by A newly developed artificial intelligence (AI) model, created by scientists at Scripps Research, has the potential to greatly improve the screening process for atrial fibrillation (AFib)—a common and potentially serious heart condition. AFib is characterized by an irregular and fast heartbeat, and it is associated with an increased risk of stroke and heart failure. The AI model is able to detect tiny variations in a person’s normal heartbeat that are indicative of AFib, which cannot be detected by standard screening tests.
The findings of this study, which were published in the journal npj Digital Medicine on December 12, 2023, utilized data from nearly half a million individuals who wore an electrocardiogram (ECG) patch to record their heart rhythms for a period of two weeks. The ECG patch is a routine screening test for AFib and other heart conditions. The AI model analyzed the collected data to identify patterns that distinguished individuals with AFib from those without the condition. This innovative model has the potential to better detect individuals at risk for AFib, ultimately helping to prevent serious complications such as stroke and heart failure.
The irregular heartbeat associated with AFib can cause blood to pool in the heart, leading to the formation of blood clots, which in turn increase the risk of strokes. AFib is also linked to an increased risk of heart failure and death. Diagnosing AFib can be challenging as many individuals with the condition may only experience occasional episodes of irregular heartbeat or have few noticeable symptoms. While some individuals with AFib may experience heart palpitations, lightheadedness, shortness of breath, and chest pain, others may not exhibit any symptoms at all.
To diagnose AFib, cardiologists typically conduct a detailed ECG test with electrodes placed on the body for about ten seconds. If nothing abnormal is detected, patients may be recommended to use a wearable ECG patch for at-home monitoring for one or two weeks. However, even over a two-week period, individuals with infrequent episodes of AFib may not have their condition captured by the wearable device.
In an effort to identify other patterns in ECG data that could indicate AFib, Giorgio Quer, Ph.D., in collaboration with iRhythm Technologies, analyzed data from 459,889 individuals who wore the company’s at-home ECG patch for two weeks. The machine learning model developed by the team was able to distinguish individuals who later developed AFib from those who did not, even when various known risk factors for AFib were integrated into the model. In fact, the AI model outperformed existing manual models that incorporated demographic data and ECG measures.
The accuracy of the AI model persisted across both older individuals, who are at higher risk for AFib, and individuals under the age of 55, who are typically excluded from general AFib screening due to their lower risk. While the model is not intended for diagnosing AFib, it is a crucial first step towards the development of a screening test for individuals at increased risk or those experiencing symptoms. With the AI model, individuals could wear an ECG patch for just one day to determine if further testing is necessary. Alternatively, the model could analyze one- or two-weeks’ worth of ECG data to identify patients who should undergo a repeat test, even if no AFib is detected during that time frame.
The next steps for Quer and his colleagues involve planning a prospective study. Additionally, they aim to incorporate other sources of data, such as electronic medical records, to further improve the performance of the AI model. By expanding and refining the model, researchers hope to enhance its accuracy and readiness for clinical application.
*Note:
1. Source: Coherent Market Insights, Public sources, Desk research
2. We have leveraged AI tools to mine information and compile it
Money Singh is a seasoned content writer with over four years of experience in the market research sector. Her expertise spans various industries, including food and beverages, biotechnology, chemical and materials, defense and aerospace, consumer goods, etc.
