Sana AudioPulse
Massachusetts Institute of
Technology (MIT)
Students:
Advisors:
Federal University of Rio
Grande do Norte (UFRN)
Marzyeh Ghassemi Kenneth Paik Talis Lincoln Barbalho Ana Gizelle
(Harvard) Lauren Scanlon
Ribeiro
(NorthEastern) Andrew Schwartz
Dr. Leo Anthony Celi Dr. Ikaro
Silva Dr. Peter Szolovits
Dr. Ana Guerreiro
Abstract
Detecting hearing loss in the first 6 months of life is essential to prevent permanent
communication deficits. An estimated 588 million people had mild or severe hearing loss
globally in 2000. In Brazil, an estimated 26.1% of the population had some level of hearing
impairment in 2003. For rural and/or poor populations, access to an audiology clinic is costly
and difficult. Sana AudioPulse presents a low-cost, portable method for ubiquitous national
detection of hearing loss. Our application uses cellphones interfacing with specialized
hardware to objectively screen newborns for hearing impairment with a suite of hearing
standard pediatric screening tests. The audiological data is securely transmitted wirelessly
to a remote centralized medical database for certification and validation by professional
audiologists. Our one and three-year projections estimate a net income of $39,000 and
$575,000 USD during the first and third year (10 then 50 healthcare facilities).
1. Background
Global Issue
In 2000, the WHO estimated 588 million people had mild or moderate-toseverehearing loss. (Mathers 2000) In 2003, an estimated 26.1% of the Brazilian population
showed some level of hearing impairment, and 6.8% were classified in the disabling hearing
impairment group. (Beria, 2003) The consequences of untreated hearing loss are high. After
6 months, the cognitive ability to process sounds significantly degrades in an irreversible
manner. (Harney 2000) Even minimal degrees of hearing loss result in significant
permanent degradation of speech and linguistic abilities. (Olusanya, 2011) Delayed and
insufficient language development can negatively affect academic achievement, social skills
and future employment. The estimated lifetime cost of hearing loss for a child with
prelingual onset exceeds $1 million USD, and many people who are deaf have difficulty
gaining employment. (Mohr, 2000)
Focus on Brazil
Brazil has an estimated 5.7 million people with some degree of hearing loss. The
situation for children is worse: 6% have hearing loss by age 4, and 6 in 1000 have hearing
loss from birth. (Botelho, 2010) Given an estimated 3,023,000 births in Brazil, more than
22,000 infants and 222,000 children are impacted each year. (UNICEF 2010) The Brazilian
government recognized the severity of this problem, and instituted a federal law (2794/01)
obligating a hearing test for every newborn within the first week of birth.
2. Problem
Many infants go untested, despite Brazil’s government mandate, at great cost to the
family and society. (Ashoka Foundation) The Brazilian health care system has only 87
audiology equipped centers that are located only in major cities. These centers distribute
around 150,000 hearing aids per year at an average cost of $300 each. These are primarily
in the south and southeast regions of the country, as the cost of the hearing-test equipment
alone is over $18,000 USD in Brazil (usually imported from U.S.).
For rural and/or poor populations, access to an audiology clinic is costly and difficult.
Patients are referred to private hospitals, where the test is $60 USD per patient and
possibly in another city. The $60 USD price represents almost 3 days of income for an
average Brazilian making only $9,000 USD per year. The few public hospitals with the
necessary equipment tend to have long wait-times that can last for hours or days; even
with this wait, the test is still $15 USD per patient. Thus it is not uncommon for hearing
problems to go undetected and untreated, resulting in academic deficits, safety risks,
difficulties maintaining employment, and interpersonal challenges.
3. Solution
Brazil needs a low-cost, portable method for ubiquitous detection of hearing loss.
Reducing cost and increasing portability is similarly crucial to any hearing loss solution in
the developing world. Brazil’s high penetration of the cell phone market makes a mobile
solution ideal. With over 90% coverage, 4 major competing companies, and 132 cellphones
per 100 inhabitants, mobile phones could provide inexpensive and ubiquitous auditory
screening.
The Sana AudioPulse team has developed an application, using cellphones interfacing
with specialized hardware, to screen for hearing impairment in newborns with a suite of
hearing screening tests. A detailed walkthrough of the Sana AudioPulse application can be
seen in the Application Screenshots section. Data is securely transmitted wirelessly to a
centralized medical database for validation by a trained audiologist and storage in a medical
record database. Our application is developed for the Android mobile operating system, and
specifically for Android 2.1 or later. Using mobile technology allows for wide-spread
deployment of hearing screenings across a broad span of Brazilian communities. Sana
AudioPulse also captures GPS information from cellphones. This data will be used to
determine several broad public health questions regarding the future allocation of screening,
detection and treatment services. These data will also be helpful to visualize geographical
areas which are being covered by the program versus those which we have potential for a
new market.
