LEST (3-6) with 31 test items, was developed at CDC Kerala and validated in a community sample of children. The detailed methodology was reported in a previous article on LEST (0-3) [2].The additional test items and the ranges were selected from different existing developmental/speech and language assessment scales, tools or guidelines like, Hearing check list by American Speech-Language-Hearing Association [6], Speech and language development in babies by Abby Deliz [7], Gilman and Gorman’s Speech Language Development Chart [8], Speech Sound Development Chart by Sander’s [9], and Ages and stages milestones for receptive and expressive language acquisition by Caroline Bowen [10].

Children between 3-6 years belonging to 11 Anganwadis from an urban ward, 20 Anganwadis from a rural Panchayat and 3 Anganwadis from tribal area, participated in this study. LEST (3-6), was applied by two trained persons having similar educational qualification as that of an ICDS Supervisor, and "extended REELS" by two Speech and Language therapists. For LEST [3-6], a vertical line was assumed by keeping a scale vertically, at the chronological age in months given horizontally in the ‘X’ axis (Web Fig. 1). All items falling short on the left side of the age line was expected to be done by the child. If not done, it was taken as that item delay. First preference was given for observation of the child and testing of the items; if it was not possible, for some of the items, parental reporting was considered valid.

The test re-test reliability (intra-class correlation of 0.61; 95% CI 0.41-0.76) and inter-rater reliability (intra-class correlation of 0.96; 95% CI 0.93-0.97) were acceptable in valid samples of 50 children.

Test results for both LEST (3–6) and the reference standard were available for a sample of 606 children (292 boys); 247 (3-4 years), 221 (4-5 years), 138 (5-6 years). Table I shows the results with one-item delay and two-item delay as test positive, against extended REELS. One-item delay and two-item delay showed a sensitivity of 81% and 47%; specificity 68% and 94%: positive predictive value 12% and 31%; negative predictive value 98% and 97%; and accuracy 68.5% and 92%, respectively. Changing the tool positivity from one-to-two item delay, resulted in decrease of sensitivity from 81% to 47% though specificity increased from 68% to 94%.

TABLE I	Test Characteristics with Two Different Criteria 

For LEST(3-6) as a screening tool for Language delay, we suggest delay in two item as test positive, because of the relatively higher positive predictive value and lesser false positives in the screened sample, with an excellent negative predictive value (97%). Since speech and language development is a continuous process without a definite line for normal/abnormal, and with limited resources for language stimulation/therapy available, a lower sensitivity (47%) may be accepted. The choice between one-item and two-item delay as "LEST delay" depends on the need, whether to have a highly sensitive test with a low positive predictive value or to have a more specific test having a higher positive predictive value, and a lower, but acceptable sensitivity. This depends also on resources available for further evaluation of screen positives, having larger (with one-item delay) or smaller (with two-item delay) false positives in screened sample.

We conclude that LEST (3-6 years) is a simple, reliable and valid Indian tool for identifying children of 3-6 years with language delay in the community.

Acknowledgments: Paul Russell and CDC Kerala Research team including Indulekha, Letha, Abhiram, Princly, Vinod, Bincymol, Leena and Asokan. We also thank Kunnampallilgejo John for sharing extended REELS for this study. Funding: NRHM and CDC, Kerala.

1. Owens RE. Language Development: An Introduction. Allyn & Bacon, Inc., 1996. Available from: http://raisingchildren.net.au/articles/language_development. html. Accessed August 14, 2015.

2. Nair M K C, Nair G H, Mini A O, Indulekha S, Letha S, Russell P S. Development and validation of language evaluation scale Trivandrum for children aged 0-3 years-LEST (0-3). Indian Pediatr. 2013;50:463-7.

3. Nair M K, Mini A O, Bhaskaran D, Harikumaran Nair G S, George B, Leena ML, et al. CDC Kerala 6: Validation of Language Evaluation Scale Trivandrum (0-3 y) against receptive expressive emergent language scale in a developmental evaluation clinic population. Indian J Pediatr. 2014;81:99-101.

4. Shriberg LD, Tomblin JB, Mc Sweeny JL. Prevalence of speech delay in 6-year-old children and comorbidity with language impairment. J Speech Hearing Res. 1999;42 (6):1461-81.

5. Extended REELS. Available from: www.docstoc.com/docs/160379205/EXTENDED-REELS. Accessed July 30, 2015.

6. How Does Your Child Hear and Talk? Three to Four Years. Available from: www.asha.org/public/speech/development/34.htm. Accessed March 31, 2015.

7. Speech and Language Development in Babies. Available from: https://suite.io/abby-deliz/146p2hf. Accessed August 14, 2015.

8. Gard A, Gilman L, Gorman J.　 Language Development Chart. In: Speech and Language Development Chart. 2nd　ed. Austin TX: ProEd; 1993. Available from: http://www.ccr4kids.org/Portals/0/pdfs/Language%20 Development%20Chart.pdf.　Accessed September 30, 2015.

9. Sanders E K. When are speech sounds learned. J speech hearing disord, 1972;37:55-63. Available from: http://www.bisd.us/sharp/english_developmental_ sound_chart%5B1%5D.pd. Accessed March 4, 2015.

10. Bowen C. Ages and Stages Summary - Language Development 0-5 years. Available from: www.speech-language-therapy.com/index.php?option=com_ content&view=article&id=34:ages&catid=11: admin&Itemid=117. Accessed September 30, 2015.