Abstract
The coronavirus disease 2019 (COVID-19) pandemic had an unprecedented global impact, resulting in both positive and negative consequences. The virus not only affected millions of lives worldwide but also caused long-term harm to multiple organ systems in many survivors, thereby substantially impairing quality of life. This persistent condition is now referred to as long COVID (LC). The aim of this study is to raise awareness of LC-related organ system impacts and to highlight the key role of artificial intelligence (AI) in mitigating these effects. The present research conducts a narrative review focusing on LC-related impacts. In this context, unstructured searches were conducted to identify a total of 69 relevant studies indexed in Embase, PubMed, Web of Science, or Scopus, each of which was reviewed by at least 2 experts with sufficient domain knowledge in health sciences. Based on the authors’ perspectives and insights, the review narratively examines damage to human organ systems attributable to LC and explores the role of AI in addressing LC-related challenges. Significant ethical, practical, and societal concerns arising from the extensive use of AI, particularly major issues such as data privacy and algorithmic bias, are also discussed. LC has caused lasting impacts on human organ systems, while AI is offering substantial potential for LC-related care.
-
Keywords: Artificial-intelligence-role; Concerns; Damage; Human-organ-systems; Long-COVID
Introduction
Coronavirus disease 2019 (COVID-19), caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) virus, triggered a global health emergency of unprecedented scale, claiming millions of lives and profoundly disrupting societies worldwide [
1]. Importantly, prior to the development of COVID-19 vaccines and effective treatments, lockdown measures were the primary strategy available to limit disease transmission. According to reports from the World Health Organization (WHO), COVID-19 vaccination began in August 2020 to reduce viral spread and associated mortality, and approximately 11,811,627,599 vaccine doses had been administered globally as of May 23, 2022. Of these, 5,193,581,059 individuals had received at least 1 dose, while 4,710,246,957 were fully vaccinated [
2]. Owing to mass vaccination efforts, the global situation gradually normalized toward the end of 2023. Assuming that most survivors are vaccinated, millions are now experiencing a range of post-COVID-19 conditions collectively referred to as long COVID (LC). LC is known by several terms, including post-COVID-19 conditions, long-haul COVID, and post-acute COVID-19 syndrome. LC is essentially a chronic condition arising as an after-effect of SARS-CoV-2 infection and/or vaccine administration, persisting for at least 3 months after acute infection, and gradually affecting 1 or more organ systems in survivors. One study reported that approximately 10% to 12% of vaccinated individuals experienced LC [
3]. More than 200 post-COVID-19 symptoms have been documented. Previous studies have identified key LC symptoms, including brain fog, fatigue, fever, insomnia, dyspnea, musculoskeletal pain, diabetes, headaches, skin rashes, neurological disturbances, gastrointestinal issues, and emotional symptoms such as sadness and anxiety, all of which reflect underlying organ system involvement [
4–
6]. In addition, some researchers have raised concerns that COVID-19 may act similarly to environmental stressors—such as air and water pollution, temperature variation, radiation exposure, pesticide exposure, and lifestyle factors including tobacco use, alcohol consumption, and fast-food intake—that increase long-term disease risks, including cancer.
At present, there is no definitive statistical evidence demonstrating an increased post-COVID-19 mortality rate; however, reports of deaths appear more frequent than during the pre-COVID-19 era. This observation highlights the need for further investigation. Encouragingly, ongoing research is actively exploring the causes of post-COVID-19 symptoms and leveraging artificial intelligence (AI)-based approaches to mitigate these impacts. Although numerous studies addressing AI applications in COVID-19 and LC exist in the literature, relatively few have focused on the ethical, practical, and societal concerns associated with the extensive use of AI.
This narrative review aims to provide a broad overview of current understanding of LC, with particular emphasis on its multiorgan system effects and underlying pathophysiological mechanisms. Drawing on the authors’ perspectives and insights, this review incorporates emerging AI applications—including machine learning (ML), deep learning (DL), and natural language processing (NLP)—to address complexities in LC diagnosis, prognosis, symptom monitoring, and therapeutic development. Furthermore, the study examines key ethical, practical, and societal challenges associated with widespread AI deployment in LC care, emphasizing issues such as data privacy, algorithmic bias, and the necessity of human oversight. Through this approach, the study seeks to raise awareness of the complex, multisystemic nature of LC-induced organ damage.
Research Questions
The specific review questions addressed here are: (1) What are the impacts of COVID-19 on the human body?; (2) Is AI playing important roles in mitigating the impacts? If so, how is AI doing it?; (3) What challenges arise from the extensive use of AI?
Contributions of the Study
Raising awareness about COVID-19–induced damage to human organ systems and the global burden. Exploring the potential of AI in addressing LC-related health challenges. Highlighting significant ethical, practical, and societal concerns related to the extensive use of AI. Indicating the growing volume of global research on LC.
Organization of the Paper
In addition to the introduction (Section 1), the materials (research articles) and search methods, including inclusion and exclusion criteria for article selection, are described in Section 2. Section 3 provides an overview of proposed pathophysiological mechanisms underlying LC. Section 4 discusses the effects of COVID-19 infection on major human organ systems in survivors. The overall global impact of LC is examined in Section 5. Section 6 addresses AI-based techniques that were applied during COVID-19 and are now being used to mitigate post-COVID-19 impacts. Ethical concerns, including data privacy and broader challenges associated with AI use, are discussed in Section 7. The limitations of the present study are outlined in Section 8, and Section 9 concludes the article. The overall discussion framework of this study is illustrated pictorially in
Figure 1.
Materials and Methods
LC is currently a broad and evolving topic that requires deeper theoretical insight; therefore, a narrative study design is considered more appropriate. In this context, the present narrative study aims to integrate valuable information from multiple sources without adhering to a standardized methodology for literature selection and analysis, in order to provide a comprehensive understanding of LC-related scenarios. In addition, the study discusses the role of AI in mitigating the identified issues based on the authors’ perspectives and insights. Although a standardized methodology for article selection was not applied, the authors nevertheless collected relevant literature from established and credible data sources using predefined search strategies to support discussion and analysis.
We searched the PubMed, ScienceDirect, Embase, Scopus, and Google Scholar databases for studies published after January 2020, using keywords such as “long COVID,” “post-COVID,” “human organ systems,” “role of AI in mitigating post-COVID,” “AI in patient monitoring,” health justice,” “health data privacy,” and “ethical considerations in AI.” The study includes relevant research articles, review papers, case studies, and materials obtained from web-based sources related to LC. In total, 69 relevant studies were included based on careful screening of titles and abstracts, with each article assessed by at least 2 reviewers possessing sufficient expertise in AI, ML, COVID-19, and data privacy and security. Importantly, reviewers initially examined the title and abstract of each article, followed by full-text evaluation to determine its relevance for inclusion in this study.
Data Sources and Search Strategy
The sources of the included articles were indexed in PubMed, Scopus, or Web of Science (WoS) databases and were ultimately selected for inclusion in this narrative review. The inclusion and exclusion criteria applied to article selection are described below. (1) Included articles comprised research articles, review articles, or book chapters published in scientific journals or presented at standard academic conferences, irrespective of geographic location. (2) Letters to the editor, or articles with unclear or incomplete data and methods were excluded from the study. Additionally, manuscripts not written in English were excluded directly.
Search Strategy
For retrieving articles from the PubMed database, the [tiab] (Title and Abstract) tag was used (e.g., “Long COVID”[tiab] AND “Human Body Impact”[tiab] AND “Role of AI”[tiab]). For searching Scopus-indexed manuscripts, the Scopus platform was accessed directly and the TITLE-ABS-KEY tag was applied (e.g., TITLE-ABS-KEY((“Long COVID” OR “Post-COVID”) AND “Human Organ Impact”)). For articles indexed in the WoS database, searches were conducted using the topic search (TS) function, such as TS=(“Long COVID” OR “Post-COVID”) AND TS=(“Human Organ Impact”) AND TS=(“Role of AI”). The Medical Subject Headings (MeSH) terms applied in the search process included “Long COVID,” “Human body,” “Immune system,” “SARS-CoV-2,” and “pathophysiological mechanisms,” among others.
