Symptom clusters in COVID-19: A potential clinical prediction tool from the COVID Symptom Study app.
Ontology highlight
ABSTRACT:
SUBMITTER: Sudre CH
PROVIDER: S-EPMC7978420 | biostudies-literature | 2021 Mar
REPOSITORIES: biostudies-literature
Publications
Symptom clusters in COVID-19: A potential clinical prediction tool from the COVID Symptom Study app.
Sudre Carole H CH Lee Karla A KA Lochlainn Mary Ni MN Varsavsky Thomas T Murray Benjamin B Graham Mark S MS Menni Cristina C Modat Marc M Bowyer Ruth C E RCE Nguyen Long H LH Drew David A DA Joshi Amit D AD Ma Wenjie W Guo Chuan-Guo CG Lo Chun-Han CH Ganesh Sajaysurya S Buwe Abubakar A Pujol Joan Capdevila JC du Cadet Julien Lavigne JL Visconti Alessia A Freidin Maxim B MB El-Sayed Moustafa Julia S JS Falchi Mario M Davies Richard R Gomez Maria F MF Fall Tove T Cardoso M Jorge MJ Wolf Jonathan J Franks Paul W PW Chan Andrew T AT Spector Tim D TD Steves Claire J CJ Ourselin Sébastien S
Science advances 20210319 12
As no one symptom can predict disease severity or the need for dedicated medical support in coronavirus disease 2019 (COVID-19), we asked whether documenting symptom time series over the first few days informs outcome. Unsupervised time series clustering over symptom presentation was performed on data collected from a training dataset of completed cases enlisted early from the COVID Symptom Study Smartphone application, yielding six distinct symptom presentations. Clustering was validated on an ...[more]
Similar Datasets
Project description:The app-based COVID Symptom Study was launched in Sweden in April 2020 to contribute to real-time COVID-19 surveillance. We enrolled 143,531 study participants (≥18 years) who contributed 10.6 million daily symptom reports between April 29, 2020 and February 10, 2021. Here, we include data from 19,161 self-reported PCR tests to create a symptom-based model to estimate the individual probability of symptomatic COVID-19, with an AUC of 0.78 (95% CI 0.74-0.83) in an external dataset. These individual probabilities are employed to estimate daily regional COVID-19 prevalence, which are in turn used together with current hospital data to predict next week COVID-19 hospital admissions. We show that this hospital prediction model demonstrates a lower median absolute percentage error (MdAPE: 25.9%) across the five most populated regions in Sweden during the first pandemic wave than a model based on case notifications (MdAPE: 30.3%). During the second wave, the error rates are similar. When we apply the same model to an English dataset, not including local COVID-19 test data, we observe MdAPEs of 22.3% and 19.0% during the first and second pandemic waves, respectively, highlighting the transferability of the prediction model.
| S-EPMC9023535 | biostudies-literature
Project description:BackgroundMental health issues have been reported after SARS-CoV-2 infection. However, comparison to prevalence in uninfected individuals and contribution from common risk factors (e.g., obesity, comorbidities) have not been examined. We identified how COVID-19 relates to mental health in the large community-based COVID Symptom Study.MethodsWe assessed anxiety and depression symptoms using two validated questionnaires in 413,148 individuals between February and April 2021; 26,998 had tested positive for SARS-CoV-2. We adjusted for physical and mental pre-pandemic comorbidities, BMI, age, and sex.FindingsOverall, 26.4% of participants met screening criteria for general anxiety and depression. Anxiety and depression were slightly more prevalent in previously SARS-CoV-2 positive (30.4%) vs. negative (26.1%) individuals. This association was small compared to the effect of an unhealthy BMI and the presence of other comorbidities, and not evident in younger participants (≤40 years). Findings were robust to multiple sensitivity analyses. Association between SARS-CoV-2 infection and anxiety and depression was stronger in individuals with recent (<30 days) vs. more distant (>120 days) infection, suggesting a short-term effect.InterpretationA small association was identified between SARS-CoV-2 infection and anxiety and depression symptoms. The proportion meeting criteria for self-reported anxiety and depression disorders is only slightly higher than pre-pandemic.FundingZoe Limited, National Institute for Health Research, Chronic Disease Research Foundation, National Institutes of Health, Medical Research Council UK.
