Nonverbal School-Aged Children with Autism

Goals of the Workshop

  • What do we know about this population?
  • What are the gaps in our knowledge?
  • What are the research opportunities that will address these gaps in the field?

Format

Presentations; Invited and General Discussions. Three major topics:

  • Who are these children?
  • How can we assess their skills and knowledge across different domains, with special reference to those abilities relevant to language acquisition? Novel measures; Sensory and Motor Skills; Apraxia
  • What treatments/interventions are effective in improving spoken language and communication in these children? Non-augmentative; Augmentative

Workshop Summary

1.  Who are these children?

Key points:

  • This is a highly variable population – there is no single set of defining characteristics or pattern of strengths and weaknesses;
  • It is a very significant challenge to assess these individuals – our current measurement tools have relatively low reliability and validity for this population;
  • Even having just one word or some echolalic speech may be a significant factor in predicting who might acquire some spoken language skills after age 5;
  • There is an important distinction between nonverbal, i.e. no receptive or expressive spoken language, preverbal i.e., younger children who have not yet responded to interventions; nonspeaking, i.e., not using speech as a primary means for communication and non-communicative i.e., no communication skills, even nonverbal or AAC.

Cathy Lord: Who remains nonverbal? What are the characteristics of this population?
There are multiple explanations for why some children with ASD never learn to speak. These explanations have emerged from prospective longitudinal studies, which show that there are cohort effects (more recent studies find a smaller percentage of children remaining nonverbal), that it is more common among children with ASD than other IQ-matched populations, but that even after age 5 some children do begin acquiring words, and in a small number of cases, increases in simple sentences. In adolescence, about 25% of nonverbal individuals with ASD show increased social withdrawal as measured on the ABC social lethargy scale. Factors predicting who remains nonverbal after age 5 may include:

  • Intellectual disability
  • Little desire or motivation to communicate
  • Poor socialization scores (Vineland)
  • Presence of challenging behaviors
  • Impaired joint attention
  • Impaired imitation of sounds and movements
  • Specific language and other motor factors

Barry Gordon: What are the characteristics and predictors of school-aged children who move from nonverbal to verbal?
A retrospective review of the literature on cases and studies of individuals who began speaking at age 5 or older (Pickett et al., 2009) highlighted the following factors:

  • Age (most children who started speaking were between 5 and 7)—none over age 13.
  • Some conceptual and semantic abilities.
  • Strong motivation to communicate via oral speech.
  • Specific training in the formation of sounds and words. Such training needs to be very intensive (often ABA approaches) and highly flexible, taking into consideration individual child abilities.
  • With intervention, 70% acquired the production of words, and only 30% developed phrase speech.

Although very few children begin to speak after middle childhood, there are a small number of these cases in the literature. Dr. Gordon presented evidence suggesting that there may be unpublished cases, and that progression to speech might be possible in more individuals than may be currently thought to be the case. Dr. Gordon and his group in the Neurology Department of The Johns Hopkins Medical Institution have one such case. This individual, A.I., was a non-verbal, low-functioning individual with autism when, at age 12, he was started on a full-time home-based intervention program. The intervention program was multi-faceted, initially focused on improving his conceptual abilities, then on his speech-articulation skills. Throughout, the program was highly flexible, and adjusted frequently over time to address A.I.’s changing goals and needs. A.I. began acquiring meaningful, voluntary sounds at age 13-14, recognizable single-word production at age 16, and now at age 22 uses spontaneous phrases and sentences (up to 7 words long) to communicate, including novel word sequences not previously taught. He continues to show signs of progress. Evidence that A.I.’s potential may not be unique comes also from studies comparing his comprehension abilities using methods that do not require explicit responses, in particular pupillometry, eye movements, and event-related potentials, and from extensive audio-video monitoring of other non-verbal subjects. A.I.’s comprehension abilities appear to be on par with those of other, currently non-verbal individuals with autism, as were, initially, some of his abilities with communicative intent and vocalic expression. Preliminary results suggest that these implicit measures of comprehension are valid in both verbal and non-verbal individuals, and therefore may be helpful adjuncts for assessments and therapeutic interventions.

