpose of this programme is therefore screening (and health promotion) rather than surveillance. In this context, any investigation of the accuracy of a screening method needs to take into consideration the performance of the current method, before reaching any conclusions about its utility. In the case of screening for speech and language disorders, there is a paucity of evidence regarding the performance of the child health surveillance programme as a method of detection of those children in need of speech and language therapy services (SALTS). Data collected during the Cambridge Language and Speech Project (3) has allowed us to investigate the relationship between psychometric assessment of language (at 45 months) and SALTS referrals. The findings show that, even using a lenient case definition (1.5 SD below the mean on any language test, without reference to measures of non-verbal IQ) 20% of the non-cases had been seen by SALTS. Furthermore 3% of those seen by SALTS had scored above the mean on all language measures. At the same time, 14% of children meeting the most stringent case criterion (ICD-10 criteria) had not been seen by SALTS (unpublished). These findings demonstrate that current practice is not targeting the group of children who are most likely to benefit from SALTS intervention, with both under-referral of children with, and over-referral of children without, significant language difficulties. A second concern about this study relates to the choice of case definition. Application of the authors criteria for assigning case status led to 21% and 29% of children being defined as having ‘severe language disorder’ or being ‘in need of therapy’ respectively. The authors point out, in their discussion, the possibility that their sample was biased in favour of children with language difficulties. Inspection of the flow chart for the study certainly supports this view. However, it is difficult to accept that these biases can entirely account for this level of prevalence given that a frequency of 2 – 8% would be expected at this age (4). Another explanation for the observed prevalence is that the norms for the reference test do not apply to the population studied, resulting in misclassification of non-cases as cases. If this occured randomly with respect to performance on the screening instrument, the effect would be to decrease sensitivity whilst leaving specificity largely unchanged. This can therefore explain the observed finding in the present study of a low sensitivity and high specificity when the reverse is more usually observed in studies of screening for speech and language disorders (3, 4). It is not difficult to make a case for early identification of speech and language disorders. What is needed, however, is a randomised control trial of the use of a systematic screening instrument compared to current practice, with careful attention paid to choice of reference test on whose basis diagnostic status is assigned. 1. Laing GJ, Law J, Levin A, Logan S. Evaluation of a structured test and a parent led method for screening for speech and language problems. British Medical Journal 2002;325:1152-1156. 2. Hall D, Hill P, Elliman D. The child surveillance handbook. Oxford: Radcliffe Medical Press; 1994. 3. Stott CM, Merricks MJ, Bolton PF, Goodyer IM. Screening for Speech and Language Disorders: the reliability, validity and accuracy of the General Language Screen. International Journal of Language and Communication Disorders 2002;37(2):133-151. 4. Law J, Boyle J, Harris F, Harkness A, Nye C. Screening for speech and language delay: a systematic review of the literature. Health Technology Assessment 1998;2(9):1-184. Competing interests:   None declared