In this issue, Emily Pfaff and colleagues show that machine learning analysis of electronic health records could be crucial in diagnosing patients with long COVID. This is the latest in a plethora of studies to use big data to determine prevalence, symptoms, or risk factors for long COVID. But is big data helping to treat patients with this heterogeneous condition, or should we be redirecting our efforts elsewhere? Big data studies have been essential in gathering crucial information that could help reduce the prevalence of long COVID. Of note, Pfaff and colleagues report that COVID-19 vaccination after acute disease decreased the likelihood of patients being identified as having persistent symptoms after acute COVID-19, a result confirmed by a community-based cohort study in the UK. Using a deidentified electronic health record dataset from more than 90 000 patients with confirmed COVID-19 in the USA to train the algorithm, Pfaff and colleagues identified about 100 000 potential long COVID cases in the National COVID Cohort Collaborative database (representing nearly 5 million COVID-19-positive cases) as of October, 2021. This estimation of the prevalence of long COVID cases in the USA differs from other predictions and might present the most accurate to date. ... During a pandemic, it is unsurprising that researchers’ scientific goals sometimes differ from those of patients; however, there has been discomfort with the poor engagement of patients in long COVID research. In an open letter, patient advocacy groups have highlighted that meaningful involvement of patient advocates and post-viral illness experts must be integrated into the RECOVER initiative to produce clinically meaningful research. As the need to find treatments for long COVID becomes increasingly urgent in the USA, patient engagement is crucial in the development of the next generation of COVID treatments. With such a heterogeneous range of patient characteristics, persistent symptoms, disease severities, management approaches, and SARS-CoV-2 strain infection and vaccination histories, teasing out the underlying mechanisms of disease and identifying treatment targets and strategies for management continues to be a challenge. Progress requires input from a range of health-care professionals, researchers from a variety of disciplines, and people living with long COVID. https://www.thelancet.com/journals/landig/article/PIIS2589-7500(22)00113-3/fulltext