Proposed Pathophysiological Mechanisms
Pathophysiological mechanisms refer to the complex biological processes within the body that are responsible for the development of disease symptoms and the progression of disease states. These mechanisms disrupt normal physiological functioning and contribute to sustained morbidity. Well-accepted pathophysiological mechanisms relevant to LC are summarized in
Table 1. AI-based approaches assist in exploring these mechanisms by identifying patterns associated with symptom development and disease progression.
Long COVID and Its Impact on Human Organs
This section discusses the impact of LC on various human organ systems.
Long COVID and Its Impact on the Major Human Organ Systems
LC is characterized by a wide spectrum of symptoms that often persist for several months or even years following the initial SARS-CoV-2 infection. The major human organ systems affected, along with the associated damage attributable to COVID-19, are reviewed below.
The nervous system
The nervous system includes the brain, vertebral structures, and an intricate network of nerves, and it plays a central role in regulating bodily functions. It coordinates voluntary actions and sensory input throughout the body by transmitting electrical signals. Commonly reported neurological and neuropsychiatric manifestations include cognitive difficulties (often described as brain fog), memory loss, diminished sense of taste (ageusia), loss of smell (anosmia), headaches, sleep disorders, panic, and anxiety [
7–
10]. According to the WHO clinical definition of LC, these symptoms are more prevalent among individuals who were hospitalized, with 53% of hospitalized individuals reporting symptoms compared with 35% of non-hospitalized individuals. In addition, a higher incidence was reported among women (approximately 49%) than men (37%). Brain fog was identified as the most frequently reported neuropsychiatric symptom, affecting 74% of individuals. Other commonly reported symptoms included memory problems (65%), disturbed sleep (62%), altered sense of smell or taste—referred to as hyposmia and hypogeusia, respectively (57%)—headaches (56%), and depression (50%) [
11].
Brain fog, also referred to as mental fog, describes a cognitive state characterized by confusion, difficulty concentrating, learning challenges, and impaired recall of past events. Research indicates that approximately 20% to 35% of individuals experience brain fog following COVID-19 infection [
12]. Sleep disturbances represent another prevalent neurological complaint in LC, affecting approximately 26% of patients, and are closely associated with mental health conditions [
13–
15]. These disturbances may include insomnia as well as vivid dreams or nightmares [
16]. Recently, Hermans et al. [
17] reported a prevalent symptom cluster—including sleep disorders, joint pain, fatigue, and headaches—among individuals with LC in lower-middle-income countries.
Overall, mental health conditions such as psychological strain, anxiety, and depression significantly affected daily life during the COVID-19 pandemic, and these effects persist in the post-pandemic period. These ongoing issues have placed additional pressure on global healthcare systems due to the increasing number of individuals experiencing mental health challenges. AI-driven monitoring and decision-support systems offer potential tools for managing and mitigating these long-term neurological and psychological consequences of the pandemic.
The cardiovascular system
The cardiovascular system consists of the heart, blood, and an extensive network of vessels, including arteries, veins, and capillaries. Its primary function is to circulate blood throughout the body, deliver oxygen and essential nutrients to all cells, and remove metabolic waste products such as carbon dioxide. Research indicates that approximately 7% to 40% of individuals develop some form of cardiovascular disorder following COVID-19 infection. Common symptoms that may indicate cardiovascular impairment include chest discomfort, palpitations or irregular heartbeats, dizziness, and an elevated resting heart rate [
18].
According to a study by Barman et al. [
19], approximately 20% to 30% of patients hospitalized due to COVID-19 exhibit elevated troponin levels. Troponin is a biochemical marker associated with type 2 myocardial infarction, which occurs when an imbalance develops between myocardial oxygen supply and oxygen demand. In individuals with COVID-19, troponin levels exceeding the 99th percentile are associated with a 3- to 6-fold increased risk of developing coronary artery disease (CAD) [
20]. Additional findings suggest that individuals infected with COVID-19 may continue to face an increased risk of myocardial infarction and stroke for up to 3 years following the initial infection [
21].
The respiratory system
The respiratory system regulates the process of breathing, which is essential for sustaining life. It enables the body to absorb oxygen and eliminate carbon dioxide. This system comprises the lungs, airways (including the trachea, bronchi, and bronchioles), diaphragm, larynx (voice box), throat, nose, and mouth, all of which work together to facilitate gas exchange. Its primary function is to ensure adequate oxygen delivery to body tissues while removing carbon dioxide as a metabolic waste product.
COVID-19 is known to severely affect the respiratory system, producing symptoms such as chest pain, choking sensations, persistent coughing, chest discomfort, dyspnea, and various pulmonary complications. Evidence indicates that approximately 80% of infections remain confined to the upper respiratory tract, whereas around 20% progress to the lower respiratory system, reaching the alveoli and resulting in lung infiltrates [
22].
A study conducted by Long et al. [
23], which analyzed data from 4,478 participants across 16 cohort studies, reported that 20% of individuals infected with COVID-19 experienced some degree of lung function impairment. Reduced diffusion capacity was the most frequently observed abnormality, followed by restrictive ventilatory patterns.
The renal system
The renal system comprises the kidneys, ureters, and urethra, and it plays a crucial role in removing liquid toxins and excess ions from the human body. By filtering approximately 200 L of fluid from the blood each day, the renal system helps maintain internal homeostasis and physiological balance.
Following COVID-19 infection, renal function may be adversely affected, leading to complications such as reduced renal perfusion, hypotension, ischemia, increased salt excretion (salt diuresis), sympathetic nervous system overactivation, and enhanced salt and water retention. Both direct and indirect effects of SARS-CoV-2 may persist after recovery from the acute phase, potentially resulting in long-term outcomes such as sepsis, recurrent episodes of acute kidney injury, and progression toward chronic kidney disease.
In a notable study, Ramamoorthy et al. [
24] were the first to propose that renal fibrosis may represent a long-term consequence of viral infection. Their findings, derived from experimental studies involving mice infected with mouse hepatitis virus-1 following acute infection, provide supporting evidence for this hypothesis.
The immune system
The immune system consists of the thymus, bone marrow, spleen, lymph nodes, tonsils, and liver, and its primary function is to protect the body against harmful bacteria and viruses. When immune function is impaired, 3 major categories of health conditions may arise: allergies, autoimmune disorders, and immunodeficiencies. Diseases such as type 1 diabetes, lupus, and rheumatoid arthritis exemplify disorders associated with immune system dysfunction.
COVID-19 has also been shown to affect immune system integrity. A 2024 study by Kratzer et al. [
25], published in the journal Allergy and conducted by a research group in Vienna, reported that SARS-CoV-2 infection may induce long-lasting immune alterations, even among individuals who experienced only mild illness. These findings enhance current understanding of the prolonged immune consequences associated with COVID-19.
The gastrointestinal system
The gastrointestinal tract consists of the mouth, pharynx, esophagus, stomach, small and large intestines, rectum, and anus, and is commonly referred to as the digestive system. Its primary function is to break down ingested food to facilitate nutrient absorption and eliminate waste products from the body.
Gastrointestinal complications associated with illness may present as diarrhea, nausea, vomiting, abdominal pain, loss of appetite, and altered taste perception. These manifestations may arise from various underlying conditions, including acid-related disorders, impaired gastrointestinal motility, belching, hepatobiliary disease, functional bowel disorders, rectal bleeding, and liver injury. A study by Ma et al. [
26] reported that individuals with severe COVID-19 infections tend to experience more pronounced gastrointestinal symptoms.
