Collaborating Authors

Can AI Identify Patients With Long COVID?


Long COVID refers to the condition where people experience long-term effects from their infection with the SARS CoV-2 virus that is responsible for the COVID-19 disease (Coronavirus disease 2019) pandemic according to the U.S. Centers for Disease Control and Prevention (CDC). A new study published in The Lancet Digital Health applies artificial intelligence (AI) machine learning to identify patients with long COVID-19 using data from electronic health records with high accuracy. "Patients identified by our models as potentially having long COVID can be interpreted as patients warranting care at a specialty clinic for long COVID, which is an essential proxy for long COVID diagnosis as its definition continues to evolve," the researchers concluded. "We also achieve the urgent goal of identifying potential long COVID in patients for clinical trials." Globally there have been over 510 million confirmed cases of COVID-19 and more than 6.2 million deaths according to April 2022 statistics from Johns Hopkins University.

Coronavirus long haulers often face fatigue, neurologic symptoms months later: study

FOX News

Fox News medical contributor Dr. Marc Siegel responds to Sen. Rand Paul and Dr. Fauci clashing over coronavirus origins Some young, otherwise healthy patients who experienced mild coronavirus reported experiencing fatigue, respiratory issues and neurologic symptoms months after recovering from the illness, according to a new study. While imaging studies appeared normal, patients reported "debilitating symptoms," building on the complexities of post-COVID-19 syndrome . Results were published in Mayo Clinic Proceedings on Tuesday, stemming from the first 100 patients treated from June to late December at the hospital's COVID-19 Activity Rehabilitation program in Rochester, Minn. The patients were 45 years old on average and 68% were women. The vast majority had not been hospitalized for COVID-19, though months later they presented with "fatigue (80%), respiratory complaints (59%) and neurologic complaints (59%)" and mood disorders, per the study.

Coronavirus 'long haulers' most often battle fatigue, 'brain fog,' study suggests

FOX News

Dr. Zijian Chen, medical director at the Mount Sinai Center for Post-COVID Care weighs in on'America's News HQ.' Most non-hospitalized coronavirus "long haulers," or patients facing symptoms over a month later, report more than four neurologic symptoms, with "brain fog" being the most predominant, according to what researchers say is the first study of its kind. A team of experts at Northwestern Medicine published peer-reviewed findings in the Annals of Clinical and Translational Neurology on Tuesday. The study intended to characterize the range of neurologic manifestations "long haulers" endure. Researchers analyzed 100 non-hospitalized "long haulers," 50 of which had confirmed positive virus tests and 50 did not. The average age of participants was 43, and the majority of participants were female.

