Plots show a significantly shorter time between NHS and ABR testing in the patients with private insurance or patients of white race. The wider range of outliers for the patients with public insurance and patients with race/ethnicity other than white indicates that many of these children are experiencing lengthy delays and longer intervals of loss to follow-up. Horizontal lines inside boxes indicate medians; top and bottom edges of boxes, 75th and 25th percentiles; error bars, 1.5 multiplied by the interquartile range; and circles beyond the error bars, data outliers.
aIndicates a statistically significant difference between patients with public vs private insurance or between white vs other (nonwhite) race/ethnicity.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Zhang L, Links AR, Boss EF, White A, Walsh J. Identification of Potential Barriers to Timely Access to Pediatric Hearing Aids. JAMA Otolaryngol Head Neck Surg. 2020;146(1):13–19. doi:10.1001/jamaoto.2019.2877
What demographic, socioeconomic, and clinical factors are associated with pediatric access to hearing aids?
In this retrospective cohort study of 90 pediatric patients diagnosed with hearing impairment and fitted with hearing aids, insurance type, race/ethnicity, and primary language appeared to be significantly associated with the timeliness of pediatric access to hearing aids, with additional associations observed with the interval from newborn screening to auditory brainstem response testing.
Identifying potential barriers to timely access to pediatric hearing aids may further the discussion of policies to eliminate disparities in access for at-risk children.
Despite various barriers identified to early pediatric access to cochlear implantation, barriers to timely access to pediatric hearing aids are not well characterized.
To identify socioeconomic, demographic, and clinical factors that may be associated with pediatric access to hearing aids.
Design, Setting, and Participants
This retrospective cohort study included 90 patients aged 1 to 15 years who were referred for auditory brainstem response (ABR) testing and evaluation for hearing aids at a single tertiary care academic medical center from March 2004 to July 2018. Children who did not receive both ABR testing and hearing aids at the same center were excluded from analysis.
Main Outcomes and Measures
Associations of insurance type (private vs public), race/ethnicity (white vs other), primary language (English vs other), cause of hearing loss (complex vs not complex), zip code, hearing aid manufacturer, and severity of hearing loss (in decibels) with the duration of intervals from newborn hearing screening to ABR testing, from ABR testing to ordering of hearing aids, and from ABR testing to dispensing of hearing aids.
Of the 90 patients, mean (SD) age was 5.6 (3.7) years, 56% were female, and 77 (86%) were non-Hispanic. Results of χ2 tests indicated significant assocations existed between public insurance and race/ethnicity and between public insurance and primary language other than English. Variables associated with the interval from newborn hearing screening to ABR testing included insurance type (mean difference, 7.4 months; 95% CI, 2.6-12.2 months) and race/ethnicity (mean difference, 6.9 months; 95% CI, 2.7-11.1 months). Increased delays between birth and a child’s first ABR test were associated with public insurance (mean difference, 6.0 months; 95% CI, 1.8-10.2 months) and race/ethnicity other than white (mean difference, 6.0 months; 95% CI, 2.3-9.7 months). The mean time from birth to initial ABR testing was a mean of 6 months longer for patients from non–English-speaking families than for those from English-speaking families (mean [SD] interval, 14.9 [16.3] months vs 9.0 [8.5] months), although the difference was not statistically significant. Severity of hearing loss was associated with a decrease in the interval from ABR testing to ordering of hearing aids after accounting for other potential barriers (odds ratio, 0.6; 95% CI, 0.4-0.9). Zip code and complexity of the child’s medical condition did not appear to be associated with timely access to pediatric hearing aids.
Conclusions and Relevance
This study’s findings suggest that insurance type, race/ethnicity, and primary language may be barriers associated with pediatric access to hearing aids, with the greatest difference observed in time to initial ABR testing. Clinical severity of hearing loss appeared to be associated with a significant decrease in time from ABR testing to ordering of hearing aids. Greater efforts to assist parents with ABR testing and coordination of follow-up may help improve access for other at-risk children.
