1. “Standards for NHS Equipment”
aka
“What Are We Measuring in NHS
Programs?”
Barbara L. Kurman, Au.D., FAAA
Vice-President,
Northeastern Technologies Group, Inc., NY
Managing Member,
Midlantic Technologies Group, LLC., PA
Judy Gravel, Ph.D.
Director, Center for Childhood Communication
The Children’s Hospital of Philadelphia, PA
2. Faculty Disclosure Information
As a Special Instrument Distributor (SID),
Barbara Kurman has contractual relationships with
most manufacturers whose products will be referred
to in this presentation.
In the past 12 months, Judith Gravel
has not had a significant financial interest or other
relationship with the manufacturer(s) of the
product(s) or provider(s) of the service(s) that will be
discussed in this presentation.
This presentation will not include discussion of
pharmaceuticals or devices that have not been
approved by the FDA. There will be no discussion or
unapproved or "off-label" uses of pharmaceuticals or
devices.
3. Agenda
• Background: review of newborn hearing
screening programs – who are we
missing?
• Technologies and current devices –
what is commercially available?
• Should We Have Standards for NHS?
• Other NHS product considerations
4. NIH 1993
Consensus Conference
“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.”
5. What hearing losses are we missing with
newborn screening?
• The best screening tests are not 100%
accurate
• Evidence suggests that at best, our
screening tests are identifying about 80%
- 90% of infants who have hearing loss in
the newborn period (e.g., Norton et al., 2000;
Davis et al., 1997; Lutman et al., 1997)
6. What hearing losses are we missing with
newborn screening?
• Mild hearing loss < 30 - 40 dB HL
• Some unusual configurations of hearing loss
– Isolated low-frequency hearing loss (both OAE and
ABR technologies)
– Oddly-shaped hearing loss
• Steeply sloping high-frequency hearing loss
• Mid-frequency hearing loss
• Delayed onset and progressive hearing losses
• AN and genetic IHC loss where OHCs are intact
(if only use OAE screening technology)
7. What hearing losses are we missing with
newborn screening?
• Missing ‘deaf’ ears
– Estimates of false-negative (false pass) rate:
• 1-2% of ‘deaf’ ears will pass ABR
screening (Roger Marsh, CHOP 2006
• Repeated screening of a ‘failed’ ear to
achieve a ‘pass’ will increase the
likelihood of a pass response in a ‘deaf’
ear
8. Centers for Disease Control and
Prevention (CDC)
Request for Proposals - January 2000
“Concerns have been raised about infants who
fail OAE but pass ABR and are then dismissed
from follow-up. These infants may have a mild
loss that was missed by ABR”
9. “A Multi-Center Evaluation of How Many
Infants with Permanent Hearing Loss
Pass a Two-Stage OAE/A-ABR Newborn
Hearing Screening Protocol”
Johnson, White, Widen, Gravel, James,
Kennalley, Maxon, Spivak, Sullivan-Mahoney,
Vohr, Weirather, and Holstrum
Pediatrics 116(3) Sept 2005; 663-672
10. • Prospective, cohort study
• 7 geographically-dispersed birthing
centers in U.S.
