Overview. Iowa City, IA ACKNOWLEDGEMENTS. Risk and Resilience in Children who are Hard of Hearing. Why study children who are hard of hearing?

Transcription

1 Risk and Resilience in Children who are Hard of Hearing Iowa City, IA Elizabeth Walker, PhD, CCC SLP/A Assistant Professor University of Iowa, Iowa City, IA Supported by NIDCD R01 DC EDUCATIONAL AUDIOLOGY ASSOCIATION JUNE 22, My history Master s degree in Communication Disorders from University of Minnesota Twin Cities CFYs in Audiology and Speech Language Pathology at Riley Children s Hospital/Devault Otologic Lab at Indiana University PhD in Speech and Hearing Science from University of Iowa 2008 Outcomes of Children with Hearing Loss (OCHL) 2013 Outcomes of School Age Children who are Hard of Hearing (OSACHH) 2013 Complex Listening Skills in School Age Hard of Hearing Children ACKNOWLEDGEMENTS BTNRH: Mary Pat Moeller, Sophie E. Ambrose, Merry Spratford, Ryan McCreery U of Iowa: Bruce Tomblin, Jake Oleson, Lenore Holte U of North Carolina, Chapel Hill: Melody Harrison, Tom Page Overview Why study children who are hard of hearing? Intro to OCHL Results Hearing aid use Aided audibility Research on hard of hearing children is rare. The greatest amount of research has been conducted on deaf children although the results of such studies are useful for understanding the effects of profound hearing loss, they are not applicable to hard of hearing children. Julia M. Davis Our Forgotten Children

2 There are many challenges with past research on children who are hard of hearing Children who are deaf or HH combined into one group ~15% of children ages 6 19 years have a significant hearing loss Small sample sizes NHANES II & III (Niskar et al., 1998) What guided the research goals of this multicenter study? Lack of prospective studies New generation of children who are hard of hearing Did not take hearing aids into account 10 Are these children at risk for delays? What are factors that lead to success? Goal of OCHL: to address research gaps Goal: Explain individual variability HH EHDI CI UNHS Disability Speech Language 2

3 New practices new outcomes? Previous outcomes model Universal newborn hearing screening (UNHS) Birth to three early childhood education programs Technological advances in amplification Degree of HL (PTA) Outcomes Historical Perspective: Ambiguity about risk posed by Note: mild to severe Small sample sizes School age children hearing Little loss consideration of Delayed relative to peers n Age (yr) malleable factors in clinical intervention Like typical peers n Age (yr) Davis et al. (1986) Briscoe et al. (2001) Elfenbein et al. (1994) Gilberston & Kamhi (1995) Blair et al. (1985) Norbury et al. (2001) OCHL outcomes model: cumulative auditory experience Degree of HL (PTA) Audiological Intervention Audibility Hearing aid use Outcomes Delage & Tuller (2007) Wolgemuth et al. (1998) Halliday et al. (2017) Halliday et al. (2017) Even mild HL has consequences Persistent risks in speech, grammar Many = hearing peers and > Language Disorders Selected children impaired Resilience in vocabulary Malleable factors that influence relationship between PTA and outcomes. Who are the OCHL participants? Study participants Inclusion criteria SUBJECTS TOTAL HH 317 NH months to 7 years at entry English primary language No major secondary disabilities No cochlear implants Permanent mild to severe bilateral hearing loss 3

4 Distribution of Better Ear Hearing Degree of hearing loss (Better ear pure tone average, BEPTA) 19 HH NH Matched sample HH NH Number of subjects Hearing (PTA) db HL < 20 db HL Age ranges Nonverbal IQ Maternal education Language use Additional disabilities 0;6 to 7;3 at entry Within the average range Matched but > US sample Spoken English in the home No autism; no major vision, cognitive, or motor disabilities Caveats for OCHL Socioeconomic status of the sample Language Outcomes background may appear English better only than Additional they disabilities would be Excluded in the general Cochlear implants Only population pre implant data We used an accelerated longitudinal design Each child followed for 3+ years. Retrospective data prior to enrollment obtained by medical records. We used an accelerated longitudinal design We conducted onsite testing of children with parents completing questionnaires BTNRH and UNC Chapel Hill tested at their medical centers (with addition travel for out of state participants) Iowa tested in vans equipped as mobile testing units Each child followed for 3+ years. Retrospective data prior to enrollment obtained by medical records. LENA project Language ENvironment Analysis Subset of 41 children who are HH and 17 children with NH, ages 12 months to 3 years Followed one day per month for one year 4

