DOES OAEAABR HEARING SCREENING MISS HEARING LOSS 2005

DOES OAE/A-ABR HEARING SCREENING MISS HEARING LOSS? 2005 EDHI Conference Jean L. Johnson, Dr. PH Center for Disability Studies Director (Interim) March 3, 2005 Atlanta, GA

International Conference on Newborn Hearing Screening, Diagnosis and Intervention Cernobbio, Italy - May 2004 Second Annual Conference of the CDC Centers on Birth Defects and Developmental Disabilities Washington, DC – July 2004 Asia Region Neonatal Screening Conference Shanghai, China - September 2004 UH College of Education Research Symposium Honolulu, HI - October 2004 American Speech-Language and Hearing Association Philadelphia, PA – November 2004 CDC Ad Hoc Group – Teleconference January 11, 2005

Research Team Principal Investigator - Jean Johnson, Dr. PH Research Coordinator - Karl R. White, Ph. D Diagnostic Evaluation Coordinator - Judith E. Widen, Ph. D Site Co-Principal Investigators Judith Gravel, Ph. D Michele James, Au. D Teresa Kennalley, MA Antonia B. Maxon, Ph. D Lynn Spivak, Ph. D Maureen Sullivan-Mahoney, MA Betty Vohr, MD Yusnita Weirather, MA

Funded by the Centers for Disease Control and Prevention CDC Consultants: June Holstrum, Ph. D Roy Ing, MD, MPH Brandt Culpepper, Ph. D Krista Biernath, MD Lee Ann Ramsey, BBA, GCPH under a Cooperative Agreement with: The Association of Teachers of Preventive Medicine with a sub-agreement to: The University of Hawai`i

Why is early identification of hearing loss so important? • Hearing loss occurs more frequently than any other birth defect. • Undetected hearing loss has serious negative consequences. • Dramatic benefits are associated with early identification of hearing loss. It is true for severe to profound loss. Is it also true for milder hearing loss?

Background ü National Institutes of Health (NIH) Consensus Panel recommended in March 1993 that: “the preferred model for screening should begin with an evoked otoacoustic emissions test and should be followed by an auditory brainstem response test for all infants who fail the evoked otoacoustic emissions test. ” ü Continuing improvement of ABR technology led to a number of hospitals in the US implementing a variation of the NIH recommendation that was based on automated ABR (AABR) ü Anecdotal reports to the Centers for Disease Control and Prevention (CDC) in the mid to late 1990’s that the two-stage OAE/AABR protocol was not identifying some infants with permanent hearing loss. ü The CDC issued a competitive Request for Proposals in 2000 to investigate whether the OAE/AABR screening protocol was not identifying babies with hearing loss.

Oto-Acoustic Emissions (OAE) Screening

Auditory Brain Stem (ABR) Screening

Research Question Are infants with permanent hearing loss not being identified when newborn hearing screening is done with a two-stage OAE/A-ABR protocol in which infants who fail OAE and pass AABR are not followed? Comparison Group OAE Screening Prior to Hospital Discharge Pass Discharge Fail AABR Screening Pass Fail Comprehensive Hearing Evaluation Before 3 Months of Age Study Sample Comprehensive Audiological Assessment at 8 -12 months of age Discharge

Criteria for Selecting Sites • 2, 000 or more births per year • Established newborn hearing screening program with at least six month history of success • Historical refer rates of less than 10% for OAE and 4% for ABR • Success in obtaining follow-up on 85% or more of referrals • Ethnic and socio-economic distribution similar to US population

Participating Sites Name of Hospital Location Arnold Palmer Hospital Orlando, Florida Good Samaritan Hospital Columbus, Ohio Jacobi Medical Center New York, New York Kapi`olani Medical Center Honolulu, Hawaii Long Island Jewish Medical System New York, New York Via Christi Regional Medical Center Kansas City, Kansas Women & Infants Hospital Providence, Rhode Island

Enrollment Process • Eligible babies (Failed OAE and Passed A-ABR) were identified following newborn hearing screening. • Parents were contacted and research study explained. • Consent was obtained from families. • Enrollment data was collected. • Contact was maintained with family at 2, 4, & 6 months of age via post cards. • Babies were seen for audiological diagnostic evaluation at 8 -12 months of adjusted age.

