Your browser doesn't support javascript.
Show: 20 | 50 | 100
Results 1 - 20 de 829
Filter
Add filters

Document Type
Year range
1.
PLoS One ; 16(12): e0261858, 2021.
Article in English | MEDLINE | ID: covidwho-1635428

ABSTRACT

As a first line of defense to the COVID-19 pandemic in 2020, people reduced social contacts to avoid pathogen exposure. Using a panel of countries, this research suggests that this was amplified in societies characterized by high social support and future orientation. People reacted more strongly in dense environments; government orders had more effect in high power distance societies. Conversely, a focus on accomplishments was associated with lower changes. Understanding people's actual behaviors in response to health threats across societies is of great importance for epidemiology, public health, international business, and for the functioning of humanity as a whole.


Subject(s)
COVID-19/epidemiology , COVID-19/prevention & control , Culture , Models, Statistical , Pandemics/prevention & control , Physical Distancing , SARS-CoV-2 , COVID-19/virology , Cross-Cultural Comparison , Government Regulation , Humans , Longitudinal Studies , Public Health/methods , Quarantine/psychology
2.
Philos Trans A Math Phys Eng Sci ; 380(2214): 20210120, 2022 Jan 10.
Article in English | MEDLINE | ID: covidwho-1621739

ABSTRACT

We describe the population-based susceptible-exposed-infected-removed (SEIR) model developed by the Irish Epidemiological Modelling Advisory Group (IEMAG), which advises the Irish government on COVID-19 responses. The model assumes a time-varying effective contact rate (equivalently, a time-varying reproduction number) to model the effect of non-pharmaceutical interventions. A crucial technical challenge in applying such models is their accurate calibration to observed data, e.g. to the daily number of confirmed new cases, as the history of the disease strongly affects predictions of future scenarios. We demonstrate an approach based on inversion of the SEIR equations in conjunction with statistical modelling and spline-fitting of the data to produce a robust methodology for calibration of a wide class of models of this type. This article is part of the theme issue 'Data science approaches to infectious disease surveillance'.


Subject(s)
COVID-19 , Disease Susceptibility , Humans , Models, Statistical , SARS-CoV-2
3.
Int J Environ Res Public Health ; 19(1)2021 12 31.
Article in English | MEDLINE | ID: covidwho-1613755

ABSTRACT

This work aimed to apply the ARIMA model to predict the under-reporting of new Hansen's disease cases during the COVID-19 pandemic in Palmas, Tocantins, Brazil. This is an ecological time series study of Hansen's disease indicators in the city of Palmas between 2001 and 2020 using the autoregressive integrated moving averages method. Data from the Notifiable Injuries Information System and population estimates from the Brazilian Institute of Geography and Statistics were collected. A total of 7035 new reported cases of Hansen's disease were analyzed. The ARIMA model (4,0,3) presented the lowest values for the two tested information criteria and was the one that best fit the data, as AIC = 431.30 and BIC = 462.28, using a statistical significance level of 0.05 and showing the differences between the predicted values and those recorded in the notifications, indicating a large number of under-reporting of Hansen's disease new cases during the period from April to December 2020. The ARIMA model reported that 177% of new cases of Hansen's disease were not reported in Palmas during the period of the COVID-19 pandemic in 2020. This study shows the need for the municipal control program to undertake immediate actions in terms of actively searching for cases and reducing their hidden prevalence.


Subject(s)
COVID-19 , Leprosy , Brazil/epidemiology , Humans , Leprosy/epidemiology , Models, Statistical , Pandemics , SARS-CoV-2
4.
PLoS One ; 17(1): e0260836, 2022.
Article in English | MEDLINE | ID: covidwho-1613339

ABSTRACT

In the era of open data, Poisson and other count regression models are increasingly important. Still, conventional Poisson regression has remaining issues in terms of identifiability and computational efficiency. Especially, due to an identification problem, Poisson regression can be unstable for small samples with many zeros. Provided this, we develop a closed-form inference for an over-dispersed Poisson regression including Poisson additive mixed models. The approach is derived via mode-based log-Gaussian approximation. The resulting method is fast, practical, and free from the identification problem. Monte Carlo experiments demonstrate that the estimation error of the proposed method is a considerably smaller estimation error than the closed-form alternatives and as small as the usual Poisson regressions. For counts with many zeros, our approximation has better estimation accuracy than conventional Poisson regression. We obtained similar results in the case of Poisson additive mixed modeling considering spatial or group effects. The developed method was applied for analyzing COVID-19 data in Japan. This result suggests that influences of pedestrian density, age, and other factors on the number of cases change over periods.