There is substantial evidence that SARS-CoV-2 negatively affects multiple components of the digestive system [
27,
28]. Furthermore, Choudhury et al. [
29] reported that 12% of individuals in the post-COVID-19 phase and 22% of those with LC exhibited symptoms indicative of gastrointestinal tract damage. Additional research suggests that individuals with LC may face an increased risk of developing chronic digestive system disorders [
30].
Dermatological complications
Dermatological disorders primarily affect the skin surface and may present as rashes, inflammation, itching, or other visible alterations in skin appearance. Research has established associations between SARS-CoV-2 infection and a range of dermatological manifestations that may emerge during recovery or in the post-infection period [
31,
32]. To date, more than 30 distinct types of COVID-19–related skin eruptions have been documented [
33], with substantial interindividual variability in clinical presentation [
31,
32]. A study conducted by McMahon et al. [
34] analyzed dermatological manifestations in patients from 41 countries, including individuals with laboratory-confirmed or suspected COVID-19, as well as cases in which skin findings were the primary or sole indicator of infection.
Table 2 summarizes the affected organ systems, key symptoms, prevalence, associated functional and biological changes, and the potential underlying mechanisms [
11,
18,
22,
23,
29,
31,
32,
35].
Prevalence, Risk Factors, and Functional Impact of Long COVID
Understanding the global burden of LC, including its variable prevalence, associated risk factors, diagnostic challenges, and profound functional impact, is essential for both public health planning and individual patient care. These dimensions are briefly described below.
Global and Regional Prevalence Estimates
LC affects millions of individuals worldwide. According to global estimates, approximately 65 million people are currently living with LC. In this context, a study by Lovaglio et al. [
36] provides the following relevant information: (1) In 2022, approximately 2.7% of the adult population in the United Kingdom was affected by LC (Islam et al.
Mymensingh Med J 2024;33:1250–7). In the United States, a report published in
MMWR Morbidity and Mortality Weekly Report (Ford et al. MMWR Morb Mortal Wkly Rep 2024;73:135–6) estimated that around 6.9% of the adult population was affected by LC in 2022. Globally, it is estimated that 10% to 20% of individuals infected with COVID-19 during the first 2 years of the pandemic developed LC (
https://www.who.int/europe/news-room/fact-sheets/item/post-covid-19-condition). (2) Reported LC prevalence among individuals infected with COVID-19 varies considerably across studies and regions, ranging from approximately 10% to 50% (
https://www.sciencealert.com/long-covid-rate-in-africa-is-almost-50-of-cases-researchers-warn). (3) The median proportion of patients reporting at least 1 persistent symptom within 60 days of COVID-19 onset was estimated to be 72.5% [
37]. (4) One study estimated the prevalence of post-acute sequelae of COVID-19 at 22.8%, which closely aligns with regional national estimates. However, a recently developed AI-based tool by researchers at Mass General Brigham suggests a higher prevalence of 22.8% of LC in the general population (Harvard University source: SciTechDaily.com, Nov. 17, 2024).
The following factors may explain the marked variation in prevalence estimates reported above. (1) Natural immunity varies across populations, communities, and geographic regions; consequently, countries experienced differing proportions of COVID-19–affected individuals. (2) No vaccine provides complete protection against infection, particularly as new viral variants emerge. However, vaccination substantially reduces the risk of developing severe disease and LC compared with remaining unvaccinated. (3) Differences in standards of living, healthcare infrastructure, and vaccination coverage further contribute to observed variability in prevalence estimates.
Identified Risk Factors
Several factors have been identified that increase an individual’s risk of developing LC.
Demographic factors
Female sex has been consistently associated with higher LC prevalence. LC predominantly affects adults younger than 65 years, and risk increases with greater severity of acute COVID-19 illness. Certain racial and ethnic populations, including American Indian communities and some multicultural societies, have demonstrated higher reported prevalence.
Vaccination status
Individuals who are unvaccinated or inadequately vaccinated against COVID-19 exhibit a higher risk of developing LC compared with fully vaccinated adults [
36].
Pre-existing health conditions
Prior disability and coexisting medical conditions significantly elevate LC risk, including cardiovascular disease, diabetes, obesity, respiratory disorders, and depression. Smoking status is also associated with increased likelihood of LC symptoms.
Severity of acute infection
LC risk increases with the severity of the initial COVID-19 illness. Individuals who required hospitalization due to COVID-19 are approximately 2 to 3 times more likely to develop LC than those who were not hospitalized.
Functional Impairment and Impact on Quality of Life
The persistent and heterogeneous symptom profile of LC results in substantial functional impairment and diminished quality of life among affected individuals [
38]. This study reports that symptoms such as chronic fatigue, post-exertional malaise (worsening of symptoms following physical or mental exertion), cognitive impairment (commonly referred to as brain fog), and autonomic dysfunction can severely compromise an individual’s capacity to work. These symptoms also interfere with academic engagement, sustained concentration, and the performance of routine daily activities for prolonged durations, often ranging from 6 months to 2 years or longer. A greater number or increased severity of LC symptoms is directly associated with poorer quality of life and reduced physical functioning (
https://www.nationalacademies.org/news/2024/06/new-report-reviews-evidence-on-long-covid-diagnosis-risk-symptoms-and-functional-impact-for-patients). These realities underscore the need for fundamental shifts in public health strategies, including the development of comprehensive long-term care and support systems. Furthermore, disparities in access to healthcare services—including COVID-19 testing, vaccination, therapeutics, and specialized rehabilitation clinics—have been documented across socioeconomic strata, geographic regions, health literacy levels, and racial and ethnic groups [
39]. These disparities were amplified during the pandemic and have since evolved into a critical health equity challenge.
Fortunately, AI is increasingly playing a valuable role in addressing many of the challenges discussed above that have emerged in the context of COVID-19 and LC.
Role of AI in Long COVID Management
AI is a field of computer science focused on enabling machines to perform intelligent behaviors analogous to human cognitive functions, including visual perception, speech recognition, image analysis, decision-making, language translation, and reasoning. AI-based models progressively improve their performance in tasks such as classification, regression, and clustering as additional data are incorporated, a process conceptually similar to human learning through experience. AI represents a broad framework encompassing subfields such as ML, DL, and generative AI. Importantly, AI played a significant role during the COVID-19 pandemic and continues to contribute to LC management by offering data-driven solutions—ranging from early screening and contact tracing to post-infection monitoring—to address major challenges in pandemic response and recovery [
40].
As evidence, advanced ML and DL techniques are increasingly used to extract actionable insights from large-scale datasets across multiple domains. The key roles of AI are outlined below. (1) Role 1: Data collection using AI-based techniques; (2) Role 2: Forecasting the spread of COVID-19 and its variants, early detection and prediction of disease progression, identification of organ systems affected by LC, and assessment of disease severity; (3) Role 3: Risk analysis and policymaking to combat COVID-19; and (4) Role 4: Drug and vaccine discovery for COVID-19.
Role 1: Data Collection Using AI-Based Techniques
To understand the dynamic and fluctuating nature of LC symptoms, extensive and continuous data collection is required. AI tools leverage electronic health records (EHRs) to gather large volumes of de-identified patient data across multiple healthcare systems. This approach supports a more comprehensive understanding of LC at the population level. A review study [
41] discusses how AI algorithms can analyze data collected from wearable devices to enable continuous patient monitoring. AI demonstrates particular strengths in this domain, as outlined below.