In 1993, the National Institutes of Health issued a consensus statement calling for universal newborn hearing screening (NHS) by the third month of age.1 Maryland introduced universal NHS in 1999, with the current minimum standard being an evoked otoacoustic emissions test, followed by an auditory brainstem response (ABR) test for patients who failed the otoacoustic emissions testing.2 Current guidelines call for completion of screening before 1 month, diagnosis before 3 months, and pediatric referral for audiologic testing and interventional services before 6 months of age.
The widespread adoption of universal NHS has been associated with earlier diagnosis and intervention for infants with hearing loss, including decreases in time intervals to hearing aid fitting or cochlear implantation.3-5 The benefits of earlier hearing amplification, especially if it occurs before a hearing-impaired child reaches age 3 months, have included increased speech intelligibility and improved language outcomes and has also led to earlier cochlear implantation, with similar improved outcomes.4,6,7 Earlier cochlear implantation has been shown to result in greater use of auditory and oral communication as well as faster rates of improvement in speech, language, and reading skills.6,8-10 In addition, children who have undergone cochlear implantation earlier have been more likely to attend mainstream schools and have reported a higher quality of life.11
Despite the clear benefits of early access to hearing amplification for child development, delays in diagnosis and intervention continue to be reported. Spivak et al12 found Medicaid to be a significant factor associated with loss to follow-up, often leading to delays in hearing aid fitting. Additional barriers identified have included problems with scheduling appointments, need for repeated testing, and financial and insurance issues.3 However, to our knowledge, a more comprehensive investigation into the potential barriers to access of pediatric hearing aids has not previously been conducted. The purpose of our study was to identify correctable factors that impede timely diagnosis of hearing impairment and delay access to pediatric hearing aids.
The Johns Hopkins University School of Medicine institutional review board approved the study. Informed consent was not required for the retrospective study based on deidentified data from patient medical records. A review of medical records was conducted to identify all patients who underwent ABR testing between March 2004 and July 2018 at a single tertiary care hospital that provides health care services to a large, both urban and rural population in Maryland. Only patients younger than 18 years who had ABR testing performed and hearing aids both fitted and dispensed at the tertiary care institution were included in the study.
The potential barriers to access that we targeted for investigation included insurance type (private vs public), race/ethnicity, primary language, cause of hearing loss, patient’s zip code, hearing aid manufacturer, and severity of hearing loss measured in decibels at ABR testing. We included as variables the dates on which NHS was performed (estimated as the date of birth), the first ABR test result was obtained, hearing aids were ordered, and hearing aids were dispensed. For analysis, the cause of each patient’s hearing impairment was categorized as a binary variable of medical complexity. The cause was considered medically complex if it was syndromic or neurologic, if the patient was delivered preterm and received care in the neonatal intensive care unit, or if the patient had a severe cardiopulmonary disorder. For the date of ABR testing, we recorded the date of the patient’s first ABR test result, including a result obtained from ABR testing at an outside institution. Table 1 indicates all of the dependent and independent variables included in the analysis with additional information on specific calculations.
We conducted 1-tailed, unpaired t tests accounting for unequal variances to assess whether the independent variables contributed to the duration of time intervals (dependent variables). Statistical significance was defined by P ≤ .05. All t test results are reported as the difference of means in months, with the corresponding 95% CI for the difference. Given the nonnormal distribution of the data, the duration of intervals (in months) between NHS, ABR testing, and hearing aid ordering and dispensing were divided into categorical quartiles for further analysis with multivariable ordered logistic regressions to investigate the association between these time variables and the variables representing possible barriers to access.
The results of χ2 analyses indicated significant associations between insurance type and race/ethnicity as well as between insurance type and primary language. Thus, these tests were repeated on 2 subsets of insurance-specific data (a subset from patients with public insurance and a subset from patients with private insurance) to assess whether either of these associations remained significant for the insurance-specific subgroups.