Subject(s)
COVID-19/epidemiology , Humans , Japan/epidemiology , Markov Chains , Models, Statistical , Monte Carlo Method , Normal Distribution , Poisson Distribution , Regression Analysis , SARS-CoV-2/pathogenicity , Spatial Analysis , Spatio-Temporal Analysis
5.
PLoS Med ; 19(1): e1003870, 2022 01.
Article in English | MEDLINE | ID: covidwho-1608093

ABSTRACT

BACKGROUND: Excess mortality captures the total effect of the Coronavirus Disease 2019 (COVID-19) pandemic on mortality and is not affected by misspecification of cause of death. We aimed to describe how health and demographic factors were associated with excess mortality during, compared to before, the pandemic. METHODS AND FINDINGS: We analysed a time series dataset including 9,635,613 adults (≥40 years old) registered at United Kingdom general practices contributing to the Clinical Practice Research Datalink. We extracted weekly numbers of deaths and numbers at risk between March 2015 and July 2020, stratified by individual-level factors. Excess mortality during Wave 1 of the UK pandemic (5 March to 27 May 2020) compared to the prepandemic period was estimated using seasonally adjusted negative binomial regression models. Relative rates (RRs) of death for a range of factors were estimated before and during Wave 1 by including interaction terms. We found that all-cause mortality increased by 43% (95% CI 40% to 47%) during Wave 1 compared with prepandemic. Changes to the RR of death associated with most sociodemographic and clinical characteristics were small during Wave 1 compared with prepandemic. However, the mortality RR associated with dementia markedly increased (RR for dementia versus no dementia prepandemic: 3.5, 95% CI 3.4 to 3.5; RR during Wave 1: 5.1, 4.9 to 5.3); a similar pattern was seen for learning disabilities (RR prepandemic: 3.6, 3.4 to 3.5; during Wave 1: 4.8, 4.4 to 5.3), for black or South Asian ethnicity compared to white, and for London compared to other regions. Relative risks for morbidities were stable in multiple sensitivity analyses. However, a limitation of the study is that we cannot assume that the risks observed during Wave 1 would apply to other waves due to changes in population behaviour, virus transmission, and risk perception. CONCLUSIONS: The first wave of the UK COVID-19 pandemic appeared to amplify baseline mortality risk to approximately the same relative degree for most population subgroups. However, disproportionate increases in mortality were seen for those with dementia, learning disabilities, non-white ethnicity, or living in London.


Subject(s)
COVID-19/epidemiology , COVID-19/mortality , Mortality/trends , Adult , Aged , Female , Humans , Male , Middle Aged , Models, Statistical , Pandemics , Risk Factors , SARS-CoV-2/pathogenicity , Time Factors , United Kingdom/epidemiology
6.
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
7.
Lancet ; 398(10299): 522-534, 2021 08 07.
Article in English | MEDLINE | ID: covidwho-1592159

