Leveraging Cough Sounds to Optimize Chest X-Ray Usage in Low-Resource Settings

Chest X-ray is a commonly used tool during triage, diagnosis and managementof respiratory diseases. In resource-constricted settings, optimizing thisresource can lead to valuable cost savings for the health care system and thepatients as well as to and improvement in consult time. We usedprospectively-collected data from 137 patients referred for chest X-ray at theChristian Medical Center and Hospital (CMCH) in Purnia, Bihar, India. Eachpatient provided at least five coughs while awaiting radiography. Collectedcough sounds were analyzed using acoustic AI methods. Cross-validation was doneon temporal and spectral features on the cough sounds of each patient. Featureswere summarized using standard statistical approaches. Three models weredeveloped, tested and compared in their capacity to predict an abnormal resultin the chest X-ray. All three methods yielded models that could discriminate tosome extent between normal and abnormal with the logistic regression performingbest with an area under the receiver operating characteristic curves rangingfrom 0.7 to 0.78. Despite limitations and its relatively small sample size,this study shows that AI-enabled algorithms can use cough sounds to predictwhich individuals presenting for chest radiographic examination will have anormal or abnormal results. These results call for expanding this researchgiven the potential optimization of limited health care resources in low- andmiddle-income countries.