THE LATEST

Researchers at Mass General Brigham have developed an innovative artificial intelligence (AI) algorithm designed to uncover previously undetected instances of long COVID-19 within patients' health records. This novel approach, termed 'precision phenotyping,' utilizes AI to identify signs of long-term COVID-19, track the evolution of symptoms over time, and rule out alternative explanations for patients' conditions.

The methodology introduced by the team suggests that as many as 22.8% of individuals may be experiencing symptoms consistent with long-term COVID-19, offering a more accurate representation of the ongoing impact of the pandemic. By longitudinally analyzing a patient's medical history, this AI tool provides a personalized healthcare approach that can help reduce the disparities and biases often present in current diagnostic methods for long COVID.

The tool developed by Mass General Brigham investigators enables clinicians to effectively sift through electronic health records, identifying cases of long COVID-19 that present a range of persistent symptoms after SARS-CoV-2 infection, including fatigue, chronic cough, and cognitive impairment. Published in the reputable journal Med, the study's results highlight that many individuals may suffer from long COVID without proper recognition, emphasizing the need for improved diagnostic tools.

Senior author Hossein Estiri, who leads AI Research at the Center for AI and Biomedical Informatics of the Learning Healthcare System (CAIBILS) at Mass General Brigham and is an associate professor of medicine at Harvard Medical School, stated, "Our AI tool could transform a confusing diagnostic process into something clear and focused, equipping clinicians to navigate the complexities of this challenging condition." The research aims to uncover the true nature of long COVID and provide insights into effective treatment strategies.

Long COVID, officially defined as the Post-Acute Sequelae of SARS-CoV-2 infection (PASC), consists of many symptoms that challenge physicians to differentiate between post-COVID symptoms and pre-existing conditions. The algorithm developed by Estiri and colleagues leverages 'precision phenotyping' to explore individual medical records, identify COVID-related symptoms, and track their progression over time, facilitating a distinction between long COVID and other underlying illnesses.

Medical residents, such as Alaleh Azhir from Brigham Women's Hospital within the Mass General Brigham system, have emphasized the potential impact of AI-powered diagnostic tools in streamlining the diagnostic process for long COVID. The patient-centered diagnoses generated by this AI tool can help correct biases present in current long COVID diagnostics, offering a more accurate depiction of the population affected by this condition.

While the researchers acknowledge limitations regarding the algorithm's integration with health record data and the regional scope of the study, they propose further investigations to evaluate the tool's efficacy across diverse patient populations. The planned release of this AI algorithm for global access represents a significant step toward enhancing diagnostic accuracy and clinical care on a broader scale.

This pioneering work by Mass General Brigham researchers lays the groundwork for a more comprehensive understanding of the long-term effects of COVID-19 and opens new avenues for future research into the genetic and biochemical underpinnings of long COVID subtypes. This remarkable AI tool has the potential to revolutionize diagnostic practices and pave the way for targeted interventions that address the complex challenges posed by COVID-19.