5 We also collected information about service provision Annual telephone questionnaire with parents Audiology service provider survey (online) Service provider survey (online) Birth to three Preschool School age Teacher survey Preschool School age Medical records ENT & pediatrician Overview Results Intro to OCHL OCHL outcomes model: Risk and resilience Language scores WNL for test norms Degree of HL (PTA) Outcomes Tomblin et al., 2015 Significant effect for degree of hearing loss. Trend for language scores to increase with age across all groups, but no difference in growth rates over time How do children who are hard of hearing compare to children with normal hearing? OCHL outcomes model: Risk and resilience Conclusion: Children who are hard of hearing are at risk for language delays..... Degree of HL (PTA) Form Content Outcomes Use * p <.0001 CHH differed significantly from SES matched age mates. Tomblin et al., Ear & Hearing, Limited perceptual salience + input frequency 5

6 Here is a profile of relative strengths 31 and vulnerabilities at 3 yrs Moderate & Mod Severe (> 45 db HL) Mild hearing loss (25 45 db HL) n = 154 CHH; 69 CNH Risk and resilience: Content areas Age = 4 years Semantics Vocabulary Morphology Scores on a task that elicited production of word endings BASIC CONCEPTS (vocabulary) SYNTAX PRAGMATICS SPEECH PRODUCTION Morphology has a specific relationship with hearing beyond that found for semantic scores. Conclusion: CHH show differential areas of vulnerability in language development Tomblin et al., 2014 Tomblin et al. (2015) 32 Hearing loss effects processing of subtle acoustic cues important for morphosyntax What about the influence of cumulative auditory experience? Bound morphemes, especially in verbs, are less salient and less frequent in the input Typically sentence medial (He needs to find ) Often involve fricatives in English Complex phonetic contexts (It s, Greg s calling ) Audibility Hearing aid use Auditory experience 33 Two questions: How do we measure? Do we see variation in auditory experience? Overview Aided audibility Results Intro to OCHL 6

7 How well do CHH access linguistic input with HAs? HA We hypothesize that access to speech will predict success for children who are HH Most previous research looked at audiometric thresholds (i.e., PTA) as a predictor of success What does PTA not tell us? Inaudible PTA = 40 db HL What is audibility? How much speech a child can hear Quantified with the Speech Intelligibility Index (SII) 25 How different configurations may impact speech perception How child will perceive speech with hearing aids (aided audibility) SII=0.0 not audible SII = 1.0 fully audible How do we quantify aided audibility?...speech Intelligibility Index (SII) (or count the dots ) How do we quantify audibility?...speech Intelligibility Index (SII) Each dot represents ~1% of information contributing to speech clarity Number of dots that are audible predict how well one understands quiet speech from a six foot distance Dots are unevenly distributed: 1000 and 3000 Hz > 250 to 500 Hz Not all frequencies are created equal some contribute more than others to the intelligibility of speech. Above line = dots inaudible; below line = dots audible More dots=more important Less dots=less important 7

8 PTA does not always = SII SII takes configuration and importance of speech cues into account 3F PTA = db HL SII =.25 3F PTA = 35 db HL SII =.40 For each band: Audibility x FIW = weighted audibility SPL o gram SII Snapshot SII = Sum of weighted audibility of all frequency bands Aided SII Unaided SII How much SII is enough? Depends on age Adults Older (9 12 years) Younger (5 8 years) How much SII is enough? Depends on degree of loss and input level From McCreery & Stelmachowicz, 2011 UWO PedAMP Protocol, 2010 Which is better? Are HAs fit appropriately? Speech intelligibility index Measured with speech ,000 Hz Unaided or aided Reflects configuration Quality of hearing aid fitting Calculated automatically Pure tone average Measured with pure tones Hz/4000 Hz Unaided or aided* Blind to configuration Does not reflect quality of hearing aid fitting Calculated manually *Limitations of aided pure tone average to be discussed Examined HA fitting by measuring SII for 288 children The characteristics of hearing aid fittings in infants and young children (McCreery, Bentler, & Roush, 2013) McCreery et al., 2013; 2015 Compare DSL target SII to measured SII 0 1, with 1 = completely audible RMS error to DSL target at 4 frequencies RMS error < 5 db = optimal HA fitting 8