• Ethnic & socio-economic
characteristics representative of
U.S. population
Johnson et al. (Pediatrics 2005 )
11. • 2-technology screening protocol:
– TEOAE (Otodynamics) or DPOAE
(Biologic)
– A-ABR (ALGO: 35 dB nHL click)
• Some used protocol in both well-baby
and NICU
Johnson et al. 2005
12. 86,634 infants: screened in 7 Centers
• 704 (0.8%) failed OAE/failed A-ABR
screening
– Comparison Group: 604 (85.8%) returned for
diagnostic tests
• 3, 362 (4%) failed OAE/passed A-ABR in
at least one ear
– Study Group: 1,524 (44%) infants were
enrolled for follow-up
Johnson et al. 2005
13. Study Group
• 973 infants (64%) returned for audiologic
follow-up
• Audiologic tests completed at mean age
9.7 months
– VRA (AC & BC);
– tympanometry;
– OAE
Johnson et al. 2005
14. Johnson et al. 2005
21 infants (30 ears) who failed OAE &
passed A-ABR in newborn period
had permanent bilateral or unilateral
HL at ~9 months of age
15. Degree (poorer ear) of PHL (VRA) in
Study & Comparison Groups
(Johnson et al. 2005)
Mild
(25-40 dB)
Mod
(41-70 dB)
Sev/Prof
(>71 dB)
Total with
PHL
Study
Grp
15
(71.4%)
5
(23.8%)
1
(4.8%)
21
(100%)
Comp
Grp
31
(19.6%)
64
(40.5%)
63
(39.9%)
158
(100%)
Total 46
(25.7%)
69
(38.5%)
64
(35.8%)
179
(100%)
16. Examples of PHL (VRA) in Infants who
fail OAE/pass A-ABR
(Johnson et al. 2005)
ID # .5 kHz 1 kHz 2 kHz 4 kHz
091 20 25 35 35
055 25 25 20 30
053 25 25 30 35
130 25 30 35 45
002 25 25 30 25
17. Examples of PHL (VRA) in Infants who
fail OAE/pass A-ABR
(Johnson et al. 2005)
ID # .5 kHz 1 kHz 2 kHz 4 kHz
131 30 35 40 45
005 45 40 40 60
122 40 40 45 40
072 50 40 30 40
003 45 45 60 55
18. Identifying Mild PHL in Infants:
Challenges
Norton et al. (2000)
• Three measures (ABR, TEOAE,
DPOAE) - able to identify majority of
ears with moderate hearing loss or
greater.
– “more difficult for any tool to distinguish
between normal hearing and mild
hearing loss” (p.533)
– “some ears with mild hearing loss will be
missed, regardless of which test is used”
(p.534)
19. Cone-Wesson et al. (2000); n = 2995
infants
• Ears with mild PHL (n=22 [30.2%] of 86 ears
with PHL) confirmed at 8-12 months (VRA)
• Outcomes (neonatal ABR and OAE [DPOAE
and TEOAE])
– 10 ears failed both OAE and A-ABR tests,
– 4 ears passed both OAE and A-ABR tests,
– 4 passed ABR and failed both OAE measures
– 2 failed ABR and passed OAE tests.
– 2 ears that failed ABR passed DPOAE but
failed TEOAE.
Identifying Mild PHL in Infants:
Challenges
20. Issues Impacting NHS Data
• No ANSI standards for use in
calibration of OAE or ABR devices
• Variability among screening devices
test parameters
• Individual variability of SPL at TM –
issue most significant for ABR
22. Variability
“Based on the estimate of the RETSPL value,
the results indicate that the stimulus levels
in current equipment [automated ABR] are
considerably above the stated stimulus
level of 35 dB HL.”
Estimated: 38.5 to 44.9 dB HL
Stevens 2004
23. Eight Manufacturers of Hearing Screening
Devices Commercially Available in U.S.
• Bio-logic
• GN Otometrics –
Madsen
• GSI – Grason
Stadler: VIASYS
• Otodynamics, Ltd
• Maico
Diagnostics
• Natus
• Interacoustics
USA
• Intelligent Hearing
Systems
• SonaMed Corp
24. Devices:
Types available currently:
• TEOAE only (1 manufacturer only)
• DPOAE only (1 manufacturer only)
• TEOAE and/or DPOAE (5
manufacturers)
• Screening OAE and/or screening ABR
in one device (6 manufacturers)
• Screening ABR only (5 manufacturers
offer ABR as stand-alone device)
26. What the Products Have in Common
• Designed for ease of use - automation
• Parameters cannot be changed by screening
personnel
• All provide only “pass” or “refer” outcome
• Probe tips or couplers (circumaural cushions)
are disposable (“single use”)
• All identify some violation of proper test
conditions: “low stimulus level”, “too noisy”.