ABSTRACT

BACKGROUND: The COVID-19 pandemic and efforts to reduce SARS-CoV-2 transmission substantially affected health services worldwide. To better understand the impact of the pandemic on childhood routine immunisation, we estimated disruptions in vaccine coverage associated with the pandemic in 2020, globally and by Global Burden of Disease (GBD) super-region. METHODS: For this analysis we used a two-step hierarchical random spline modelling approach to estimate global and regional disruptions to routine immunisation using administrative data and reports from electronic immunisation systems, with mobility data as a model input. Paired with estimates of vaccine coverage expected in the absence of COVID-19, which were derived from vaccine coverage models from GBD 2020, Release 1 (GBD 2020 R1), we estimated the number of children who missed routinely delivered doses of the third-dose diphtheria-tetanus-pertussis (DTP3) vaccine and first-dose measles-containing vaccine (MCV1) in 2020. FINDINGS: Globally, in 2020, estimated vaccine coverage was 76·7% (95% uncertainty interval 74·3-78·6) for DTP3 and 78·9% (74·8-81·9) for MCV1, representing relative reductions of 7·7% (6·0-10·1) for DTP3 and 7·9% (5·2-11·7) for MCV1, compared to expected doses delivered in the absence of the COVID-19 pandemic. From January to December, 2020, we estimated that 30·0 million (27·6-33·1) children missed doses of DTP3 and 27·2 million (23·4-32·5) children missed MCV1 doses. Compared to expected gaps in coverage for eligible children in 2020, these estimates represented an additional 8·5 million (6·5-11·6) children not routinely vaccinated with DTP3 and an additional 8·9 million (5·7-13·7) children not routinely vaccinated with MCV1 attributable to the COVID-19 pandemic. Globally, monthly disruptions were highest in April, 2020, across all GBD super-regions, with 4·6 million (4·0-5·4) children missing doses of DTP3 and 4·4 million (3·7-5·2) children missing doses of MCV1. Every GBD super-region saw reductions in vaccine coverage in March and April, with the most severe annual impacts in north Africa and the Middle East, south Asia, and Latin America and the Caribbean. We estimated the lowest annual reductions in vaccine delivery in sub-Saharan Africa, where disruptions remained minimal throughout the year. For some super-regions, including southeast Asia, east Asia, and Oceania for both DTP3 and MCV1, the high-income super-region for DTP3, and south Asia for MCV1, estimates suggest that monthly doses were delivered at or above expected levels during the second half of 2020. INTERPRETATION: Routine immunisation services faced stark challenges in 2020, with the COVID-19 pandemic causing the most widespread and largest global disruption in recent history. Although the latest coverage trajectories point towards recovery in some regions, a combination of lagging catch-up immunisation services, continued SARS-CoV-2 transmission, and persistent gaps in vaccine coverage before the pandemic still left millions of children under-vaccinated or unvaccinated against preventable diseases at the end of 2020, and these gaps are likely to extend throughout 2021. Strengthening routine immunisation data systems and efforts to target resources and outreach will be essential to minimise the risk of vaccine-preventable disease outbreaks, reach children who missed routine vaccine doses during the pandemic, and accelerate progress towards higher and more equitable vaccination coverage over the next decade. FUNDING: Bill & Melinda Gates Foundation.


Subject(s)
COVID-19 , Diphtheria-Tetanus-Pertussis Vaccine , Measles Vaccine , Vaccination Coverage/statistics & numerical data , Child , Global Health , Humans , Models, Statistical
8.
PLoS One ; 16(12): e0261424, 2021.
Article in English | MEDLINE | ID: covidwho-1599330

ABSTRACT

The COVID-19 outbreak has caused two waves and spread to more than 90% of Canada's provinces since it was first reported more than a year ago. During the COVID-19 epidemic, Canadian provinces have implemented many Non-Pharmaceutical Interventions (NPIs). However, the spread of the COVID-19 epidemic continues due to the complex dynamics of human mobility. We develop a meta-population network model to study the transmission dynamics of COVID-19. The model takes into account the heterogeneity of mitigation strategies in different provinces of Canada, such as the timing of implementing NPIs, the human mobility in retail and recreation, grocery and pharmacy, parks, transit stations, workplaces, and residences due to work and recreation. To determine which activity is most closely related to the dynamics of COVID-19, we use the cross-correlation analysis to find that the positive correlation is the highest between the mobility data of parks and the weekly number of confirmed COVID-19 from February 15 to December 13, 2020. The average effective reproduction numbers in nine Canadian provinces are all greater than one during the time period, and NPIs have little impact on the dynamics of COVID-19 epidemics in Ontario and Saskatchewan. After November 20, 2020, the average infection probability in Alberta became the highest since the start of the COVID-19 epidemic in Canada. We also observe that human activities around residences do not contribute much to the spread of the COVID-19 epidemic. The simulation results indicate that social distancing and constricting human mobility is effective in mitigating COVID-19 transmission in Canada. Our findings can provide guidance for public health authorities in projecting the effectiveness of future NPIs.


