Spirometry, the measurement of breath, is one of the commonly used tests in the diagnosis of pulmonary and respiratory illnesses. Today, patients can only have spirometry tests conducted at a clinic or a hospital's Pulmonary Function Test Lab. Since accurate spirometry requires an expert technician's guidance, all diagnostic spirometries can only be conducted at these locations. As a result, the wait times for spirometries are often long, e.g. 3 months in Houston and nearly impossible even in large towns in India. This research team will develop SmartSpiro, the first clinical grade portable spirometer that brings spirometry to the masses.
The innovation in SmartSpiro is that it performs a role of a trained technician by algorithmically finding errors in spirometry, using a combination of signal processing and machine learning. Current prototype can already detect all the major spirometry errors with 94% accuracy and it will be improved as algorithms are further refined. In addition to error correction, to improve ease of use SmartSpiro for non-expert users, the team will be adding local language customization and spoken word feedback capabilities. They will also be designing customizable incentive-driven displays on the spirometer hardware to support better patient compliance. SmartSpiro aims to commercialize a device which caters to under-served patient segments and pulmonary conditions, e.g. developing countries in India, or rural areas in the developed countries like Houston. In addition, SmartSpiro's automated test capability will allow for use of spirometry outside of testing labs, and empower point-of-care pulmonary testing (e.g. at the patient?s bedside) which is currently infeasible using current spirometry technology.
