Your browser doesn't support javascript.
Show: 20 | 50 | 100
Results 1 - 20 de 210
Filter
1.
Cochrane Database Syst Rev ; 8: CD015061, 2021 08 23.
Article in English | MEDLINE | ID: covidwho-1813447

ABSTRACT

BACKGROUND: Individuals dying of coronavirus disease 2019 (COVID-19) may experience distressing symptoms such as breathlessness or delirium. Palliative symptom management can alleviate symptoms and improve the quality of life of patients. Various treatment options such as opioids or breathing techniques have been discussed for use in COVID-19 patients. However, guidance on symptom management of COVID-19 patients in palliative care has often been derived from clinical experiences and guidelines for the treatment of patients with other illnesses. An understanding of the effectiveness of pharmacological and non-pharmacological palliative interventions to manage specific symptoms of COVID-19 patients is required. OBJECTIVES: To assess the efficacy and safety of pharmacological and non-pharmacological interventions for palliative symptom control in individuals with COVID-19. SEARCH METHODS: We searched the Cochrane COVID-19 Study Register (including Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE (PubMed), Embase, ClinicalTrials.gov, World Health Organization International Clinical Trials Registry Platform (WHO ICTRP), medRxiv); Web of Science Core Collection (Science Citation Index Expanded, Emerging Sources); CINAHL; WHO COVID-19 Global literature on coronavirus disease; and COAP Living Evidence on COVID-19 to identify completed and ongoing studies without language restrictions until 23 March 2021. We screened the reference lists of relevant review articles and current treatment guidelines for further literature. SELECTION CRITERIA: We followed standard Cochrane methodology as outlined in the Cochrane Handbook for Systematic Reviews of Interventions. We included studies evaluating palliative symptom management for individuals with a confirmed diagnosis of COVID-19 receiving interventions for palliative symptom control, with no restrictions regarding comorbidities, age, gender, or ethnicity. Interventions comprised pharmacological as well as non-pharmacological treatment (e.g. acupressure, physical therapy, relaxation, or breathing techniques). We searched for the following types of studies: randomized controlled trials (RCT), quasi-RCTs, controlled clinical trials, controlled before-after studies, interrupted time series (with comparison group), prospective cohort studies, retrospective cohort studies, (nested) case-control studies, and cross-sectional studies. We searched for studies comparing pharmacological and non-pharmacological interventions for palliative symptom control with standard care. We excluded studies evaluating palliative interventions for symptoms caused by other terminal illnesses. If studies enrolled populations with or exposed to multiple diseases, we would only include these if the authors provided subgroup data for individuals with COVID-19. We excluded studies investigating interventions for symptom control in a curative setting, for example patients receiving life-prolonging therapies such as invasive ventilation.  DATA COLLECTION AND ANALYSIS: We used a modified version of the Newcastle Ottawa Scale for non-randomized studies of interventions (NRSIs) to assess bias in the included studies. We included the following outcomes: symptom relief (primary outcome); quality of life; symptom burden; satisfaction of patients, caregivers, and relatives; serious adverse events; and grade 3 to 4 adverse events. We rated the certainty of evidence using the GRADE approach.  As meta-analysis was not possible, we used tabulation to synthesize the studies and histograms to display the outcomes.  MAIN RESULTS: Overall, we identified four uncontrolled retrospective cohort studies investigating pharmacological interventions for palliative symptom control in hospitalized patients and patients in nursing homes. None of the studies included a comparator. We rated the risk of bias high across all studies. We rated the certainty of the evidence as very low for the primary outcome symptom relief, downgrading mainly for high risk of bias due to confounding and unblinded outcome assessors. Pharmacological interventions for palliative symptom control We identified four uncontrolled retrospective cohort studies (five references) investigating pharmacological interventions for palliative symptom control. Two references used the same register to form their cohorts, and study investigators confirmed a partial overlap of participants. We therefore do not know the exact number of participants, but individual reports included 61 to 2105 participants. Participants received multimodal pharmacological interventions: opioids, neuroleptics, anticholinergics, and benzodiazepines for relieving dyspnea (breathlessness), delirium, anxiety, pain, audible upper airway secretions, respiratory secretions, nausea, cough, and unspecified symptoms.  Primary outcome: symptom relief All identified studies reported this outcome. For all symptoms (dyspnea, delirium, anxiety, pain, audible upper airway secretions, respiratory secretions, nausea, cough, and unspecified symptoms), a majority of interventions were rated as completely or partially effective by outcome assessors (treating clinicians or nursing staff). Interventions used in the studies were opioids, neuroleptics, anticholinergics, and benzodiazepines.  We are very uncertain about the effect of pharmacological interventions on symptom relief (very low-certainty evidence). The initial rating of the certainty of evidence was low since we only identified uncontrolled NRSIs. Our main reason for downgrading the certainty of evidence was high risk of bias due to confounding and unblinded outcome assessors. We therefore did not find evidence to confidently support or refute whether pharmacological interventions may be effective for palliative symptom relief in COVID-19 patients. Secondary outcomes We planned to include the following outcomes: quality of life; symptom burden; satisfaction of patients, caregivers, and relatives; serious adverse events; and grade 3 to 4 adverse events. We did not find any data for these outcomes, or any other information on the efficacy and safety of used interventions. Non-pharmacological interventions for palliative symptom control None of the identified studies used non-pharmacological interventions for palliative symptom control. AUTHORS' CONCLUSIONS: We found very low certainty evidence for the efficacy of pharmacological interventions for palliative symptom relief in COVID-19 patients. We found no evidence on the safety of pharmacological interventions or efficacy and safety of non-pharmacological interventions for palliative symptom control in COVID-19 patients. The evidence presented here has no specific implications for palliative symptom control in COVID-19 patients because we cannot draw any conclusions about the effectiveness or safety based on the identified evidence. More evidence is needed to guide clinicians, nursing staff, and caregivers when treating symptoms of COVID-19 patients at the end of life. Specifically, future studies ought to investigate palliative symptom control in prospectively registered studies, using an active-controlled setting, assess patient-reported outcomes, and clearly define interventions. The publication of the results of ongoing studies will necessitate an update of this review. The conclusions of an updated review could differ from those of the present review and may allow for a better judgement regarding pharmacological and non-pharmacological interventions for palliative symptom control in COVID-19 patients.


