Reading and communication skills after universal newborn screening for permanent childhood hearing impairment
- 1School of Psychology, University of Southampton, Southampton, UK
- 2School of Medicine, Southampton General Hospital, University of Southampton, Southampton, UK
- 3UCL Institute of Child Health, London, UK
- 4Public Health Sciences and Medical Statistics/RDSU, School of Medicine, University of Southampton, Southampton, UK
- 5National Perinatal Epidemiology Unit, University of Oxford, Oxford, UK
- Colin Kennedy, Mailpoint 21, Child Health, Southampton General Hospital, Southampton SO16 6YD, UK;
- Accepted 14 October 2008
- Published Online First 17 November 2008
Background: Birth in periods with universal newborn screening (UNS) for permanent childhood hearing impairment (PCHI) and early confirmation of PCHI have been associated with superior subsequent language ability in children with PCHI. However their effects on reading and communication skills have not been addressed in a population-based study.
Methods: In a follow-up study of a large birth cohort in southern England, we measured reading by direct assessment and communication skills by parent report in 120 children with bilateral moderate, severe or profound PCHI aged 5.4–11.7 years, of whom 61 had been born in periods with UNS, and in a comparison group of 63 children with normal hearing.
Results: Compared with birth during periods without UNS, birth during periods with UNS was associated with better reading scores (inter-group difference 0.39 SDs, 95% CI 0.02 to 0.76, p = 0.042) and communication skills scores (difference 0.51 SDs, 95% CI 0.06 to 0.95, p = 0.026). Compared with later confirmation, confirmation of PCHI by age 9 months was also associated with better reading (difference 0.51 SDs, 95% CI 0.15 to 0.87, p = 0.006) and communication skills (difference 0.56 SDs, 95% CI 0.12 to 1.00, p = 0.013). In the children with PCHI, reading, communication and language ability were highly correlated (r = 0.62–0.84, p<0.001).
Conclusion: Birth during periods with UNS and early confirmation of PCHI predict better reading and communication abilities at primary school age. These benefits represent functional gains of sufficient magnitude to be important in children with PCHI.
Previously we reported that universal newborn screening (UNS) and early detection of permanent childhood hearing impairment (PCHI) were associated with higher scores for language at primary school age.1 A recent systematic review by the US Preventative Services Task Force (USPSTF)2 assessed that study as providing good-quality evidence of benefit to language development and was accompanied by a USPSTF recommendation to screen for hearing loss in all newborn infants.3 In comparison with their hearing peers, children with PCHI also have poorer academic outcomes, particularly reading attainment with an average reading age of 9 years when aged 17 years.4 5 In this report, we extend our previously reported findings on language acquisition to examine the impact of early detection of PCHI on applied practical skills further along the causal pathway, such as reading and communication skills.
Communication and reading skills were assessed at the same home visit and in the same children with bilateral PCHI ⩾40 dB hearing level, identified in a 1992–1997 birth cohort in eight districts of southern England, whose language abilities we reported previously.1 In addition to the assessments previously reported, “basic reading” (BR) and “reading comprehension” (RC) scores were derived on the Weschler Objective Reading Dimensions6 (WORD). The communication domain of the Vineland Adaptive Behaviour Scales (VABS)7 was also completed, using information provided by the principal caregiver, usually the mother. Reading and communication scores provided by a group of normally hearing children, born in the same hospitals and of similar age at assessment, were used to derive z scores (including z aggregate reading = zBR + zRC) for the children with PCHI. Consistent with our previous trial of UNS,8 9 early confirmation of PCHI was pre-specified as confirmation by 9 completed months of age.
Multiple linear regression (using Stata version 810) was used to examine the effects of birth in periods with UNS and of early confirmation of PCHI on reading and communication scores and to adjust for other explanatory variables, namely severity of hearing impairment, non-verbal ability and maternal education. The South and West UK Multi-centre Research Ethics Committee approved this study (MREC/99/6/77) and participating principal caregivers provided written informed consent.
Of 168 identified children with bilateral PCHI in the cohort of 156 733 births, 120 were enrolled and assessed at a mean (range) age of 7.9 (5.4–11.7) years as were 63 children with normal hearing.1 The characteristics of the families and children born in periods with and without UNS were similar except that the mothers of babies born in periods without UNS tended, presumably by chance, to have completed education to a higher level (table 1). The median age of confirmation of PCHI was 10 months (interquartile range 2–25) with a skewed distribution (mean 17.5 months). The children in the normally hearing comparison group were selected to be English speaking. They had fewer health problems and were more likely to have parents in “higher” occupations than the children with PCHI but were in other respects similar to them (table 1).