Miao et al. [
42] developed an NLP framework to track LC symptoms across demographic groups, geographic regions, and time. They also created an online dashboard that integrates LC symptom text with NLP algorithms, including recurrent neural networks (RNNs) and convolutional neural networks (CNNs), to support symptom analysis. NLP-based AI algorithms are also capable of analyzing data collected from wearable devices for continuous patient monitoring, as reported in this study [
43]. By tracking patient information in real time, clinicians can tailor treatment strategies to individual symptom profiles, while patients can use these insights to better manage their health based on changes in symptoms and physical condition. As a result, wearable devices offer substantial potential for continuous patient monitoring. Collectively, these technologies represent a transition toward more comprehensive and real-time patient data collection, which is essential for understanding the fluctuating and evolving nature of LC symptoms. The review study provides a broad discussion demonstrating that NLP algorithms are highly effective at extracting valuable symptom information from unstructured text in electronic medical records, including clinical notes [
44]. This capability enables the identification of a wide range of symptoms, such as fever, cough, headache, fatigue, and dyspnea, which may not be captured through structured data sources alone [
44].
Role 2: Forecasting the Spread of COVID-19 and Its Variants, Early Detection and Prediction of Disease Progression, Identification of Organ Systems Affected by Long COVID, and Assessment of Disease Severity
Using dynamically expanding COVID-19 datasets, researchers have sought to control emerging cases through predictive modeling. De Araujo Morais and da Silva Gomes [
45] introduced an effective forecasting model that combines autoregressive integrated moving average methods with neural network approaches to predict daily COVID-19 case counts worldwide. This hybrid model was able to capture both linear and nonlinear patterns within the data. New variants of SARS-CoV-2 emerged rapidly earlier in the pandemic and continue to arise, making effective variant prediction critically important. To address this challenge, Ullah et al. [
46] developed a CNN-based model. This approach integrates a variational autoencoder–decoder network with the Basic Local Alignment Search Tool to determine whether a detected variant originates from existing viral strains or represents a novel mutation.
Recent advancements in AI have covered various applications to understand the progression and outcomes in patients with COVID-19. Specifically, DL-based models, taking chest computed tomography (CT) images and clinical data, have been developed to predict progression of COVID-19 disease [
47]. Researchers have also developed early warning systems using DL algorithms with collected clinical data and patient CT scans to predict deterioration processes in patients with COVID-19 [
43,
48]. For early detection and diagnosis of the disease, crucial tools based on chest X-ray have been developed [
49].
AI-based models have also been introduced to detect LC in affected individuals. Ensemble learning approaches, such as XGBoost, have been applied in recent studies to identify LC cases [
50–
52]. Patel et al. [
53] proposed a random forest–based model that utilizes blood protein levels to distinguish affected individuals. Similarly, Sengupta et al. [
54] introduced a hybrid architecture combining bidirectional long short-term memory networks with 1-dimensional CNNs to classify LC patients. In contrast, a model developed by Subramanian et al. [
55] using VGG-16, ResNet-50, and U-Net architectures demonstrated superior predictive accuracy.
Role 3: Risk Analysis and Policymaking to Combat COVID-19
AI algorithms are increasingly capable of dynamically assessing COVID-19 transmission and infection risk to support healthcare professionals. Cloud- and fog-computing-based models have proven particularly useful for large-scale data collection, real-time risk assessment, and efficient monitoring and control of disease spread [
56,
57]. Furthermore, AI has emerged as a powerful tool for pandemic response by enabling data-driven policy decision-making. Bhatia et al. [
56] developed a decision tree–based model to assess infection severity and classify individuals according to symptom profiles, achieving a reported accuracy of 96.68% and supporting timely monitoring and intervention.
Role 4: Drug and Vaccine Discovery for COVID-19
In the absence of definitive curative treatments and virus-specific medications for COVID-19, and given the urgent need for rapid solutions due to high transmissibility and fatality rates compared with other viral infections, researchers increasingly turned to AI-based approaches. As traditional self-care measures became insufficient, technology-driven alternatives gained prominence. By analyzing large-scale biological and clinical datasets, AI has demonstrated the capacity to identify potential drug targets, predict drug efficacy, and optimize formulation parameters.
Conventional drug and vaccine discovery processes are widely recognized as costly and time-intensive, often requiring many years or even decades; however, the pandemic necessitated accelerated alternatives. A major challenge in drug repurposing lies in identifying robust drug–disease relationships. To overcome these limitations, drug repurposing or repositioning and AI-assisted vaccine discovery emerged as prominent strategies, with AI contributing substantially by optimizing candidate selection and formulation parameters. For example, Pfizer employed AI-driven tools during clinical trials to efficiently process large datasets and meet regulatory requirements. Using multiple network-based strategies, Zeng et al. [
58] developed an integrative DL framework, COVID-19 Knowledge Graph Embeddings (CoV-KGE), to identify adaptable drugs for COVID-19.
Lv et al. [
59] applied a multimodal approach to the virus–host interactome by integrating graph convolutional networks, network diffusion, and network proximity methods, achieving a reported success rate of 62% and identifying 6 drugs that reduced viral infection. The development of AI-driven vaccines for viruses such as Zika and Ebola has also been discussed in the context of pandemic preparedness [
59,
60].
Importantly, multiple AI algorithms significantly facilitated drug repurposing and therapeutic deployment during the COVID-19 era. In 2020, Mohanty et al. [
61] provided a comprehensive review of Repurpose Drug Databases and Open Chemical/Drug Databases and developed a model using ML, RNNs, CNNs, and deep belief networks to rapidly and accurately screen candidate drugs. Gupta et al. [
62] applied ML techniques to datasets derived from PubChem, achieving an approximate prediction accuracy of 73%. More recently, in 2024, Serrano et al. [
63] reported that AI played a crucial role in accelerating the identification of reliable biomarkers, diagnostics, and novel therapeutic candidates for LC.
Overall Discussion about the Uses of AI in COVID-19 and Long COVID
Thus, AI has increasingly replaced many manually managed traditional systems that are time-consuming, error-prone, and costly. More specifically, current AI-driven monitoring systems—often integrating AI with the Internet of Things (IoT), including sensors and connected devices—are capable of continuously monitoring patients’ vital parameters, including early symptom changes, to generate actionable insights for both patients and healthcare providers. This capability supports earlier diagnosis and contributes to improved patient outcomes. AI algorithms are designed by researchers to collect live data for disease diagnosis, which are subsequently analyzed using statistical tools such as SPSS to identify target trends based on specific categories and geographic locations. Based on these analyses, appropriate predictive models are developed to enable accurate and rapid diagnosis, thereby assisting clinical practitioners. Recent AI-based algorithms, including CNNs, RNNs, and generative adversarial networks, aim to enhance decision quality by selecting the most relevant features for a given decision problem. This feature selection improves prediction accuracy and reduces error or misclassification rates on unseen data. Furthermore, removing irrelevant features from datasets decreases learning time and facilitates the generation of compact, informative knowledge representations, ultimately reducing decision-making time for end users and organizations.
AI-based technologies are playing a central role in redefining how modern medicine addresses the complexities of LC. In particular, AI is bridging multiple gaps in medical science by facilitating cross-organ diagnostics, uncovering latent correlations across heterogeneous symptoms, and enabling scalable population-level surveillance. Collectively, these capabilities reduce clinical workload, enhance patient engagement, and support evidence-informed policymaking.
Based on the discussion presented in Section 6, a comprehensive overview of key AI applications in LC—highlighting the techniques employed, data sources utilized, and associated clinical impacts—is provided in
Table 3 [
42,
47,
48,
50,
52,
55,
58,
62].
Concerns and Challenges in AI Deployment for Long COVID
AI has the potential to revolutionize healthcare, particularly in the management of complex conditions such as LC; however, its integration into clinical practice raises substantial ethical and practical concerns. Addressing these concerns requires the establishment of clear guidelines and robust governance frameworks [
64]. Several key challenges arising from the generation of large-scale data through LC-related sources—such as patient records, medical imaging, and genomic information—and from the application of AI-based tools to these data are highlighted below.
