The potential of machine learning to improve medical diagnosis
According to the Society to Improve Diagnosis in Medicine, misdiagnosis is among the most common, catastrophic, and costly medical errors. He also estimates that errors affect more than 12 million Americans each year, with associated costs likely in excess of $100 billion. An accurate medical diagnosis is an essential first step in patient care and greatly improves a patient’s overall chances of achieving positive health outcomes.
In recent years, machine learning, a type of artificial intelligence (AI), has become a powerful tool for improving medical diagnosis. Machine learning technologies are trained to identify patterns that may be hidden or complex. For example, after feeding large amounts of data to a computer, machine learning can identify structure and patterns in the data. Then it can use these patterns to predict answers to problems or group the information into useful groups for comparison, such as similar images of cancerous lesions. Machine learning can also be used to identify patterns that may be hidden or complex, such as details from X-ray, ultrasound, and magnetic resonance imaging (MRI) imaging.
Today’s WatchBlog article examines our recent work on the use of machine learning to provide new medical diagnostic capabilities, as well as some of the challenges this technology faces in becoming mainstream.
How Could Machine Learning Affect Medical Diagnosis?
Although still in the early stages of implementation, machine learning has the potential to provide more accuracy in diagnostic results, as well as save time and money, and most importantly, save lives.
For example, machine learning could detect diseases earlier. Six out of 10 Americans live with a chronic disease, such as cancer or heart disease, at least once. Machine learning is able to aid in cancer diagnosis by using medical imaging data to detect, measure and analyze tumours. By applying its advantage in computing power to perform data and image analysis faster than human medical professionals can alone, machine learning could perform screenings in less time. This could reduce referral wait times for high-risk patients and ease the burden on clinics facing understaffing or other challenges.
Machine learning technologies could also improve diagnostic consistency and accuracy by removing situations that contribute to human error. For example, human specialists performing a diagnosis are affected by factors such as fatigue and may vary in their interpretation of data and images.
Machine learning could also expand access to healthcare. Certain regions and populations in the United States have limited access to healthcare professionals. This emerging technology could automate certain tasks, which could reduce clinical workloads and allow non-specialists to perform complex tasks, such as cardiac imaging and analysis. This could allow healthcare professionals to reach wider segments of the population in home care settings or smaller clinical settings, and provide more patients with access to care.
How widespread is machine learning in healthcare and what might limit its use?
Several machine learning technologies are used by healthcare professionals in the United States, with most technologies relying on imaging data such as X-rays or MRIs. Our recent works examined how machine learning was used to diagnose five common diseases: certain cancers, diabetic retinopathy, Alzheimer’s disease, heart disease and COVID-19. Cancer was the most common imaging-based application today, and machine learning was used to help detect, measure, and analyze tumors and lesions.
While researchers continue to expand AI and machine learning capabilities in medical diagnostics, these technologies have generally not been widely adopted and face a number of challenges limiting more widespread use. For example, some medical providers may be reluctant to use machine learning in their clinics until its performance is more widely proven in various clinical settings. The lack of knowledge among some healthcare professionals about how machine learning would fit into and improve their workflow, as well as gaps in guidance and regulatory requirements, as well as the cost of implementing implementation and maintenance, may also limit its development and use.
Our recent report details these challenges and offers policy options for lawmakers to consider. These include policies to encourage or require evaluation of machine learning diagnostic technologies in a range of real-world settings, expand access to high-quality medical data, and promote collaboration between developers, vendors and regulators. Learn more about machine learning and AI in medical diagnostics by consulting our report.