The road to addressing Long Covid


The risk of COVID-19 has been largely communicated only in terms of deaths and hospital capacity, with recovery and survival conflated with each other. Around one in three people with symptomatic COVID-19 still experience symptoms 12 weeks after onset ([ 1 ][1]). Long Covid can be experienced by all age groups and not only those with acute severe disease. The debilitating symptoms are wide-ranging, multisystemic, and predominantly fluctuating or relapsing. There is still much to understand about Long Covid, but what is not well understood should not be ignored. Long Covid is likely the first illness in history that has been defined by patients through social media platforms such as Twitter and Facebook. People with Long Covid formed a movement that demanded recognition of what was happening to them. During the first wave of the pandemic in 2020, online testimonials of prolonged symptoms following severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection were the only source of reassurance to others with a similar experience, including this author ([ 2 ][2]). In the absence of any guidance or recognition about the possibility of a persistent illness, peer support is all that people with Long Covid had. Many previously healthy and active people described persistent symptoms of the acute illness that fluctuated, with new symptoms appearing weeks later. In many countries, most nonhospitalized people did not have lab confirmation of SARS-CoV-2 infection owing to lack of access to community testing, so their symptoms remained without a diagnosis. By summer 2020, thousands were joining social media support groups. A common theme started to emerge: lack of recognition by the medical profession. Patients, including doctors, with Long Covid were consulting their health care providers, and their symptoms were commonly minimized, dismissed, or labeled as anxiety ([ 3 ][3]). A narrative emerged of people struggling to make sense of their symptoms and forming their own groups to understand and research what was happening to them in an international citizen science movement ([ 4 ][4]). The testimonials of people living with Long Covid demonstrated themes of stigma and discrimination. Those whose initial disease was characterized as “mild” commonly experience Long Covid. There is some indication that having more symptoms at the start of the illness is linked to the development of Long Covid and more multisystem involvement later ([ 5 ][5], [ 6 ][6]). The most prevalent symptom of Long Covid is commonly called “fatigue.” This is often mistaken for tiredness, but it is better described as a feeling of utter exhaustion, energy drain, or bodily dysfunction that is not necessarily triggered by exertion and is not always relieved by rest. The prevalence of fatigue is followed closely by symptoms of cognitive dysfunction, including poor memory or concentration, confusion, and “brain fog”. Chest pain or heaviness, breathlessness, headache, muscle aches, dizziness, and palpitations are also common ([ 5 ][5]). A wide range of other symptoms have also been reported, affecting the cardiopulmonary, neurocognitive, and gastrointestinal systems, as well as effects on skin and eyes, and general pain, making Long Covid a multisystem disease ([ 4 ][4], [ 5 ][5]). Triggers of symptoms include physical activity, stress, sleep disturbance, and cognitive tasks ([ 5 ][5]). Most prevalence estimates to date are based on follow-up of hospitalized patients. Follow-up of discharged COVID-19 patients in Wuhan, China, 6 months after symptom onset showed that 76% were still symptomatic ([ 7 ][7]). In nonhospitalized COVID-19 patients, prevalence estimates are variable depending on the study design, applied definitions, population sample, and duration of follow-up. A UK community prevalence study of over half a million people [Real-time Assessment of Community Transmission 2 (REACT-2) study] reported a prevalence among those who reported having COVID-19 of 38% (33% in males, 42% in females) with at least one symptom lasting 12 weeks or more, and 15% having at least three symptoms lasting 12 weeks or more ([ 1 ][1]). The Office for National Statistics (ONS) estimates that in June 2021, almost 1 million people who reported having COVID-19 had symptoms for more than 4 weeks in the UK (1.5% of the population), of which 385,000 were estimated to have had COVID-19 at least a year ago ([ 8 ][8]). All age groups were affected by Long Covid, including children, with an estimated 33,000 aged 2 to 16 years with Long Covid, of which 26,000 had symptoms for at least 12 weeks and 9000 for at least 1 year ([ 8 ][8]). Whole-population prevalence of self-reported Long Covid of any duration as estimated by the ONS was highest in working-age adults (1.6% in 25- to 34-year-olds and 2.1% in 35- to 69-year-olds), particularly those in frontline professions ([ 8 ][8]). In those aged 2 to 11 and 12 to 16, the estimated population prevalence was 2 in 