| S-EPMC8282115 | biostudies-literature
Project description:Background and objectiveCOVID-19 manifests with a broad range of symptoms. This study investigates whether clusters of respiratory, gastrointestinal, or neurological symptoms can be used to diagnose COVID-19.MethodsWe surveyed symptoms of 483 subjects who had completed COVID-19 laboratory tests in the last 30 days. The survey collected data on demographic characteristics, self-reported symptoms for different types of infections within 14 days of onset of illness, and self-reported COVID-19 test results. Robust LASSO regression was used to create 3 nested models. In all 3 models, the response variable was the COVID-19 test result. In the first model, referred to as the "main effect model," the independent variables were demographic characteristics, history of chronic symptoms, and current symptoms. The second model, referred to as the "hierarchical clustering model," added clusters of variables to the list of independent variables. These clusters were established through hierarchical clustering. The third model, referred to as the "interaction-terms model," also added clusters of variables to the list of independent variables; this time clusters were established through pairwise and triple-way interaction terms. Models were constructed on a randomly selected 80% of the data and accuracy was cross-validated on the remaining 20% of the data. The process was bootstrapped 30 times. Accuracy of the 3 models was measured using the average of the cross-validated area under the receiver operating characteristic curves (AUROCs).ResultsIn 30 bootstrap samples, the main effect model had an AUROC of 0.78. The hierarchical clustering model had an AUROC of 0.80. The interaction-terms model had an AUROC of 0.81. Both the hierarchical cluster model and the interaction model were significantly different from the main effect model (α = .04). Patients with different races/ethnicities, genders, and ages presented with different symptom clusters.ConclusionsUsing clusters of symptoms, it is possible to more accurately diagnose COVID-19 among symptomatic patients.
| S-EPMC9811485 | biostudies-literature
Project description:BackgroundCOVID-19 symptom reports describe varying levels of disease severity with differing periods of recovery and symptom trajectories. Thus, there are a multitude of disease and symptom characteristics clinicians must navigate and interpret to guide care.ObjectiveTo find natural groups of patients with similar constellations of post-acute sequelae of COVID-19 (PASC) symptoms.DesignCohort SETTING: Outpatient COVID-19 recovery clinic with patient referrals from 160 primary care clinics serving 36 counties in Texas.PatientsAdult patients seeking COVID-19 recovery clinic care between November 15, 2020, and July 31, 2021, with laboratory-confirmed mild (not hospitalized), moderate (hospitalized), or severe (hospitalized with critical care) COVID-19.Main measuresDemographics, COVID illness onset, and duration of persistent PASC symptoms via semi-structured medical assessments.Key resultsFour hundred forty-one patients (mean age 51.5 years; 295 [66.9%] women; 99 [22%] Hispanic, and 170 [38.5%] non-White, racial minority) met inclusion criteria. Using a k-medoids algorithm, we found that PASC symptoms cluster into two distinct groups: neuropsychiatric (N = 186) (e.g., subjective cognitive dysfunction) and pulmonary (N = 255) (e.g., dyspnea, cough). The neuropsychiatric cluster had significantly higher incidences of otolaryngologic (X2 = 14.3, p < 0.001), gastrointestinal (X2 = 6.90, p = 0.009), neurologic (X2 = 441, p < 0.001), and psychiatric sequelae (X2 = 40.6, p < 0.001) with more female (X2 = 5.44, p = 0.020) and younger age (t = 2.39, p = 0.017) patients experiencing longer durations of PASC symptoms before seeking care (t = 2.44, p = 0.015). Patients in the pulmonary cluster were more often hospitalized for COVID-19 (X2 = 3.98, p = 0.046) and had significantly higher comorbidity burden (U = 20800, p = 0.019) and pulmonary sequelae (X2 = 13.2, p < 0.001).ConclusionsHealth services clinic data from a large integrated health system offers insights into the post-COVID symptoms associated with care seeking for sequelae that are not adequately managed by usual care pathways (self-management and primary care clinic visits). These findings can inform machine learning algorithms, primary care management, and selection of patients for earlier COVID-19 recovery referral.Trial registrationN/A.
| S-EPMC9663188 | biostudies-literature
Project description:BackgroundIn the Netherlands, the prevalence of post COVID-19 condition is estimated at 12.7% at 90-150 days after SARS-CoV-2 infection. This study aimed to determine the occurrence of fatigue and other symptoms, to assess how many patients meet the Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) criteria, to identify symptom-based clusters within the P4O2 COVID-19 cohort and to compare these clusters with clusters in a ME/CFS cohort.MethodsIn this multicentre, prospective, observational cohort in the Netherlands, 95 post COVID-19 patients aged 40-65 years were included. Data collection at 3-6 months after infection included demographics, medical history, questionnaires, and a medical examination. Follow-up assessments occurred 9-12 months later, where the same data were collected. Fatigue was determined with the Fatigue Severity Scale (FSS), a score of ≥ 4 means moderate to high fatigue. The frequency and severity of other symptoms and the percentage of patients that meet the ME/CFS criteria were assessed using the DePaul Symptom Questionnaire-2 (DSQ-2). A self-organizing map was used to visualize the clustering of patients based on severity and frequency of 79 symptoms. In a previous study, 337 Dutch ME/CFS patients were clustered based on their symptom scores. The symptom scores of post COVID-19 patients were applied to these clusters to examine whether the same or different clusters were found.ResultsAccording to the FSS, fatigue was reported by 75.9% of the patients at 3-6 months after infection and by 57.1% of the patients 9-12 months later. Post-exertional malaise, sleep disturbances, pain, and neurocognitive symptoms were also frequently reported, according to the DSQ-2. Over half of the patients (52.7%) met the Fukuda criteria for ME/CFS, while fewer patients met other ME/CFS definitions. Clustering revealed specific symptom patterns and showed that post COVID-19 patients occurred in 11 of the clusters that have been observed in the ME/CFS cohort, where 2 clusters had > 10 patients.ConclusionsThis study shows persistent fatigue and diverse symptomatology in post COVID-19 patients, up to 12-18 months after SARS-CoV-2 infection. Clustering showed that post COVID-19 patients occurred in 11 of the clusters that have been observed in the ME/CFS cohort.