Subject(s)
COVID-19/prevention & control , COVID-19/transmission , Epidemics/prevention & control , SARS-CoV-2 , Travel/statistics & numerical data , Basic Reproduction Number/statistics & numerical data , COVID-19/epidemiology , Canada/epidemiology , Humans , Incidence , Models, Statistical , Physical Distancing , Quarantine/methods
9.
PLoS One ; 16(12): e0260051, 2021.
Article in English | MEDLINE | ID: covidwho-1596447

ABSTRACT

OBJECTIVES: To model the risk of COVID-19 mortality in British care homes conditional on the community level risk. METHODS: A two stage modeling process ("doubly latent") which includes a Besag York Mollie model (BYM) and a Log Gaussian Cox Process. The BYM is adopted so as to estimate the community level risks. These are incorporated in the Log Gaussian Cox Process to estimate the impact of these risks on that in care homes. RESULTS: For an increase in the risk at the community level, the number of COVID-19 related deaths in the associated care home would be increased by exp (0.833), 2. This is based on a simulated dataset. In the context of COVID-19 related deaths, this study has illustrated the estimation of the risk to care homes in the presence of background community risk. This approach will be useful in facilitating the identification of the most vulnerable care homes and in predicting risk to new care homes. CONCLUSIONS: The modeling of two latent processes have been shown to be successfully facilitated by the use of the BYM and Log Gaussian Cox Process Models. Community COVID-19 risks impact on that of the care homes embedded in these communities.


Subject(s)
COVID-19/epidemiology , Nursing Homes/statistics & numerical data , Residence Characteristics , Geography , Humans , Models, Statistical , Risk Factors
10.
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
11.
CMAJ Open ; 9(4): E1223-E1231, 2021.
Article in English | MEDLINE | ID: covidwho-1593829

ABSTRACT

BACKGROUND: The COVID-19 pandemic has led to an increased demand for health care resources and, in some cases, shortage of medical equipment and staff. Our objective was to develop and validate a multivariable model to predict risk of hospitalization for patients infected with SARS-CoV-2. METHODS: We used routinely collected health records in a patient cohort to develop and validate our prediction model. This cohort included adult patients (age ≥ 18 yr) from Ontario, Canada, who tested positive for SARS-CoV-2 ribonucleic acid by polymerase chain reaction between Feb. 2 and Oct. 5, 2020, and were followed up through Nov. 5, 2020. Patients living in long-term care facilities were excluded, as they were all assumed to be at high risk of hospitalization for COVID-19. Risk of hospitalization within 30 days of diagnosis of SARS-CoV-2 infection was estimated via gradient-boosting decision trees, and variable importance examined via Shapley values. We built a gradient-boosting model using the Extreme Gradient Boosting (XGBoost) algorithm and compared its performance against 4 empirical rules commonly used for risk stratifications based on age and number of comorbidities. RESULTS: The cohort included 36 323 patients with 2583 hospitalizations (7.1%). Hospitalized patients had a higher median age (64 yr v. 43 yr), were more likely to be male (56.3% v. 47.3%) and had a higher median number of comorbidities (3, interquartile range [IQR] 2-6 v. 1, IQR 0-3) than nonhospitalized patients. Patients were split into development (n = 29 058, 80.0%) and held-out validation (n = 7265, 20.0%) cohorts. The gradient-boosting model achieved high discrimination (development cohort: area under the receiver operating characteristic curve across the 5 folds of 0.852; validation cohort: 0.8475) and strong calibration (slope = 1.01, intercept = -0.01). The patients who scored at the top 10% captured 47.4% of hospitalizations, and those who scored at the top 30% captured 80.6%. INTERPRETATION: We developed and validated an accurate risk stratification model using routinely collected health administrative data. We envision that modelling such risk stratification based on routinely collected health data could support management of COVID-19 on a population health level.