Data Privacy and Security Concerns
This category includes the following major concerns.
Confidentiality breaches
Significant risks exist to the privacy of sensitive patient information, particularly through cyberattacks, hacking incidents, or vulnerabilities in data security infrastructures. Such events can lead to serious breaches of confidentiality and unauthorized data disclosure [
65].
Data ownership and control
Patients may not always be fully informed about who owns or controls their health data when such data are processed by algorithms or managed by external organizations. Clear and transparent policies are therefore required to define data ownership and to protect patients’ rights regarding the use, sharing, and governance of their medical information.
Regulatory compliance
Compliance with data protection regulations—such as the Health Insurance Portability and Accountability Act in the United States and the General Data Protection Regulation in the European Union—is essential to protect patient data and maintain public trust in healthcare systems. Regulatory compliance also helps safeguard developers and users of AI-enabled systems from legal liability. Federated learning approaches can support compliance by enabling model training without direct sharing of raw patient data; however, broader legal and ethical standards are still evolving to adequately address these complex challenges.
Algorithmic Bias and Its Implications for Health Equity
Accurate and representative data are fundamental to the development of reliable AI and ML models. When training datasets are incomplete, imbalanced, or biased, AI systems may reinforce or exacerbate existing healthcare disparities [
65]. Key forms of algorithmic bias relevant to LC applications are outlined below.
Bias in training data
When algorithms are trained on datasets that do not adequately represent population diversity, their performance may be reduced for certain demographic groups, particularly those from minority or underserved communities. For example, a model developed primarily using data from Caucasian populations may yield less accurate diagnostic or prognostic results for African, Hispanic, or Asian individuals.
Disparities linked to healthcare access
Efforts to optimize algorithmic performance may unintentionally reinforce pre-existing inequalities in healthcare access. For instance, models that prioritize high-cost or high-frequency healthcare users may overlook individuals in rural or low-income settings, who often have limited access to healthcare services and diagnostic resources [
65].
Ethical considerations
Algorithmic bias raises substantial ethical concerns related to fairness and equity in healthcare delivery. A key risk is that certain population groups may receive lower-quality care, thereby widening existing health disparities and producing unjust outcomes [
65]. This highlights that algorithmic bias is not merely a technical issue but also a social determinant that can intensify structural health inequalities. Sustained and deliberate efforts are therefore required to ensure the development and use of inclusive, representative datasets that promote equitable healthcare outcomes. It is also critical to recognize that bias can emerge at multiple stages, including data collection, data access, preprocessing, model development, and variations in the expertise of system designers.
Transparency, Accountability and Human Oversight
The design of advanced models (particularly DL architectures) can create a black box problem, where it becomes difficult to understand how conclusions are reached. A few key issues are outlined below:
Opacity of AI decisions
Limited transparency in complex AI algorithms makes their internal decision-making processes difficult to interpret. This limitation is particularly concerning in clinical contexts, where understanding the rationale behind diagnostic outputs or treatment recommendations is essential for safe and accountable care.
Need for explainability
Progress in explainable approaches is crucial to improve understanding of how models arrive at their outputs, thereby enhancing trust and supporting clinical adoption.
Maintaining human oversight
Human oversight, through human-in-the-loop or human-on-the-loop frameworks, remains essential to ensure clinical accuracy, prevent adverse outcomes, and uphold professional integrity, particularly in high-stakes healthcare environments [
66]. AI systems are limited in their ability to generate novel clinical insights or integrate experiential judgment and may hallucinate or overgeneralize; therefore, expert review and intervention remain necessary.
Implementation Challenges
Integrating AI technologies into existing healthcare systems presents substantial practical challenges, as outlined below.
Technical challenges
Robust architectures for handling ML models and associated data, together with high-performance computing systems, are required for the effective integration of AI technologies into healthcare settings.
Knowledge gaps and resistance to change
An often overlooked challenge is the natural human resistance to adopting new ways of working. Increasing awareness within the healthcare community regarding the benefits of AI integration therefore represents a significant psychological and organizational challenge.
Lack of standard guidelines
At present, there are few well-defined and widely accepted guidelines governing the ethical use of AI and ML in healthcare. This lack of standardization remains a significant barrier to AI adoption in the healthcare domain.
Table 4 summarizes these ethical and practical challenges, along with proposed mitigation strategies [
65–
67].
Limitations of the Current Review and Future Directions
This review provides an integrated overview of LC and examines the role of AI in its management. Afterall, the review studies [
68,
69] assisted the authors for shaping the present article. Although the analysis is based on 69 peer-reviewed publications, some relevant studies may not have been captured. In addition, heterogeneity in research designs, study populations, symptom definitions, and follow-up durations may limit the comparability of reported findings.
Future Directions
To improve the understanding, diagnosis, and management of LC, further progress is required in the following areas.
Unified data practices and system interoperability
Establishing standardized data-collection practices and improving interoperability across EHRs systems are essential. Consistent, high-quality, and interoperable datasets will support the development of reliable AI tools and help address limitations related to data variability.
Well-designed clinical studies
Future research should prioritize prospective, controlled clinical studies evaluating AI tools for LC diagnosis, prognosis, and treatment monitoring. Strong clinical evidence is necessary to demonstrate the safety, accuracy, and real-world utility of AI-based applications.
Biomarker identification
The identification of clear and reliable biomarkers for LC remains a major gap. AI-driven analytical approaches may play an important role in detecting biological indicators that enhance diagnostic accuracy and support personalized treatment strategies.
Model clarity and explainability
Advancing explainable AI approaches is important for improving transparency in AI-supported clinical decision-making. Greater interpretability may increase confidence among healthcare professionals and patients and promote responsible AI use.
Ethical and regulatory governance
As AI adoption in healthcare accelerates, there is a growing need for clear ethical standards and regulatory frameworks. Such policies should ensure fair and responsible AI deployment—particularly for complex, long-term conditions such as LC—while protecting privacy and addressing bias and accountability concerns.
Conclusion
The long-term consequences of the COVID-19 pandemic continue to affect human health, with many survivors experiencing persistent, multiorgan complications. This review highlights LC as a multisystem condition that can substantially impair daily functioning and quality of life, underscoring the need for improved care strategies and innovative interventions.
In recent years, AI has gained increasing attention in LC research. AI techniques are being used to develop structured digital data resources that enhance analysis and knowledge sharing. When combined with IoT technologies, AI-enabled monitoring systems support real-time patient observation and earlier detection of health deterioration.
AI offers multiple applications for LC management, including support for clinical decision-making through pattern recognition, early detection, and outcome prediction. Intelligent systems also facilitate ongoing symptom tracking via wearable devices and language-based platforms, while accelerating the identification of potential therapeutic candidates. Together, these applications demonstrate the potential of AI to address the complex challenges associated with LC.
Beyond scientific and clinical benefits, this review also highlights broader implications of AI use in LC care, including data protection, mitigation of algorithmic bias, and the need to preserve human judgment in clinical decision-making. Continued global research is essential to deepen understanding of LC and to strengthen future management strategies.
HIGHLIGHTS
• This review raises awareness about coronavirus disease 2019 (COVID-19)–induced damage to human organ systems and the global burden.
• The potential of artificial intelligence (AI) for addressing long COVID–related health challenges is explored.
• Significant ethical, practical, and societal challenges are posed by the extensive use of AI.
• More research is necessary on long COVID throughout the world.
Article information
Ethics Approval
Not applicable.
Conflicts of Interest
The authors have no conflicts of interest to declare.
Availability of Data
All data (published articles) analyzed during this study are included in this published article. For other relevant data, these may be requested through the corresponding author.
Acknowledgements
The authors acknowledge the review studies [68,69] for shaping the present article. The authors acknowledge the support of an AI-generated tool to rephrase a few sentences and correct grammatical errors.
