The effects of bilingualism on stuttering during late childhood

Transcription

1 Additional information is published online only at ad.bmj.om/ontent/vol93/ issue11 1 Division of Psyhology and Language Sienes, University College London, London, UK; 2 Department of Language and Communiation Siene, City University, London, UK Correspondene to: Peter Howell, Division of Psyhology and Language Sienes, University College London, Gower Street, London WC1E 6BT, UK; p.howell@ul.a.uk Aepted 24 June 2008 Published Online First 9 September 2008 This paper is freely available online under the BMJ Journals unloked sheme, see ad.bmj.om/info/unloked.dtl The effets of bilingualism on stuttering during late hildhood P Howell, 1 S Davis, 1 R Williams 2 ABSTRACT Objetives: To examine stuttering by hildren speaking an alternative language exlusively (LE) or with English (BIL) and to study onset of stuttering, shool performane and reovery rate relative to monolingual speakers who stutter (MONO). Design: Clinial referral sample with supplementary data obtained from speeh reordings and interviews. Setting: South-East England, Partiipants: Children aged 8 12 plus who stuttered (monolingual and bilingual) and fluent bilingual ontrols (FB). Main outome measures: Partiipants stuttering history, SATS sores, measures of reovery or persistene of stuttering. Results: 69 (21.8%) of 317 hildren were bilingual. Of 38 hildren who used a language other than English at home, 36 (94.7%) stuttered in both languages. Fewer LE (15/38, 39.5%) than BIL (23/38, 60.5%) hildren stuttered at first referral to lini, but more hildren in the fluent ontrol sample were LE (28/38, 73.7%) than BIL (10/38, 26.3%). The assoiation between stuttering and bilingual group (LE/BIL) was signifiant by x 2 test; BIL speakers have more hane of stuttering than LE speakers. Age at stuttering onset and male/female ratio for LE, BIL and MONO speakers were similar (4 years 9 months, 4 years 10 months and 4 years 3 months, and 4.1:1, 4.75:1 and 4.43:1, respetively). Eduational ahievement was not affeted by bilingualism relative to the MONO and FB groups. The reovery rate for the LE and MONO ontrols together (55%) was signifiantly higher by x 2 test than for the BIL group (25%). Conlusions: BIL hildren had an inreased risk of stuttering and a lower hane of reovery from stuttering than LE and MONO speakers. Bilingualism has been regarded as a risk fator for stuttering. 1 2 However, there is little information about how a seond language affets the hanes of stuttering onset and of reovery. Consequently, a study was onduted on all referrals for stuttering for hildren first seen when they were between 8 and 10 years of age who attended linis that speialised in the diagnosis and treatment of stuttering. Children who used a seond language in the pre-shool years either (a) exlusively (these learned English at shool, termed LE) or (b) along with English in the home (bilingual from birth, termed BIL) were seleted. The majority of the hildren were seen again when they were aged 12 plus. Prevalene rates of stuttering for LE and BIL hildren as ompared with all referrals were examined to determine if a disproportionate number of speakers of a language other than English is referred to stuttering linis. Cheks What is already known on this topi Popular oneption and early data suggest an inreased risk of stuttering for bilingual speakers, although the data are sparse and this laim has been questioned reently. The Lidombe treatment programme has been reported to be an effetive treatment for young bilingual hildren who stutter. What this study adds Children who are bilingual usually stutter in both their languages (rather than just one). If a minority language alone is used in the home up to age 5, the hane of starting to stutter is lower and the reovery rate is higher than for hildren who aquire English as well as a minority language during this period. Learning English onurrently with or after a minority language does not affet eduational attainment at key stages one (age 7) and two (age 11). were made to verify that the LE and BIL hildren were stuttering by omparing age at stuttering onset and gender ratio for these speaker groups against monolingual speakers from a referral sample who stuttered (MONO). Early shool performane of LE and BIL hildren who stuttered was ompared with that of a MONO group and bilingual hildren who were fluent (the fluent bilingual group, termed FB). The patterns of onset and reovery in the LE, BIL and MONO groups were ompared. METHODS Partiipants and sub-groups A total of 317 hildren who stuttered partiipated. They all (a) started shool in the UK at age 4 or 5, (b) first presented at a lini when aged between 8 and 10 years and () lived in the greater London area. Stuttering was onfirmed by a speialist speeh-language therapist at the lini. Reported onset of stuttering usually ours before age 6. The attendane at lini 2 or more years later is partly due to the time needed to proess hildren in the health system; these are seondary referrals to the speialist lini and devolved budgets to loal area health authorities an ause delay. Reordings were taken to estimate the perentage of stuttered 42 Arh Dis Child 2009;94: doi: /ad