Date Collected for Each Participating Baby Birthdate Bronchio-pulmonary Dsplasia Gender Mechanical Ventilation >7 Days Birth Weight ECMO Gestational Age Number of Children in Home APGAR Scores Number of Adults in Home Days in NICU Total Household Income Malformations of the Head and Neck Child’s Race/Ethnicity Syndrome Associated with Hearing Loss Health Insurance In-utero Infections Family History of Hearing Loss

Study Sample § 1, 524 Infants Enrolled § 973 (63. 8%) Returned for Evaluation § 1, 432 Ears Evaluated

Enrollment of Study Participants Enrollment Period Births During Enrollment Site #1 May 1, 2001 to Dec 31, 2002 16, 608 6. 3% 0. 8% WB/NICU Site #2 June 1, 2001 to Jan 31, 2003 9, 393 4. 5% 0. 9% WB/NICU Site #3 May 1, 2001 to Jan 31, 2003 24, 032 2. 4% 0. 8% WB/NICU Site #4 Sep 20, 2001 to Jan, 2003 4, 509 8. 0% 1. 0% WB Site #5 May 15, 2001 to Jan 31, 2003 9, 252 3. 1% 0. 8% WB Site #6 May 1, 2001 to Jan 31, 2003 16, 623 5. 3% 1. 2% WB/NICU Site #7 May 1, 2001 to Jan 31, 2003 6, 217 9. 6% 2. 8% WB/NICU 86, 634 4. 8% 1. 0% Total Referral Rate OAE ABR Recruitment From

Enrollment of Study Participants (continued) Number of Babies: Site # 1 Site # 2 Site # 3 Site # 4 Site # 5 Site # 6 Site # 7 Total Births During Enrollment Eligible for Enrollment Total Enrolled Number Not Recruited Number of Refusals 16, 608 1, 044 191 418 435 6. 3% 18. 3% 40. 0% 41. 7% 421 370 18 33 4. 5% 87. 9% 4. 3% 7. 8% 456 170 11 275 1. 9% 37. 3% 2. 4% 60. 3% 285 84 186 15 6. 3% 29. 5% 65. 3% 209 147 30 32 2. 3% 70. 3% 14. 4% 15. 3% 433 266 50 117 7. 0% 61. 4% 11. 5% 27. 0% 614 296 71 247 3. 7% 48. 2% 11. 6% 40. 2% 3, 462 1, 524 784 1, 154 4. 0% 44. 0% 22. 6% 33. 3% 9, 393 24, 032 4, 509 9, 252 6, 217 16, 623 86, 634

Audiological Diagnostic Evaluation • Visual reinforcement audiometry (VRA) • Tympanometry • OAE - Either TOAE or DPOAE

VRA Protocol • Protocol patterned after National institutes of Health Study (Norton, Univ. of Washington) • Responses at 500, 1 K, 2 K, 4 K Hz – Order of testing 2 K, . 5 K, 4 K, 1 K – Aiming for minimal response level of 15 d. B HL • Multiple visits often necessary to complete testing – 68% completed in 1 visit – 24% required 2 visits – 8% required 3 or more visits