9 Are HAs fit appropriately? Target vs. Actual REAR (RMS error) The characteristics of hearing aid fittings in infants and young children (McCreery, Bentler, & Roush, 2013) Compare DSL target SII to measured SII 0 1, with 1 = completely audible RMS error to DSL target RMS error < 5 db = optimal HA fitting RMS error < 5 db is a good fit Fitting data compared to DSL targets Calculate RMS error of deviations from target at 5., 1, 2, and 4 khz SII = speech intelligibility index RMS = root mean square How can you measure RMS error? Confidence intervals for SII when hearing aids are fit appropriately Below dashed line = poor fit For PTA of 50 db HL, SII could range from ~ Bagatto, et al., 2011 Actual hearing aid fit quality What impacts quality of fitting? Below dashed line = poor fit Take home message: Hearing aids are not appropriately fit for all children 35% below normative mean, 10% below 95% confidence intervals McCreery et al., 2015 Audiologist s degree Level of specialization with children Online survey Techniques for HA verification 9

10 Accuracy of Verification methods Probe microphone real ear measures RMS error= 5.67 db (SD = 3.95 db) Functional gain (aided soundfield) RMS error=7.92 db (SD = 4.67 db) Functional gain: Not functional, nothing to gain Audiogram with hearing aids is NOT verification. No information about speech audibility. Cannot assess maximum output. Represents a stimulus and level that are not encountered by children. No estimation of advanced features McCreery, Bentler, Roush, 2013 Audibility Conclusions Fitting to in close proximity to prescriptive targets provides consistent audibility Greater degree of hearing loss associated with poorer audibility 35% of the children in the study had poorer audibility than predicted by SII normative mean (from Bagatto et al., 2011). 10% were below the 95% confidence intervals of the normative mean. Does it matter if hearing aids are optimally fit? Degree of HL (PTA) Davis et al. (1986) Elfenbein et al. (1994) Norbury et al. (2001) Audiological Intervention Aided Audibility Outcomes Delage & Tuller (2007) Halliday et al. (2017) Aided Audibility Contributes to Language GROWTH Quartiles of Aided Benefit, after controlling for degree of loss Benefit holds for mild to severe degrees of HL Tomblin, Harrison, Ambrose, Walker, Oleson, & Moeller, E&H (2015); McCreery et al., E &H (2015) 59 Children who receive the most benefit from HAs show steeper growth in language skills 10

11 Overview Do children who are HH vary into the amount of time they wear their HAs? Hearing aid use Which factors predict daily HA use time in children who are hard of hearing? Results Aided audibility 272 children with hearing aids How consistently do children wear HAs in different settings? Intro to OCHL Walker et al., 2013 Are parents accurate at estimating average daily hearing aid use time? How did we measure amount of daily HA use? Measuring hearing aid use consistency Subjective Objective Put an X in the boxes below to indicate how consistently your child uses HAs in the situations listed: Situation Never (0) Rare (1) Sometimes (2) Often (3) Always (4) N/A Car Pre School/School Day Care Hearing aid questionnaire average # of hours per day consistency of use across contexts: in the car, meal times, book sharing, etc Hearing aid data logging Meal Time Playing Alone Book Sharing Playground Public (store, zoo) Walker et al., What factors predict the amount of time children wear HAs on a daily basis? Predictors pure tone average gender recruitment site (Iowa, Boys Town, UNC) maternal educational level Outcome Parent report of daily hearing aid use What factors predict the amount of time children wear HAs on a daily basis? Parent report of use (hours) Weekday Weekend Hours of Use (n=272) Chronological age Age (years) 11