2 syllables and to allow the stuttering severity instrument measure SSI-3 3 to be applied (see below for details of how SSI-3 was alulated). When eah hild was seen initially, all 317 aregivers were asked whether they used a language other than English exlusively or sometimes in the home. In ases where use of a seond language other than English was reported, the aregivers had a further interview. When the aregiver did not speak English, help with translation was obtained from a lose friend who spoke the same language as well as English. The interview obtained bakground details about the hild, the family and information about languages used in the home. Details were olleted about the biologial relationship of the aregiver to the hild (all were parents of the hild they ared for). Data on gender and age at onset of stuttering were obtained. The hild s Standard Attainment Test eduational sores (SATS) at ages 7 and 11 (key stages one and two) were obtained for English, mathematis and siene. It was also established whether the aregiver spoke English to the hild from birth up until the hild went to shool, and whether are was given by this person exlusively or some of the time up to age 5 years. The aregivers indiated whether the hild spoke English when they entered shool. Information was obtained about where the hild was born and, in ases where the hild was an immigrant, the age at whih they had entered the UK and their language use in their former ountry. This language information was used to lassify hildren as LE or BIL. There were 69 ases where at least one language other than English was spoken in the home. Caregivers of 38 of those 69 hildren reported that they primarily or exlusively used a language other than English in the home and had done so sine the birth of their hild. This riterion exluded people who opted to speak a language other than English in the home for their hild s eduational/soial advanement. All 38 families have ontinued to partiipate. Nine of the group of 38 have not yet reahed 12 years, leaving 29 who were ategorised as reovered from stuttering or not. The 38 seleted hildren were divided into LE and BIL groups. The LE hildren may be regarded as not being bilingual until the age of shool onset. The LE group onsisted of 15 hildren (39.5% of the 38 hildren) who did not speak English when they started shool. During this period, English was not used by the aregivers, English-language media were not aessed in the home and if there were siblings, they did not speak English up to the age of shool entry. Twenty three (60.5%) of the hildren spoke English in addition to the primary language spoken in the home (BIL group). They were all exposed to English in the home from birth. A group of monolingual English speaking hildren who stuttered (MONO) was seleted in order to assess the impat stuttering has on epidemiology and eduation and to determine how the stuttering severity of the MONO group ompared with that of the experimental groups. The MONO hildren were seleted at random from the entire sample of suh hildren referred to the lini, subjet to the restritions that the group had to math the LE and BIL groups for age and gender. They had similar soio-eonomi bakgrounds and attended similar shools to the LE and BIL hildren. When hildren in all groups passed 12 years of age, they and their families were interviewed about their stuttering and shool reord, and the hildren were reorded so that a stuttering severity measure ould be obtained (to onfirm whether or not their stuttering had ontinued). A seond ontrol group of fluent hildren who used a language other than English in their pre-shool years was reruited (fluent bilingual, termed FB). These hildren were agemathed to those hildren who stuttered at seond attendane, so that SATS sores ould be ompared. The hildren were reruited from shools in the same athment areas as the linis and they had similar eduational and soio-eonomi bakgrounds as the LE and BIL hildren, and reported no history of speeh/hearing problems. They were divided into LE and BIL groups using the same riteria as above. Data from the LE and BIL groups of FB hildren were used to ompare eduational attainment levels with those obtained for LE and BIL hildren who stuttered. Classifiation