Utility of the WHO Ten Questions Screen for Disability Detection in a Rural Community the North Indian Experience by Pratibha Singhi, a Munish Kumar, b Prabhjot Malhi, c and Rajesh Kumar d a Department of Pediatrics b Department of Pediatrics c Department of Pediatrics d Department of Community Medicine Summary The utility of the WHO Ten Questions Screen (TQS) was studied in a rural community of North India. The study was done in three villages, in two phases. In phase 1, the TQS was administered to parents of aged between 2 and 9 years, during a house-to-house survey. In phase 2, all screened positive and a random sample of 110 screened negative were clinically evaluated in detail. The total population of the three villages was 5830 with 1763 aged between 2 and 9 years. Seventy-six were positive on the TQS, of these, 38 were found to have significant disability, 18 had protein energy malnutrition and 19 were found normal on clinical evaluation. All the 110 screen-negative were normal. Significantly larger numbers of boys were positive on TQS as compared to girls [Odd Ratio (OR) 1.5]. The sensitivity of the TQS for significant disability was 100%; the positive predictive value was 50% and was higher for boys than for girls. Of the 50% classified as false positive 23% had mild delays due to malnutrition. The estimated prevalence of disability was 16/1000. The TQS was found to be a sensitive tool for detection of significant disabilities among 2 9 years of age. The low-positive predictive value would lead to over referrals but a large number of these would benefit from medical attention. Introduction Childhood disability is a major problem in developing countries. There is paucity of data on prevalence rates mainly because of the lack of systems for data collection. Detection of disabilities in communities is difficult; lack of public awareness about disabilities and rehabilitation, lack of trained professionals and lack of screening tools are some of the causes. To be used on a large population, with limited resources, any screening tool should be simple, inexpensive, and at the same time, sensitive and specific. The WHO Ten Questions Screen (TQS) was devised as a simple questionnaire for the detection of common disabilities (physical, mental, speech, hearing, visual and epilepsy) among 2 9 years of age, at the Correspondence: Pratibha Singhi, Advanced Pediatric Center, Postgraduate Institute of Medical Education and Research, Sector 12, Chandigarh 110 012, India. Tel.: 91-0172-2755304. E-mail <psinghi@glide.net.in>. community level. It contains 10 questions and probe questions that follow each of the questions, and is intended to be used across cultures. The validity of the questionnaire has been reported from some countries such as Bangladesh, Jamaica and Pakistan, [1 3] with variable results. There is little information on its use in the Indian population. We therefore used the screen in a rural setting in North India to study its utility in the detection of disabilities. Method The TQS was first translated into the two commonly used local languages Hindi as well as Punjabi and was pilot tested on 30 each. The study was done in three selected villages around Chandigarh, which were already being visited by workers from the community medicine department of our institute and where a good rapport had already been established. The study was conducted in two phases. In phase 1, a house-to-house survey was done and the WHO 10-point questionnaire was administered as a personal interview to parents/guardians of between the ages of 2 and 9 completed years. ß The Author [2007]. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org 383 doi:10.1093/tropej/fmm047 Advance Access Published on 7 June 2007
