NIDCD Outcomes Research in Children with Hearing Loss
December 12 and 13, 2006
Following are the recommendations in each of the five areas (Outcomes Research—General Issues; Outcomes Research—Hearing Loss in Children Design and Child Factors; Sources of Variance: Family, Child, Treatment, Social; Complementary/Additional Research Needs; and Other Considerations). Consensus statements of participant discussions are incorporated as appropriate.
Section 1. Outcomes Research—General Issues
A general issue concerned funding mechanisms. Basic and applied research with strong theoretical rationales concerning biological, psychological, and/or social mechanisms relevant to outcomes might be funded via an R01 mechanism using CSR standing study sections. Applied research that is more narrowly targeted toward demographic or bigger picture clinical outcomes, however, might need alternate review mechanisms that allow a more focused assessment of clinical relevance and proficiency. A consensus of the Workshop was that both basic and applied research should be encouraged.
Another general issue concerned selection of appropriate experimental cohorts and control groups. Discussions concerned the rationales for targeted issues such as 1) characteristics of loss, i.e., degree, configuration, age of onset, stability, and bilateral or unilateral; 2) presence or absence of multiple handicaps, and 3) characteristics of appropriate control groups, e.g., if children with hearing loss have multiple handicaps, similar handicapping conditions in children without hearing loss are needed. A consensus of the Workshop was that initial research projects might benefit from focusing on outcomes in children with hearing loss without co-morbidity prior to attempting to understand outcomes in children with hearing loss in the presence of other complicating factors.
For national or multi-site clinical studies, a consensus of the Workshop is that experimental cohorts and control groups must be nationally representative. To the extent possible, existing longitudinal datasets should be scrutinized for their value in identifying important confounds and co-variates, sensitive screening markers, patterns of development after identification/intervention, and multidimensional measures that could yield a more sensitive characterization of outcomes “in the whole child” after identification/ intervention. A consensus of the Workshop was that sensitive characterization will require appropriate assessment of multiple dimensions across time.
Another general issue is how to ensure access to a sufficient population given the low incidence of children with hearing loss. This concern is related to the sample size necessary to achieve adequate statistical power. Studies concerned with descriptions of outcomes of hearing loss without regard to intervention will need to find ways of identifying children from population samples rather than clinical referral sources. This may require the development of sensitive screening markers that could be applied to large scale population surveys or the use of ongoing epidemiologic studies of child health where hearing screening is being done. In short, a representative and sufficient sample of children will require appropriate infrastructure. Some members of the Workshop thought the infrastructure of a system that provides comprehensive access to primary service providers might be helpful and were enthusiastic about the possibility of cooperating with State systems to address this possibility. Other members were more guarded in their assessment of the latter's potential. It should also be noted that some existing State systems may not allow researchers to identify children with hearing loss who are successful and don't require assistance.
Another general issue is that outcome studies of health services using observational methods rather than random assignment will need to incorporate a thoughtful model of intervening and confounding variables. Measures of these variables may be challenging. There is a current lack of measures of speech and language services that would reflect quality. A sensitive array of multidimensional variables and appropriate data analytic approaches are critical to the success of any outcome study.
A final general issue is that outcomes themselves need to be well rationalized and viewed within a complex causal structure. What are the multidimensional factors underlying outcomes in communication and quality of life? How is success defined and measured for outcomes? What is the impact of intervention on outcome as a function of child/parental/psychosocial factors and type of intervention, and how is quality of service quantified for different types of intervention? It should be noted that some outcomes such as improved speech and language development may be viewed as more proximal to the basic hearing loss whereas communication functions such as conversational participation or classroom performance may be viewed as more distal and dependent upon speech and language.
Section 2. Outcomes Research: Key Issues in Design and Child Factors
Design. Consideration of the following as relevant
- Targeted age range
- Age-dependent multifaceted outcome measures
- Focus on child and family
- Stability of outcome measures
- Definition of success
- Cohort size
- Quality of life
- Recruitment, retention, and loss to follow-up
- Statistical approaches
- Complexities and challenges of multi-center research including individual center compliance, timely submission of quality data, and methods of direct data entry
Child Factors Between Birth to 5 years. Consideration of subset of following as relevant
- What are the effects of hearing loss on development for children with and without other disabilities?
- What are the effects of socioeconomic and related factors, such as poverty, on outcomes?
- What are different developmental trajectories from birth to 3 years of age for language and cognition in children with varying levels of hearing and with unilateral vs bilateral losses compared to typical hearing children?
- What is the incidence, family-child characteristics, and developmental course for children with progressive hearing loss?
- What is the incidence, family-child characteristics, age of onset, and developmental course for children with delayed onset hearing loss?