Portia Iversen: What has been learned from the recent Autism Speaks workshops?
Videos of cases such as Tito, who communicates via computer, demonstrate the efficacy of alternative/augmentative communication (AAC) for some children. Key elements include the need to tap the individual child’s motivation and preferences (e.g., visual vs. verbal/auditory) and adapt approaches and the child’s basic behaviors to effectively use AAC devices. In some cases, basic skills like pointing or touching a keypad may need to be explicitly taught.

Conclusions from the AS workshops:

  • We know little about the nonverbal phenotype; there is no current standard terminology or taxonomy.
  • There is very little research on this portion of the ASD population; they are often the “non-responders” in a treatment protocol but are rarely defined any further.
  • A major barrier is the lack of a “best practices” approach to assessment. There is a clear need for developing reliable and valid assessment methods; measuring skills in all areas that are implicated in language/communication development.
  • It is often assumed that “nonverbal” equals “intellectual disability,” or lack of receptive language skills, but this assumption may be based on the lack of valid assessments for this population.

2a)  How can we assess these children’s skills and knowledge across different domains, with special reference to those abilities relevant to language acquisition?

Key Points:

  • Standard assessment approaches are not useful in most cases with this group of individuals. However implicit measures of behavior (using eye-tracking) or brain activity (using MEG or EEG/ERP) may reveal more about a child’s knowledge of the language and how he/she may differ from others in the way they process sounds and speech. Sensitive assessments may be critical in helping to guide treatment targets for this population.
  • Work on these newer methods is just beginning and there is a dearth of research demonstrating the validity or reliability of these novel approaches. Nevertheless, they hold significant promise for the future. Their use may require significant prior “training” of the individual with ASD in many cases (e.g., motion control).
  • Consideration needs to be given to extending the application of implicit behavioral and brain imaging methods to other key aspects of cognition that are known to be important for language acquisition (e.g., IQ; understanding intentionality).
  • Novel methods of assessment could be integrated into interventions – e.g., taking response to intervention (RTI) or dynamic assessment approach. This is especially important in considering brain measures, as it may well be that treatments lead to changes in the brain before they emerge in observable behavior.
You Might Also Read  Alternative Communication and Brain-Computer Interface Synergy

Presentations addressed the following questions: What is the methodology and what are its strengths? Why is it or why might it be useful for this population? What abilities can be assessed with this methodology?

Helen Tager-Flusberg: Eye-tracking
Numerous studies have found that assessment of receptive language often suggests that it is more significantly impaired in children with ASD than expressive language skills (and in comparison to other populations), but this might be a problem with standardized tests. Ongoing studies are using eye-tracking methods for assessing language skills (e.g., vocabulary, phrases, grammatical knowledge) in children with ASD. Initial data suggest that children with ASD look significantly longer at pictures that match a spoken word that they know (based on parent reports) – demonstrating the validity of this method; and that the measure is reliable (based on multiple assessment protocols). Eye-tracking measures also can assess online measures of language processing – i.e., time is taken to match words to pictures. Nonverbal children take significantly longer to shift their gaze to the matching picture, which may significantly impact their acquisition of language.

Current eye trackers have several key advantages:

  • They are completely non-invasive (no hat or goggles are required for eye-trackers that have infrared receptors embedded in the computer screen). The child needs to be able to sit (may include some restraint if needed) in front of the computer screen but without the need to remain perfectly still.
  • Calibration can be achieved with no verbal instructions needed.
  • The methods for assessment only require the child to look at the computer screen, which can be enhanced by using motivating visual/auditory attention grabbers that could be tailored to the preferences of individual children.
  • Eye-tracking measures could be used to assess not only receptive language, but also other skills relevant to language acquisition (e.g., nonverbal IQ, joint attention, etc.).