Subject(s)
COVID-19/therapy , Palliative Care , Aged , Aged, 80 and over , Bias , COVID-19/diagnosis , Humans , Male , SARS-CoV-2 , Systematic Reviews as Topic
2.
Cochrane Database Syst Rev ; 8: CD014962, 2021 08 05.
Article in English | MEDLINE | ID: covidwho-1813444

ABSTRACT

BACKGROUND: Remdesivir is an antiviral medicine with properties to inhibit viral replication of SARS-CoV-2. Positive results from early studies attracted media attention and led to emergency use authorisation of remdesivir in COVID-19.  A thorough understanding of the current evidence regarding the effects of remdesivir as a treatment for SARS-CoV-2 infection based on randomised controlled trials (RCTs) is required. OBJECTIVES: To assess the effects of remdesivir compared to placebo or standard care alone on clinical outcomes in hospitalised patients with SARS-CoV-2 infection, and to maintain the currency of the evidence using a living systematic review approach. SEARCH METHODS: We searched the Cochrane COVID-19 Study Register (which comprises the Cochrane Central Register of Controlled Trials (CENTRAL), PubMed, Embase, ClinicalTrials.gov, WHO International Clinical Trials Registry Platform, and medRxiv) as well as Web of Science (Science Citation Index Expanded and Emerging Sources Citation Index) and WHO COVID-19 Global literature on coronavirus disease to identify completed and ongoing studies without language restrictions. We conducted the searches on 16 April 2021. SELECTION CRITERIA: We followed standard Cochrane methodology. We included RCTs evaluating remdesivir for the treatment of SARS-CoV-2 infection in hospitalised adults compared to placebo or standard care alone irrespective of disease severity, gender, ethnicity, or setting.  We excluded studies that evaluated remdesivir for the treatment of other coronavirus diseases. DATA COLLECTION AND ANALYSIS: We followed standard Cochrane methodology. To assess risk of bias in included studies, we used the Cochrane RoB 2 tool for RCTs. We rated the certainty of evidence using the GRADE approach for outcomes that were reported according to our prioritised categories: all-cause mortality at up to day 28, duration to liberation from invasive mechanical ventilation, duration to liberation from supplemental oxygen, new need for mechanical ventilation (high-flow oxygen or non-invasive or invasive mechanical ventilation), new need for invasive mechanical ventilation, new need for non-invasive mechanical ventilation or high-flow oxygen, new need for oxygen by mask or nasal prongs, quality of life, adverse events (any grade), and serious adverse events. MAIN RESULTS: We included five RCTs with 7452 participants diagnosed with SARS-CoV-2 infection and a mean age of 59 years, of whom 3886 participants were randomised to receive remdesivir. Most participants required low-flow oxygen (n=4409) or mechanical ventilation (n=1025) at baseline. We identified two ongoing studies, one was suspended due to a lack of COVID-19 patients to recruit. Risk of bias was considered to be of some concerns or high risk for clinical status and safety outcomes because participants who had died did not contribute information to these outcomes. Without adjustment, this leads to an uncertain amount of missing values and the potential for bias due to missing data. Effects of remdesivir in hospitalised individuals  Remdesivir probably makes little or no difference to all-cause mortality at up to day 28 (risk ratio (RR) 0.93, 95% confidence interval (CI) 0.81 to 1.06; risk difference (RD) 8 fewer per 1000, 95% CI 21 fewer to 7 more; 4 studies, 7142 participants; moderate-certainty evidence). Considering the initial severity of condition, only one study showed a