Criteria for Categorizing Hearing Loss Category Criteria Not Permanent Hearing Loss Using the ”best” results from all assessments, MRL thresholds of < 20 d. B at 1 K, 2 K, and 4 K. Probable Not Permanent Hearing Loss Complete MRL data not available at 1 K, 2 K, and 4 K, BUT * All frequencies had MRLs < 20 d. B OR OAEs within normal limits* OR Tone burst ABR data < 25 d. B. Permanent Hearing Loss (PHL) Sensorineural MRLs > 25 d. B at 1 K, 2 K, or 4 K (tested with good confidence) OR ABR threshold > 30 d. B; AND if tested, OAEs below normal limits at the frequencies with elevated MRLs; AND normal middle ear functioning based on tympanometry or bone conduction. Permanent Hearing Loss (PHL) Conductive MRLs > 25 d. B at 1 K, 2 K, or 4 K (tested with good confidence); AND if tested, OAEs below normal limits; AND bone conduction thresholds < 20 d. B with an Air/Bone gap > 15 d. B at frequencies with MRLs > 25 d. B. Increased Suspicion of PHL High Suspicion MRLs > 25 d. B at 1 K, 2 K, or 4 K, BUT OAEs within normal limits for those frequencies OR only fair confidence in VRA testing. Some Suspicion * MRLs > 30 d. B at 1 frequency or > 25 d. B at more than one frequency, BUT abnormal tympanometry AND no bone conduction. * Sound field thresholds > 25 d. B (with fair confidence) AND normal tympanometry AND OAEs below normal limits. Not Sufficient Data to Rule Out PHL * No MRLs or OAEs within normal limits for 1 K, 2 K, or 4 K and none of the above criteria for permanent hearing loss are met. * OAEs within normal limits were defined as > 3 -6 d. B at 1 K and > 6 d. B at 2 K and 4 K.

Hearing Status of Study Ears from All Hospitals # of Infants with Dx Data Percent of Infants w/ Dx Data Total Ears Not PHL Permanent Loss Hearing (PHL) Increased of Suspicion PHL: SNHL PC High Some Probable Not PHL Not Sufficient Data Site #1 81 42. 4% 148 131 0 0 6 11 Site #2 299 80. 8% 478 432 7 0 0 0 35 4 Site #3 42 50. 0% 59 40 0 2 6 11 Site #4 109 74. 1% 165 82 10 5 17 12 24 15 Site #5 86 50. 6% 111 58 2 0 0 5 16 30 Site #6 184 69. 2% 241 202 4 0 2 8 8 17 Site #7 172 58. 1% 230 195 2 0 0 1 5 27 Total 973 63. 8% 1432 1140 100% 79. 6% 25 1. 7% 5 0. 3% 19 1. 3% 28 2. 0% 100 7. 0% 115 8. 0%

PHL in Comparison Group Sites (Fail OAE/Fail A-ABR)

PHL in Ears of Study Infants that Passed Initial OAE Total Ears Not PHL Permanent Loss Hearing (PHL) Increased of Suspicion PHL: SNHL PC High Some Probable Not PHL Not Sufficient Data Site #1 13 11 0 0 0 2 Site #2 112 107 0 0 1 0 3 1 Site #3 25 3 0 0 15 6 Site #4 53 19 0 0 2 6 15 11 Site #5 53 30 0 0 1 2 5 15 Site #6 127 60 0 0 3 1 38 25 Site #7 113 30 0 0 22 61 496 100% 260 52. 4% 0 0. 0% 8 1. 6% 9 1. 8% 98 19. 8% 121 24. 4% Total

Degree of Hearing Loss in Study and Comparison Group Babies 28. 6% Mild Moderate (25 -40 d. B) (41 -70 d. B) Study Group Comparison Group Total Severe through Profound (>70 d. B) Total Infants 15 5 1 21 71. 4% 23. 8% 4. 8% 100. 0% 31 64 63 158 19. 6% 40. 5% 39. 6% 100. 0% 46 69 64 179 25. 7% 38. 5% 35. /% 100. 0% As measured in the worse ear 80. 3%

Comparability of Study and Comparison Groups What Percent of “Referred” Babies Did Sites Try to Follow? What Percent of “Followed” Babies Were Diagnosed? Study Group (Fail OAE/ Pass A-ABR) 44% 64% Comparison Group (Fail OAE/ Fail A-ABR) 100% 87% ü Reasonable to adjust prevalence rates for those who were not recruited ü Adjusting prevalence rates for differences in the percent of diagnostics completed is problematic ü Families who think their child has a hearing loss are more likely to return ü Families that are poor, single heads of household, transient , etc are less likely to return and these variables may be correlated with the incidence of hearing loss