Nicole Gage: Magnetoencephalography (MEG)
MEG records the dipolar magnetic field produced by thousands of cells firing in synchrony – considered to be a direct signal of neural activity. It has excellent temporal resolution and good spatial resolution. Multiple dependent measures can be derived from MEG. Studies investigating auditory processing in ASD find prolonged latency in the M100 that doesn’t change with age and reduced dynamic range for feature contrasts (relative to age-matched controls) in the right hemisphere. Very unusual patterns were seen in a child with ASD and sensory reactivity.

Strengths and weaknesses for nonverbal children with ASD:

  • It is safe, non-invasive, and silent; the child can be accompanied by a parent.
  • Assessment can be conducted without explicit task requirements or compliance.
  • Individual subject analyses that discriminate between hemispheres allow for case study approaches.
  • Multiple regions and stages in processing can be acquired in one assessment.
  • But – it is expensive and not widely available.
  • Motion compliance is required – so can be quite challenging for some ASD individuals.
  • There are still methodological limitations in inferences that can be made about brain activity.
  • There is currently very little data on the use of MEG for assessing this population, especially over time, and in key domains that are relevant for language acquisition.

April Benasich: Event-related potentials
An ongoing feasibility study is exploring the adaptation of protocols developed for infants and children with other language disorders for use in nonverbal ASD children. The goal is to identify whether these children have the cognitive capacities and potential to understand spoken and/or written language. For this group (many of whom use some form of AAC), EEG/ERPs are used to assess language information processing to evaluate receptive skills in the absence of spoken language. Tasks include (1) lexical priming – to test word meaning and association. In TD children, mismatches between pictures and words elicit an N400 response – an implicit measure of processing a ‘difference’; (2) grammatical morphology – assessing noun-verb agreement(P600 over anterior sites); and (3) discourse semantics – discriminating grammatical vs. meaningless discourse (left vs. right hemisphere activity).

ERPs can be used to assess the child’s linguistic knowledge, however, a major goal of this project is to develop the methods and detailed procedures for evaluating the children’s individuals needs (e.g., reinforcers; operant approaches) for how best to teach them to tolerate the EEG net and complete the protocols. This method builds on substantial literature on ERPs to different aspects of linguistic knowledge.

2b) What are the sensory and motor areas to be assessed? What methodologies are appropriate?

Grace Beranek: Sensory abilities
Atypical responses to sensory stimuli are highly prevalent in ASD and show significant developmental changes. Three patterns have been identified: hyper-responsiveness; hypo-responsiveness, and sensory-seeking. These are not mutually exclusive and not exclusive to ASD, though hypo-responsiveness is more common in ASD and this type of behavior predicts best later social communication and language abilities. Preliminary analyses of an ongoing study dividing children on language outcomes showed that hypo-responsiveness and sensory seeking were higher in children who didn’t make gains in language.

There are important gaps in our knowledge:

  • No published studies focus directly on nonverbal school-aged children with ASD, so we know very little about this important area of functioning in this population; this reflects the bias toward studying younger and higher-functioning children, as well as a bias toward studying hyper-responsiveness.
  • There are significant methodological challenges. Most measures, which are almost all parent reports, were not developed for the ASD population, and it is hard to interpret the sensory preferences of the nonverbal child so we do not have valid and reliable instruments for the assessment of children at different ages and ability levels. Multi-trait, multi-method (including physiological and behavioral measures) approaches are needed.
  • Research in this area needs to address underlying mechanisms; heterogeneity; and variation based on developmental stage and context, and needs to incorporate multidisciplinary partnerships.