beneficial effect of remdesivir in patients who received low-flow oxygen at baseline (RR 0.32, 95% CI 0.15 to 0.66, 435 participants), but conflicting results exists from another study, and we were unable to validly assess this observations due to limited availability of comparable data. Remdesivir may have little or no effect on the duration to liberation from invasive mechanical ventilation (2 studies, 1298 participants, data not pooled, low-certainty evidence). We are uncertain whether remdesivir increases or decreases the chance of clinical improvement in terms of duration to liberation from supplemental oxygen at up to day 28 (3 studies, 1691 participants, data not pooled, very low-certainty evidence).   We are very uncertain whether remdesivir decreases or increases the risk of clinical worsening in terms of new need for mechanical ventilation at up to day 28 (high-flow oxygen or non-invasive ventilation or invasive mechanical ventilation) (RR 0.78, 95% CI 0.48 to 1.24; RD 29 fewer per 1000, 95% CI 68 fewer to 32 more; 3 studies, 6696 participants; very low-certainty evidence); new need for non-invasive mechanical ventilation or high-flow oxygen (RR 0.70, 95% CI 0.51 to 0.98; RD 72 fewer per 1000, 95% CI 118 fewer to 5 fewer; 1 study, 573 participants; very low-certainty evidence); and new need for oxygen by mask or nasal prongs (RR 0.81, 95% CI 0.54 to 1.22; RD 84 fewer per 1000, 95% CI 204 fewer to 98 more; 1 study, 138 participants; very low-certainty evidence). The evidence suggests that remdesivir may decrease the risk of clinical worsening in terms of new need for invasive mechanical ventilation (67 fewer participants amongst 1000 participants; RR 0.56, 95% CI 0.41 to 0.77; 2 studies, 1159 participants; low-certainty evidence).  None of the included studies reported quality of life. Remdesivir probably decreases the serious adverse events rate at up to 28 days (RR 0.75, 95% CI 0.63 to 0.90; RD 63 fewer per 1000, 95% CI 94 fewer to 25 fewer; 3 studies, 1674 participants; moderate-certainty evidence). We are very uncertain whether remdesivir increases or decreases adverse events rate (any grade) (RR 1.05, 95% CI 0.86 to 1.27; RD 29 more per 1000, 95% CI 82 fewer to 158 more; 3 studies, 1674 participants; very low-certainty evidence). AUTHORS' CONCLUSIONS: Based on the currently available evidence, we are moderately certain that remdesivir probably has little or no effect on all-cause mortality at up to day 28 in hospitalised adults with SARS-CoV-2 infection. We are uncertain about the effects of remdesivir on clinical improvement and worsening. There were insufficient data available to validly examine the effect of remdesivir on mortality in subgroups depending on the extent of respiratory support at baseline.  Future studies should provide additional data on efficacy and safety of remdesivir for defined core outcomes in COVID-19 research, especially for different population subgroups. This could allow us to draw more reliable conclusions on the potential benefits and harms of remdesivir in future updates of this review. Due to the living approach of this work, we will update the review periodically.


Subject(s)
Adenosine Monophosphate/analogs & derivatives , Alanine/analogs & derivatives , Antiviral Agents/therapeutic use , COVID-19/drug therapy , Adenosine Monophosphate/therapeutic use , Alanine/therapeutic use , Bias , COVID-19/mortality , Cause of Death , Confidence Intervals , Disease Progression , Humans , Middle Aged , Oxygen/administration & dosage , Randomized Controlled Trials as Topic , Respiration, Artificial , SARS-CoV-2 , Ventilator Weaning
3.
BMJ Open ; 12(4): e057885, 2022 Apr 11.
Article in English | MEDLINE | ID: covidwho-1784830