| S-EPMC10880228 | biostudies-literature
Project description:BackgroundSome of the most vexing issues with the COVID-19 pandemic were the inability of facilities and events, such as schools and work areas, to track symptoms to mitigate the spread of the disease. To combat these challenges, many turned to the implementation of technology. Technology solutions to mitigate repercussions of the COVID-19 pandemic include tools that provide guidelines and interfaces to influence behavior, reduce exposure to the disease, and enable policy-driven avenues to return to a sense of normalcy. This paper presents the implementation and early evaluation of a return-to-work COVID-19 symptom and risk assessment tool. The system was implemented across 34 institutions of health and education in Alabama, including more than 174,000 users with over 4 million total uses and more than 86,000 reports of exposure risk between July 2020 and April 2021.ObjectiveThis study aimed to explore the usage of technology, specifically a COVID-19 symptom and risk assessment tool, to mitigate exposure to COVID-19 within public spaces. More specifically, the objective was to assess the relationship between user-reported symptoms and exposure via a mobile health app, with confirmed COVID-19 cases reported by the Alabama Department of Public Health (ADPH).MethodsThis cross-sectional study evaluated the relationship between confirmed COVID-19 cases and user-reported COVID-19 symptoms and exposure reported through the Healthcheck web-based mobile application. A dependent variable for confirmed COVID-19 cases in Alabama was obtained from ADPH. Independent variables (ie, health symptoms and exposure) were collected through Healthcheck survey data and included measures assessing COVID-19-related risk levels and symptoms. Multiple linear regression was used to examine the relationship between ADPH-confirmed diagnosis of COVID-19 and self-reported health symptoms and exposure via Healthcheck that were analyzed across the state population but not connected at the individual patient level.ResultsRegression analysis showed that the self-reported information collected by Healthcheck significantly affects the number of COVID-19-confirmed cases. The results demonstrate that the average number of confirmed COVID-19 cases increased by 5 (high risk: β=5.10; P=.001), decreased by 24 (sore throat: β=-24.03; P=.001), and increased by 21 (nausea or vomiting: β=21.67; P=.02) per day for every additional self-report of symptoms by Healthcheck survey respondents. Congestion or runny nose was the most frequently reported symptom. Sore throat, low risk, high risk, nausea, or vomiting were all statistically significant factors.ConclusionsThe use of technology allowed organizations to remotely track a population as it is related to COVID-19. Healthcheck was a platform that aided in symptom tracking, risk assessment, and evaluation of status for admitting individuals into public spaces for people in the Alabama area. The confirmed relationship between symptom and exposure self-reporting using an app and population-wide confirmed cases suggests that further investigation is needed to determine the opportunity for such apps to mitigate disease spread at a community and individual level.
| S-EPMC10410382 | biostudies-literature
Project description:ObjectivesThis study aims to describe the incidence, symptom clusters and determinants of post-acute COVID symptoms using data from the COVID RADAR app in the Netherlands.DesignProspective cohort.SettingGeneral population in the Netherlands from April 2020 to February 2022.ParticipantsA total of 1478 COVID RADAR app users, with data spanning 40 days before to 100 days after positive SARS-CoV-2 test.Outcome measuresIncidence and duration of 10 new symptoms that developed during acute infection, defined as 10 days prior and 10 days after positive test. Clustering of these post-acute COVID symptoms and associations between factors known in the acute phase and 100-day symptom persistence.ResultsThe most frequent post-acute symptoms were cough, loss of smell or taste and fatigue. At 100 days postinfection, 86 (8%) participants still experienced symptoms. Three post-acute COVID symptom clusters were identified: non-respiratory (headache and fatigue; 49% of participants with post-acute COVID symptoms); olfactory (15%) and respiratory (8%). Vaccination was associated with a lower risk of post-acute COVID symptoms 100 days after infection, although CIs were wide (OR: 0.5; 95% CI: 0.2 to 1.5), but not with non-respiratory symptoms (OR: 1.0; 95% CI: 0.3 to 4.4). Severe acute disease increased the risk of post-acute COVID symptoms (OR: 1.4; 95% CI: 1.2 to 1.5; per additional acute symptom).ConclusionsIn this cohort of infected community-dwelling app users, 5%-10% experienced post-acute COVID symptoms. The symptoms cluster in several distinct entities, which differ in incidence, patient characteristics and vaccination effects. This suggests multiple mechanisms underlying the development of post-acute COVID symptoms.