as persistent/reovered Stuttering was reassessed when the hild was seen at age 12 plus by (a) the hild, (b) the aregiver and () a researher (the same for all hildren). All the sales used in these assessments have been normed (a statistial onept in psyhometris) (see supplementary data available online). The hild questionnaire and the researher report forms are given as supplementary material (the parent questionnaire was the same as the hild questionnaire exept that statements were hanged from the first to the third person). For the hild and aregiver, seven of the 15 questions on Boberg and Kully s questionnaire 4 were employed, and a further question was onstruted whih ombined three more of their questions. For eah of the eight questions, the hild or aregiver indiated the extent of agreement on a five-point sale. Eah question was sored 1 5, where 1 represented fluent behaviour and 5 dysfluent behaviour. The sores aross all questions were summed, the maximum sore being 40. Sores lower than 21 were onsidered reovered and sores greater than 21 were onsidered persistent. The sores on these questionnaires were orrelated with SSI-3 sores as a validation riterion (part of the norming proedure desribed in the supplementary data). This showed that these ut-offs divided stutterers at the low end of the moderate sale (stuttering ould not be moderate to have reovered and had to be at least moderate to be designated persistent). The researher visited eah hild s home and reorded an interview that lasted approximately 90 min. During this visit, the researher talked with a aregiver and the hild about the speeh problem and their experiene in the lini. He also sought their views about ommuniation style and selfonfidene in a range of typial environments. These inluded home and soial gatherings with adults and hildren in and out of shool. Performane and experiene in shool were assessed in terms of inter-personal relationships with staff and other hildren (inluding bullying). General health issues were also examined, inluding frequent absene from shool and hildhood illnesses. The researher subsequently assessed speeh flueny, soial-onversational skills, and whether the hild had a positive self-image/onfidene about speeh, using the reordings and notes taken at the home visit. Eah of the three assessments was sored on a sale of 0 (good) to 3 (poor). The sores for the three fators were summed to give one sore of between 0 and 9. A sore of 5 or above indiated still stuttering. To be designated as persistent, the aregiver, hild and researher all had to rate the hild as still stuttering. To be designated as reovered, the aregiver, hild and researher all had to designate the hild as not stuttering. All ases were unambiguously lassified on all three riteria. 5 8 All partiipants have been followed up for a minimum of 12 months and substantially longer in some ases (the mean length of follow-up for all speakers is 31.5 months with Arh Dis Child 2009;94: doi: /ad

3 Figure 1 The SSI-3 sores of the reovered and persistent groups over these two ages. an SD of 24 months). The reovered partiipants showed no relapse and none of the partiipants designated as persistent reovered during this period. The perentage of syllables stuttered out of the total syllables spoken in a 2 min reording of spontaneous speeh made on the seond oasion when the hild was aged 12 plus was used to further hek persistene/reovery. All reovered stutterers had fewer than 4% stuttered syllables, whereas all persistent stutterers had more than 4% stuttered syllables. Yairi and Ambrose reported that 3% stuttered syllables distinguished most speakers who stutter from fluent ontrols. 9 From this, the reovered stutterers an be onsidered to be lose to fluent whereas the persistent stutterers annot. Stuttering severity assessment (SSI-3) The SSI-3, a standard measure of stuttering severity, was administered when the hild was first seen and again at age 12 plus. 3 The assessments were always onduted on samples spoken in English, inluding samples from two hildren who did not stutter in English but did in their first language. This was beause SSI-3 is not available for those hildren s first languages and no norms exist. For SSI-3, a monologue, a dialogue and a text, all of whih ontained at least 200 syllables, were reorded using a Sennheiser K6 mirophone and a Sony DAT reorder. Assoiated physial onomitants suh as tis and twithes were noted. SSI-3 sores were obtained by qualified personnel. Statistial analysis SPSS 11.0 (release ; SPSS, Chiago, IL) was used for the desriptive analysis and for the non-parametri x 2 tests. For parametri measures either independent t tests (SATS performane ategories) or ANOVA (age at onset, SATS absolute sores and