For patients with multiple ABR test results, the date of the first ABR test result was used in measuring the time from NHS to ABR, and the date of the actionable ABR test result (the result immediately prior to the ordering and dispensing of hearing aids) was used in calculating the times from the ABR test to ordering and to dispensing of hearing aids. All statistical calculations were conducted with Stata, version 14.2 (StataCorp LLC).
Ninety patients were included in the study. Fifty (56%) were female, 36 (40%) were white, and their mean (SD) age was 5.6 (3.7) years. Eighty-five of the 90 patients (94%) had bilateral hearing impairment at NHS. Other patient demographic data are listed in Table 2.
Results of t tests showed associations with the delay from NHS to first ABR test both for insurance type (mean interval difference, 7.4 months; 95% CI, 2.6-12.2 months) and for race/ethnicity (mean interval difference, 6.9 months; 95% CI, 2.7-11.1 months). The association between primary language and delay from NHS to ABR test (mean interval difference, 10.1 months; 95% CI, –1.3 to 21.6 months) was difficult to assess because of the wide CI. Mean differences (95% CIs) for variables associated with delay from NHS to first ABR test were 6.0 (1.8-10.2) months for insurance type (private vs public) and 6.0 (2.3-9.7) months for race/ethnicity (white vs other) (Figure). The mean (SD) age of patients at the first ABR test was 9.7 (6.9) months. After the ABR test result was received, it took a mean (SD) of 2.6 (4.9) months for hearing aids to be ordered and 3.5 (5.0) additional months for hearing aids to be dispensed to the family. Twenty-one patients (23%) underwent multiple ABR tests prior to the ordering and dispensing of hearing aids.
The results of t tests indicated that the type of insurance was significantly associated with the time interval between NHS and the first ABR test. For patients with public insurance, the interval to the first ABR test was a mean (SD) of 13.0 (12.6) months compared with 7.0 (6.1) months for those with private insurance (mean difference, 6.0 months; 95% CI, 1.8-10.2 months). In an ordered logistic regression analysis including complexity of the medical condition, severity of hearing loss, patient’s zip code, and hearing aid manufacturer, we found increased odds that public insurance contributed to a delay of the first ABR test (odds ratio [OR], 1.9; 95% CI, 0.8-4.4). However, the wide 95% CI indicated that the results were not statistically significant. In this model, severity of hearing loss was associated with a decrease in the time interval from actionable ABR test to ordering of hearing aids (OR, 0.6; 95% CI, 0.4-0.9). No association was found between public and private insurance and the interval between the ABR test and dispensing of hearing aids when ordered logistic regression included the variables listed.
The results of t tests indicated an association between race/ethnicity and the time interval between NHS and the first ABR test. For white patients, the mean (SD) interval was 6.3 (5.6) months compared with 12.3 (11.8) months for patients of other race/ethnicity (mean difference, 6.0; 95% CI, 2.3-9.7 months).
In ordered logistic regression analyses of the associations between both primary language and race/ethnicity that accounted for the complexity of the medical condition, severity of hearing loss, patient’s zip code, and hearing aid manufacturer, a significant difference was found between white patients and patients of other race/ethnicity for the time interval from NHS to ABR test (OR, 2.7; 95% CI, 1.2-6.2). When race/ethnicity was included in an ordered logistic regression analysis with the above variables and insurance type, there was no longer a statistically significant difference (OR, 0.8; 95% CI, 0.3-1.8). However, again, in the multivariable ordered logistic regression analysis, severity of hearing loss was associated with a decrease in the time to ordering of hearing aids (OR, 0.6; 95% CI, 0.4-0.9) but not in that for dispensing them (OR, 0.7; 95% CI, 0.5-1.1). In the subset of individuals with either public or private insurance, race/ethnicity was not associated with the time interval to first ABR (patients with public insurance: 6.6; 95% CI, 0.8 to 12.4 vs those with private insurance: 2.5, 95% CI, −1.1 to 6.3).
