Smart algorithms for spirometry

Spirometry is the foundational pulmonary function test (PFT) to monitor general respiratory health. It is used for evaluating abnormal respiratory symptoms, for diagnosing of obstructive airway diseases like chronic obstructive lung disease (COPD) and asthma, for monitoring response to therapeutic intervention and disease management, for assessing preoperative risks, and for measuring endpoints in clinical and epidemiological studies. In the last several years, spirometry has been advocated as one of the standards of primary care in light of the tremendous socio-economic burden on our society due to chronic respiratory diseases. In 2015, chronic obstructive lung (COPD) was the third leading cause of mortality accounting for 5% of all deaths while asthma was prevalent in 5% of the global population. Direct (healthcare), indirect (lost production) and disability-adjusted life years (DALYs) due to COPD and asthma have resulted in an annual economic burden of 210 billion Euros in the EU alone. Therefore, the role of primary care spirometry is critical in addressing the global burden of chronic respiratory diseases. Primary care spirometry allows general practitioners (GPs) to diagnose COPD and asthma that are often presented with undifferentiated symptoms, to determine the efficacy of treatment and management or to decide if a patient should be referred to a specialist. Well-performed spirometry, when interpreted together with medical history and symptoms is the key to timely diagnosis and intervention. This further reduces overhead costs related to redundant diagnostic tests and consultation at speciality centres. However, spirometry remains underused in primary care despite the availability of portable spirometers at affordable prices. Most cases in primary care are still diagnosed and managed purely on the basis of history and physical findings. Many studies have shown that underuse of spirometry has resulted in a disturbing trend of misdiagnoses and inappropriate management. The quality of primary care spirometry also remains a cause of concern, and it is considerably poorer than standardized lung function laboratories. Since spirometry quality assurance involves both subjective and quantitative criteria, it suffers from a large inter-operator variation primarily dependent on training and experience. The main reasons underutilization and unreliability of spirometry in primary care include a lack of expertise in conducting and in interpreting spirometry. While several authors have emphasised the importance of intensive training for those involved with the performance and interpretation of spirometry, most primary care practitioners lack the time and money to attend such workshops. This necessitates the need for new measures that could potentially boost the usage of spirometry in primary care. Thus, this doctoral project is an attempt to address the issues related to insufficient expertise through algorithmic innovation. Our central hypothesis is that the application of AI and other mathematical models on the complete spirometry signal (flow-volume data measured during spirometric manoeuvre), in addition to classical spirometry indices, may help in identifying characteristics crucial to improving clinical outcomes related to spirometry testing. Our goal is to develop an artificial intelligence (AI)-based smart spirometry software that would assist primary care practitioners in performing high-quality spirometry, and would enable them in making superior clinical interpretations by providing detailed diagnostic and phenotyping information. Thus, its capabilities would extend beyond state-of-the-art spirometry software that provides a rudimentary feedback on manoeuvre quality, and generates a report that is left to the reader's expertise for interpretation. This project adopts a multidisciplinary approach that encompasses respiratory medicine, physics, statistics and AI. These technological innovations would render the spirometer into a smart device that would boost its usage in primary care and improve the overall quality of primary care

Publication year:2021

Accessibility:Open