ABSTRACT

INTRODUCTION: Postviral syndromes (PVS) describe the sustained presence of symptoms following an acute viral infection, for months or even years. Exposure to the SARS-CoV-2 virus and subsequent development of COVID-19 has shown to have similar effects with individuals continuing to exhibit symptoms for greater than 12 weeks. The sustained presence of symptoms is variably referred to as 'post COVID-19 syndrome', 'post-COVID condition' or more commonly 'Long COVID'. Knowledge of the long-term health impacts and treatments for Long COVID are evolving. To minimise overlap with existing work in the field exploring treatments of Long COVID, we have only chosen to focus on non-pharmacological treatments. AIMS: This review aims to summarise the effectiveness of non-pharmacological treatments for PVS, including Long COVID. A secondary aim is to summarise the symptoms and health impacts associated with PVS in individuals recruited to treatment studies. METHODS AND ANALYSIS: Primary electronic searches will be performed in bibliographic databases including: Embase, MEDLINE, PyscINFO, CINAHL and MedRxiv from 1 January 2001 to 29 October 2021. At least two independent reviewers will screen each study for inclusion and data will be extracted from all eligible studies onto a data extraction form. The quality of all included studies will be assessed using Cochrane risk of bias tools and the Newcastle-Ottawa grading system. Non-pharmacological treatments for PVS and Long COVID will be narratively summarised and effect estimates will be pooled using random effects meta-analysis where there is sufficient methodological homogeneity. The symptoms and health impacts reported in the included studies on non-pharmacological interventions will be extracted and narratively reported. ETHICS AND DISSEMINATION: This systematic review does not require ethical approval. The findings from this study will be submitted for peer-reviewed publication, shared at conference presentations and disseminated to both clinical and patient groups. PROSPERO REGISTRATION NUMBER: The review will adhere to this protocol which has also been registered with PROSPERO (CRD42021282074).


Subject(s)
COVID-19 , Bias , COVID-19/complications , COVID-19/therapy , Humans , Meta-Analysis as Topic , Research Design , SARS-CoV-2 , Syndrome , Systematic Reviews as Topic
4.
Am J Public Health ; 111(12): 2167-2175, 2021 12.
Article in English | MEDLINE | ID: covidwho-1760043

ABSTRACT

High-quality data are accurate, relevant, and timely. Large national health surveys have always balanced the implementation of these quality dimensions to meet the needs of diverse users. The COVID-19 pandemic shifted these balances, with both disrupted survey operations and a critical need for relevant and timely health data for decision-making. The National Health Interview Survey (NHIS) responded to these challenges with several operational changes to continue production in 2020. However, data files from the 2020 NHIS were not expected to be publicly available until fall 2021. To fill the gap, the National Center for Health Statistics (NCHS) turned to 2 online data collection platforms-the Census Bureau's Household Pulse Survey (HPS) and the NCHS Research and Development Survey (RANDS)-to collect COVID-19‒related data more quickly. This article describes the adaptations of NHIS and the use of HPS and RANDS during the pandemic in the context of the recently released Framework for Data Quality from the Federal Committee on Statistical Methodology. (Am J Public Health. 2021;111(12):2167-2175. https://doi.org/10.2105/AJPH.2021.306516).


Subject(s)
COVID-19/epidemiology , Health Surveys/methods , Internet , National Center for Health Statistics, U.S./organization & administration , Bias , Cross-Sectional Studies , Data Collection/methods , Data Collection/standards , Health Surveys/standards , Humans , Interviews as Topic , Pandemics , SARS-CoV-2 , Telephone , United States/epidemiology
6.
Epidemiol Infect ; 150: e48, 2022 02 21.
Article in English | MEDLINE | ID: covidwho-1758088

ABSTRACT

SARS-CoV-2 serological tests are used to assess the infection seroprevalence within a population. This study aims at assessing potential biases in estimating infection prevalence amongst healthcare workers (HCWs) when different diagnostic criteria are considered. A multi-site cross-sectional study was carried out in April-September 2020 amongst 1.367 Italian HCWs. SARS-CoV-2 prevalence was assessed using three diagnostic criteria: RT-PCR on nasopharyngeal swab, point-of-care fingerprick serological test (POCT) result and COVID-19 clinical pathognomonic presentation. A logistic regression model was used to estimate the probability of POCT-positive result in relation to the time since infection (RT-PCR positivity). Among 1.367 HCWs, 69.2% were working in COVID-19 units. Statistically significant differences in age, role and gender were observed between COVID-19/non-COVID-19 units. Prevalence of SARS-CoV-2 infection varied according to the criterion considered: 6.7% for POCT, 8.1% for RT-PCR, 10.0% for either POCT or RT-PCR, 9.6% for infection pathognomonic clinical presentation and 17.6% when at least one of the previous criteria was present. The probability of POCT-positive result decreased by 1.1% every 10 days from the infection. This study highlights potential biases in estimating SARS-CoV-2 point-prevalence data according to the criteria used. Although informative on infection susceptibility and herd immunity level, POCT serological tests are not the best predictors of previous COVID-19 infections for public health monitoring programmes.