Subject(s)
COVID-19/epidemiology , Decision Trees , Hospitalization/statistics & numerical data , Risk Assessment , Adult , Aged , COVID-19/therapy , Female , Humans , Male , Middle Aged , Models, Statistical , Ontario/epidemiology , Risk Assessment/methods , Risk Factors
12.
Gac Sanit ; 35 Suppl 2: S604-S609, 2021.
Article in English | MEDLINE | ID: covidwho-1587737

ABSTRACT

OBJECTIVE: Global society pays huge economic toll and live loss due to COVID-19 (Coronavirus Disease 2019) pandemic. In order to have a better management of this pandemic, many institutions develop their own models to predict number of COVID-19 cases, hospitalizations and mortalities. These models, however, are shown to be unreliable and need to be revised on a daily basis. METHODS: Here, we develop a Bose-Einstein (BE)-based statistical model to predict daily COVID-19 cases up to 14 days in advance. This fat-tailed model is chosen based on three reasons. First, it contains a peak and decaying phase. Second, it also has both accelerated and decelerated phases which are similarly observed in an epidemic curve. Third, the shape of both the BE energy distribution and the epidemic curve is controlled by a set of parameters. The BE model daily predictions are then verified against simulated data and confirmed COVID-19 daily cases from two epidemic centres, i.e. New York and DKI Jakarta. RESULT: Over- predictions occur at the earlier stage of the epidemic for all data sets. Models parameters for both simulated and New York data converge to a certain value only at the latest stage of the epidemic progress. At this stage, model's skill is high for both simulated and New York data, i.e. the predictability is greater than 80% with decreasing RMSE. On the other hand, at that stage, the DKI's model's predictability is still fluctuating with increasing RMSE. CONCLUSION: This implies that New York could leave the stay-at-home order, but DKI Jakarta should continue its large-scale social restriction order. There remains a great challenge in predicting the full course of an epidemic using small data collected during the earlier phase of the epidemic.


Subject(s)
COVID-19 , Humans , Models, Statistical , New York/epidemiology , Pandemics , SARS-CoV-2
13.
J Med Internet Res ; 23(2): e26081, 2021 02 09.
Article in English | MEDLINE | ID: covidwho-1575190

ABSTRACT

BACKGROUND: The COVID-19 pandemic has had profound and differential impacts on metropolitan areas across the United States and around the world. Within the United States, metropolitan areas that were hit earliest with the pandemic and reacted with scientifically based health policy were able to contain the virus by late spring. For other areas that kept businesses open, the first wave in the United States hit in mid-summer. As the weather turns colder, universities resume classes, and people tire of lockdowns, a second wave is ascending in both metropolitan and rural areas. It becomes more obvious that additional SARS-CoV-2 surveillance is needed at the local level to track recent shifts in the pandemic, rates of increase, and persistence. OBJECTIVE: The goal of this study is to provide advanced surveillance metrics for COVID-19 transmission that account for speed, acceleration, jerk and persistence, and weekly shifts, to better understand and manage risk in metropolitan areas. Existing surveillance measures coupled with our dynamic metrics of transmission will inform health policy to control the COVID-19 pandemic until, and after, an effective vaccine is developed. Here, we provide values for novel indicators to measure COVID-19 transmission at the metropolitan area level. METHODS: Using a longitudinal trend analysis study design, we extracted 260 days of COVID-19 data from public health registries. We used an empirical difference equation to measure the daily number of cases in the 25 largest US metropolitan areas as a function of the prior number of cases and weekly shift variables based on a dynamic panel data model that was estimated using the generalized method of moments approach by implementing the Arellano-Bond estimator in R. RESULTS: Minneapolis and Chicago have the greatest average number of daily new positive results per standardized 100,000 population (which we refer to as speed). Extreme behavior in Minneapolis showed an increase in speed from 17 to 30 (67%) in 1 week. The jerk and acceleration calculated for these areas also showed extreme behavior. The dynamic panel data model shows that Minneapolis, Chicago, and Detroit have the largest persistence effects, meaning that new cases pertaining to a specific week are statistically attributable to new cases from the prior week. CONCLUSIONS: Three of the metropolitan areas with historically early and harsh winters have the highest persistence effects out of the top 25 most populous metropolitan areas in the United States at the beginning of their cold weather season. With these persistence effects, and with indoor activities becoming more popular as the weather gets colder, stringent COVID-19 regulations will be more important than ever to flatten the second wave of the pandemic. As colder weather grips more of the nation, southern metropolitan areas may also see large spikes in the number of cases.


