Diagnosis of two of the most common ear diseases in children may someday rely less on specialized medical equipment and more on widely used modern technology available to more providers and patients, particularly in resource-limited communities.
Investigators at the University of Washington recently published a study showing the feasibility of using smartphones to detect middle ear fluid in children with suspected acute otitis media and acute otitis media with effusion.
The study also included data from a clinical study of more than 98 patient ears testing the sensitivity and specificity of smartphone use in this setting, which demonstrated a diagnostic potential of smartphones comparable to and even better than the traditional diagnostic tools of pneumatic otoscopy or tympanometry.
“In the future, being able to perform diagnostic tests using technology that is readily available could increase the availability of healthcare in resource-limited regions and decrease the cost of healthcare if we can avoid emergency room visits and patients do some initial triage at home,” said study coauthor Sharat C. Raju, MD, a resident with the department of otolaryngology–head and neck surgery at the University of Washington in Seattle.
How It Works
With an array of sensors that includes microphones and speakers, smartphones have the technological components needed to assess tympanic membrane mobility to detect middle ear fluid. Using these two basic components of smartphones, the investigators improved on the technology by using machine learning methods to include additional data processing and signal processing systems to improve on the accuracy of older diagnostic techniques (pneumatic otoscopy or tympanometry).
“What happens is that the microphone [from the smartphone] picks up essentially the reflections of sound of the ear drum similar to tympanometry,” said Dr. Raju. Instead of calculating the angle of the wave form produced using the older data processing system used with tympanometry, which is very regimented and uses strict and fairly arbitrary cutoffs for what constitutes fluid and what does not, Dr. Raju said that the machine learning used with the smartphones informed what the different shapes of the curves meant based on real patients.
The investigators tested the system in two separate cohorts of patients, all of whom underwent otolaryngologic surgery. The first clinical study included patients aged 18 months to 17 years (n = 98 ears), and the second clinical study included children between nine and 18 months of age (n = 15 ears). Patients with existing tympanostomy tubes, those with tympanic membrane perforations, those who had undergone previous tympanoplasty, or those with comorbid middle ear diseases were not included in either cohort.