Mark Mahone: Motor skills
These are important to assess because motor skills provide insight (e.g., behavioral markers) into brain mechanisms that underlie language/communication. Motor skills encompass a range of measures – one key distinction in assessment is intentional vs. involuntary motor movements. Motor skills, including imitation of body movements, are impaired in ASD. Praxis is significantly correlated with symptom severity. There is a relationship between measured IQ and performance on standardized motor assessment protocols, and there are significant sex differences in developmental trajectories related to differences in brain maturation. Motor stereotypies are found in other populations but in ASD, hand/finger movements and gait deficits are more common.

Challenges and Gaps in Knowledge:

  • Standardized assessments depend on comprehension of verbal instructions or imitation.
  • Interfering behaviors may impact the assessment of motor functions.
  • Need to find reliable ranges of responses for each child (e.g., point, eye gaze).
  • There are no longitudinal studies charting developmental changes and maturational signs (e.g., overflow).
  • Should explore the use of alternative measures (e.g., EMG, eye-tracking) and assessment of movements that are not under conscious control (e.g., spontaneous movements).
You Might Also Read  Bronchitis; Causes, Symptoms, Diagnosis, Treatment

2c) What is verbal apraxia and how might it be assessed in this population?

Larry Shriberg: Apraxia
Apraxia is defined as impairments in the transcoding (planning/programming) of stored memory of encoded speech into an executed output. Childhood apraxia of speech (CAS) is one of several different types of speech sound disorders. It is domain-specific (i.e. independent of other apraxias), marked by delayed speech onset and deficits on challenging word repetition tasks. Research on CAS is now multi-disciplinary but plagued with the lack of clear operational definitions used across studies and poor differential diagnosis. Promising phenotypic markers specific to CAS include the production of vowels, stress, and rate–reflecting spatiotemporal accuracy and consistency in speech production. Analyses of speech from verbal children with ASD suggest that they may have (in some cases) motor speech disorder – NOS rather than frank CAS.

Challenges and Gaps in Knowledge:

  • For nonverbal children, there is very limited data. We need repertoires of speech using perceptual and instrumental methods for nonverbal children with ASD.
  • Analyses should focus on the segmental (vowel and consonant systems) and suprasegmental (prosody, laryngeal, and resonance values) aspects of speech collected in diverse environments/discourse contexts.
  • The use of dynamic assessments would be useful, including those evoked by emerging technologies (e.g., computer media and robots).

Discussion Key Points: John Connolly, Rhea Paul, Stewart Mostofsky; the entire working group

  • There are different approaches to the definition of apraxia – reflecting some differences between disciplines (e.g., neurology and speech/language pathology).
  • There is still little understanding of how sensory and repetitive/stereotyped behaviors directly interfere with language/communication in this population.
  • We don’t know whether one problem in nonverbal individuals is their ability to connect perceptual experience with action models.
  • Evidence for deficits in timing – e.g., speech; motor plans need to be further explored.
  • Methods for evaluating comprehension can be combined.
  • Major gaps in knowledge include – relations among different underlying factors associated with nonverbal ASD as well as the role of attentional (internal and external) factors at the level of the individual child.
  • Multiple areas of assessment are needed for each child to help define the population (vocal repertoire, receptive language; symbolic capacities; intentional communication/nonverbal means of communication; nonverbal cognition; sensory and motor skills; severity of social impairment; imitation; demographic variables; neural processing of basic visual and auditory/speech stimuli).

3. What treatments/interventions are effective in improving spoken language and communication in these children?

3a) Non-augmentative interventions

Presentations addressed the following questions: What is the treatment? What evidence is there for its efficacy?

Laura Schreibman: Behavioral Interventions
Applied behavioral analyses (ABA) encompasses classic discrete trial training (which, though effective, may be associated with lack of generalization, spontaneous use of language, or “robotic”‘ responding). Due to these issues, more naturalistic behavioral interventions developed (e.g., PRT, incidental teaching, milieu, etc.) which share common components (child-adult interactions in more loosely structured environments; child initiation and choice of materials; explicit prompting of target behavior; reinforcement for attempts to respond as well as correct responses). Predictive characteristics of children who are responders to PRT include those who are more engaged with objects, have some verbal behavior (including echolalia), have more approach, and have less avoidance behavior. Ongoing randomized controlled trials (RCT) with preschoolers comparing PECS (visual approach) with PRT (verbal approach) suggest that if a child had even 1 word before treatment, they made progress – no major differences between approaches.

