Hearing Loss
Conditions
Keywords
Rural, Screening, Tympanometer, Mobile health (mHealth), Otitis media, ML Algorithm
Brief summary
Children in rural and underserved communities experience a disproportionately high burden of infection-related hearing loss, which is often preventable with timely identification and follow-up care. A tympanometer is a device that evaluates the health of the middle ear but is often not used in screening programs due to barriers of high cost and need for trained professionals such as audiologists to use the device. To address these barriers, a low cost, lay friendly, mobile health (mHealth) prototype tympanometer has been developed. In addition, a machine learning (ML) algorithm has been designed to guide lay users to interpret tympanometer results, overcoming the need for audiologists. Broad implementation of the lay friendly tympanometer and ML algorithm will transform screening in areas where prevalence of infection-related hearing loss is high, and access to specialty care is limited. In this study the prototype mHealth tympanometer will be evaluated against commercial tympanometry. It will be used by audiologists, an essential step before testing the performance of the device with lay users. Approximately 20 adult patients and 10 pediatric patients with various middle ear pathologies will be enrolled in the study, with the goal of obtaining 60 ears with data from both the prototype and commercial tympanometers. Audiologists will obtain tympanometry measurements on both ears of each participant and interpret the results, first with the prototype device and then with the commercial tympanometer that is typically used during a clinical evaluation. Tympanometry interpretation will include classification into one of 3 categories (Type A, B, or C). Agreement in results between the prototype and commercial devices will be assessed. Agreement between audiologist interpretations and the ML algorithm from the prototype mHealth tympanometer will be assessed.
Interventions
DEVICEPrototype mHealth tympanometer
The prototype mHealth tympanometer functions similarly to a commercial tympanometer. Tympanometers are designed to assess middle ear function by measuring changes in acoustic tone related to subtle movement of the tympanic membrane from an applied air pressure sweep. An audiologist will review the tympanogram tracings and classify results as Type A, B, or C.
DEVICEStandard of care commercial tympanometer
A commercial tympanometer routinely used as part of standard of care will be used as the comparison. Tympanometers are designed to assess middle ear function by measuring changes in acoustic tone related to subtle movement of the tympanic membrane from an applied air pressure sweep. An audiologist will review the tympanogram tracings and classify results as Type A, B, or C.
Sponsors
University of Arkansas
Study design
Allocation
NA
Intervention model
SINGLE_GROUP
Primary purpose
SCREENING
Masking
NONE
Intervention model description
Each participant will be first tested with prototype mHealth tympanometer, and then with a commercial tympanometer routinely used in clinical settings.
Eligibility
Sex/Gender
ALL
Age
1 Years to No maximum
Healthy volunteers
No
Inclusion criteria
* Presenting to Audiology Clinic for evaluation, where tympanometry is warranted for testing at the discretion of the audiologist * Presence of various middle ear health states/pathologies that result in Type A, B, C tympanograms * English-speaking
Exclusion criteria
* Children or adults with cognitive disabilities * Unable to provide consent/assent * Individuals who are unable to sit still * Any other condition that in the opinion of the investigator, might interfere with the safe conduct of the study or place the participant at increased risk
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Classification Agreement Between Prototype mHealth Tympanometer and Commercial Tympanometer | Same day as participant visit, up to 10 minutes | Audiologist will test each ear of the participant first with the prototype device, interpret the results into one of the three tympanogram categories (Type A/B/C) and record the results. The procedure will be repeated with a commercial device. Agreement between the two devices with respect to the classification categories will be assessed. |
| Classification Agreement Between Audiologist and ML Algorithm Interpretations Using Prototype mHealth Tympanometer | Same day as participant visit, up to 10 minutes | Tympanogram results collected using the prototype mHealth tympanometer will be run through the ML algorithm for classification into Type A/B/C. Agreement between the audiologist and the ML algorithm with respect to the classification categories will be assessed. |
Participant flow
Participants by arm
| Arm | Count |
|---|---|
| Tympanometer Comparison Both ears of each participant will be tested first with the prototype mHealth tympanometer followed by standard of care commercial tympanometer | 38 |
| Total | 38 |
Withdrawals & dropouts
| Period | Reason | FG000 |
|---|---|---|
| Overall Study | Withdrawn due to caregiver consent obtained from non-legal guardian | 1 |
Baseline characteristics
| Characteristic | Tympanometer Comparison |
|---|---|
| Age, Categorical <=18 years | 16 Participants |
| Age, Categorical >=65 years | 11 Participants |
| Age, Categorical Between 18 and 65 years | 11 Participants |
| Race (NIH/OMB) American Indian or Alaska Native | 0 Participants |
| Race (NIH/OMB) Asian | 0 Participants |
| Race (NIH/OMB) Black or African American | 7 Participants |
| Race (NIH/OMB) More than one race | 0 Participants |
| Race (NIH/OMB) Native Hawaiian or Other Pacific Islander | 0 Participants |
| Race (NIH/OMB) Unknown or Not Reported | 5 Participants |
| Race (NIH/OMB) White | 26 Participants |
| Sex: Female, Male Female | 24 Participants |
| Sex: Female, Male Male | 14 Participants |
Adverse events
| Event type | EG000 affected / at risk | EG001 affected / at risk |
|---|---|---|
| deaths Total, all-cause mortality | 0 / 39 | 0 / 39 |
| other Total, other adverse events | 0 / 39 | 0 / 39 |
| serious Total, serious adverse events | 0 / 39 | 0 / 39 |
Outcome results
Primary
Classification Agreement Between Audiologist and ML Algorithm Interpretations Using Prototype mHealth Tympanometer
Tympanogram results collected using the prototype mHealth tympanometer will be run through the ML algorithm for classification into Type A/B/C. Agreement between the audiologist and the ML algorithm with respect to the classification categories will be assessed.
Time frame: Same day as participant visit, up to 10 minutes
Population: The analytic sample is limited to ear-level results that produced a classification result of A/B/C for both devices. Ears were not included in the main analysis if data from one device or both devices were missing for that ear.
| Arm | Measure | Value (NUMBER) |
|---|---|---|
| Tympanometer Comparison | Classification Agreement Between Audiologist and ML Algorithm Interpretations Using Prototype mHealth Tympanometer | 0.81 Kappa-statistic (Agreement) |
Primary
Classification Agreement Between Prototype mHealth Tympanometer and Commercial Tympanometer
Audiologist will test each ear of the participant first with the prototype device, interpret the results into one of the three tympanogram categories (Type A/B/C) and record the results. The procedure will be repeated with a commercial device. Agreement between the two devices with respect to the classification categories will be assessed.
Time frame: Same day as participant visit, up to 10 minutes
Population: The analytic sample is limited to ear-level results that produced a classification result of A/B/C for both devices. Ears were not included in the main analysis if data from one device or both devices were missing for that ear.
| Arm | Measure | Value (NUMBER) |
|---|---|---|
| Tympanometer Comparison | Classification Agreement Between Prototype mHealth Tympanometer and Commercial Tympanometer | 93.33 Percent Agreement |