TABLE 1 Age and sex distribution of the sample and percent positive on the Ten Questions (TQ) Age (years) Boys Girls Total Total no. No. (and% TQþve) Total no. No. (and% TQþve) Total no. No. (and% TQþve) 2 120 8 (6.6) 97 0 217 8 (3.6) 3 106 7 (6.6) 91 7 (7.6) 197 14 (7.1) 4 127 7 (5.5) 131 5 (3.8) 258 12 (3.4) 5 125 3 (2.4) 115 2 (1.7) 240 5 (2.5) 6 105 6 (5.7) 93 3 (3.2) 198 9 (4.5) 7 127 5 (3.9) 128 3 (2.3) 255 8 (3.1) 8 99 6 (6.0) 85 6 (7.0) 184 12 (6.5) 9 110 8 (7.2) 104 2 (1.9) 214 10 (4.6) Total 919 47 (5.1) 844 29 (3.4) 1763 76 (4.3) In phase 2, all with a positive-screening result as well as a random sample of 110 screennegative, were evaluated clinically. Clinical examinations were done by a team of a pediatrician and a psychologist, who were blinded to the screening results. A detailed pre-structured Medical and Neurodevelopment Assessment Form was used to record the results of the physical examination and developmental assessment of included in phase 2 of the study. Hearing assessment by age appropriate methods was done. IQ assessment was done by the psychologist using the Malin s adaptation of the Weschler Intelligence scale for school [4], for between 6 and 9 years of age and the WPSS1 [5] for aged between 2 and 5 years. Definitions Used Disability Inability to carry out certain activities considered normal for one s age, sex etc. Standard procedures and criteria were used to diagnose and assess the level of disability (mild, moderate and severe) in five areas: cognitive (mental retardation), motor, vision, hearing and seizures. Mental retardation Significantly sub average general intellectual functioning (IQ standard score of less than 70) in association with deficits in two or more of the following ten adaptive skill areas: communication, self care, home living, social skills, community use, self direction, health and safety, functional, academics, leisure and work. Hearing loss Moderate (32 35 db), severe (50 70 db) or profound (70 db or greater). Delayed speech No use of single word by 1.5 years or single word vocabulary of 10 words by 2 years of age; The Expressive Language Milestone Scale [6] was used for older. Impaired vision Visual acuity <3/60 (Snellen) or its equivalent. Seizure disorder Diagnosis was made on the basis of medical history and EEG other investigations were done if needed. Children diagnosed to have a disability of any kind were referred to the existing community-based rehabilitation programmes. The data were analyzed using the SPSS 10 statistical package. Results The total population of the three villages was 5830; there were 1763 aged between 2 and 9 completed years. The age and sex distribution of the screened as well as the rates of positive responses on the TQS are shown in Table 1. The percent positive on the TQS was lower for 5- and 7-year-old as compared to other ages. A larger percentage of boys had a positive screening as compared to girls [5.1% vs. 3.4%]. A total of 76 were found to be positive on any one of the 10 questions. On analysis of the individual responses to the 10 questions (Table 2), it was found that the maximum numbers of positive responses were related to learning, and the minimum to vision. Significantly larger numbers of boys screened positive for problems of delayed milestones, no speech or unclear speech and slowness (p-values 0.01, 0.007, 0.04, respectively Table 2). For the TQS as a whole, the odds ratio was 1.5 for boys vs. girls, indicating that parents were about 1.5 times more likely to report a problem for sons than for daughters. The problems of comprehension and 384 Journal of Tropical Pediatrics Vol. 53, No. 6
TABLE 2 Frequencies and odds ratio (ORs) of individual question (rows 1 10) on TQ in boys and girls Question Related disability Total no. (% positive) No. of positive Boys vs. girls Boys Girls OR 95% CL 1 Developmental Milestones 33 (1.8) 25 8 2.9 1.25, 7.07 2 Vision 5 (0.2) 3 2 1.38 0.19, 11.7 3 Hearing 10 (0.5) 8 2 3.70 0.73, 25.2 4 Comprehension 7 (0.4) 6 1 5.54 0.67, 122 5 Movements 9 (0.5) 7 2 3.23 0.62, 22.5 6 Seizures 7 (0.4) 6 1 5.54 0.67, 122 7 Learning 60 (3.4) 38 22 1.61 0.92, 2.84 8 No speech 15 (0.8) 13 2 6.04 1.30, 38.8 9 Unclear speech 20 (1.1) 17 8 5.28 1.46, 22.7 10 Slowness 18 (1.0) 16 3 4.97 1.36, 21.4 TABLE 3 Results of clinical evaluation of positive on the TQS Age (years) Total screen positive Significant disability PEM Normal 2 8 6 0 2 3 14 6 5 2 4 12 8 1 0 5 5 1 0 5 6 9 4 3 2 7 8 5 1 2 8 12 3 5 4 9 10 4 3 2 Total 76 38 18 19 a a One child had stuttering. learning were significantly more common in older as compared to younger (p-values 0.01 and 0.02, respectively). The age wise results of clinical evaluation for screen-positive are shown in Table 3. Thirty-eight of 76 evaluated had significant (comprehension, vision, hearing, seizures and motor disorder) disabilities, 18 had mild delays with protein energy malnutrition (PEM) and 19 were found to be normal. One child was found to have mild stuttering. Of the 110 screen-negative clinically evaluated, none had significant disability. A total of 79 disabilities were identified in the 38, many of whom had multiple disabilities. Of the 79, there were 32 cognitive, 29 motor, 7 seizures, 2 vision and 9 hearing disabilities. The estimated prevalence was calculated as the weighted average of prevalence estimated in those screened positive and those screened negative. These estimates were based upon a sample of 186 clinically evaluated and were adjusted for the fact that with positive-screening results were over represented within this sample. Based on the numbers of true and TABLE 4 Distribution of parameters in 186 clinically evaluated Parameter Boys Girls Younger Older All Trueþve (a) 29 9 21 17 38 Falseþve (b) 18 20 10 28 38 False ve (c) 0 0 0 0 0 True ve (d) 60 5 50 60 110 false positives and negatives (Table 4), the validity of the TQS for screening significant disability was calculated (Table 5). The sensitivity and negative predictive value were found to be 100%, indicating that all 38 found to have significant disabilities were positive on the TQS. The positive predictive value of the screen for significant disabilities was 50%. The overall estimated prevalence of disabilities was about 16/1000 and was significantly (p < 0.05) higher for boys (16.4/1000) than girls (5.1/1000), but not significantly different among older vs. younger. A total of 38 were classified as false positives on the TQS. Although none of them had significant disability, 50% had mild delays with malnutrition or a minor speech problems. If these milder conditions are also included, the positive predictive value of the screen goes up to 75%. Discussion The TQS was found to be a sensitive tool for detection of significant disabilities in aged between 2 and 9 years. The positive-predictive value of the screen was only 50% implying that if used as a case finding tool for epidemiological purposes, it would lead to a number of over referrals. However, of those classified as false positives for significant disability, 23% had mild delays due to malnutrition Journal of Tropical Pediatrics Vol. 53, No. 6 385
TABLE 5 Estimates of the validity of the TQs for screening serious disabilities, and estimates of prevalence of serious disability in rural areas around Chandigarh based on the sub-sample of 186 clinically evaluated All Boys Girls Older (5 9) years Younger (2 4) years Sensitivity ¼ a/a þ c x 100 1.00 1.00 1.00 1.00 1.00 Specificity ¼ d/b þ d 100 0.74 0.76 0.71 0.68 0.83 Positive predictive value ¼ a/a þ b 100 0.50 0.61 0.31 0.37 0.67 Negative predictive value ¼ d/c þ d 100 0 0 0 0 0 and 25% were normal. This observation suggests that referring all with positive results will not necessarily result in an inefficient use of professional resources, since early identification and treatment of malnutrition and other health problems is also important and helps in prevention of further significant disabilities. The TQS has been used in other developing countries. The overall screen positive percentages were 15.6, 14.7 and 6.9 in Jamaica, Pakistan and Bangladesh, respectively. [1 3] An analysis of the validity of the TQS, [7] found an acceptable (80 100%) sensitivity for serious cognitive, motor and seizure disabilities but not for vision and hearing disabilities. The positive predictive value of the TQS has been found to be low, and was 22% in the Bangladesh study [1], but 70% of classified as false positive had mild disabilities or conditions for which early detection and intervention could be beneficial. In a study from Kenya [8], the overall sensitivity of the TQS was found to be 98.6% while the positive predictive value was only 29% and was highest (55.6%) for seizures and low for speech and learning difficulty. Although the positive-predictive value of the TQS is low, it must be remembered that the TQS is a low-cost-quick screening tool that can be used by community workers, and not an assessment tool. As such, false positivity is not unexpected particularly if it is validated against a detailed clinical