Gaps

  • Why is there heterogeneity in response to treatment? Need to individualize treatment approaches to child characteristics and specific aspects of language/communication skills.
  • Increased flexibility in using multiple protocols to achieve more comprehensive outcomes.

Connie Kasari: Preverbal, Nonverbal? The Effects of Intervention
There is a key need to develop targeted interventions, with special attention paid to the content of interventions that should address underlying skills. Toddler/early intervention studies have demonstrated the efficacy of interventions that target joint attention skills or social engagement/symbolic play. Both led to significant increases in language, though joint attention was more effective for children with fewer than 5 words.

An ongoing multi-site CCNIA intervention is being conducted by Kasari, with Landa and Kaiser comparing two interventions (joint attention+milieu therapy and joint attention+AAC) in 96 5-8-year-old nonverbal children with ASD (minimal nonverbal cognitive mental age on the Leiter of 24 months) who have already had a least 2 years of intervention. A sequential multiple assignment randomized trial (SMART) design is being used to allow for more flexible changes in intensity or treatment assignment after 12 of the 24 weeks of intervention.

Challenges:

  • Establishing fundamental joint attention and social communication skills, especially when the child is not engaged with the materials and/or exhibits significant interfering sensory or other avoidance behaviors.
  • How to measure the meaningful change in spoken language (responding vs. initiating; beyond single words; functions expressed; rate of speaking and turn-taking).
  • Measuring effectiveness beyond spoken language.

3b) Augmentative interventions (AAC)

Presentations addressed the following questions: What is the treatment? What evidence is there for its efficacy? Can speech and literacy result from the use of augmentative devices? Can useful communication via augmentative devices be obtained for this population?

Nancy Brady: AAC for children with autism
AAC encompasses any non-speech means for expressive/receptive communication. Research dating back to the 1970s has demonstrated the ability of children with ASD to learn different forms of AAC (PECS is the most popular) using both single-subject and randomized group designs. Current reviews suggest that PECS can be effective in improving communication and perhaps also decreasing challenging behaviors, but there is little evidence for maintenance and generalization, and limited effects on speech.

An ongoing observational study of preschoolers in classrooms and homes is exploring the use and predictors (child and environmental factors) of success with AAC (mostly PECS; some sign and/or speech generating device). The majority of the children who were successful users also had some speech (1-5 spoken words), which accompanied their AAC communications. Survey data indicate very limited use of AAC either in classrooms or at home and a lack of use in peer interactions.

Challenges:

  • Need for rigorous, well-controlled studies with longer-term behavioral outcomes.
  • Research on empirically validated assessment and staff/family training processing.
  • Research on promoting generalized AAC use at home and in schools with a variety of communication partners.

MaryAnn Romski: Augmentative language interventions for school-aged children with autism: Speech-generating devices (SGD)
SGDs can provide nonverbal children with ASD a means by which they can communicate, augment existing speech to increase intelligibility, and teach language.  Speech-generating technologies can compensate for use of a primarily visual AAC; provide speech feedback to conversation partners; may use different response modes (direct contact; visual scanning) and can bridge to independent speech. The System for Augmenting Language (SAL) long-term research program demonstrates different developmental patterns (slow, particularly children with lower comprehension skills, and more rapid).  Now, this has been extended to younger children (toddlers), comparing speech, AC as input, and AC as output using a  parent coaching model embedded into routines.  Preliminary findings from a retrospective analysis suggest that for the children with ASD, AC as output led to the greatest increases, including symbol-infused joint engagement.  This intervention can transition children to modest spoken language skills, though research on older children is needed.