Subject(s)
COVID-19 Nucleic Acid Testing , COVID-19 Serological Testing , COVID-19/diagnosis , COVID-19/epidemiology , Health Personnel , Point-of-Care Testing , SARS-CoV-2 , Adult , Bias , Cross-Sectional Studies , Female , Humans , Italy/epidemiology , Male , Middle Aged , Occupational Exposure , Prevalence , Probability , Seroepidemiologic Studies
7.
AACN Adv Crit Care ; 33(1): 111-118, 2022 Mar 15.
Article in English | MEDLINE | ID: covidwho-1744839

Subject(s)
Racism , Bias , Humans
8.
Am J Epidemiol ; 190(11): 2280-2283, 2021 11 02.
Article in English | MEDLINE | ID: covidwho-1722199

ABSTRACT

Dimitris and Platt (Am J Epidemiol. 2021;190(11):2275-2279) take on the challenging topic of using "shocks" such as the severe acute respiratory system coronavirus 2 (SARS-CoV-2) pandemic as instrumental variables to study the effect of some exposure on some outcome. Evoking our recent lived experiences, they conclude that the assumptions necessary for an instrumental variable analysis will often be violated and therefore strongly caution against such analyses. Here, we build upon this warranted caution while acknowledging that such analyses will still be pursued and conducted. We discuss strategies for evaluating or reasoning about when such an analysis is clearly inappropriate for a given research question, as well as strategies for interpreting study findings with special attention to incorporating plausible sources of bias in any conclusions drawn from a given finding.


Subject(s)
COVID-19 , Pandemics , Bias , Humans , Pandemics/prevention & control , SARS-CoV-2
9.
Am J Epidemiol ; 190(11): 2275-2279, 2021 11 02.
Article in English | MEDLINE | ID: covidwho-1722196

ABSTRACT

Epidemiologists sometimes use external sources of variation to explore highly confounded exposure-outcome relationships or exposures that cannot be randomized. These exogenous sources of variation, or natural experiments, are sometimes proposed as instrumental variables to examine the effects of given exposures on given outcomes. Previous epidemiologic studies have applied this technique using famines, earthquakes, weather events, and previous pandemics as exogenous sources of variation for other exposures; interest in applying this technique using the current severe acute respiratory system coronavirus 2 (SARS-CoV-2) pandemic is already documented. Yet large-scale events like these likely have broad and complicated impacts on human health, which almost certainly violates the exclusion restriction assumption of instrumental variable analyses. We review the assumptions of instrumental variable analyses, highlight previous applications of this method with respect to natural experiments with broad impacts or "shocks," and discuss how these relate to our current observations of the SARS-CoV-2 pandemic. While we encourage thorough investigation of the broad impacts of the SARS-CoV-2 pandemic on human health, we caution against its widespread use as an instrumental variable to study other exposures of interest.


Subject(s)
COVID-19 , Pandemics , Bias , Epidemiologic Studies , Humans , Pandemics/prevention & control , SARS-CoV-2
10.
J Gerontol B Psychol Sci Soc Sci ; 77(4): e83-e94, 2022 Apr 01.
Article in English | MEDLINE | ID: covidwho-1704272

ABSTRACT

OBJECTIVES: Optimistic bias refers to the phenomenon that individuals believe bad things are less likely to happen to themselves than to others. However, whether optimistic bias could vary across age and culture is unknown. The present study aims to investigate (a) whether individuals exhibit optimistic bias in the context of coronavirus disease 2019 (COVID-19) pandemic, and (b) whether age and culture would moderate such bias. METHOD: 1,051 participants recruited from China, Israel, and the United States took the online survey. Risk perceptions consist of 3 questions: estimating the infected probability of different social distance groups (i.e., self, close others, and nonclose others), the days that it would take for the number of new infections to decrease to zero and the trend of infections in regions of different geographical distances (i.e., local place, other places inside participants' country, and other countries). Participants in China and the United States also reported their personal communal values measured by Schwartz's Value Survey. RESULTS: Results from Hierarchical Linear Modeling generally confirmed that (a) all participants exhibited optimistic bias to some extent, and (b) with age, Chinese participants had a higher level of optimistic bias than the Israeli and U.S. participants. Compared to their younger counterparts, older Chinese are more likely to believe that local communities are at lower risk of COVID-19 than other countries. DISCUSSION: These findings support the hypothesis that age differences in risk perceptions might be influenced by cultural context. Further analysis indicated that such cultural and age variations in optimistic bias were likely to be driven by age-related increase in internalized cultural values.