How Many Additional Babies with Permanent Hearing Loss (PHL) were Identified? Comparison Group Study Group (Fail OAE/ Fail A-ABR) (Fail OAE/ Pass A-ABR) 158 1. 82 21. 24 Number of Babies Prevalence per 1, 000 *Adjusted for proportion of OAE fails that enrolled Represents 12% of all babies with PHL in birth cohort Total 179 2. 06

Is it important that 21 Babies (30 ears) with PHL were found? • How many does it add to what would have been identified otherwise? • How many ears with hearing loss were found among those that passed the initial screen? • How many babies would you have to follow to find 21 babies PHL? • Is this congenital or late-onset hearing loss?

How many babies must be screened to find 21 with PHL? The obvious answer is 973, but…. • This ignores that most screening programs that use OAE also do second stage OAE screen (usually following hospital discharge) • Such outpatient screening is less expensive than the diagnostic protocol used in this study • Difficulty of getting babies to return for outpatient screening must be considered

Screening Failures True Positives False Positives

Cost of Screening • Direct cost • Indirect cost • Follow-up cost

Cost-Benefit of Screening Benefit COST Good follow-up Mediocre follow-up Poor follow-up

Were any of these ears lateonset losses? • This study was not designed to answer that question. • We do know that IF all of the ears with risk factors had been followed and identified, 9 of 21 babies would still have been missed • Little is known about the incidence or what predicts late-onset hearing loss • Most (>3/4 ths) of the hearing losses “missed” were mild as was expected

Different Criteria for Determining Permanent Hearing Loss Comparison Group Study Group Total Based only on those meeting criteria for PHL 1. 82 . 24 2. 06 Including those categorized as high suspicion of PHL 1. 82 (21 babies) . 43 (33 babies) 2. 49

Different Criteria for Determining Permanent Hearing Loss Comparison Group Study Group Total Based only on those meeting criteria for PHL 1. 82 . 24 2. 06 X 1. 82 Including those categorized as high suspicion of PHL (21 babies) X X X . 43 (33 babies) 2. 49

Variation Among Sites ü The study design assumed that sites are all equally well implemented ü To the degree that this isn’t true, data from some sites may be a better estimate of the number of babies being missed

Indicators of Implementation Quality at Each Site # 1 Site # 2 Site # 3 Site # 4 Site # 5 Site # 6 Site # 7 # of Study Group Infants with PHL Average “Rank” for Implementat ion Quality of Site Births During Enrollment Period % and N of Eligible Infants Enrolled % and N of Refusals During Recruitment % Returning for Diagnostic Evaluations % with “Not Sufficient Data” 0 5. 8 16, 608 18. 3% 41. 7% 42. 4% 7. 4% 191 435 87. 9% 7. 8% 80. 8% 370 33 29. 5% 5. 3% 50. 0% 18. 6% 84 15 70. 3% 15. 3% 74. 1% 9. 1% 147 32 37. 3% 60. 3% 50. 6% 27. 0% 170 275 61. 4% 27. 0% 69. 2% 7. 1% 266 117 48. 2% 40. 2% 58. 1% 11. 7% 296 247 5 0 10 2 3 2 PHL=permanent hearing loss 1. 3 4. 8 2. 8 6. 0 3. 0 4. 5 9, 393 4, 509 9, 252 24, 032 6, 217 16, 623

Presumed Curve of Health Care

Bell-Shaped Curve of Health Care

Best Estimate of Amount of PHL Missed by OAE/AABR protocol 12% of children with 23% of children with PHL in birth cohort Babies Who Failed OAE / Failed AABR All Sites with Best Implementation 1. 82 2. 27 PHL in birth cohort Babies who failed OAE/ Passed AABR Based on 44% that participated Adjusted for those who did not participate 2. 06 2. 37 (. 24 increase) (. 55 increase) 2. 75 2. 95 (. 48 Increase) (. 68 increase) 17% of children with 23% of children with PHL in birth cohort