SSI-3 sores) were used for assessing differenes between groups. For the parametri measures, 95% onfidene intervals were alulated and p,0.05 was onsidered to be signifiant. Ethis approval Ethis approval was obtained from the UCL Committee on the Ethis of Non-NHS Human Researh (projet ref 0754/003). RESULTS Prevalene of LE and BIL in a sample of speakers who stutter The prevalene of a seond language in the sample was 69/317 or 21.8%, whih ompared to 28.4% of bilingual hildren in general reported by the London Eduation Authority. A goodness of fit between the obtained data and the estimates using 28.4% as the expeted value for bilingualism was not signifiant (x 2 =3.702, df = 1, p = 0.054). Thus, there does not appear to be a differene between the number of hildren who used a minority language in this sample and in London shools in general. Two (5.3%) of the 38 LE and BIL hildren stuttered in only one language. They were both BIL hildren who did not stutter in English. This is in agreement with reports that show that for bilingual speakers who stutter, it is usual to stutter in one language only. The number of LE hildren who stuttered (15) was lower than the number of BIL hildren who stuttered (23). However, when FB hildren were divided into LE and BIL groups using the same riteria, the reverse was true, with 28 LE and 10 BIL hildren. The assoiation between LE and BIL and flueny group (hildren who stuttered/hildren who were fluent) was signifiant (x 2 = 9.051, df = 1, p = 0.003). Thus, a relatively high proportion of BIL hildren stuttered ompared to those who were fluent. Table 1 Summary of statistial tests performed (A) Comparisons between LE and BIL speakers Stutterers M/F ratio % Stutter in one/ both languages Age at stuttering onset SATS SSI-3 LE, n = 15 4:1 0% 4 years 9 months Same as all peers Same as other stutterers BIL, n = :1 5.3% 4 years 10 months Same as all peers Same as other stutterers NS NS NS NS (B) Comparison between all bilingual stutterers and bilingual non-stutterers Prevalene of a seond language in respetive sample Bilingual stutterers 21.8% Bilingual non-stutterers 28.4% NS (C) Comparisons between stuttering and non-stuttering groups Inidene of LE and BIL Reovery rate of LE and BIL Stuttering sample Fewer LE than BIL Reovery rate of LE better than that of BIL Non-stuttering sample More LE than BIL Signifiant Signifiant BIL, bilingual from birth; LE, learned English at shool; M/F, male/female; NS, not signifiant; SATS, Standard Attainment Test eduational sore; SSI-3, Stuttering Severity Instrument, Third Edition. 44 Arh Dis Child 2009;94: doi: /ad

4 Individual profiles and family struture of LE and BIL hildren who stuttered ompared to monolingual ontrols who stuttered The groups of hildren who stuttered were ompared on other harateristis to determine whether there were any differenes between the LE, BIL and MONO hildren. Stuttering affets more males than females. 8 9 This was true of all three groups of hildren who stutter, with male/female ratios of 4:1, 4.75:1 and 4.43:1 for LE, BIL and MONO groups, respetively. A x 2 test showed that there was no assoiation between these groups and gender, indiating that all three groups had the same gender imbalane towards males. Stuttering usually starts early in development but some time after language onset. 8 9 The mean age at onset for the LE, BIL and MONO groups was 4 years 9 months, 4 years 10 months and 4 years 3 months, respetively. Again this orresponds with estimates obtained on other hildren in this age range, 8 although hildren younger than those examined here report earlier onset A between-groups ANOVA showed that age at onset of stuttering did not differ between the three groups. When onset and gender were ompared, LE and BIL hildren were similar to MONO speakers who stutter. Early shool performane of LE and BIL hildren relative to MONO and FB ontrols The aregivers reported SATS sores for English, mathematis and siene for the hildren at key stage one (age 7) and key stage two (age 11). By the age of 7 the hildren had not yet attended lini but they had had a minimum of 2 years shooling. SATS performane sores are published for eah key stage and are lassified as exeptional, beyond expetation, at the level expeted and below expetation for the partiular age group. Children were assigned to one of these ategories separately for eah subjet examined (English, mathematis and language). Although performane levels would be expeted to be lower for key stage one than for key stage two, the distributions an, nevertheless, still be ompared. This was done for the LE, BIL, MONO and FB LE and FB BIL groups. The distribution of performane levels of the hildren aross these groups was not signifiantly different by x 2 test (there was no assoiation between SATS performane ategory aross the five speaker groups, p = 0.865). Separate ANOVAs were onduted for English, mathematis and siene on absolute sores. For eah of these analyses there was one between-group fator (speaker group with five levels: LE who stuttered, BIL who stuttered, MONO who stuttered, FB who were LE, FB who were BIL) and one within-groups fator (assessment stage with two levels: key stage one, key stage two). For all three assessment types there was a signifiant improvement aross assessment stages (English, F(1,44) = , p,0.001; mathematis, F(1,44) = , p,0.001; siene, F(1,44) = , p,0.001) but no effet of speaker group. The interation between speaker group and assessment stage was marginally signifiant for mathematis (F(4,44) = 2.647, p = 0.046). This arose beause the FB who were LE sored higher in mathematis at key stage one but fell bak to the same level as the other speaker groups at key stage two. Severity assessment and reovery rates of LE and BIL hildren ompared to MONO ontrols SSI-3 severity estimates 3 were examined at first attendane and at age 12 plus. The hildren were first divided into those who reovered and those who persisted in stuttering at age 12 plus, as severity measures were expeted to differ aross ages for the two types. The SSI-3 sores were examined in a three-way ANOVA with two between-groups fators (fator 1, speaker groups, has three levels: LE, BIL and MONO; fator 2, reovery type, has two levels: persistent or reovered) and one withingroup fator (age with two levels: 8 10 and 12 plus). There was a signifiant derease over ages: F(1,44) = , p, There was also a signifiant main effet of reovery type, F(1,44) = , p,0.001, but no signifiant main effet of speaker group, F(2,44) = 0.186, p = There was also a signifiant interation between age and reovery type, F(1,44) = , p, Figure 1 shows this interation on SSI-3 sores for persistent and reovered stutterers at the two ages. The SSI-3 sores of the persistent groups were about the same aross the two ages (sores around 30 points).the SSI-3 sores of the reovered group dereased from about 25 points at the first test age to around 15 at the seond test age. Stuttering tends to persist in BIL speakers and there is a higher risk of suh hildren starting to stutter. The LE and MONO groups were ollapsed together beause the numbers in the LE group were low when these speakers were divided into persistent and reovered ases, and the patterns in the ollapsed groups were similar. There was a signifiant assoiation between speaker group (BIL and LE-MONO) and reovery outome (persistent/reovered): x 2 = 4.381, df = 1, p = The BIL group had a 25% reovery rate, whereas the LE and MONO groups together had a 55% reovery rate. These and other statistis are summarised in table 1. DISCUSSION The inidene of bilingualism in the linial sample was 21.8%. This was roughly omparable with the reported inidene of bilingualism in the same geographial area (28.4% of pupils in the area overed by London Eduational Authority are bilingual). Stuttering in one language by bilingual hildren is rare (only 5.3% of the sample of 38 bilingual hildren stuttered in just one of their languages). 10 The similar gender ratios (all around 4:1 for the LE, BIL and MONO groups) and the similar reported age at stuttering onset (around 4K years for all groups) supported the view that all groups stuttered. The BIL group was partiularly prone to starting to stutter. Thus, at the age at whih the hildren first attended lini, there were more BIL than LE hildren, whereas the reverse was the ase in a ontrol fluent sample. The statistis supported the view that there was a higher hane of BIL hildren starting to stutter as the assoiation between stuttering/not stuttering and BIL/LE was signifiant. This showed that BIL and LE groups were distributed differently in the stuttering, ompared to the non-stuttering, sample (60.5% of the stuttering group were BIL, whereas only 26.3% of the non-stuttering group were BIL). The BIL group also had a lower hane of reovery. The reovery rate at 12 plus was higher for the LE and MONO groups ombined than for the BIL group. The statistis supported the position that BIL hildren had a lower hane of reovery as there was a signifiant assoiation between the BIL versus the LE-MONO group and reovery outome. Inspetion of the data revealed that only 25% of the BIL group reovered whereas 55% of the LE-MONO group reovered. Together, these findings suggest that if a hild uses a language other than English in the home, deferring the time when they learn English redues the hane of starting to stutter and aids the hanes of reovery later in hildhood. A final fator of note is that shool performane was not affeted with respet to whether the hild stuttered or not. Arh Dis Child 2009;94: doi: /ad