Outcome Expectations: Birth to 3 years. Consideration of subset of following as relevant
- Joint attention, communicative gestures, words, symbolic play, vocabulary, word combinations
- Morphological inflections, conversational interactions, triadic attention
- Communication Sample (Computerized Language Analysis (CLAN))
- Developmental achievement
- Adaptive behavior
- Quality of life, parenting stress and impact on family
- Auditory skill development- parent report, cortical, behavioral
Outcome Expectations: 3–5 years. Consideration of subset of following as relevant
- Adequacy of amplification and consistency of use
- Syntax and grammatical morphology, discourse, spoken word recognition
- Developmental achievement, adaptive behavior
- Quality of life, parenting stress and impact on family
- Standardized tests, language sample analyses either in Computerized Language Analysis (CLAN) or Systematic Analysis of Language Transcripts (SALT): videotape or audiotape
- Unimodal and multimodal speech recognition
Outcome Expectations: School Age Children. Consideration of subset of following as relevant
- All of above for 3–5 years, with addition of tests for academic achievement and literacy
Section 3. Sources of Variance in Children with Hearing Loss: Family, Child, Treatment, Social
Multidimensional factorial and data analytic approaches should consider a subset of the following as relevant.
- Caregiver/parental interaction
- Home environment
- Social support
- Parental expectations and values
- Degree of hearing loss
- Age of onset
- Educational experience
- Auditory experience
- Auditory access
- Multiple disabilities
- Technology and approach
- Skill of provider
- Quality of program
- Quality of services
- Time course (age of identification/ intervention entry)
- Setting (childcare/preschool/ classroom/home)
- Parental compliance
- Frequency of specialized services
- Societal Expectations
Section 4. Addressing Research Needs and Gaps in Information
Research projects addressing the needs and gaps below are considered of high priority.
1. Developmental research focused on in infants and children with mild to severe hearing loss
- Taxonomic characterization of functional hearing status beyond the audiogram
- Refinement of audiometric and electrophysiological auditory responses
- Speech processing in presence of competing messages, particularly ones with semantic content
- Multi-modal perceptual- linguistic- cognitive development
- Multi-modal processing
- Foundational skill assessment of percept to symbol to concept continuum
- Attentional skill development
- Social-cognitive development
- Cognitive foundations of learning
- Joint attention
- Foundational skill assessment of gesture to symbol continuum
- Caregiver-child interactions (interpersonal communications as they relate to communication development)
2. Development of Measurement Tools for Children less than 36 months
- Perceptual cognitive development
- Borrow from effective paradigms in cognitive psychology to assess infant and early toddlers
- Develop valid and reliable measurement tools, both behavioral and physiological, for children less than 36 months
- Understand perception-production interrelationships within speech processing system
- Auditory Input and Access
- Auditory experience in quiet and in noise-corrupted, challenging, reverberant conditions
- Auditory signal input and effective device usage (wear time, caregiver placement and head control, classroom listening)
- Auditory signal input and assistive device (classroom listening, personal entertainment and their relevance, if any, for intervention (e.g., iPod, MP3) Managing discourse demands of multi-talker environments
- Effectiveness of amplification in unilateral hearing loss (e.g., Contralateral Routing of Signal (CROS))
- Assessment of Rehabilitative and Educational Intervention
- Refined taxonomy characterization of intervention to describe content and candidacy
- Multidisciplinary assessment of intervention—draw on examples from child language
- Assessment of outcomes with multivariate models—develop objective characterization of expected outcome that informs parents
- Parental support
- Engagement with intervention
- Level of parent/teacher support generated by the intervention
- Patient-reported, subjective measures and development of metrics
- Based on existing measures, such as those required by the Food and Drug Administration (FDA) HR-QoL (Hearing Quality of Life)
- Compliance with World Health Organization models of performance assessment
- Correlation with clinical metrics of performance
Section 5. Other Considerations
Standardize testing across agencies and intervention programs
All relevant parties should be encouraged to consider instruments (early language, speech, achievement) that could be recommended for administration at the state level. Considerations should include likely state participants and programs, data collections systems, individual vs. aggregate data, etc.
Facilitate access to data
- All relevant parties should be encouraged to create compatible data sets and to make current data sets available as allowed.
- Clinical researchers should be encouraged to consider attending Early Hearing Detection and Intervention (EHDI) meetings and developing relationships at the state level with Part C administrators.
- All relevant parties should be encouraged to consider mechanisms for following and tracking children across agencies (Health Research Services Administration, Center for Disease Control and Department of Education) over time (State EHDI and Part C birth–3 and Part B 3–21).
Additional considerations for randomized clinical trials and well designed quasi-experimental studies
- In addition to the NIDCD, the Department of Education Institute of Educational Sciences also supports and funds studies on intervention.
- All relevant parties should be encouraged to consider the desirability of common protocols.
- All relevant parties should be encouraged to consider measures similar to those used in other large data sets, such as Early Childhood Longitudinal Study, the Early Childhood Outcomes Study or the National Health and Nutrition Examination Survey.
Federal and State Government
All parties should continue to interact and coordinate at the national, state and local levels.