Reading (WORD) and communication skills (VABS)
Scores for WORD and for VABS were available in 102 and 98 of 120 (85% and 82%), respectively, of the children with PCHI, including 16 and 11 children who used British Sign Language (BSL). The unadjusted scores for these measures, standardised for age, were significantly higher in the normal hearing comparison group than in the children with PCHI (table 2). Comparing the hearing-impaired group with the normally hearing group, we found larger deficits in language comprehension than in communication and reading (table 2). Among children with PCHI, outcome scores were converted to z scores. There was a consistent trend to higher unadjusted scores in those who were born during periods with UNS compared with those born in periods without UNS or whose PCHI was confirmed by 9 months of age compared with those whose PCHI was confirmed later.
Effect of birth in periods with UNS
After adjustment for severity of PCHI, mother’s educational level and the child’s non-verbal ability, group mean z scores for WORD aggregate reading and for the VABS communication scale for those children born in periods with UNS were significantly higher than for children born in periods without UNS by 0.39 and 0.51, respectively (table 3). These adjusted scores were positively correlated with previously reported1 receptive and expressive language z scores (r = 0.62–0.84, p<0.001).
Effect of early detection of PCHI
Compared with scores in those whose PCHI was confirmed later, adjusted group mean scores were also significantly higher in those whose PCHI was confirmed by 9 months by 0.51 on WORD aggregate reading and 0.56 on the VABS communication scale (Table 4). Compared to the size of effect of birth in periods with UNS and of early detection of PCHI observed in all children with PCHI, the effect observed in children that used BSL was similar with respect to reading scores and larger with respect to communication scores, but this was based on few observations.
Mediating role of language
Adding the aggregate language z scores to the regression models substantially reduced the differences between the group mean z scores, whether comparing children born in periods with and without UNS (to 0.16, 95% CI −0.10 to 0.42, p = 0.24 for reading and 0.31, 95% CI −0.08 to 0.70, p = 0.11 for communication) or children whose PCHI was detected early and late (to 0.09, 95% CI −0.17 to 0.36, p = 0.49 for reading and 0.19, 95% CI −0.20 to 0.59, p = 0.33 for communication), suggesting that the benefits to reading and communication were partially mediated by greater language ability.
Analysis in the Wessex subgroup
In the Wessex subgroup, whose exposure to UNS (or not) was allocated in a controlled trial, the chance of unknown confounding should have been reduced. In this subgroup, the best estimate of the size of the effect of birth in periods with UNS on WORD aggregate reading scores and VABS communication scale scores (table 5) was greater than that observed in the study population as a whole. This was also true with early confirmation of the PCHI (cf. tables 3 and 4 with tables 5 and 6). In the case of aggregate reading scores, these inter-group differences remained statistically significant (p = 0.04 and 0.004) in spite of reduced numbers of observations (tables 5 and 6).
This population-based sample of children showed a significant association between both birth in periods with UNS and also early detection of PCHI on the one hand and higher mean scores at primary school age for reading and communication (by 0.4–0.6 SDs) on the other. These outcomes demonstrate gains in skills that are important in daily living and of sufficient magnitude to be of functional significance.
This study may have underestimated the size of the benefit of UNS for reading and communication skills for children with PCHI from birth because this birth cohort was the first in the UK to which UNS was applied so that intervals between screening positive, confirmation of PCHI and fitting of hearing aids were longer than is the current standard of care.11 12 The data obtained in the Wessex subgroup were experimental since their exposure to UNS was the intervention in a controlled trial,8 9 suggesting a cause-and-effect relationship between UNS and higher subsequent reading and communication scores.
The only previous population-based study to examine the effect of age at first diagnosis on reading in children with PCHI of primary school found no significant benefit.13 The setting of that study differed from the present report in that the newborn screening for PCHI had been undertaken in high-risk but not in standard-risk newborns. Other limitations of that study were the inclusion of children with only mild PCHI (21% of the sample), an ascertainment rate of 89 of 132 eligible children (67%), and a relatively low rate of early confirmation of PCHI with only 11 (12%) cases with PCHI detected by age 6 months.
We have previously reported that confirmation of hearing impairment prior to age 9 months was associated with higher scores on language tests at age 8 years.1 We chose language as the outcome measure because it can be precisely estimated by direct assessment and is on the causal pathway from early intervention to better functional outcomes. For both theoretical and practical reasons, however, those functional outcomes cannot be inferred from the finding of higher language scores, as explained below.