Subject(s)
COVID-19 , Aged , Bias , COVID-19/epidemiology , China/epidemiology , Humans , Israel/epidemiology , Surveys and Questionnaires , United States/epidemiology
11.
J Exp Psychol Appl ; 27(4): 632-656, 2021 Dec.
Article in English | MEDLINE | ID: covidwho-1649845

ABSTRACT

At the onset of the coronavirus disease (COVID-19) global pandemic, our interdisciplinary team hypothesized that a mathematical misconception-whole number bias (WNB)-contributed to beliefs that COVID-19 was less fatal than the flu. We created a brief online educational intervention for adults, leveraging evidence-based cognitive science research, to promote accurate understanding of rational numbers related to COVID-19. Participants from a Qualtrics panel (N = 1,297; 75% White) were randomly assigned to an intervention or control condition, solved health-related math problems, and subsequently completed 10 days of daily diaries in which health cognitions and affect were assessed. Participants who engaged with the intervention, relative to those in the control condition, were more accurate and less likely to explicitly mention WNB errors in their strategy reports as they solved COVID-19-related math problems. Math anxiety was positively associated with risk perceptions, worry, and negative affect immediately after the intervention and across the daily diaries. These results extend the benefits of worked examples in a practically relevant domain. Ameliorating WNB errors could not only help people think more accurately about COVID-19 statistics expressed as rational numbers, but also about novel future health crises, or any other context that involves information expressed as rational numbers. (PsycInfo Database Record (c) 2022 APA, all rights reserved).


Subject(s)
COVID-19 , Adult , Bias , Humans , Mathematics , Pandemics , SARS-CoV-2
12.
Stat Med ; 41(10): 1735-1750, 2022 May 10.
Article in English | MEDLINE | ID: covidwho-1653345

ABSTRACT

We propose a modified self-controlled case series (SCCS) method to handle both event-dependent exposures and high event-related mortality. This development is motivated by an epidemiological study undertaken in France to quantify potential risks of cardiovascular events associated with COVID-19 vaccines. Event-dependence of vaccinations, and high event-related mortality, are likely to arise in other SCCS studies of COVID-19 vaccine safety. Using this case study and simulations to broaden its scope, we explore these features and the biases they may generate, implement the modified SCCS model, illustrate some of the properties of this model, and develop a new test for presence of a dose effect. The model we propose has wider application, notably when the event of interest is death.


Subject(s)
COVID-19 Vaccines , COVID-19 , Bias , COVID-19/epidemiology , COVID-19/prevention & control , COVID-19 Vaccines/adverse effects , Humans , Research Design , Vaccination
13.
Acad Emerg Med ; 29(2): 206-216, 2022 02.
Article in English | MEDLINE | ID: covidwho-1642593

ABSTRACT

BACKGROUND: Throughout 2020, the coronavirus disease 2019 (COVID-19) has become a threat to public health on national and global level. There has been an immediate need for research to understand the clinical signs and symptoms of COVID-19 that can help predict deterioration including mechanical ventilation, organ support, and death. Studies thus far have addressed the epidemiology of the disease, common presentations, and susceptibility to acquisition and transmission of the virus; however, an accurate prognostic model for severe manifestations of COVID-19 is still needed because of the limited healthcare resources available. OBJECTIVE: This systematic review aims to evaluate published reports of prediction models for severe illnesses caused COVID-19. METHODS: Searches were developed by the primary author and a medical librarian using an iterative process of gathering and evaluating terms. Comprehensive strategies, including both index and keyword methods, were devised for PubMed and EMBASE. The data of confirmed COVID-19 patients from randomized control studies, cohort studies, and case-control studies published between January 2020 and May 2021 were retrieved. Studies were independently assessed for risk of bias and applicability using the Prediction Model Risk Of Bias Assessment Tool (PROBAST). We collected study type, setting, sample size, type of validation, and outcome including intubation, ventilation, any other type of organ support, or death. The combination of the prediction model, scoring system, performance of predictive models, and geographic locations were summarized. RESULTS: A primary review found 445 articles relevant based on title and abstract. After further review, 366 were excluded based on the defined inclusion and exclusion criteria. Seventy-nine articles were included in the qualitative analysis. Inter observer agreement on inclusion 0.84 (95%CI 0.78-0.89). When the PROBAST tool was applied, 70 of the 79 articles were identified to have high or unclear risk of bias, or high or unclear concern for applicability. Nine studies reported prediction models that were rated as low risk of bias and low concerns for applicability. CONCLUSION: Several prognostic models for COVID-19 were identified, with varying clinical score performance. Nine studies that had a low risk of bias and low concern for applicability, one from a general public population and hospital setting. The most promising and well-validated scores include Clift et al.,15 and Knight et al.,18 which seem to have accurate prediction models that clinicians can use in the public health and emergency department setting.