27. TEOAE: Examples Product Variability
Manuf Stimulus Stim Level
(dB SPL)
Pass
Criteria
1 6 freq bands:
1.5 – 4 kHz
83 4 dB S/N
2 Primary resp.
band: 1.6 – 3.2
kHz
84 6 dB S/N
3 Primary resp.
band:
1.2 – 3.5 kHz
80 6 dB S/N
28. DPOAE: Examples Product Variability
Manuf Stimulus
f1-f2
Levels
L1- L2
Pass Criteria
1 4 tonal pairs
(f1-f2)
(2 – 5 kHz)
65/55 6 dB SNR in 3
frequencies = pass (not
consecutive)
2 5 tonal pairs:
2-6 kHz
65/55 3 out of 5 frequencies =
pass which meet: Min
-5 dB SPL amplitude &
min 8 dB SNR, or min
noise floor amplitude -17
dB SPL
29. Automated ABR:
Examples Product Variability
Manuf Change
levels?
Ear Canal
Calibration?
Transducer
Choice?
1 yes yes yes
2 yes yes No –
probe only
3 yes** ? No –
circumaural
only
30. Calibration
• Manufacture Recommends:
– Some recommend a yearly calibration:
(e.g., on site by a NASED –certified
audiometric technician or ship device back
to factory)
– Some recommend no annual calibration is
needed
• NASED (National Association of Special
Equipment Distributors) – examining
calibration standards for OAE devices where
no current national standards exist
32. Problems Associated with the
Lack of National Standards for
NHS Programs
• Prevalence data may not be
comparable across programs
– Impacts our understanding of differences in
prevalence among populations and
geographic regions
– Impacts public health, education, fiscal
policy planning
33. Comparing Within & Among
NHS Data Sets
• NYS Demonstration Project – same protocol
and pass/fail criteria
• English Screening program – same protocol
and pass/fail criteria
• Ontario screening program – same protocol
and pass/fail criteria
• U.S. screening program - variability of
screening protocols & pass/fail criteria among
hospitals & states and across the nation
34. Problems Associated with the
Lack of National Standards for
NHS Programs
• Manufacturer algorithms designed to
provide best performance (low fail rates)
along with high sensitivity (most hearing
losses identified)
• Pass-fail criteria differ among devices
• Few data from manufacturers on the
statistical false-negative (false-pass)
rate
35. Screening outcome by ABR compared to Four Different
DPOAE Pass Criteria
(Barker, Lesperance & Kileny, 2000)
• 1184 ears examined: all ears passed ABR
screening (35 dB nHL)
• DPOAE screening
• Depending on four different pass-refer
criterion applied, between 64% and 89%
of ears passed screening
38. JCIH 2000
• In-hospital screening
– variety of protocols useful
– one technology and two technology
protocols
• Well-baby nursery (WBN) versus
Neonatal Intensive Care Unit (NICU)
• Re-screen (out-patient)
– within 1 month of discharge
– minimizes number of referrals audiologic
assessment: reduces over-referral rate
(false-positive outcomes)
39. • Physiologic screen of hearing only
• Screening Equipment (OAE – ABR)
• not specified
• required to be fully automated (only
providing pass-refer outcome) when
used by technical personnel
The Law:
NY State UNHS Regulations
40. Solutions:
Use published parameters from clinical
trials that provided data about accuracy of
screening tests (sensitivity & specificity)
Only one study available:
Norton et al. 2000
NIH-NIDCD
41. Multi-center Investigation
Sponsored by NIH-NIDCD
Investigators: Norton, Gorga, Widen, Folsom, Sininger,
Cone-Wesson, Vohr et al., 2000
“To determine the accuracy of three
measures of peripheral auditory
system status (TEOAE, DPOAE and
ABR) applied in the perinatal period
for predicting behavioral hearing
status at 8-12 months corrected age.”