| S-EPMC11409358 | biostudies-literature
Project description:BackgroundAs a counter-cluster measure to prevent the spread of the infectious novel coronavirus disease (COVID-19), an efficient system for health observation outside the hospital is urgently required. Personal health records (PHRs) are suitable for the daily management of physical conditions. Importantly, there are no major differences between the items collected by daily health observation via PHR and the observation of items related to COVID-19. Until now, observations related to COVID-19 have been performed exclusively based on disease-specific items. Therefore, we hypothesize that PHRs would be suitable as a symptom-tracking tool for COVID-19. To this end, we integrated health observation items specific to COVID-19 with an existing PHR-based app.ObjectiveThis study is conducted as a proof-of-concept study in a real-world setting to develop a PHR-based COVID-19 symptom-tracking app and to demonstrate the practical use of health observations for COVID-19 using a smartphone or tablet app integrated with PHRs.MethodsWe applied the PHR-based health observation app within an active epidemiological investigation conducted by Wakayama City Public Health Center. At the public health center, a list is made of individuals who have been in close contact with known infected cases (health observers). Email addresses are used by the app when a health observer sends data to the public health center. Each health observer downloads the app and installs it on their smartphone. Self-observed health data are entered daily into the app. These data are then sent via the app by email at a designated time. Localized epidemiological officers can visualize the collected data using a spreadsheet macro and, thus, monitor the health condition of all health observers.ResultsWe used the app as part of an active epidemiological investigation executed at a public health center. During the investigation, 72 close contacts were discovered. Among them, 57 had adopted the use of the health observation app. Before the introduction of the app, all health observers would have been interviewed by telephone, a slow process that took four epidemiological officers more than 2 hours. After the introduction of the app, a single epidemiological officer can carry out health observations. The app was distributed for free beginning in early March, and by mid-May, it had been used by more than 20,280 users and 400 facilities and organizations across Japan. Currently, health observation of COVID-19 is socially recognized and has become one of the requirements for resuming social activities.ConclusionsHealth observation by PHRs for the purpose of improving health management can also be effectively applied as a measure against large-scale infectious diseases. Individual habits of improving awareness of personal health and the use of PHRs for daily health management are powerful armaments against the rapid spread of infectious diseases. Ultimately, similar actions may help to prevent the spread of COVID-19.
| S-EPMC7340163 | biostudies-literature
Project description:Background: Early prediction of symptoms and mortality risks for COVID-19 patients would improve healthcare outcomes, allow for the appropriate distribution of healthcare resources, reduce healthcare costs, aid in vaccine prioritization and self-isolation strategies, and thus reduce the prevalence of the disease. Such publicly accessible prediction models are lacking, however. Methods: Based on a comprehensive evaluation of existing machine learning (ML) methods, we created two models based solely on the age, gender, and medical histories of 23,749 hospital-confirmed COVID-19 patients from February to September 2020: a symptom prediction model (SPM) and a mortality prediction model (MPM). The SPM predicts 12 symptom groups for each patient: respiratory distress, consciousness disorders, chest pain, paresis or paralysis, cough, fever or chill, gastrointestinal symptoms, sore throat, headache, vertigo, loss of smell or taste, and muscular pain or fatigue. The MPM predicts the death of COVID-19-positive individuals. Results: The SPM yielded ROC-AUCs of 0.53-0.78 for symptoms. The most accurate prediction was for consciousness disorders at a sensitivity of 74% and a specificity of 70%. 2,440 deaths were observed in the study population. MPM had a ROC-AUC of 0.79 and could predict mortality with a sensitivity of 75% and a specificity of 70%. About 90% of deaths occurred in the top 21 percentile of risk groups. To allow patients and clinicians to use these models easily, we created a freely accessible online interface at www.aicovid.net. Conclusion: The ML models predict COVID-19-related symptoms and mortality using information that is readily available to patients as well as clinicians. Thus, both can rapidly estimate the severity of the disease, allowing shared and better healthcare decisions with regard to hospitalization, self-isolation strategy, and COVID-19 vaccine prioritization in the coming months.
| S-EPMC8262614 | biostudies-literature