In the Journals

Cough-sound analyzer app shows potential as diagnostic tool for pediatric respiratory diseases

Paul Porter, MD
Paul Porter

A novel automated cough-sound analyzer technology incorporated in a smartphone app shows potential as a high-level diagnostic aid in the assessment of common pediatric respiratory diseases, according to results from a prospective, multicenter study published in Respiratory Research.

“We were able to differentiate the common important childhood respiratory diseases this way without the need for an examination or any testing,” Paul Porter, MD, researcher in the department of pediatrics in the School of Nursing, Midwifery and Paramedicine at Curtin University in Bentley, Western Australia, told Healio Pulmonology. “The technology can be used in situations where the ideal diagnostic workup cannot occur, such as with telehealth, in resource-poor environments, and in areas with differing levels of experience and funding — the test can be provided very cheaply and objectively.”

The differential diagnosis of pediatric respiratory diseases is difficult and suboptimal, as existing algorithms are associated with error rates. Consequently, this results in misdiagnoses, inappropriate use of antibiotics and unacceptable morbidity and mortality, according to background information provided in the study.

However, advancements in artificial intelligence and acoustic engineering in recent years appear promising in identifying respiratory conditions based on sound analysis.

For this reason, Porter and colleagues sought to evaluate the diagnostic accuracy of a novel automated cough-sound analyzer for pediatric respiratory disease among 585 pediatric patients aged 29 days to 12 years.

Researchers used a trained algorithm to analyze a prospective data set of cough sounds and compared diagnoses made between the automated cough analyzer vs. consensus clinical diagnoses from a panel of pediatricians after review of hospital charts and available research to determine positive and negative percent agreement values for various pediatric respiratory conditions.

Results showed the positive and negative percent agreement values for asthma were 97% with the automated analyzer compared with 91% with clinical reference.

This was followed by an 87% value for pneumonia with the automated analyzer vs. 85% with clinical reference, an 85% value for croup with the analyzer vs. 82% with clinical reference, an 84% value for bronchitis with the analyzer vs. 81% with clinical reference, and an 83% value for lower respiratory tract disease with the analyzer vs. 82% with clinical reference.

“We have a lot of additional research underway,” Porter told Healio Pulmonology. “We are very interested in disease severity, particularly asthma severity, and incorporating the tool in the asthma action plan. There are also triage studies that are underway in the emergency department and telehealth settings to show the effect of the automated analyzer on treatment times, appropriateness and outcomes. Not to mention, there is an adult program in which we have announced top-line results showing the technology works in diagnosing COPD and pneumonia, and we are writing this workup now.” – by Jennifer Southall

Disclosures: Porter reports he has scientific adviser and shareholder roles with ResApp Health, the developer of the technology. Please see the study for all other authors’ relevant financial disclosures. The study was funded by ResApp Health; Joondalup Health Campus provided office space, IT services and consumables in kind.

Paul Porter, MD
Paul Porter

A novel automated cough-sound analyzer technology incorporated in a smartphone app shows potential as a high-level diagnostic aid in the assessment of common pediatric respiratory diseases, according to results from a prospective, multicenter study published in Respiratory Research.

“We were able to differentiate the common important childhood respiratory diseases this way without the need for an examination or any testing,” Paul Porter, MD, researcher in the department of pediatrics in the School of Nursing, Midwifery and Paramedicine at Curtin University in Bentley, Western Australia, told Healio Pulmonology. “The technology can be used in situations where the ideal diagnostic workup cannot occur, such as with telehealth, in resource-poor environments, and in areas with differing levels of experience and funding — the test can be provided very cheaply and objectively.”

The differential diagnosis of pediatric respiratory diseases is difficult and suboptimal, as existing algorithms are associated with error rates. Consequently, this results in misdiagnoses, inappropriate use of antibiotics and unacceptable morbidity and mortality, according to background information provided in the study.

However, advancements in artificial intelligence and acoustic engineering in recent years appear promising in identifying respiratory conditions based on sound analysis.

For this reason, Porter and colleagues sought to evaluate the diagnostic accuracy of a novel automated cough-sound analyzer for pediatric respiratory disease among 585 pediatric patients aged 29 days to 12 years.

Researchers used a trained algorithm to analyze a prospective data set of cough sounds and compared diagnoses made between the automated cough analyzer vs. consensus clinical diagnoses from a panel of pediatricians after review of hospital charts and available research to determine positive and negative percent agreement values for various pediatric respiratory conditions.

Results showed the positive and negative percent agreement values for asthma were 97% with the automated analyzer compared with 91% with clinical reference.

This was followed by an 87% value for pneumonia with the automated analyzer vs. 85% with clinical reference, an 85% value for croup with the analyzer vs. 82% with clinical reference, an 84% value for bronchitis with the analyzer vs. 81% with clinical reference, and an 83% value for lower respiratory tract disease with the analyzer vs. 82% with clinical reference.

“We have a lot of additional research underway,” Porter told Healio Pulmonology. “We are very interested in disease severity, particularly asthma severity, and incorporating the tool in the asthma action plan. There are also triage studies that are underway in the emergency department and telehealth settings to show the effect of the automated analyzer on treatment times, appropriateness and outcomes. Not to mention, there is an adult program in which we have announced top-line results showing the technology works in diagnosing COPD and pneumonia, and we are writing this workup now.” – by Jennifer Southall

Disclosures: Porter reports he has scientific adviser and shareholder roles with ResApp Health, the developer of the technology. Please see the study for all other authors’ relevant financial disclosures. The study was funded by ResApp Health; Joondalup Health Campus provided office space, IT services and consumables in kind.