42. NIH Multi-Center Study
(Norton et al. Ear & Hearing: 5; 2000)
• 7 institutions
• 7,179 infants evaluated
– 2,348 = WBN babies
– 4,478 = NICU babies
– 355 = well babies with high risk
indicators
• Targeted for VRA @ 8-12 months : NICU, WBN
with HRI, and 80 WBN (no HRI) infants who
failed one or more neonatal testing
• 3,134 (64%) returned for VRA
43. Screening Test Accuracy
for detecting HL > 30 dB HL
using VRA MRL, SAT and PTA2 (2 + 4 kHz) and PTA3
(1, 2, 4 kHz) at 8-12mos as ‘gold standard’
Progressive HL and ME pathology excluded (Norton et al., 2000)
MRL @
1 kHz
MRL @
2 kHz
MRL @
4 kHz
SAT PTA2 PTA3
DP
75/75
0.75 0.89 0.81 0.87 0.86 0.84
DP
65/50
0.70 0.92 0.89 0.92 0.88 0.83
TEOAE
@ 80
0.74 0.92 0.88 0.94 0.90 0.84
ABR @
30
0.90 0.89 0.82 0.91 0.88 0.87
44. Conclusions: Screening Tests
(Norton et al., 2000)
• Screening test failure rate increased as
degree of hearing loss increased
• DP 65/50, TEOAE and ABR similar in
identifying HL of moderate degree and
greater (PTA2: 2.0 + 4.0 kHz)
• Overall poorer performance of DPOAE
75/75 condition
45. Could We Do It With Current
Commercially Available Devices?
46. Automated ABR Test Parameters
Norton et al. 2000
• Custom-designed system; not available
commercially
• 30 dB nHL clicks
– Automated Pass Criterion: Fsp > 3.1
– Visual confirmation of wave V (Observer
Based Criteria: OBC) in addition
• Stimuli calibrated using OAE system &
probe
– Tester could adjust level with resultant voltage
setting for click
– Click transduced via OAE probe
47. Commercially-Available
Automated ABR Screening Products
• 5 of 6 manufacturers allow probe/ear
canal transducer option
– Verify level within tolerance parameters in
order to run the test
• 2 of 6 manufacturers allow either
probe/ear canal transducer or
circumaural cushion option
– Verify level within tolerance parameters in
order to run the test
48. Commercially-Available
Automated ABR Screening Products
• 1 manufacturer uses circumaural
cushion only
– ? Verify level
• 6 of 6 manufactures allow user
(administrator) to set stimulus level
• 3 manufacturers employ some version
of Fsp stopping criteria to determine
pass/fail
• 3 manufacturers can meet Norton et al.
2000
49. Distortion Product OAE Test Parameters
Norton et al. 2000
• Commercially available product (not
screening device)
• Evoking stimuli: 65/50 dB SPL
– Stimulus level (SPL) measured in ear canal
prior to test
• Pass criteria: 3 dB S/N +2 SD
(equivalent to 8-10 dB S/N) at 2, 3, & 4
kHz (f2)
50. Commercially-Available
DPOAE Screening Products
• 6 of 9 manufactures allow either the
user or administrator to set stimulus
levels and/or pass-fail criteria
• 3 manufacturers ship to purchaser
without a way to change levels or pass-
fail criteria
• 6 manufacturers can meet Norton et al.
2000 criteria
51. Transient OAE Test Parameters
Norton et al. 2000
• Commercially available product (not
screening device)
• Evoking stimulus: 80 dB pSPL click
– Stimulus level (SPL) measured in ear canal
prior to test
• Pass criteria: S/N = 3 dB at 1.5 kHz; 6
dB at 2, 3, & 4 kHz bandwidths
52. Commercially-Available
TEOAE Screening Products
• 5 manufactures have available
screening products
• Products cannot be changed re:
stimulus level or pass/fail criteria
• 5 manufacturers can meet Norton et al.
2000 criteria
53. Lacking A National or Agreed-Upon
Standard –
What Else Should You Look for in
Selecting Screening Devices?
54. Considerations for Product Purchase
• Stand alone device: 1 technology
– With true database management
– Without true database managment
• Combination device: OAE/ABR
– With database management
– Without database management
• Ease of use
• Cost of instrument & disposables
• Flexibility
• Durability
• Maintenance
• Service, support & training
55. Suggestions for How Screening Programs
Can select an Appropriate Device
• Decide what you want to screen for (sensory,
neural, mild HL, etc.)
• Consider the screening environment and
screening personnel
• Require manufacturers specifications to
reflect all stimulus parameters and pass/fail
criteria
• Have a evidence-base for selection of
screening parameters and pass/fail criteria