Figure 1.
Conceptual diagram of the present study.
COVID, coronavirus disease; AI, artificial intelligence.
Table 1.Proposed pathophysiological mechanisms
Table 1.
| Organ system |
Functional and biochemical changes in the body and the mechanisms thereof |
| Nervous system |
Immune dysregulation: This occurs due to breakdown or maladaptive changes in molecular immune-control processes, which may lead to autoimmune diseases and some cancers. COVID-19 can cause long-lasting alterations in immune system function. Some patients show persistent inflammation, whereas others exhibit reduced or ineffective immune responses. This imbalance can affect multiple organs and may contribute to fatigue, reduced exercise tolerance, and difficulties with concentration or memory. Prolonged immune activation may also impair the body’s ability to heal and return to baseline health. |
| Autoimmune diseases: These occur when the immune system attacks the body’s own tissues instead of protecting them. Emerging evidence suggests that COVID-19 may trigger autoimmune processes in certain individuals, potentially contributing to symptoms affecting the nervous system, circulation, or other organs. |
| Viral persistence: This refers to the long-term presence of viral components within the body, potentially causing chronic infection and diminished metabolic function. In some individuals, fragments of the virus appear to persist for months after the initial infection. |
| Microvascular inflammation: Inflammation resulting from immune dysregulation or infection may activate the microcirculation and affect blood pressure regulation. Small-vessel inflammation, clotting abnormalities, and impaired circulation may persist after the acute illness. |
| Neuroinflammation: This refers to inflammation of nervous tissue, often involving microglial activation. It may contribute to memory impairment and disruption of daily functioning. |
| Cardiovascular system |
Endothelial dysfunction condition: This condition arises when the inner lining of blood vessels fails to function properly, contributing to cardiovascular complications such as microvascular inflammation or thrombo-inflammation. |
| Direct viral impact on cardiac tissue: SARS-CoV-2 may directly damage cardiac tissue, potentially leading to long-term or permanent cardiac injury. |
| Respiratory system |
Viral damage: Direct viral injury to lung tissue can impair healing and recovery. |
| Pulmonary infiltrates: Accumulation of substances denser than air in the lungs can lead to infections such as pneumonia or tuberculosis. |
| Inflammation: Chronic inflammation may affect energy metabolism, digestion, cognition, and overall physiological balance. |
| Fibrosis: Excessive accumulation of fibrous connective tissue causes scarring and impaired lung function and may also affect the heart, liver, skin, and other organs. |
| Renal system |
Viral illness of kidney cells: Viral infection may hijack renal cellular pathways, disrupting normal kidney function. Inflammation and reduced renal blood flow can contribute to conditions such as atherosclerosis, diabetes, and obesity. |
|
Sepsis: Systemic infection may lead to widespread inflammation and multiorgan dysfunction. |
| Immune system |
Immune dysregulation and viral persistence: Persistent immune imbalance and viral remnants may impair immune responses. |
| Oxidative stress: Infection may reduce host immune defense capacity. |
| Autoantibody production: Breakdown of immune self-tolerance may trigger B-cell-mediated autoantibody generation. |
| GI system |
Viral infection of GI epithelial cells, intestinal microbiota dysbiosis (associated with bacterial infections such as Helicobacter pylori and Clostridioides difficile), and inflammation involving the liver. |
| Dermatological system |
Immune responses play a critical role in neutralizing pathogens. Impaired immune responses may result in frequent infections, delayed wound healing, and skin inflammation. |
| Direct viral effects on skin cells may cause lesions and inflammatory skin manifestations. |
Table 2.Summary of long COVID symptoms and the affected human organ systems
Table 2.
| Organ system |
Key symptoms |
Prevalence/severity notes |
| Nervous system |
Brain fog, memory loss, anosmia/ageusia, headaches, sleep disorders, anxiety, depression |
Higher in hospitalized patients (53%) and women (49%); brain fog (74%), memory problems (65%), disturbed sleep (62%); impaired work/school for 6 months–2+ years [11] |
| Cardiovascular system |
Chest pain/pressure, irregular heartbeat (palpitations), light-headedness, increased resting heart rate, increased risk of heart attacks |
7%–40% experience condition; 20%–30% hospitalized exhibit elevated troponin; risk of heart attacks/strokes persists up to 3 years [18] |
| Respiratory system |
Chest pain, shortness of breath, persistent coughing, lung dysfunction (diffusion impairment, restrictive defects) |
20% lung function abnormalities; potential for pulmonary fibrosis [22] |
| Renal system |
Reduced renal perfusion, hypotension, ischemia, increased salt excretion, recurrent AKI, onset of CKD |
Renal fibrosis as long-term consequence [23] |
| Immune system |
Lasting alterations to immune system, chronic fatigue, autoimmunity, reactivation of latent viruses (EBV, HHV-6, CMV), bacterial diseases (TB) |
Affects even mild cases; linked to fatigue, POTS, neurocognitive dysfunction [35] |
| GI system |
Diarrhea, nausea, vomiting, abdominal pain, loss of appetite, altered taste perception, liver damage, motility problems |
Increased likelihood in long COVID patients; 12%–22% exhibit GI tract damage [29] |
| Dermatological system |
Rashes, inflammation, itching, over 30 types of skin rashes |
Varies widely among individuals [31,32] |
Table 3.Key AI applications for long COVID: techniques, data sources, and health consequences
Table 3.
| Medical issues |
AI applications |
AI techniques |
Health consequences |
| Diagnosis and prognosis |
Detecting long COVID ML models for risk prediction for lung CT images. |
XGBoost (accuracy, 98%; precision, 99%; sensitivity, 99%) for detecting LC [51]; methodological rigour, superior; relevance to review, superior; clarity of reporting, superior; validity of findings, superior |
Improved diagnostic accuracy, reduced diagnostic bias, early risk warning, personalized care |
| Detection of new variants of COVID-19 |
Random Forest (precision, 0.85; recall, 0.86) for detecting LC [49]; methodological rigour, superior; relevance to review, superior; clarity of reporting, superior; validity of findings, superior, CNN-based VGG-16 with LR, 0.001 (accuracy, 45.87%; precision, 45.8%; recall, 100%; F1-score, 62.8%), ResNet-50 with LR, 0.001 (accuracy, 94.24%; precision, 92.42%; recall, 95.31; F1-score, 93.85) [54]; methodological rigour, average; relevance to review, average; clarity of reporting, average; validity of findings, superior |
|
Variational autoencoder-decoder networks achieves 0.72 AUC value for detection of new variants of COVID-19 [45]; methodological rigour, superior; relevance to review, superior; clarity of reporting, average; validity of findings, average |
| Lung damage assessment |
Imaging analysis |
Deep Neural Network, U-Net (segmentation, accuracy, 97.132%) [46]; methodological rigour, average; relevance to review, average; clarity of reporting, superior; validity of findings, superior
|
Automated diagnosis, precise lung opacity segmentation, quantification of affected lung volume, reduced diagnostic inconsistency |
| Symptom tracking and patient monitoring |
NLP, wearable data analysis |
Developed interface, best accuracy, 89% for understanding of long COVID symptoms [41]; methodological rigour, superior; relevance to review, superior; clarity of reporting, superior; validity of findings, superior |
Timely symptom identification, comprehensive symptom characterization, continuous patient monitoring, personalized treatment guidance |
| Predicting COVID-19 malignant progression |
AUC of 0.920 in the single-centre study and an average AUC of 0.874 in the multicentred study [47]; methodological rigour, superior; relevance to review, superior; clarity of reporting, average; validity of findings, average |
|
| Treatment discovery |
Drug remodeling and novel compound identification |
DL-based model (accuracy, 73%) in drug prediction [61]; methodological rigour, superior; relevance to review, superior; clarity of reporting, superior, 62% success rate in drug-design [57]; methodological rigour, superior; relevance to review, superior; clarity of reporting, average; validity of findings, average |
Accelerated discovery of biomarkers, diagnostics, repurposed drugs, and novel therapies; optimized formulation parameters |
Table 4.Ethical and practical challenges of AI in healthcare for long COVID
Table 4.