5 Aknowledgements: We thank Mariam Abbas and Katie Gunn for assistane in data olletion. We also thank the liniians and their staff who gave freely of their time and assistane, and the families and hildren who partiipated in the study. Funding: The Wellome Trust, London, UK Competing interests: None. Ethis approval: Ethis approval was obtained from the UCL Committee on the Ethis of Non-NHS Human Researh (projet ref 0754/003). Patient onsent: Parental onsent obtained. Contributors: PH, SD, and RW were involved in the oneption and design of the study and interpretation of the data. SD and RW undertook reruitment and linial assessments. SD supervised aspets of data olletion. PH drafted the artile and all authors revised and approved the submitted version. PH is guarantor. REFERENCES 1. Stern E. A preliminary study of bilingualism and stuttering in four Johannesburg shools. J Logop 1948;1: Travis LE, Johnson W, Shover J. The relation of bilingualism to stuttering: a survey in the East Chiago, Indiana, shools. J Speeh Disord 1937;2: Riley GD. Stuttering Severity Instrument for Children and Adults, Third Edition. Austin, TX: Pro-Ed, Arhivist Paraetamol and atopi diseases 4. Boberg E, Kully D. Long-term results of stuttering treatment program. J Speeh Hear Res 1994;37: Davis S, Shisa D, Howell P. Anxiety in speakers who persist and reover from stuttering. J Commun Disord 2007;40: Howell P. Signs of developmental stuttering up to age eight and at 12 plus. Clin Psyhol Rev 2007;27: Howell P, Davis S, Williams SM. Auditory abilities of speakers who persisted, or reovered, from stuttering. J Flueny Disord 2006;31: Howell P, Davis S, Williams R. Late hildhood stuttering. J Speeh Lang Hear Res 2008;51(3): Yairi E, Ambrose, N. Early hildhood stuttering. Austin, TX: Pro-Ed, Nwokah EE. The imbalane of stuttering behavior in bilingual speakers. J Commun Disord 1988;13: Van Riper C. The nature of stuttering. Englewood Cliffs, NJ: Prentie-Hall, Mannson H. Childhood stuttering: inidene and development. J Flueny Disord 2000;25: Yaruss JS, La-Salle LR, Conture EG. Evaluating stuttering in young hildren: diagnosing data. Am J Speeh Lang Pathol 1998;7: Shenker RC, Roberts P. Long term outome of the Lidombe Program in bilingual hildren. In: Au-Yeung J, Leahy MM, eds. Proeedings of the International Flueny Assoiation 5th World Congress; July 2006, Dublin. Dublin: International Flueny Assoiation, 2007: Evidene has aumulated over the last few years that the use of paraetamol during pregnany, in hildhood and as an adult may inrease the risk of asthma. Phase three of the International Study of Asthma and Allergies in Childhood (ISAAC) has provided more data (Rihard Beasley and olleagues. The Lanet 2008;372: ; see also Comment, ibid: ). The study inluded hildren aged 6 7 years at 73 entres in 31 ountries on six ontinents. Information was olleted from parents or guardians by questionnaire. Around the world, most hildren were given paraetamol in infany. Multivariate analysis of data from hildren with ovariate data showed that paraetamol use for fever in infany inreased the risk of asthma symptoms, rhinoonjuntivitis and ezema at age 6 7 years by 46%, 48% and 35%, respetively. The inreased risk of asthma symptoms varied from 22% in the Indian subontinent to 82% in the Asia-Paifi region. The results were similar for boys and girls. The population attributable risk for asthma symptoms due to paraetamol use for fever in infany (data from 47 entres ombined) was 21%. For severe asthma symptoms, the population attributable risk was 22%. For rhinoonjuntivitis and ezema at age 6 7 years, the population attributable risks were 22% and 17%, respetively. Current use of paraetamol was assoiated with dose-dependent inreases in risk of all three onditions. With urrent use of paraetamol at least one a month, the risk of asthma symptoms at age 6 7 was inreased by a fator of 3.2, and by 61% with use one or more in the past year. The risks of rhinoonjuntivitis and ezema were also inreased with urrent use. This is the largest study of paraetamol and asthma yet reported, but onfounding fators, reall bias and reporting bias annot be exluded. Both these researhers and The Lanet ommentator all for a large randomised ontrolled trial. They agree that present evidene does not justify a hange in pratie. The biologial basis for a link between paraetamol and atopi diseases is unertain but redued ability to ombat oxidative stress and promotion of the T- helper-ell-2 pathway have been suggested. (NB. The September 20th, 2008 issue of The Lanet inludes four original papers and four review artiles, all on the subjet of asthma.) 46 Arh Dis Child January 2009 Vol 94 No 1