What’s the Best Estimate of the Number of Babies Missed by the OAE/AABR Screening Protocol? • Depends on the criteria used for determining PHL • Variation among sites • Adjustments for Differences Between Study and Comparison Groups

Conclusions ü The OAE/AABR protocol, as implemented at these sites, failed to identify a substantial number of infants with PHL. § Best estimate is. 55 per thousand or 24% of all infants with PHL. § Most were mild sensorineural hearing loss § Impossible from this study to determine how many are late-onset losses ü About 41% might be identified if all infants with risk factors or opposite refer ears were followed, but this is not likely.

Recommendations ü Screening for permanent hearing loss should extend into early childhood (e. g. physician’s offices, early childhood programs). ü Emphasize to families and physicians that passing hospital-based hearing screening does not eliminate the need to vigilantly monitor language development. ü Screening program administrators should ensure that the stimulus levels of equipment used are consistent with the degree of hearing loss they want to identify. ü The relative advantages and disadvantages of the twostage (OAE/A-ABR) protocol need to be carefully considered for individual programs.

Further Research Recommendations ü Prevalence and methods of identifying late-onset hearing loss ü Identification and monitoring of progressive hearing loss ü Ongoing investigation of sensitivity of various screening protocols and equipment (including what level of hearing loss is targeted) ü Practicality and cost-efficiency of alternative or additional “continuous” screening and surveillance techniques, especially in early childhood

Questions to Ponder • What degree of hearing loss do you want to identify? • What cost will be required for that identification? • Can you assure that follow-up will occur? • What interventions can you provide? • How can you provide continuing surveillance of late onset or progressive loss? • How comfortable are you with the quality of services being provided?

Key Large-Scale Newborn Hearing Screening Studies in the United States Location/Dates Cohort Size Nurseries Screening Techniques Refer Rates Follow Up Rate Prevalence Per 1000 of Hearing Loss RIHAP 3 (8/90 – 2/91) 1, 850 NICU & WBN OAE / ABR 26. 9% Colorado (1/92 -12/96 27, 938 NICU & WBN AABR Rhode Island (1/93 – 12/95) 53, 121 NICU & WBN OAE / ABR 14. 7% New Jersey (1/93 – 12/95) 15, 749 NICU & WBN ABR 3% Hawai‘i (1/94 – 12/95) 9, 605 WBN OAE 89% 4. 15 Texas (1/94 – 6/97) 54, 228 NICU & WBN OAE & AABR 3. 5% 82. 3% 3. 14 NIDCD (1994 -1997) 4, 478 2, 701 NICU & WBN ABRTOAEDPOAE- 64. 4% (Research) 56. 0 New York (1/96 – 12/96) 69, 761 NICU & WBN OAE & AABR 72% 8. 00 (NICU) 0. 9 (WBN) Washington, DC (2/97 – 12/02) 39, 437 NICU & WBR TOAE 1. 6% 82. 7 2. 3 overall 1. 9 WBN 6. 8 - NICU ATPM/CDC (5/2001 – 1/2003) 86, 634 NICU & WBN OAE: 4. 8 AABR: 1. 0 63. 8(Research) 85. 8 2. 05 5. 95 73% 2. 56 74 --88%a 2. 00 3. 3 overall 2. 00 – WBN 13. 0 NICU

Other Dissemination • Article has been selected for publication in Pediatrics • Five articles prepared for publication in American Journal of Audiology • Presentation at Council on Exceptional Children Conference in Baltimore in April 2005.

Additional Information on Newborn Hearing Screening www. infanthearing. org www. babyhearing. org jeanj@hawaii. edu

Deafness separates people from people. --- Helen Keller