Measure Lung Function on a Mobile Phone
Home spirometry is gaining acceptance in the medical community because of its ability to detect pulmonary exac-erbations and improve outcomes of chronic lung ailments. However, cost and usability are significant barriers to its widespread adoption. To this end, we present SpiroSmart, a low-cost mobile phone application that performs spirometry sensing using the built-in microphone. We evaluate Spi-roSmart on 52 subjects, showing that the mean error when compared to a clinical spirometer is 5.1% for common measures of lung function. Finally, we show that pul-monologists can use SpiroSmart to diagnose varying de-grees of obstructive lung ailments.
Spirometry is the most widely employed objective measure of lung function [37] and is central to the diagnosis and management of chronic lung diseases, such as asthma, chronic obstructive pulmonary disease (COPD), and cystic fibrosis. During a spirometry test, a patient forcefully ex-hales through a flow-monitoring device (a tube or mouth-piece), which measures instantaneous flow and cumulative exhaled volume (Figure 1). Spirometry is generally per-formed in medical offices and clinics using conventional spirometers, but home spirometry with portable devices is slowly gaining acceptance [6,26]. Measurement of spirome-try at home allows patients and physicians to more regular-ly monitor for trends and detect changes in lung function that may need evaluation and/or treatment. Home spirome-try has the potential to result in earlier treatment of exacer-bations, more rapid recovery, reduced health care costs, and improved outcomes [15,23,34,35]. However, challenges currently faced by home spirometry are cost, patient com-pliance and usability, and an integrated method for upload-ing results to physicians [9,12]. Importantly, while office-based spirometry is coached by a trained technician, current home spirometers have no coaching, feedback, or quality control mechanisms to ensure acceptable measurements.
In this paper, we present SpiroSmart, a smartphone-based approach that measures lung function using the phone’s built-in microphone (i.e., a complete software-enabled solu-tion). SpiroSmart requires the user to hold the smartphone at approximately arm’s length, breathe in their full lung volume, and forcefully exhale at the screen of the phone until the entire lung volume is expelled. The phone’s mi-crophone records the exhalation and sends the audio data to a server, which calculates the exhaled flow rate by estimat-ing models of the user’s vocal tract and the reverberation of sound around the user’s head. Flow rate is estimated by calculating the envelope of the sound in the time domain; performing resonance tracking in the frequency domain; while measuring white noise gain through linear prediction. SpiroSmart is able to compute and provide flow rates and graphs similar to those found in home or clinical spirome-ters (Figure 1).
There are important advantages to developing a smartphone-based solution as compared to current commer-cial spirometers. Firstly, the low-cost and inherent portabil-ity of the smartphone allows much greater uptake of home spirometry. The relative low cost of smartphones as com-pared to spirometers can also help in lowering access barri-ers to medical devices in the developing world. Secondly, a smartphone spirometer can have built-in coaching and feedback—mechanisms to maximize measurement accepta-bility that are critically lacking in current home spirometers. Thirdly, smartphones provide the capability of easy data uploading, enabling longitudinal tracking of results and instantaneous alerts. Finally, with the smartphone, spirome-try can easily be coupled with evaluations such as symptom scores, cough sensing, or oximetry to provide a comprehen-sive disease self-management tool.
Our research team closely collaborated with experienced pulmonologists from two different hospitals to help inform the design of SpiroSmart and to critically compare the accu-racy of SpiroSmart to a clinical spirometer (endorsed by the American Thoracic Society or ATS). Through a study that included 52 subjects, we show that SpiroSmart has a mean error of 5.1% for the most common measures of lung func-tion and that SpiroSmart can be used directly out-of-the-box, without any user-specific training or calibration. How-ever, our results indicate that the performance of Spi-roSmart improves if we calibrate this generic system for a particular user, decreasing the mean error in estimation of lung function to 4.6%. Lastly, we compare the differences in diagnosis from five pulmonologists using measures and graphs generated from SpiroSmart and from a clinical spi-rometer. We show that SpiroSmart is effective for diagnos-ing not only abnormal lung function but also the degree of obstruction.