1000 and 5 in 1000, respectively ([ 8 ][8]). These figures mean that in other countries that also experienced high rates of infection, such as Brazil, the United States, and India, millions of economically active people may be disabled by Long Covid. ![Figure][9] Meeting the need of Long Covid The public health response to the COVID-19 pandemic needs to adequately address the direct long-term effects of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection in the context of the ongoing pandemic. An adequate response should incorporate the 4 Rs: Reporting, Recognition (including Rehabilitation), and Research. GRAPHIC: N. CARY/ SCIENCE Experiencing Long Covid can result in considerable disability, functional limitations, and loss of productivity and resources. Two-thirds of people with Long Covid say it limits their ability to undertake their day-to-day activities ([ 8 ][8]). It substantially affects not only leisure and social activities, but also being able to self-care, care for children or older adults, and carry out domestic chores ([ 5 ][5]). Also, it often affects the ability to work, commonly resulting in taking sick leave and losing income ([ 5 ][5]). This will likely accentuate the socioeconomic disparities that are reflected by rates of SARS-CoV-2 infection and mortality. Both the ONS and REACT-2 data show highest prevalence of Long Covid in those living in the most economically deprived areas ([ 1 ][1], [ 8 ][8]). The effect of Long Covid on mental health is also concerning. In this context, anxiety can be caused by multiple factors, including the uncertainty of prognosis and treatment, as well as being denied recognition, employment benefits, and support because of being disbelieved, particularly if there is no lab confirmation of SARS-CoV-2 infection. Anxiety may be secondary to not recovering rather than being the primary manifestation of the illness. Being diagnosed with anxiety with no adequate attention to other symptoms can be isolating and detrimental to the patient's well-being ([ 3 ][3]). Health inequities are likely to widen in such scenarios because some groups who already suffer structural disparities may be stereotyped as less credible in interpreting their own health. The diagnostic criteria of Long Covid are still not standardized. Indeed, even the name varies by country and institution (e.g., post-COVID-19 syndrome, post-COVID-19 condition, postacute sequelae of COVID-19). Without universal criteria that do not exclude those without lab-confirmed infection, health and social care systems will not be able to accurately track the prevalence and address the impact of Long Covid. One important issue is whether “Long Covid,” as a label, will include organ pathology diagnosed weeks or months after COVID-19, or whether these cases move out into an alternative diagnostic category, leaving only those with “unexplained” symptoms as having Long Covid. In this scenario, follow-up of Long Covid patients with thorough clinical assessment and investigations if symptoms continue would still be needed to avoid missing treatable pathology and prevent neglect and stigmatization. Long-term sequelae have been reported with other viral infections. Most relevant are other coronavirus diseases, with a quarter to a third of those with SARS and Middle East respiratory syndrome (MERS) having lingering lung function abnormalities, reduced exercise capacity, and psychological manifestations ([ 9 ][10]). Autonomic dysfunction after viral illness, which has been observed in Long Covid but is also a feature of similar conditions such as myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS), can cause disabling syndromes including postural orthostatic tachycardia syndrome (POTS). Exact case definitions of what is collectively termed “postviral syndrome” are needed. Long Covid research may also be applicable to a wider range of chronic illnesses, including ME/CFS, which similarly lack adequate understanding and recognition. Tissue scarring, and organ damage, could be the cause of persistent symptoms. Patients surviving COVID-19 in the short term have higher rates of organ-specific pathology, including of the heart, liver, kidneys, and pancreas, and higher rates of hospital read-mission and mortality a few months after infection ([ 10 ][11]). This may not fully explain the range of symptoms and disability experienced by many “long-haulers” with no obvious evidence of organ damage. However, mechanisms such as vascular damage, hypercoagulability, and microthrombosis have also been implicated ([ 11 ][12]). These could go undetected because a clinical diagnosis is dependent on the extent of clinical investigations. Another potential mechanism involves immune system dysregulation and autoantibodies, consistent with the cyclical nature of the symptoms. People with Long Covid have increased concentrations of cytokines in serum at 8 months after infection, indicating up-regulation of immune and inflammatory mediators ([ 12 ][13]). Even viral reservoirs as a potential cause of Long Covid cannot currently be excluded. SARS-CoV-2 nucleocapsid protein has been detected in extrapulmonary tissues, including gut, liver, gall bladder, and lymph nodes, up to 6 months after infection ([ 13 ][14]). Responses to COVID-19 vaccines may help identify the underlying mechanisms. Vaccination could help the immune system tackle residual virus, or, if the mechanism is autoimmune, immunization might “reset” the immune system ([ 14 ][15]). Therefore, it is imperative that vaccinated people with Long Covid are systematically followed-up. The mechanisms underlying Long Covid may be different in different groups of patients or may coexist in the same patient. Currently, Long Covid is an umbrella term that may include multiple conditions. To move forward to a more systematic response to the crisis of Long Covid, better reporting, recognition, and research are needed ([ 15 ][16]) (see the figure). Reporting involves systems that can measure Long Covid. This can be achieved through agreeing on specific diagnostic criteria, establishing disease registers, and following up those with acute infection or a positive test using contact tracing infrastructures. It is unknown if or when many of those with Long Covid will recover, particularly given the relapsing nature of the illness. Surveillance systems must start assessing recovery and medium-term survival (1, 2, and 5 years after SARS-CoV-2 infection). Without knowing how many people remain ill following acute infection, the pandemic and postpandemic responses will always be deficient because they will not account for the full impact of COVID-19. Recognition requires listening and believing patient testimonies, thorough clinical assessment and investigations, personalized treatment, and rehabilitation. It must include equitable clinical and social care pathways, addressing financial support, and employment rights. Long Covid is a multi-organ condition that necessitates a multidisciplinary clinical approach ([ 11 ][12]). Patients should not feel that they have to prove their own lived experience in a doctor-patient context. This includes parents and carers of children with Long Covid. It is not the patient's duty to convince; it is the doctor's duty to listen and not prejudge. Rigorous research to understand the mechanisms, risk factors, prognosis, and subgroup characteristics, and to identify potential therapeutics for Long Covid, is desperately needed. Other chronic conditions, such as ME/CFS, fibromyalgia, and some connective tissue disorders, are largely under-researched in terms of underlying mechanisms, diagnostics, therapeutics, and management options. The understanding of Long Covid offers an opportunity to pave the way toward better outcomes for all patients experiencing similar conditions. In addition to mechanistic studies and trials of potential therapies, multiple research methods should be applied, including representative community-based population studies, clinical and health care studies, and qualitative analysis of lived experiences ([ 5 ][5]). Each have their advantages and biases, but the true picture of Long Covid can only be ascertained with a multidisciplinary approach. Sole dependence on recruiting patients through health care services will exclude many people, making such studies unrepresentative and nongeneralizable. Heath records studies are likely to be biased toward people with more access to testing and care in the acute phase of their illness, and/or to those with high health literacy and resources to navigate health care systems. In the absence of specific guidelines to limit clinical variation in diagnosis due to differences in knowledge and beliefs, research must also recruit directly from the community, harnessing the power of citizen science. The road to properly addressing Long Covid is long and must be traveled with humility, open mindedness, compassion, and scientific rigor. This is relevant not only to COVID-19, but also future pandemics and to other neglected chronic conditions. Science, policy, and society as a whole seem to acknowledge and address immediate impacts much better than the subsequent effects. Let this pandemic be the time to change that. 1. [↵][17]1. M. Whitaker et al ., Persistent symptoms following SARS-CoV-2 infection in a random community sample of 508 707 people (Imperial College London. 24 June 2021); . 2. [↵][18]1. N. A. Alwan , Nature 584, 170 (2020). [OpenUrl][19][CrossRef][20] 3. [↵][21]1. T. Kingstone et al ., BJGP Open. 4, bjgpopen20X101143 (2020). 4. [↵][22]1. H. E. Davis et al ., EClinicalMedicine 10.1016/j.eclinm.2021.101019 (2021). 5. [↵][23]1. N. Ziauddeen et al ., medRxiv, 10.1101/2021.03.21.21253968 (2021). 6. [↵][24]1. C. H. Sudre et al ., Nat. Med. 27, 626 (2021). [OpenUrl][25] 7. [↵][26]1. C. Huang et al ., Lancet 397, 220 (2021). [OpenUrl][27][CrossRef][28][PubMed][29] 8. [↵][30]Office for National Statistics, “Prevalence of ongoing symptoms following coronavirus (COVID-19) infection in the UK: 1 July 2021,”; . 