Subject(s)
COVID-19 , Bias , Cohort Studies , Humans , Prognosis , SARS-CoV-2
14.
Med Phys ; 49(2): 978-987, 2022 Feb.
Article in English | MEDLINE | ID: covidwho-1631696

ABSTRACT

PURPOSE: Over the last 2 years, the artificial intelligence (AI) community has presented several automatic screening tools for coronavirus disease 2019 (COVID-19) based on chest radiography (CXR), with reported accuracies often well over 90%. However, it has been noted that many of these studies have likely suffered from dataset bias, leading to overly optimistic results. The purpose of this study was to thoroughly investigate to what extent biases have influenced the performance of a range of previously proposed and promising convolutional neural networks (CNNs), and to determine what performance can be expected with current CNNs on a realistic and unbiased dataset. METHODS: Five CNNs for COVID-19 positive/negative classification were implemented for evaluation, namely VGG19, ResNet50, InceptionV3, DenseNet201, and COVID-Net. To perform both internal and cross-dataset evaluations, four datasets were created. The first dataset Valencian Region Medical Image Bank (BIMCV) followed strict reverse transcriptase-polymerase chain reaction (RT-PCR) test criteria and was created from a single reliable open access databank, while the second dataset (COVIDxB8) was created through a combination of six online CXR repositories. The third and fourth datasets were created by combining the opposing classes from the BIMCV and COVIDxB8 datasets. To decrease inter-dataset variability, a pre-processing workflow of resizing, normalization, and histogram equalization were applied to all datasets. Classification performance was evaluated on unseen test sets using precision and recall. A qualitative sanity check was performed by evaluating saliency maps displaying the top 5%, 10%, and 20% most salient segments in the input CXRs, to evaluate whether the CNNs were using relevant information for decision making. In an additional experiment and to further investigate the origin of potential dataset bias, all pixel values outside the lungs were set to zero through automatic lung segmentation before training and testing. RESULTS: When trained and evaluated on the single online source dataset (BIMCV), the performance of all CNNs is relatively low (precision: 0.65-0.72, recall: 0.59-0.71), but remains relatively consistent during external evaluation (precision: 0.58-0.82, recall: 0.57-0.72). On the contrary, when trained and internally evaluated on the combinatory datasets, all CNNs performed well across all metrics (precision: 0.94-1.00, recall: 0.77-1.00). However, when subsequently evaluated cross-dataset, results dropped substantially (precision: 0.10-0.61, recall: 0.04-0.80). For all datasets, saliency maps revealed the CNNs rarely focus on areas inside the lungs for their decision-making. However, even when setting all pixel values outside the lungs to zero, classification performance does not change and dataset bias remains. CONCLUSIONS: Results in this study confirm that when trained on a combinatory dataset, CNNs tend to learn the origin of the CXRs rather than the presence or absence of disease, a behavior known as short-cut learning. The bias is shown to originate from differences in overall pixel values rather than embedded text or symbols, despite consistent image pre-processing. When trained on a reliable, and realistic single-source dataset in which non-lung pixels have been masked, CNNs currently show limited sensitivity (<70%) for COVID-19 infection in CXR, questioning their use as a reliable automatic screening tool.


Subject(s)
COVID-19 , Deep Learning , Artificial Intelligence , Bias , Humans , Radiography , SARS-CoV-2
16.
Nat Med ; 27(12): 2079-2081, 2021 12.
Article in English | MEDLINE | ID: covidwho-1612197
17.
Philos Trans A Math Phys Eng Sci ; 380(2214): 20210121, 2022 Jan 10.
Article in English | MEDLINE | ID: covidwho-1603979

ABSTRACT

We develop a statistical model for the testing of disease prevalence in a population. The model assumes a binary test result, positive or negative, but allows for biases in sample selection and both type I (false positive) and type II (false negative) testing errors. Our model also incorporates multiple test types and is able to distinguish between retesting and exclusion after testing. Our quantitative framework allows us to directly interpret testing results as a function of errors and biases. By applying our testing model to COVID-19 testing data and actual case data from specific jurisdictions, we are able to estimate and provide uncertainty quantification of indices that are crucial in a pandemic, such as disease prevalence and fatality ratios. This article is part of the theme issue 'Data science approach to infectious disease surveillance'.


Subject(s)
COVID-19 Testing , COVID-19 , Bias , False Positive Reactions , Humans , Models, Statistical , SARS-CoV-2 , Selection Bias , Sensitivity and Specificity
18.
Nat Microbiol ; 7(1): 97-107, 2022 01.
Article in English | MEDLINE | ID: covidwho-1596437