Patients who identified English as their primary language waited a mean of approximately 6 months less from birth or NHS to ABR testing compared with patients with a primary language other than English (mean [SD] interval of 9.0 [8.5] months vs 14.9 [16.3] months). However, t tests indicated a large CI that could signal imprecise measurement (mean difference, 5.9 months; 95% CI, −3.7 to 15.4). When analyses controlled for race/ethnicity, complexity of medical condition, severity of hearing loss, patient’s zip code, and hearing aid manufacturer, no significant differences owing to primary language were found for either the time interval to first ABR test or intervals from ABR test result to ordering and dispensing hearing aids. However, as occurred in analyses of insurance type and race/ethnicity, an ordered logistic regression analysis including the above independent variables and primary language also indicated that severity of hearing loss was associated with a decrease in the interval from ABR test to ordering of hearing aids (OR, 0.6; 95% CI, 0.3-0.9).
Multiplicity of ABR tests was associated with a decrease in the total interval between NHS and dispensing of hearing aids compared with the interval for patients with 1 ABR test result (mean difference, 8.6 months; 95% CI, 4.9-12.3 months), with the decrease seen predominantly between NHS and the first ABR test (mean difference, 8.1 months; 95% CI, 5.1-11.0 months). There was no association between number of ABR tests and the interval between the ABR test and ordering of hearing aids (OR, 0.02; 95% CI, −2.7 to 2.6) or dispensing of hearing aids (OR, 0.3; 95% CI, −2.7 to 2.1). Results of χ2 tests indicated no association between multiple ABRs and insurance type or race/ethnicity.
Previous research in this field has demonstrated substantial delays in the diagnosis of pediatric hearing impairment owing to the loss of patients to follow-up, which is more common in rural populations with difficulty accessing specialty care.13,14 Our study results suggest that insurance type, race/ethnicity, and primary language as potential barriers to timely access to pediatric hearing aids. Our data also showed that severity of hearing loss may be a factor contributing to a decrease in the delay between hearing tests and the ordering of hearing aids.
As expected, public vs private insurance was associated with differences in diagnosis and intervention for infant hearing loss. Previous research done in other pediatric diseases, such as obstructive sleep apnea, indicated that children with public insurance experience much longer intervals from initial evaluation to confirmatory testing and surgical intervention than those with private insurance.15
We found that insurance type had a clinically important association with time to first ABR test, with patients covered under public insurance taking almost twice as long as those with private insurance, although most public insurance options cover the cost of hearing aids. This finding is consistent with those reported in the literature on barriers to cochlear implantation; for example, it has been reported that, for children with public insurance, diagnosis and implantation are delayed.10,11,16 Children who have public insurance are also reported to be more likely to undergo cochlear implantation after 2 years of age, which has been linked with poorer long-term outcomes in speech and language.7,17
Research suggests that there is no evidence that improvements in reimbursement result in higher rates of cochlear implantation.8 Potential barriers may be difficulties in prior authorization by insurers, restriction of services to specific facilities, and the costs of multiple appointments.
A previous study by Chang et al8 indicates that patients with Medicaid are more likely than privately insured patients to miss follow-up appointments after cochlear implantation and have higher complication rates after implantation. These findings can shed light on the delay in access to hearing aids as well because the process for ordering and dispensing hearing aids following the ABR test requires scheduling multiple clinic visits. Whereas the use of pediatric hearing aids is associated with a very low risk of complications, the maintenance of hearing aids has been cited as a concern for some families and may lead some parents to delay their child’s clinical evaluation.18 Other factors in such decisions may be the potential cost of and availability of additional resources for speech and language development associated with pediatric hearing amplification.19
Stern et al20 showed differences in rates of cochlear implantation by race/ethnicity, with a rate of implantation in white children up to 10 times that in black children. Similarly, our results suggested that race/ethnicity may be significantly associated with the time to ABR testing, with a decreased interval for white patients.