VABS communication domain scores assess what the individual understands, says, reads and writes as reflected in behaviours as they occur in an individual’s environment rather than assessing responses to standardised auditory stimuli in a direct testing situation (ie, language testing). VABS communication scores are higher in children with auditory impairments than in those with visual impairments whereas the reverse is true for language tests.7 Communication requires linguistic skills and the ability to decode a message that may also require auditory (as in spoken language), visual (as in gesture, lip movement and reading) or tactile (as in Braille) skills. Compared with that of normally hearing children, the communication of children with hearing impairment depends to a greater extent on gesture, on awareness of social cues and on other skills that compensate for their reduced auditory acuity. These skills are assessed by communication scores but not by standard language tests that are standardised on normally hearing children.
Reading requires visual–perceptual skills to access a cognitive representation of linguistic information whereas in the case of comprehension and expression of spoken language this access is dependent on the ability to process an auditory input and to generate a verbal utterance. Reading is thus less dependent than oral language on auditory acuity which is reduced in hearing impairment. The existence of a relationship between the development of reading and language skills is well established in normally hearing children14 but poorly understood in hearing-impaired children. In the present study, the greater deficits in language than in reading relative to those of the normally hearing comparison group suggest that the language scores underestimated communicating ability in the hearing-impaired children. This limitation on the use of language scores as a measure of the benefit of UNS is circumvented by showing the benefits to reading and communication reported here.
The processes and mechanisms that underlie the development of reading ability in these children are not clear and may differ from those known to play a strong predictive role in the early development of reading skills in normally hearing children. These include language skills such as vocabulary and phonological awareness (the ability to segment and blend the sounds of a language eg, identifying the rhyme that is common to “hat”, “bat” and “cat”).14 – 19 Reading for 7–8-year-old children with a severe to profound PCHI, for example, was reported to be predicted by expressive vocabulary and speech reading (ie, the child’s ability to attend to the shape of the lips and mouth in understanding oral speech) but not by phonological awareness.19 It is, however, possible that UNS followed by early confirmation of PCHI may allow children to follow a path to literacy more similar to that of hearing children.
Our observation of higher scores for reading and communication in the setting of UNS provides evidence that provision of UNS and confirmation of PCHI by 9 months can be expected to lead not only to higher scores on formal tests of language ability, but also to practical benefits to the reading and communication skills of these children, including the subgroup that uses BSL. Further assessment of the participating children after longer follow-up is planned to establish whether these children show improvement at high school age in language, reading, social communication, and school attainment.
What is already known on this topic
Bilateral permanent childhood hearing impairment (PCHI) >40 dB hearing level impairs language and school achievement.
Universal newborn screening and confirmation of PCHI by age 9 months improves language at primary school age.
What this study adds
Universal newborn screening and early confirmation of permanent childhood hearing impairment improve reading ability at primary school age and adaptive behaviour with respect to communication skills.
The best estimate of the size of these benefits is around 0.5 SD.
We are grateful for the funding support of The Wellcome Trust (Ref. 061839).
We gratefully acknowledge the help and assistance received from children, families, school staff, Specialist Teachers of the Hearing Impaired, particularly Jan Nanor in London, Speech and Language Pathologists, and Audiology professionals in the local area teams. We particularly thank the following senior Audiology staff: Dr Margaret Baldwin, Dr Joy Bhattacharya, Dr Alyson Bumby, Iris Curtis, Dr Carol Hunter, Dr David Reed, Scott Richards, Sue Robinson, Dr Peter Savundra, Mr Huw Thomas, Dr Peter Watkin and Dr Tim Williamson. We thank Ros Herman for the assessments of British Sign Language. We also thank the team of researchers (Helen Davis, Shirley Golden, Eleanor Lutman, Kristen Paul and Helen Ryder) for their excellent work. We thank Julie Brinton and Hazel Blythe at the University of Southampton for their advice to the authors.
Funding: This study was funded by a grant from the Wellcome Trust (reference 061839). The authors’ work was independent of the funders and the funders had no involvement. The sponsors of the study had no role in the study design, data collection, data analysis, data interpretation or writing of the report.
Competing interests: None.
Ethics approval: The South and West UK Multi-centre Research Ethics Committee approved this study.
Patient consent: Parental consent obtained.
The design and application for funding was developed from an idea by CRK with the help of CML, MM, SP, JS and SW. DCM oversaw the conduct of the study with help from all the other authors. Statistical analysis was undertaken by HMY assisted by MM. All authors contributed to preparation of the manuscript and saw the final version.
This paper is freely available online under the BMJ Journals unlocked scheme, see http://adc.bmj.com/info/unlocked.dtl