Subject(s)
COVID-19/epidemiology , Communicable Disease Control , COVID-19/prevention & control , COVID-19/transmission , Health Policy , Humans , Longitudinal Studies , Models, Statistical , Pandemics , Public Health , Public Health Surveillance , Registries , SARS-CoV-2 , United States/epidemiology
14.
J Med Internet Res ; 23(2): e23458, 2021 02 26.
Article in English | MEDLINE | ID: covidwho-1574596

ABSTRACT

BACKGROUND: During a pandemic, it is important for clinicians to stratify patients and decide who receives limited medical resources. Machine learning models have been proposed to accurately predict COVID-19 disease severity. Previous studies have typically tested only one machine learning algorithm and limited performance evaluation to area under the curve analysis. To obtain the best results possible, it may be important to test different machine learning algorithms to find the best prediction model. OBJECTIVE: In this study, we aimed to use automated machine learning (autoML) to train various machine learning algorithms. We selected the model that best predicted patients' chances of surviving a SARS-CoV-2 infection. In addition, we identified which variables (ie, vital signs, biomarkers, comorbidities, etc) were the most influential in generating an accurate model. METHODS: Data were retrospectively collected from all patients who tested positive for COVID-19 at our institution between March 1 and July 3, 2020. We collected 48 variables from each patient within 36 hours before or after the index time (ie, real-time polymerase chain reaction positivity). Patients were followed for 30 days or until death. Patients' data were used to build 20 machine learning models with various algorithms via autoML. The performance of machine learning models was measured by analyzing the area under the precision-recall curve (AUPCR). Subsequently, we established model interpretability via Shapley additive explanation and partial dependence plots to identify and rank variables that drove model predictions. Afterward, we conducted dimensionality reduction to extract the 10 most influential variables. AutoML models were retrained by only using these 10 variables, and the output models were evaluated against the model that used 48 variables. RESULTS: Data from 4313 patients were used to develop the models. The best model that was generated by using autoML and 48 variables was the stacked ensemble model (AUPRC=0.807). The two best independent models were the gradient boost machine and extreme gradient boost models, which had an AUPRC of 0.803 and 0.793, respectively. The deep learning model (AUPRC=0.73) was substantially inferior to the other models. The 10 most influential variables for generating high-performing models were systolic and diastolic blood pressure, age, pulse oximetry level, blood urea nitrogen level, lactate dehydrogenase level, D-dimer level, troponin level, respiratory rate, and Charlson comorbidity score. After the autoML models were retrained with these 10 variables, the stacked ensemble model still had the best performance (AUPRC=0.791). CONCLUSIONS: We used autoML to develop high-performing models that predicted the survival of patients with COVID-19. In addition, we identified important variables that correlated with mortality. This is proof of concept that autoML is an efficient, effective, and informative method for generating machine learning-based clinical decision support tools.


Subject(s)
COVID-19/mortality , Machine Learning , COVID-19/virology , Female , Humans , Male , Middle Aged , Models, Statistical , Pandemics , Retrospective Studies , SARS-CoV-2/isolation & purification , Survival Analysis
15.
Proc Natl Acad Sci U S A ; 118(51)2021 12 21.
Article in English | MEDLINE | ID: covidwho-1569346