examination. Interestingly, in our study, the positive-predictive value was found to be higher for boys (61%) as compared to girls (31%). This is consistent with the observed tendency for parents in India to display more concern for the health of sons than for daughters. The higher number of false positives in older may be due to the fact that parents become more aware of the child s problems as the age of the child advances. In the present study, question 7 (related to hearing) elicited a much larger positive response than any other question whereas in the studies from Jamaica and Pakistan, a higher-positive response rate was reported for question 10 (retardation) and for questions 1, 5, 6 and 9 (delayed milestones, movement, fits, learning and unclear speech), respectively. In our study, the frequency of positive responses to questions 8 and 9, related to speech and question 10 related to slowness was somewhat different between boys and girls which is similar to that reported in the Bangladesh study. Problems in comprehension and learning (questions 4 and 7) were expectedly reported more frequently in older than in younger. However, there were no significant age or gender differences in the probability of being positive on the TQS on any one or more questions. Thus, the screen does not appear to grossly over identify or under identify any one or the other of these groups. None of the who screened positive or screened negative and randomly evaluated, had autism or attention deficit hyperactivity disorder (ADHD). The TQS is not specifically designed to detect these disabilities, although with speech delays and motor problems associated with these disabilities could theoretically be detected with this screen. Only 8% of parents/guardians were aware about rehabilitation programmes for disabled being run in their area. A population-based survey of childhood disability in Eastern Jeddah using the TQS and further information found a prevalence rate of 36.7/1000 and the families reported lack of services for their disabled [9]. We found that the TQS is an effective screening tool for detection of significant disabilities in our culture. This is in accordance with the reported crosscultural reliability of the screen [10]. Minor variations in culturally divergent population are acceptable for questionnaires designed for large-scale epidemiological research. The over-referrals associated with the screen could perhaps be minimized by the use of further specific probe questions. A couple of questions related to screening for autism and ADHD could perhaps increase the scope and completeness of the questionnaire. References 1. Zaman S, Khan N, Islam S, et al. Validity of the ten questions for screening serious childhood 386 Journal of Tropical Pediatrics Vol. 53, No. 6
disabilities: result from urban Bangladesh. Int J Epidemiol 1990;19:613 20. 2. Thorburn M, Desai P, Paul TJ, et al. Identification of childhood disability in Jamaica: evaluation of the ten-question screen. Int J Rehabil Res 1992;15:262 70. 3. Durkin MS, Hasan ZM, Hasan KZ. The ten questions screen for childhood disabilities: it uses and limitations in Pakistan. J Epidemiol Commun H 1995;49:431 6. 4. Malin AJ. Malin s Intelligence Scale for Indian Children Manual Nagpur. Child Guidance Centre, 1971. 5. Wecholer D. Wechster Preschoel and Primary Scale of Intelligence. 3rd edn. San Antonio, Texas, Psychological Corperation, 2002. 6. Leptan J. The Early Language Milestone Scale. Autism. Pro-ed inc. Austin, Texas, USA, 1987. 7. Durkin MS, Davidson LL, Hasan ZM, et al. Validity of ten question screen for childhood disability: results from population based studies in Bangladesh, Jamaica and Pakistan. Epidemiology 1994;5:283 9. 8. Muga E. Screening for disability in a community: ten questions screen for, in Bondo, Kenya. Afr Health Sci 2003;3:33 9. 9. Milaat WA, Ghabrah TM, Al-Bar HM, et al. Population-based survey of childhood disability in eastern Jeddah using the ten-question tool. Disabil Rehabil 2001;23:199 203. 10. Durkin MS, Wang W, Shrout PE, et al. Evaluation of the Ten Questions screen for childhood disability: reliability and internal structure in different cultures. J Clin Epidemiol 1995;48:657 66. Journal of Tropical Pediatrics Vol. 53, No. 6 387