You Might Also Read  MND; Causes, Symptoms, Diagnosis, Treatment

Key Points/Discussion: Rebecca Landa, Janice Light, and the entire working group

  • There is a wide range of AAC interventions including those that use unaided AAC systems (e.g., signs) and those that use aided AAC systems (e.g., PECS, communication boards, high-tech speech-generating devices using various language representations and outputs). Research is required to determine which interventions work best with which individuals and conditions to achieve desired outcomes.
  • Research on AAC generally has not provided detailed descriptions of participants; relies on single-subject designs; most of the participants are younger than 10; too often requesting is the focus of teaching; little is known about long-term effects.
  • Emerging literature suggests benefits on social interaction and behavior; can also be used to teach literacy.
  • AAC does not preclude the possibility of spoken language gains, based on current studies.
  • Question of whether developing language can lead to literacy or vice versa in this population.

General Discussion of Workshop Goals

Very little research over the past couple of decades has directly investigated the nonverbal school-aged child with ASD; almost none have been included in research studies. This has left us with a dearth of knowledge about who these children are, what their underlying skills and impairments are, and how we might be able to provide interventions that support the development of fundamental communication skills.

Recommendations:

  • A working group should be established to develop a set of recommended measures and benchmarks for nonverbal school-aged children with ASD that can be used for providing basic descriptive information about participants in research studies (including treatment/intervention research) as well as potential outcome measures. This will also facilitate defining this population on multiple dimensions. In some cases, creative solutions for assessment may be needed for children who do not readily comply with standardized approaches and dynamic assessment procedures should also be considered. Domains to be considered include:
    • Vocal repertoire including echolalic speech
    • Intentional communication skills
    • Level of language/communication, including both receptive and expressive in speech and non-speech modes
    • Imitation skills
    • Object play skills
    • Joint attention skills
    • Nonverbal IQ
    • Social impairment severity
    • Other behaviors (e.g., sensory; atypical; motor)
    • Communicative environment
    • Communication partners
  •  Identified areas of gaps in research/targets for future investigations
    • Studies investigating the gap between the children’s measured behavior and cognitive abilities and knowledge that is not readily apparent on standard assessments. Development of novel methods for assessing cognition using a range of neural and behavioral approaches for this population.
    • Studies on the underlying mechanisms explain why these children do not acquire spoken language by the age of 5, even when provided with ample standard intervention opportunities. Potential areas of investigation include control over voluntary vs. involuntary motor skills (oral-motor skills); processing auditory/speech stimuli; social attention mechanisms; fundamental impairments in intentional communication etc.
    • Comprehensive treatment studies focusing on this population should address active ingredients of intervention (children have mostly not responded to current best practice interventions; thus still not speaking at age 5 years). An important active ingredient may be the content of the treatment (communication and adjunct skills); a range of tools and techniques; studies in the home, clinic, and schools; environmental context factors; use of creative research designs such as RTI and sequential treatment designs; profile-based treatment studies, etc.; incorporating novel implicit and brain imaging measures to assess changes that may occur before observable behavior changes.
  • There is currently nothing in the literature on this population, which suggests there is a need for a comprehensive review of the state of knowledge (perhaps based on the workshop). This review would summarize what we know about these children, what we do not know, how we can evaluate them, and potentially effective treatments and treatment designs.

References

 

1www.nidcd.nih.gov/health/statistics/quick-statistics

2 Collins, JG. (1997) Prevalence of selected chronic conditions: United States 1990-1992. National Center for Health Statistics. Vital Health Stat 10(194).

3www.nidcd.nih.gov/health/statistics/text-description-use-hearing-aids-2006

4 Davis, A, Smith, P, Ferguson, M, Stephens, D, and Gianopoulos, I, (2007). Acceptability, Benefit and Costs of Early Screening for Hearing Disability: A Study of Potential Screening Tests and Models. Health Technol Assess. Oct; 11(42): 1-294.