12 Datalogging by age groups Consistency of use is lower for children with milder degrees of hearing loss Walker, et al 2015 Walker et al, 2013 Longitudinal trends in HA use: Infants and toddlers Longitudinal trends in HA use: School age children Routine users: Mothers with a bachelor s or postgrad degree Routine users: Mothers with a bachelor s or postgrad degree Worse hearing loss (higher BEPTA) Limited users: Mothers with HA school education Limited users: Mother s with some college Milder hearing loss (lower BEPTA) Walker, et al., 2015 Walker, et al., 2015 Are parents accurate at estimating daily hearing aid use time? n = 133, r =.76 Parent report = hours Data logging = 8.3 hours As children get older, parents become more accurate reporters ages Underestimated Average difference = 2.6 hours As children get older, parents become more accurate reporters Overestimated Walker et al, 2013 Walker et al.,

13 Hearing Aid Use Conclusions Younger children and children with milder hearing losses wore HAs less consistently than older children and those with more severe hearing loss. Parents overestimated the amount of time their child wore their HAs, but the correlation between parents estimates and datalogging was very high. Clinicians may rely on parental self report of HA use time as a general estimate of how much the child wears HAs. Caveat: HA datalogging and consistency ratings are preferred with parents of younger children when monitoring HA compliance. Does it matter if hearing aids are worn consistently? Degree of HL (PTA) Audiological Intervention Hearing aid use Outcomes Consistent HA Use Benefits Growth Tomblin, Harrison, Ambrose, Walker, Oleson, & Moeller, E&H (2015) 75 Children who wear HAs during all waking hours show steeper growth in language skills Tomblin et al., 2015 Special populations: Are there differences in outcomes for children with mild hearing loss, as a function of amount of hearing aid use? Vocabulary Articulation Grammar Phonological processing Speech recognition in noise Situation of clinical equipoise regarding benefits of HAs for children with mild hearing loss Current evidence base Children with mild HL are at risk for delays (Bess et al., 1985; Blair et al., 1985; Davis et al., 1981; Yoshinaga Itano et al., 2008) Other studies show minimal impact of mild HL on outcomes, with ambiguity re. HA benefit (Porter et al., 2014; Wake et al., 2006) Limitations of past studies Mild often grouped with unilateral/minimal HL OR moderate to severe/profound Most studies do not describe influence of both aided audibility and amount of daily HA use on outcomes However, >33% of children with mild HL do not wear HAs consistently (Fitzpatrick et al., 2010) Does consistency in HA use matter?

14 HA use groups n= Average HA use (hrs) Significant differences: Better ear pure tone average (nonusers > parttime, full time) Benefits of Consistent Hearing Aid Use Extends to Children with Mild Losses Full time (>8.7 hrs) Part time (2 8.3 hrs) Nonusers (<2 hrs) No significant differences between the three groups: maternal education levels nonverbal IQ level of audibility 1 SD 2.5 SD 79 Walker et al., JSLHR (2015) There were no significant differences between groups for articulation Form Morphology Phonological processing Content Use Cumulative auditory experience affects structural aspects of language development 81 Walker et al., 2015 There were no differences between groups for speech recognition in noise What are the implications? Percent correct Phoneme +10 SNR Phoneme 5 SNR Word +10 SNR Phoneme 5 SNR CASPA condition Non users (n = 4) Part time (n = 8) Full time (n = 6) Take home message: Do not rely solely on audiological outcome measures to determine benefit from HAs Traditional word recognition tests may not be sensitive to individual differences for children with mild hearing loss Protective factors include: early and consistent use of well fit hearing aids 14

15 Complex listening Project 1 st and 3 rd grade Areas of interest: Classroom acoustics measurements Perception performance in noisy/reverberant conditions Subjective ratings of ease of listening in classroom and home NIH NIDCD R01 DC OSACHH study Coming soon! 2 nd and 4 th grade Areas of interest: Literacy and academic (math, spelling) achievement Advanced social cognition Working memory, executive function NIH NIDCD R01 DC

16 one page results summary Thank you! elizabeth Free access to OCHL supplement in Ear & Hearing Pediatric Audiology Biostatistics, Linguistics, & Psychology Questions? Child Language Project Management 94 16