9. [↵][31]1. H. Ahmed et al ., J. Rehabil. Med. 52, jrm00063 (2020). [OpenUrl][32][CrossRef][33][PubMed][29] 10. [↵][34]1. D. Ayoubkhani et al ., BMJ 372, (2021). 11. [↵][35]1. A. Nalbandian et al ., Nat. Med. 27, 601 (2021). [OpenUrl][36][CrossRef][37][PubMed][38] 12. [↵][39]1. C. Phetsouphanh et al ., medRxiv 10.1101/2021.06.01.21257759 (2021). 13. [↵][40]1. C. C. L. Cheung et al ., Gut gutjnl-2021-324280 (2021). 14. [↵][41]1. K. Katella , “Why Vaccines May Be Helping Some With Long COVID,” Yale Medicine, 2021; . 15. [↵][42]1. N. A. Alwan , “We must pay more attention to covid-19 morbidity in the second year of the pandemic,” BMJ Opinion, 2021; . Acknowledgments: N.A.A. has experienced Long Covid and has contributed in a nonfunded advisory role to World Health Organization post-COVID-19 condition definition and outcomes meetings. She receives research support from the National Institute for Health Research (NIHR), NIHR Southampton Biomedical Research Centre, and NIHR Applied Research Collaboration Wessex. [1]: #ref-1 [2]: #ref-2 [3]: #ref-3 [4]: #ref-4 [5]: #ref-5 [6]: #ref-6 [7]: #ref-7 [8]: #ref-8 [9]: pending:yes [10]: #ref-9 [11]: #ref-10 [12]: #ref-11 [13]: #ref-12 [14]: #ref-13 [15]: #ref-14 [16]: #ref-15 [17]: #xref-ref-1-1 "View reference 1 in text" [18]: #xref-ref-2-1 "View reference 2 in text" [19]: {openurl}?query=rft.jtitle%253DNature%26rft.volume%253D584%26rft.spage%253D170%26rft_id%253Dinfo%253Adoi%252F10.1038%252Fd41586-020-02335-z%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [20]: /lookup/external-ref?access_num=10.1038/d41586-020-02335-z&link_type=DOI [21]: #xref-ref-3-1 "View reference 3 in text" [22]: #xref-ref-4-1 "View reference 4 in text" [23]: #xref-ref-5-1 "View reference 5 in text" [24]: #xref-ref-6-1 "View reference 6 in text" [25]: {openurl}?query=rft.jtitle%253DNat.%2BMed.%26rft.volume%253D27%26rft.spage%253D626%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [26]: #xref-ref-7-1 "View reference 7 in text" [27]: {openurl}?query=rft.jtitle%253DLancet%26rft.volume%253D397%26rft.spage%253D220%26rft_id%253Dinfo%253Adoi%252F10.1016%252FS0140-6736%252820%252932656-8%26rft_id%253Dinfo%253Apmid%252Fhttp%253A%252F%252Fwww.n%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [28]: /lookup/external-ref?access_num=10.1016/S0140-6736(20)32656-8&link_type=DOI [29]: /lookup/external-ref?access_num=http://www.n&link_type=MED&atom=%2Fsci%2F373%2F6554%2F491.atom [30]: #xref-ref-8-1 "View reference 8 in text" [31]: #xref-ref-9-1 "View reference 9 in text" [32]: {openurl}?query=rft.jtitle%253DJ.%2BRehabil.%2BMed.%26rft.volume%253D52%26rft.spage%253Djrm00063%26rft_id%253Dinfo%253Adoi%252F10.2340%252F16501977-2694%26rft_id%253Dinfo%253Apmid%252Fhttp%253A%252F%252Fwww.n%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [33]: /lookup/external-ref?access_num=10.2340/16501977-2694&link_type=DOI [34]: #xref-ref-10-1 "View reference 10 in text" [35]: #xref-ref-11-1 "View reference 11 in text" [36]: {openurl}?query=rft.jtitle%253DNat.%2BMed.%26rft.volume%253D27%26rft.spage%253D601%26rft_id%253Dinfo%253Adoi%252F10.1038%252Fs41591-021-01283-z%26rft_id%253Dinfo%253Apmid%252F33753937%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [37]: /lookup/external-ref?access_num=10.1038/s41591-021-01283-z&link_type=DOI [38]: /lookup/external-ref?access_num=33753937&link_type=MED&atom=%2Fsci%2F373%2F6554%2F491.atom [39]: #xref-ref-12-1 "View reference 12 in text" [40]: #xref-ref-13-1 "View reference 13 in text" [41]: #xref-ref-14-1 "View reference 14 in text" [42]: #xref-ref-15-1 "View reference 15 in text"

Researchers use machine learning to identify US patients with long COVID


A group of Northeastern researchers is tapping into the power of machine learning to develop new models for identifying patients who may have post-acute sequelae of SARS-CoV-2 infection, or so-called "long COVID." Using electronic health records from the National COVID Cohort Collaborative, a federal database that compiles medical information about COVID-19 patients, researchers were able to develop models that helped identify COVID long haulers across a range of features--from past COVID diagnosis, to the types of medications they've been prescribed, according to new research published in Lancet Digital Health. The data harmonization effort drew from a variety of information sources to construct a picture of what long COVID looks like in the U.S.--and who is most likely to have it. Those sources include demographic data, healthcare visit details, diagnoses and medications for 97,995 adults with COVID-19, the study says. Patients most likely suffering from the post-infection illness, which is estimated to plague between 10-30% of people who contract COVID-19, are often characterized as having new or lingering symptoms that are present 90 days after being diagnosed with the viral infection--a criteria researchers also used to determine their base population in their analysis.