| Challenge class |
Specific risk |
Implication for long COVID care |
Suggested mitigation strategy |
| Data privacy and security |
Confidentiality breaches and unauthorized access to sensitive patient data [64] |
Loss of patient trust, legal liability, misuse of medical information |
Federated learning, robust encryption, clear data-ownership policies, strict HIPAA/GDPR compliance |
| Algorithmic bias |
AI models trained on biased datasets that perpetuate healthcare disparities [64] |
Inaccurate diagnoses in marginalized populations, suboptimal care, worsening health inequalities, unfair resource allocation |
Diverse training datasets, bias-detection algorithms, continuous auditing, fairness-aware evaluation metrics |
| Transparency and accountability |
Black-box decision-making processes that are difficult to interpret [66] |
Reduced clinician trust, difficulty validating AI outputs, limited error detection |
Adoption of explainable AI (XAI), transparent documentation of model logic |
| Human oversight |
Over-reliance on AI with insufficient human review [65] |
Misdiagnosis, adverse patient outcomes, erosion of clinician critical thinking |
Human-in-the-loop and human-on-the-loop frameworks, clear review and override protocols, positioning AI as assistive |
| Implementation |
Inadequate resources and unforeseen technical barriers |
Delayed adoption, widening knowledge gaps, underutilization of AI potential |
Robust data architecture, staff-training programs, interdisciplinary collaboration, clearly defined implementation roadmaps |
References
- 1. World Health Organization (WHO) (CH). Coronavirus disease (COVID-19) [Internet]. WHO; 2022 [cited 2025 Nov 13]. Available from: https://www.who.int/news-room/fact-sheets/detail/coronavirus-disease-%28covid-19%29.
- 2. World Health Organization (WHO) (CH). COVID-19 cases, world [Internet]. WHO; 2025 [cited 2025 Nov 13]. Available from: https://covid19.who.int.
- 3. Al-Aly Z, Bowe B, Xie Y. Long COVID after breakthrough SARS-CoV-2 infection. Nat Med 2022;28:1461−7.
- 4. Garrigues E, Janvier P, Kherabi Y, et al. Post-discharge persistent symptoms and health-related quality of life after hospitalization for COVID-19. J Infect 2020;81:e4−6.
- 5. Greenhalgh T, Knight M, A'Court C, et al. Management of post-acute COVID-19 in primary care. BMJ 2020;370:m3026.
- 6. Peluso MJ, Deeks SG. Early clues regarding the pathogenesis of long-COVID. Trends Immunol 2022;43:268−70.
- 7. Nasserie T, Hittle M, Goodman SN. Assessment of the frequency and variety of persistent symptoms among patients with COVID-19: a systematic review. JAMA Netw Open 2021;4:e2111417.
- 8. Reiss AB, Greene C, Dayaramani C, et al. Long COVID, the brain, nerves, and cognitive function. Neurol Int 2023;15:821−41.
- 9. Lippi G, Sanchis-Gomar F, Henry BM. Covid-19 and its long-term sequelae: what do we know in 2023? Pol Arch Intern Med 2023;133:16402.
- 10. O'Mahoney LL, Routen A, Gillies C, et al. The prevalence and long-term health effects of long COVID among hospitalised and non-hospitalised populations: a systematic review and meta-analysis. EClinicalMedicine 2023;55:101762.
- 11. Soriano JB, Murthy S, Marshall JC, et al. A clinical case definition of post-COVID-19 condition by a Delphi consensus. Lancet Infect Dis 2022;22:e102−7.
- 12. Panagea E, Messinis L, Petri MC, et al. Neurocognitive impairment in long COVID: a systematic review. Arch Clin Neuropsychol 2025;40:125−49.
- 13. Huang C, Huang L, Wang Y, et al. 6-month consequences of COVID-19 in patients discharged from hospital: a cohort study. Lancet 2021;397:220−32.
- 14. Premraj L, Kannapadi NV, Briggs J, et al. Mid and long-term neurological and neuropsychiatric manifestations of post-COVID-19 syndrome: a meta-analysis. J Neurol Sci 2022;434:120162.
- 15. Efstathiou V, Stefanou MI, Demetriou M, et al. Long COVID and neuropsychiatric manifestations (review). Exp Ther Med 2022;23:363.
- 16. Scarpelli S, Nadorff MR, Bjorvatn B, et al. Nightmares in people with COVID-19: did coronavirus infect our dreams? Nat Sci Sleep 2022;14:93−108.
- 17. Hermans LE, Wasserman S, Xu L, et al. Long COVID prevalence and risk factors in adults residing in middle- and high-income countries: secondary analysis of the multinational Anti-Coronavirus Therapies (ACT) trials. BMJ Glob Health 2025;10:e017126.
- 18. Dixit NM, Churchill A, Nsair A, et al. Post-acute COVID-19 syndrome and the cardiovascular system: what is known? Am Heart J Plus 2021;5:100025.
- 19. Barman HA, Atici A, Sahin I, et al. Prognostic significance of cardiac injury in COVID-19 patients with and without coronary artery disease. Coron Artery Dis 2021;32:359−66.
- 20. Nai Fovino L, Cademartiri F, Tarantini G. Subclinical coronary artery disease in COVID-19 patients. Eur Heart J Cardiovasc Imaging 2020;21:1055−6.
- 21. Hilser JR, Spencer NJ, Afshari K, et al. COVID-19 is a coronary artery disease risk equivalent and exhibits a genetic interaction with ABO blood type. Arterioscler Thromb Vasc Biol 2024;44:2321−33.
- 22. Visco V, Vitale C, Rispoli A, et al. Post-COVID-19 syndrome: involvement and interactions between respiratory, cardiovascular and nervous systems. J Clin Med 2022;11:524.
- 23. Long Q, Li J, Hu X, et al. Follow-ups on persistent symptoms and pulmonary function among post-acute COVID-19 patients: a systematic review and meta-analysis. Front Med (Lausanne) 2021;8:702635.
- 24. Ramamoorthy R, Hussain H, Ravelo N, et al. Kidney damage in long COVID: studies in experimental mice. Biology (Basel) 2023;12:1070.
- 25. Kratzer B, Gattinger P, Trapin D, et al. Differential decline of SARS-CoV-2-specific antibody levels, innate and adaptive immune cells, and shift of Th1/inflammatory to Th2 serum cytokine levels long after first COVID-19. Allergy 2024;79:2482−501.
- 26. Ma Y, Zhang L, Wei R, et al. Risks of digestive diseases in long COVID: evidence from a population-based cohort study. BMC Med 2024;22:14.
- 27. Rokkas T. Gastrointestinal involvement in COVID-19: a systematic review and meta-analysis. Ann Gastroenterol 2020;33:355−65.
- 28. Bogariu AM, Dumitrascu DL. Digestive involvement in the long-COVID syndrome. Med Pharm Rep 2022;95:5−10.
- 29. Choudhury A, Tariq R, Jena A, et al. Gastrointestinal manifestations of long COVID: a systematic review and meta-analysis. Therap Adv Gastroenterol 2022;15:17562848221118403.
- 30. Ren Z, Wang H, Cui G, et al. Alterations in the human oral and gut microbiomes and lipidomics in COVID-19. Gut 2021;70:1253−65.
- 31. Gottlieb M, Long B. Dermatologic manifestations and complications of COVID-19. Am J Emerg Med 2020;38:1715−21.