Sana AudioPulse will be used by non-audiologist medical staff at the field hospitals to
administer a hearing test using the a portable hardware capable of performing Distortion
Product Otoacoustic Emissions (DPOAE) and Tone Burst Otoacoustic Emissions
recordings(Etymotic ER-10C). The ER-10C (Figure 8) is a well-known specialized piece of
hardware designed for psychoacoustical measurements; using ear probes, the evoked
response to two test frequencies is used to generate assess information about the auditory
system. This auditory data is transmitted the ER-10C via a cable (or Bluetooth) to a cell
phone running our customized AudioPulse software. Furthermore, the cell phones equipped
with Sana AudioPulse, additionally guide the medical staff members through standard
clinical questionary to assess the infant’s risk factors (such as family history, infections
during pregnancy, or otoacoustic drugs delivered). Sana AudioPulse uploads the risk-factor
information along with the ER-10C evoked recordings to a remote hospital electronic
medical record system, where authorized audiologists certify and authenticate the results.
Once results are interpreted, audiologist can send requests for further on-site information
(pictures, patient history, etc.), or prescribe a visit to a follow-up audiology center via the
phone’s SMS (text messaging) service.
4. Business Model
The Sana AudioPulse platform benefits from significant commercial advantages in
cost and portability, while facilitating access to highly trained medical professionals. Our
business model is to roll out the system to public and private hospitals and clinics that lack
existing audiology equipment. The Sana AudioPulse hardware will be made available at low
or no cost, with revenue generated via a transaction fee per test and reading.
Consumer Benefit
Hospitals with already overburdened resources have a difficult time justifying the
capital costs of purchasing the old audiology equipment and hiring the required on staff
audiologist. These hospitals are unable to cover these capital costs because they can’t
adequately directly bill their patients as many of these include the rural poor, while the
government reimburses at an inadequate $3 USD per test. The Sana AudioPulse platform
will be offered to hospitals for a low startup cost, well below the $18,000 USD for existing
hardware. Furthermore, hospitals will not need to retain an audiologist on staff, as the
Sana AudioPulse will upload the hearing test data to be interpreted by a remote on
audiologist. The hospital will be charged a transaction fee of $3 per upload, which is
completely offset by the government reimbursement rate, resulting in a net zero additional
cost.
Financial Analysis
Our first-year projections are based on an average 20 births per day in 10
community hospitals. Significant costs include server hardware and maintenance, device
hardware, and staff audiologists. By retaining trained audiologists to only interpret the test
results, we are able to use their time more efficiently, allowing a single audiologist to cover
the output of up to five or more hospitals. The three-year projections estimate scaling the
platform to 50 total clinics in Brazil, resulting in a net income of $575,000 USD. See the
financial projection in Section 8.
Initial Setup
The Sana AudioPulse pilot will be implemented over the course of 6 months in Natal,
Brazil. A detailed timeline is shown in the section 5 timeline table. As students at MIT
develop the mobile application, UFRN students will begin server and cellular network setup
in Brazil. Before deploying the application in Brazil, there will be a local pilot at Children's
Waltham Hospital to compare with state of the art hearing screening methods and the
remote analysis of the system. Once the systems have been validated, Sana AudioPulse will
be integrated with the Audiology Clinic of the UFRN University Hospital as a system field
test. This will then lead to data collection from maternity university hospital in Natal for
approximately a month, where our system will be compared with standards and commercial
methods.
Future Directions
In addition to community hospitals, the platform could be offered to alternative
community access points to healthcare, particularly the widespread vaccination clinics.
These generally publicly supported clinics interact with newborns and children to administer
vaccines throughout Brazil according to the schedule established by the National
Immunizations Program (PNI). As the Sana AudioPulse platform does not require highly
skilled labor, the nurses or healthcare workers could additionally administer the hearing test
to newborn, rather than requiring them to travel to a community hospital.