ABSTRACT

Global and national surveillance of SARS-CoV-2 epidemiology is mostly based on targeted schemes focused on testing individuals with symptoms. These tested groups are often unrepresentative of the wider population and exhibit test positivity rates that are biased upwards compared with the true population prevalence. Such data are routinely used to infer infection prevalence and the effective reproduction number, Rt, which affects public health policy. Here, we describe a causal framework that provides debiased fine-scale spatiotemporal estimates by combining targeted test counts with data from a randomized surveillance study in the United Kingdom called REACT. Our probabilistic model includes a bias parameter that captures the increased probability of an infected individual being tested, relative to a non-infected individual, and transforms observed test counts to debiased estimates of the true underlying local prevalence and Rt. We validated our approach on held-out REACT data over a 7-month period. Furthermore, our local estimates of Rt are indicative of 1-week- and 2-week-ahead changes in SARS-CoV-2-positive case numbers. We also observed increases in estimated local prevalence and Rt that reflect the spread of the Alpha and Delta variants. Our results illustrate how randomized surveys can augment targeted testing to improve statistical accuracy in monitoring the spread of emerging and ongoing infectious disease.


Subject(s)
COVID-19/epidemiology , Models, Statistical , SARS-CoV-2/isolation & purification , Basic Reproduction Number , Bias , COVID-19/diagnosis , COVID-19/transmission , COVID-19 Testing/statistics & numerical data , Forecasting , Humans , Prevalence , Reproducibility of Results , SARS-CoV-2/genetics , Spatio-Temporal Analysis , United Kingdom/epidemiology
19.
MedEdPORTAL ; 17: 11190, 2021.
Article in English | MEDLINE | ID: covidwho-1599218

ABSTRACT

Introduction: The morbidity and mortality (M&M) conference has long been a part of the education of residents of all specialties in the United States, yet its structure is variable across training programs. Recent literature has described the use of M&M as a forum for education in quality improvement methodology; however, a structure focusing on education in cognitive biases and errors has not been previously described in MedEdPORTAL. Methods: This structured M&M conference series called upon resident presenters and peers in the audience to examine cognitive biases and errors involved in specific patient cases. Associated materials included preparatory guidelines provided to faculty advisors and resident presenters, a presentation template used during the introductory session, and a handout used during the discussion portions of presentations. Results: During the 2019-2020 academic year, a total of 24 PGY 2 pediatrics residents presented M&M cases. They identified a mean of 3.7 (SD = 1.9) cognitive biases and/or errors per case and a mean of 1.7 (SD = 0.7) debiasing strategies per case. Peers in the audience were also successful in identifying potential biases and errors at play during presentations. Discussion: We found that through this M&M conference structure, residents were able to demonstrate the ability to identify cognitive errors and biases both within themselves and in peers. This provided an effective forum for the identification and discussion of debiasing strategies, even when the series was forced to transition to a virtual format due to the COVID-19 pandemic.


Subject(s)
COVID-19 , Pandemics , Bias , Child , Cognition , Curriculum , Humans , Morbidity , SARS-CoV-2 , United States
20.
Eur J Phys Rehabil Med ; 57(5): 850-857, 2021 10.
Article in English | MEDLINE | ID: covidwho-1592179

ABSTRACT

INTRODUCTION: This paper updates and summarizes the current evidence informing rehabilitation of patients with COVID-19 and/or describing the consequences of the disease and its treatment. EVIDENCE ACQUISITION: Studies published from May 1st to June 30th, 2021 were selected, excluding descriptive studies and expert opinions. Papers were categorized according to study design, research question, COVID-19 phase, limitations of functioning of rehabilitation interest, and type of rehabilitation service involved. From this edition, we improved the quality assessment using the Joanna Briggs Institute checklists for observational studies and the Cochrane Risk of Bias Tool for randomized-controlled clinical trials (RCTs). EVIDENCE SYNTHESIS: Twenty-five, out of 3699 papers, were included. They were three RCTs, 13 cross-sectional studies and nine cohort studies. Twenty studies reported data on symptom prevalence (N.=13) or disease natural history (N.=7); and five studies reported intervention effectiveness at the individual level. All study participants were COVID survivors and 48% of studies collected information on participants 6 months or longer after COVID-19 onset. The most frequent risks of bias for RCTs concerned weaknesses in allocation concealment, blinding of therapists, and lack of intention-to-treat analysis. Most analytical studies failed to identify or deal with confounders, describe or deal with dropouts or eventually perform an appropriate statistical analysis. CONCLUSIONS: Most studies in this updated review targeted the prevalence of limitations of functioning of rehabilitation interest in COVID-19 survivors. This is similar to past review findings; however, data in the new studies was collected at longer follow-up periods (up to one year after symptom onset) and in larger samples of participants. More RCTs and analytical observational studies are available, but the methodological quality of recently published studies is low. There is a need for good quality intervention efficacy and effectiveness studies to complement the rapidly expanding evidence from observational studies.


Subject(s)
COVID-19 , Bias , Cohort Studies , Cross-Sectional Studies , Humans , SARS-CoV-2
SELECTION OF CITATIONS
SEARCH DETAIL