In addition, our results identified race/ethnicity as a potential barrier when modeled with primary language but not with public insurance. This discrepancy may be explained by the high level of correlation between race/ethnicity and public insurance that was shown by the χ2 tests of association between these 2 variables. In addition, differences may exist in community and cultural acceptance and awareness of infant hearing loss that are dependent on barriers that are yet to be identified. A previous study by Spivak et al12 demonstrated that parental factors, such as readiness to accept a diagnosis of hearing impairment in a child, have a significant association with the timeliness of testing and diagnosis. A study by Caballero et al18 demonstrated that unique concerns exist in the Hispanic community that could result in challenges to managing pediatric hearing aids. Additional research to better characterize these specific cultural barriers would be beneficial for providing direction and support to encourage compliance with testing and treatment guidelines.
Despite the large difference in time to ABR between primary language groups, no significant association was found. This finding may be due to the broad SD among families who identified a language other than English as their primary language (mean [SD], 14.9 [16.3] months). Only 2 families with private insurance did not speak English as their primary language compared with the 43 families that did. Thus, the association between primary language and hearing aid access could be secondary to the association between language and insurance type. This possibility seems likely in the context of previous literature regarding cochlear implantation, in which language barriers have been identified as a significant contributor to delay, largely owing to difficulty understanding complex medical information and navigating through the health care system to reach the necessary specialty care.3,10 Artières et al6 showed that navigating the health care system and communicating with health care professionals are often the biggest limitations for families regardless of their insurance type.
Caballero et al18 showed that primary language can serve as a proxy for identifying challenges specific to certain ethnic groups. For example, in the Hispanic population, additional barriers that have led to delayed access to hearing aids include families wanting to seek a community with others who have children with hearing loss as well as concerns about hearing aid maintenance and management. Some of these secondary factors may be correlated with language, and these factors must also be considered when proposing solutions to overcome the language barrier.
Severity of hearing loss was associated with a decrease in the interval between ABR testing and ordering of hearing aids. This finding is consistent with the existing literature describing significant decreases in times to diagnosis, referral, and hearing aid fitting among children with more profound hearing impairment.3,5,21 Our data indicate that this difference is robust enough to overcome statistically significant barriers, such as insurance type, race/ethnicity, and primary language, to decrease the time from the ABR test to the ordering of hearing aids. Our data also suggest that severity of hearing loss may not be associated with the duration of the interval between ABR testing and the dispensing of hearing aids.
Patients with multiple ABR tests were more likely to have a shorter interval from NHS to receipt of hearing aids than those with only 1 ABR test. One possible explanation for this difference may be that, for infants with severe hearing loss and strong clinical signs of deafness, clinicians and families may attempt multiple sleep-deprived ABR tests within a shorter period of time in order to obtain an accurate diagnosis. For patients with less obvious clinical signs or less severe hearing loss, families and clinicians may not perceive an urgent need to initiate or repeat early ABR testing. This scenario might explain why no association was found between multiplicity of ABR tests and either insurance type or race/ethnicity.
Another important possibility to consider is an association between low socioeconomic status and insurance type, race/ethnicity, and primary language. The latter 3 variables have been used as effective proxies for lower socioeconomic status in previous efforts to assess barriers to cochlear implantation, especially because the criteria for Medicaid eligibility are based on federal poverty guidelines.8,20
Previously published studies demonstrate an extensive connection between lower socioeconomic status and delays in cochlear implantation as well as worse outcomes following implantation.8,17 Families with lower socioeconomic status often have difficulty getting transportation to and from appointments, have fewer social supports, and lack access to speech therapy services.10,17 Recent interventions with multidisciplinary clinics for pediatric patients newly diagnosed with hearing impairment have proven to be convenient and effective ways to coordinate care for these higher-risk patients.22
Yehudai et al23 showed that parental educational level is independently associated with mainstream classroom placement of deaf and hearing-impaired children following cochlear implantation. Families with higher socioeconomic status or a higher educational level often have access to more information about cochlear implantation and surgery.8 These families also often have a greater willingness to undergo surgery and have higher rates of cochlear implantation.8,24 The same socioeconomic factors are likely to affect family decision-making with regard to pediatric hearing aids.