5. Timeline
April
May
June
July
August
Sept
Application
Development
Server/Cellular
Network Setup in
Brazil
Pilot Testing At
Children's Waltham
Hospital
Sana AudioPulse
Integration with
Brazilian System
Field Test of Entire
System with Initial
Hospital (Maternity of
UFRN)
In-field Data
Collection With 4
Hospitals (> 100
Patients)
6. Conclusion
Sana AudioPulse integrates an innovative medical device with ubiquitous smartphone
enabled cellphones for ubiquitous national detection of hearing loss. We require no new
medical staff to be hired on by healthcare facilities, instead using mobile technology to
extend the reach of current medical professionals. Our business model targets a vital area
of interest to the Brazilian government and care providers. Our application and overall
platform have significant commercial advantages in cost and portability. Revenue is
generated via a $3 transaction fee per test, matching the current government
reimbursement rate. Our strategy is therefore a net zero additional cost to healthcare
providers.
Our system overcomes resource and infrastructure limitations, bringing medical
expertise and important childhood screening tests to the rural and poor. After our initial
pilot of maternity hospitals this September, we plan to expand to 10 and then 50 healthcare
facilities in the first and third year of the project. Our mobile platform is focused on
improving quality of life, educational development, and societal productivity. Sana
AudioPulse is a revolution in childhood screening: an economically sustainability solution
democratizing quality care.
7. References
Botelho, F. A., Bouzada, M. C. F., Resende, L. M. D., Silva, C. F. X., & Oliveira, E. A. (2010).
Prevalence of hearing impairment in children at risk. Brazilian journal of
otorhinolaryngology, 76(6), 739-744.
Mathers C, Smith A, & Concha M. Global burden of hearing loss in the year 2000 [Internet
site].
Available
from:
http://www.who.int/healthinfo/statistics/bod_hearingloss.pdf.
Accessed March 30, 2012.
Beria J. Prevalence of deafness and hearing impairment: Preliminary results of a population
-based study in southern Brazil. Informal Consultation on Epidemiology of Deafness and
Hearing Impairment in Developing Countries and Update of the Who Protocol. WHO,
Geneva, March 2003.
Harney CL. Infant hearing loss: the necessity for early identification. Bol Assoc Med PR.
2000;92(9-12):130-2.
Olusanya BO, Bamigboye BA, & Somefun AO. The burden and management of infants with
life-long and irreversible hearing impairment in Nigeria. Niger J Med. 2011 JulSep;20(3):310-21. Review.
Mohr PE, Feldman JJ, Dunbar JL, McConkey-Robbins A, Niparko JK, Rittenhouse RK, &
Skinner. The societal costs of severe to profound hearing loss in the United States. MW. Int
J Technol Assess Health Care. 2000 Autumn;16(4):1120-35.
Weinstein H. 2008 Ashoka Fellowship, Ashoka Foundation [Internet site]. Available from:
http://www.ashoka.org/fellow/howard-weinstein. Accessed March 30, 2012.
8. Financial Projections
Year 1
Hospitals
Births per day per hospital
Transaction Fee
Year 2
Year 3
10
20
3
20
20
3
50
20
3
$219,000
$438,000
$1,095,000
Server Cost/Maintenance
Software Development
Hardware Cost/Loss
Audiologist
SG&A
$30,000
$80,000
$10,000
$40,000
$20,000
$30,000
$160,000
$20,000
$80,000
$40,000
$30,000
$160,000
$50,000
$180,000
$100,000
Net
$39,000
$108,000
$575,000
Revenue
In US Dollars
9. Application Screenshots
Figure 1- Main Menu
Figure 2-The App will allow for series of
objective hearing tests that do not require
response from the patient (ideal for infant
testing). Two of these are DPOAEs and
TBOAes (see text for detail).
Figure 3- Risk factors for the infant are collected in order for the audiologist to them into
account when assessing the final diagnosis. This list is highly customizable and easily scalable.
Figure 4- Results of the ER-10C hearing test is automatically displayed on the cell phone App for real time
assessment of the infant’s hearing for preliminary notification (until final confirmation by an audiologist).
The pink region represents a border-line loss, and a response *below* the pink region represent moderate-tosevere loss.
Figure 5- The data is transmitted to a remote electronic medical record system for professional validation.
This figure shows the queue for patients waiting for an audiologist’s confirmation of the test results.
Figure 6- The audiologist has access to more sophisticated and professional analysis at the back-end of our
applications. This graph the frequency of both stimuli and the patient’s physiological response.
Figure 7- Other detailed data of the auditory test accessible to the physician include the risk factors (left
column) and the detailed audiogram including stimulus level, physiological response, and noise floor levels.
Figure 8- The ER-10C is a $3,000 portable, specialized hardware capable of recording otoacoustic emissions.