5 Senate Report 110-410, page 111, Report of the Committee on Appropriations, U.S. Senate, on S. 3230 (making appropriations to the Departments of Labor, Health and Human Services, and Education, and Related Agencies Appropriation Bill, 2009).

6 http://www.healthypeople.gov/hp2020/

7http://www.census.gov/prod/2009pubs/p95-09-1.pdf

8http://www.statistics.gov.uk/pdfdir/iahi0809.pdf

9 Kochkin, S. (2007). MarkeTrak VII: Obstacles to Adult Non-User Adoption of Hearing Aids. Hearing Journal. Vol. 60, No. 4: 24-50.

10http://www.hearingreview.com/2005/09/why-are-hearing-instruments-so-expensive/  Source for second statistic is no longer available online (verified January 2019).

11 Kochkin, S. (2009). MarkeTrak VIII: 25-Year Trends in the Hearing Health Market. Hearing Review. Vol. 16, No. 11: 12-31.

12 Kirkwood, D. (2009). Despite Challenging Economic Condition, Practitioners in the Survey Remain Upbeat. Hearing Journal. Vol. 62, No. 4: 28-31.

13http://www.consumerreports.org/health/healthy-living/home-medical-supplies/hearing/hearing-aids/overview/hearing-aids-ov.htm

14https://www.maaudiology.org/

15www.nidcd.nih.gov/health/statistics/quick-statistics

16 US Census Bureau figures; link no longer available.

17 Davila, EP, Caban-Martinez, AJ, Muennig, P, Lee, DJ, Fleming, LE, Ferraro, KF, LeBlanc, WG, Lam, BL, Arheart, KL, McCollister, KE, Zheng, D, and Christ, SL, (2009). Sensory Impairment among Older U.S. Workers. Am J Public Health, 99:1378–1385. doi:10.2105/ AJPH.2008.141630

18http://www.ahrq.gov/QUAL/qrdr08.htm#toc

19http://www.ahrq.gov/QUAL/qrdr08.htm#toc

20 Margolis RH, Saly GL, Le C, Laurence J. (2007). Qualind: A method for assessing the accuracy of automated tests. J Am Acad Audiol 18, 78-89.

21 Smits C, Houtgast T (2006). Results from the Dutch speech-in-noise screening test by telephone. Ear Hear 26, 89-95

22 Davis, A, Smith, P, Ferguson, M, Stephens, D, and Gianopoulos, I, (2007). Acceptability, Benefit and Costs of Early Screening for Hearing Disability: A Study of Potential Screening Tests and Models. Health Technol Assess. Oct; 11(42): 1-294.

23 Estimated from: Davis, A, Smith, P, Ferguson, M, Stephens, D, and Gianopoulos, I, (2007). Acceptability, Benefit and Costs of Early Screening for Hearing Disability: A Study of Potential Screening Tests and Models. Health Technol Assess. Oct; 11(42): 1-294; Personal communication, Howard Weinstein, August 2009.

24 http://www.hearingreview.com/issues/articles/2007-04_01.asp

25 Krumm, M. (2007). Audiology telemedicine. Journal of Telemedicine and Telecare,13 (5), 224-229.

26 Margolis RH, Morgan DE (2008). Automated pure-tone audiometry: an analysis of capacity, need, and benefit. Am J Audiol 17, 109-113

27 Mehrotra, A, Liu, H, Adams, JL, Wang, MC, Lave, JR, Thygeson, NM, Solberg, LI and McGlynn, EA (2009). Comparing Costs and Quality of Care at Retail Clinics with that of Other Medical Settings for 3 Common Illnesses. Annals Internal Medicine. 151: 321-328.

28 Nachtegaal, J, Smit, J, Smits, C, Bezemer, P, van Beek, J, Festen, J, and Kramer, S. (2009). The Association between Hearing Status and Psychosocial Health Before the Age of 70 Years: Results From an Internet-Based National Survey on Hearing. Ear & Hearing, Vol. 30, No. 3, 302–312.