Our data suggest that insurance type, race/ethnicity, and primary language are associated with delays in access to pediatric hearing amplification that have significant clinical implications. Sleep-deprived ABR testing is often not possible after age 6 months. Insurance and race/ethnicity are associated with substantial delays that can cause children in the high-risk group to miss the window of opportunity for sleep-deprived ABR testing (Table 3). This delay may result in additional wait-time for scheduling of sedated ABR testing and a further delay in definitive diagnosis and treatment of hearing loss. Although ABR testing was delayed past the 6-month threshold in both linguistic cohorts, the interquartile range for the families indicating a language other than English as their primary language shows that almost 75% of patients in that cohort needed sedated ABR testing compared with 50% of patients in families identifying English as their primary language.
Our data contain a large range of outliers for patients with public insurance and those with nonwhite race/ethnicity in comparison with their low-risk counterparts (ie, those with private insurance and white race/ethnicity). Children in these higher-risk groups are not only experiencing delays in reaching key milestones for hearing amplification but also may have very extended periods of delay.
This study has limitations. Our study included only patients who had ABR testing and hearing aids dispensed at our institution in order to hold as many factors constant as possible throughout the sample. However, this restriction gives rise to the possibility that institution-specific factors could limit the generalizability of our study results.
In the subgroup analyses, some samples were very small. For example, there were only 9 black patients in the private insurance group. Despite the statistical significance of the association of race/ethnicity with the duration of the interval between NHS and ABR testing, it is important to consider the possibility that a few data outliers in the time to ABR testing could have skewed the results. It is also possible that our sample size did not have enough power to allow the detection of other factors associated with access to pediatric hearing aids; this factor could explain the surprising lack of an association between mean income (corresponding to patient zip code) and delay to ABR testing.
Another important limitation was the codependence of race/ethnicity, language, and insurance type in our χ2 analyses. Some of our models might have had the ability to detect only medium to large differences to the association between these factors.
Our hospital serves an area with a population that is 63% black and in which 22% of families are living below the poverty level.25 Given the specific demographic and socioeconomic characteristics of our patients, it may be difficult to generalize our findings to the national level. However, our study highlights some important vulnerabilities that must be addressed to ensure timely access to pediatric hearing aids in a predominantly urban population, and it could provide a stimulus for further research into health disparities in such populations.
In our institutionally controlled sample, insurance type, race/ethnicity, and primary language appeared to be important and codependent factors associated with access to pediatric hearing aids. The results of our study likely have implications for decision-making at the state and national levels because many of the potential barriers to access that we investigated are similar to those identified previously as barriers to diagnosis and treatment of other medical conditions. We hope the results of our research will further the discussion of policies to eliminate these barriers to access for children at risk.
Accepted for Publication: August 7, 2019.
Corresponding Author: Jonathan Walsh, MD, Department of Otolaryngology–Head and Neck Surgery, Johns Hopkins University School of Medicine, 601 N Caroline St, Baltimore, MD 21287 (email@example.com).
Published Online: October 10, 2019. doi:10.1001/jamaoto.2019.2877
Author Contributions: Ms Zhang and Dr Walsh had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Zhang, Boss, White, Walsh.
Acquisition, analysis, or interpretation of data: Zhang, Links, Walsh.
Drafting of the manuscript: Zhang, White.
Critical revision of the manuscript for important intellectual content: Zhang, Links, Boss, Walsh.
Statistical analysis: Zhang, Links, Walsh.
Administrative, technical, or material support: White, Walsh.
Conflict of Interest Disclosures: None reported.
Additional Contributions: We thank the Department of Otolaryngology for their support of this project.