- 32. Gao JC, Huang A, Desai A, et al. “COVID toes”: a true viral phenomenon or a diagnosis without a leg to stand on? JAAD Int 2022;9:1−6.
- 33. Singh R, Freeman EE. Viruses, variants, and vaccines: how COVID-19 has changed the way we look at skin. Curr Dermatol Rep 2022;11:289−312.
- 34. McMahon DE, Gallman AE, Hruza GJ, et al. Long COVID in the skin: a registry analysis of COVID-19 dermatological duration. Lancet Infect Dis 2021;21:313−4.
- 35. Raj V, Opie M, Arnold AC. Cognitive and psychological issues in postural tachycardia syndrome. Auton Neurosci 2018;215:46−55.
- 36. Lovaglio PG, Borgonovo F, Manzo Margiotta A, et al. Estimating long COVID-19 prevalence across definitions and forms of sample selection. Front Epidemiol 2025;5:1597799.
- 37. Davis HE, McCorkell L, Vogel JM, et al. Long COVID: major findings, mechanisms and recommendations. Nat Rev Microbiol 2023;21:133−146.
- 38. de Oliveira JF, de Avila RE, de Oliveira NR, et al. Persistent symptoms, quality of life, and risk factors in long COVID: a cross-sectional study of hospitalized patients in Brazil. Int J Infect Dis 2022;122:1044−51.
- 39. Williams AM, Clayton HB, Singleton JA. Racial and ethnic disparities in COVID-19 vaccination coverage: the contribution of socioeconomic and demographic factors. Am J Prev Med 2022;62:473−82.
- 40. Hassantabar S, Ahmadi M, Sharifi A. Diagnosis and detection of infected tissue of COVID-19 patients based on lung X-ray image using convolutional neural network approaches. Chaos Solitons Fractals 2020;140:110170.
- 41. Shajari S, Kuruvinashetti K, Komeili A, et al. The emergence of AI-based wearable sensors for digital health technology: a review. Sensors (Basel) 2023;23:9498.
- 42. Miao L, Last M, Litvak M. An interactive analysis of user-reported long COVID symptoms using Twitter data. In: Proceedings of the 2nd Workshop on Deriving Insights from User-Generated Text; 2022 May 27; Dublin. Association for Computational Linguistics; 2022. p. 10−9.
- 43. Lassau N, Ammari S, Chouzenoux E, et al. Integrating deep learning CT-scan model, biological and clinical variables to predict severity of COVID-19 patients. Nat Commun 2021;12:634.
- 44. Wieland-Jorna Y, van Kooten D, Verheij RA, et al. Natural language processing systems for extracting information from electronic health records about activities of daily living: a systematic review. JAMIA Open 2024;7:ooae044.
- 45. de Araujo Morais LR, da Silva Gomes GS. Forecasting daily COVID-19 cases in the world with a hybrid ARIMA and neural network model. Appl Soft Comput 2022;126:109315.
- 46. Ullah A, Malik KM, Saudagar AK, et al. COVID-19 genome sequence analysis for new variant prediction and generation. Mathematics 2022;10:4267.
- 47. Wang R, Jiao Z, Yang L, et al. Artificial intelligence for prediction of COVID-19 progression using CT imaging and clinical data. Eur Radiol 2022;32:205−12.
- 48. Fang C, Bai S, Chen Q, et al. Deep learning for predicting COVID-19 malignant progression. Med Image Anal 2021;72:102096.
- 49. Kumar A, Tripathi AR, Satapathy SC, et al. SARS-Net: COVID-19 detection from chest X-rays by combining graph convolutional network and convolutional neural network. Pattern Recognit 2022;122:108255.
- 50. Pfaff ER, Girvin AT, Bennett TD, et al. Identifying who has long COVID in the USA: a machine learning approach using N3C data. Lancet Digit Health 2022;4:e532−41.
- 51. Jiang S, Loomba J, Sharma S, et al. Vital measurements of hospitalized COVID-19 patients as a predictor of long COVID: an EHR-based cohort study from the RECOVER Program in N3C. In: 2022 IEEE International Conference on Bioinformatics and Biomedicine (BIBM); 2022 Dec 6-8; Las Vegas, NV. IEEE; 2022. p. 3023−30.
- 52. Jha M, Gupta R, Saxena R. A precise method to detect post-COVID-19 pulmonary fibrosis through extreme gradient boosting. SN Comput Sci 2023;4:89.
- 53. Patel MA, Knauer MJ, Nicholson M, et al. Organ and cell-specific biomarkers of Long-COVID identified with targeted proteomics and machine learning. Mol Med 2023;29:26.
- 54. Sengupta S, Loomba J, Sharma S, et al. Analyzing historical diagnosis code data from NIH N3C and RECOVER Programs using deep learning to determine risk factors for long COVID. In: 2022 IEEE International Conference on Bioinformatics and Biomedicine (BIBM); 2022 Dec 6-8; Las Vegas, NV. IEEE; 2022. p. 2797−802.
- 55. Subramanian R, Rubi R, Dedeepya M, et al. Quantitative progression analysis of post-acute sequelae of COVID-19, pulmonary fibrosis (PASC-PF) and artificial intelligence driven CT scoring of lung involvement in COVID-19 infection using HRCT-chest images. Med Res Arch 2022;10:1−10.
- 56. Bhatia M, Manocha A, Ahanger TA, et al. Artificial intelligence-inspired comprehensive framework for COVID-19 outbreak control. Artif Intell Med 2022;127:102288.
- 57. Kaur I, Behl T, Aleya L, et al. Artificial intelligence as a fundamental tool in management of infectious diseases and its current implementation in COVID-19 pandemic. Environ Sci Pollut Res Int 2021;28:40515−32.
- 58. Zeng X, Song X, Ma T, et al. Repurpose open data to discover therapeutics for COVID-19 using deep learning. J Proteome Res 2020;19:4624−36.
- 59. Lv H, Shi L, Berkenpas JW, et al. Application of artificial intelligence and machine learning for COVID-19 drug discovery and vaccine design. Brief Bioinform 2021;22:bbab320.
- 60. Black S, Bloom DE, Kaslow DC, et al. Transforming vaccine development. Semin Immunol 2020;50:101413.
- 61. Mohanty S, Harun Ai Rashid M, Mridul M, et al. Application of artificial intelligence in COVID-19 drug repurposing. Diabetes Metab Syndr 2020;14:1027−31.
- 62. Gupta R, Srivastava D, Sahu M, et al. Artificial intelligence to deep learning: machine intelligence approach for drug discovery. Mol Divers 2021;25:1315−60.
- 63. Serrano DR, Luciano FC, Anaya BJ, et al. Artificial intelligence (AI) applications in drug discovery and drug delivery: revolutionizing personalized medicine. Pharmaceutics 2024;16:1328.
- 64. Alowais SA, Alghamdi SS, Alsuhebany N, et al. Revolutionizing healthcare: the role of artificial intelligence in clinical practice. BMC Med Educ 2023;23:689.
- 65. Seh AH, Zarour M, Alenezi M, et al. Healthcare data breaches: insights and implications. Healthcare (Basel) 2020;8:133.
- 66. Frenette J. Ensuring human oversight in high-performance AI systems: a framework for control and accountability. World J Adv Res Rev 2023;20:1507−16.
- 67. Hassija V, Chamola V, Mahapatra A, et al. Interpreting black-box models: a review on explainable artificial intelligence. Cogn Comput 2024;16:45−74.
- 68. Ezazi Bojnordi T, Pourali S, Haghighat F, et al. Breast cancer detection algorithms: a comparative review of leading artificial intelligence models. InfoSci Trend 2025;2:57−77.
- 69. Alizade R, Safarian S, Geramifar P, et al. A systematic review of artificial intelligence for endoscopic and histologic assessment in ulcerative colitis. InfoSci Trend 2025;2:12−38.
