PEDIATRICS Vol. 121 No. 6 June 2008, pp. 1300 (doi:10.1542/peds.2008-0976)
LETTER TO THE EDITOR |
Should We Screen for Language Delay in Toddlers?: In Reply
Heleen M.E. van Agt, MAHarry J. de Koning, MD, PhD
Department of Public Health
Erasmus Medical Center
3000 CA Rotterdam, Netherlands
We are pleased to see that vanderPloeg et al now agree that our trial showed consistent positive effects for our primary outcome, special school attendance. Our trial was constructed and powered for exactly this purpose (namely, to investigate whether screening for and early treatment of language disorders reduced the attendance at a special school), because children who suffer from severe language problems will not be able to attend normal school but require education at a special school. As a result of earlier detection and treatment, in the Netherlands, 2500 children (of 200000 8-year-olds) per year would be able to remain at a normal school. In a randomized trial of screening, a positive result at the end provides evidence about the effect of screening, if the trial is correctly randomized (groups are comparable for baseline characteristics), drop-out and outcome measures are uniformly assessed and blindly taken in both groups, and analyses include control for possible confounders, as was the case in our trial. Supporting evidence is the fact that in the study arm more treatments related to language development were reported in the child's preschool period after age 2 when the screening was implemented.
It is well known that it is scientifically incorrect to "cherry-pick" results, as vanderPloeg et al do, and say that 3 of 7 showed significant results. It was from a theoretical point of view that results would be positive for expressive language skills such as spelling and not necessarily for receptive language skills such as reading. The evidence about effective interventions suggests that speech and language therapy is effective for children with expressive difficulties but that there is less evidence for effective interventions for children with receptive difficulties.1 However, almost all effects were in the expected direction. Because not all teachers complied, we were not able to show significance. Repeating a grade was found to be higher in the experimental group, which can be plausibly explained: as a result of the screening, children's school functioning might be improved to such an extent that instead of attending a special school the child had been repeating a grade to be able to stay at a normal school.
As explained thoroughly in our article, 1-sided testing was justified because the screening under study would only be considered for implementation if a significant result in 1 direction was found, namely, an a priori quantified reduction of language problems.2
vanderPloeg et al further assume that if a language delay is not detected at a very young age, parents will be falsely reassured in all instances. However, the complement of sensitivity (1-sensitivity) is not the same as false reassurance. Given the intensive repeated monitoring of children's development from 0 to 4 years within child health care centers, false reassurance is very unlikely.
The purpose of a randomized, controlled trial (RCT) is not to investigate the mechanism of the intervention. Therefore, we have no details of the complete trajectory of assessment, treatment, parental concern, and professional advice in the period after the screening and later. Despite that, vanderPloeg incorrectly bring up calculations based on interim results as explanatory evidence, although the calculated figure falls within the confidence limits of our results. In fact, the absolute difference in the number of children treated for language disorders is consistent with our primary outcome.
That vanderPloeg et al conclude that the observed effect on the final outcome in a well-performed RCT with added screening in the study arm was caused mainly by other factors without being able to pinpoint those factors that could have added to the effect in the study arm seems nonscientific and, therefore, highly illogical.
We are confident that our results will stand firmly and that future well-conducted systematic overviews will include this large-scale RCT. We feel that not acting now (or waiting for another trial with results in 10 years' time) is harming the children who should benefit now, and we would not like to take that responsibility.
REFERENCES
- Law J, Garrett Z, Nye C. Speech and language therapy interventions for children with primary speech and language delay or disorder. Cochrane Database Syst Rev. 2003;(3):CD004110
- Bland JM, Altman DG. One and two sided tests of significance.
BMJ. 1994;309
(6949):248
[Free Full Text]
PEDIATRICS (ISSN 1098-4275). ©2008 by the American Academy of Pediatrics
CiteULike 
Connotea 
Del.icio.us 
Digg 
Facebook 
Reddit 
Technorati 
Twitter What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