ABSTRACT

Short-term forecasts of traditional streams from public health reporting (such as cases, hospitalizations, and deaths) are a key input to public health decision-making during a pandemic. Since early 2020, our research group has worked with data partners to collect, curate, and make publicly available numerous real-time COVID-19 indicators, providing multiple views of pandemic activity in the United States. This paper studies the utility of five such indicators-derived from deidentified medical insurance claims, self-reported symptoms from online surveys, and COVID-related Google search activity-from a forecasting perspective. For each indicator, we ask whether its inclusion in an autoregressive (AR) model leads to improved predictive accuracy relative to the same model excluding it. Such an AR model, without external features, is already competitive with many top COVID-19 forecasting models in use today. Our analysis reveals that 1) inclusion of each of these five indicators improves on the overall predictive accuracy of the AR model; 2) predictive gains are in general most pronounced during times in which COVID cases are trending in "flat" or "down" directions; and 3) one indicator, based on Google searches, seems to be particularly helpful during "up" trends.


Subject(s)
COVID-19/epidemiology , Health Status Indicators , Models, Statistical , Epidemiologic Methods , Forecasting , Humans , Internet/statistics & numerical data , Surveys and Questionnaires , United States/epidemiology
16.
Comput Math Methods Med ; 2021: 2689000, 2021.
Article in English | MEDLINE | ID: covidwho-1566408

ABSTRACT

We have studied one of the most common distributions, namely, Lindley distribution, which is an important continuous mixed distribution with great ability to represent different systems. We studied this distribution with three parameters because of its high flexibility in modelling life data. The parameters were estimated by five different methods, namely, maximum likelihood estimation, ordinary least squares, weighted least squares, maximum product of spacing, and Cramér-von Mises. Simulation experiments were performed with different sample sizes and different parameter values. The different methods were compared on the generated data by mean square error and mean absolute error. In addition, we compared the methods for real data, which represent COVID-19 data in Iraq/Anbar Province.


Subject(s)
COVID-19/epidemiology , Public Health Informatics/methods , Algorithms , Computer Simulation , Humans , Iraq , Least-Squares Analysis , Likelihood Functions , Models, Statistical , Public Health Informatics/standards , SARS-CoV-2 , Statistics as Topic
17.
Comput Math Methods Med ; 2021: 4321131, 2021.
Article in English | MEDLINE | ID: covidwho-1553710

ABSTRACT

The COVID-19 pandemic has had a devastating effect on many people, creating severe anxiety, fear, and complicated feelings or emotions. After the initiation of vaccinations against coronavirus, people's feelings have become more diverse and complex. Our aim is to understand and unravel their sentiments in this research using deep learning techniques. Social media is currently the best way to express feelings and emotions, and with the help of Twitter, one can have a better idea of what is trending and going on in people's minds. Our motivation for this research was to understand the diverse sentiments of people regarding the vaccination process. In this research, the timeline of the collected tweets was from December 21 to July21. The tweets contained information about the most common vaccines available recently from across the world. The sentiments of people regarding vaccines of all sorts were assessed using the natural language processing (NLP) tool, Valence Aware Dictionary for sEntiment Reasoner (VADER). Initializing the polarities of the obtained sentiments into three groups (positive, negative, and neutral) helped us visualize the overall scenario; our findings included 33.96% positive, 17.55% negative, and 48.49% neutral responses. In addition, we included our analysis of the timeline of the tweets in this research, as sentiments fluctuated over time. A recurrent neural network- (RNN-) oriented architecture, including long short-term memory (LSTM) and bidirectional LSTM (Bi-LSTM), was used to assess the performance of the predictive models, with LSTM achieving an accuracy of 90.59% and Bi-LSTM achieving 90.83%. Other performance metrics such as precision,, F1-score, and a confusion matrix were also used to validate our models and findings more effectively. This study improves understanding of the public's opinion on COVID-19 vaccines and supports the aim of eradicating coronavirus from the world.


Subject(s)
COVID-19 Vaccines , COVID-19/prevention & control , Deep Learning , Social Media , Attitude , Attitude to Health , Databases, Factual , Humans , Language , Models, Statistical , Neural Networks, Computer , Public Opinion , Reproducibility of Results , Vaccination
18.
Gac Med Mex ; 157(3): 231-236, 2021.
Article in English | MEDLINE | ID: covidwho-1535078

ABSTRACT

INTRODUCTION: The scarcity of person-centered applications aimed at developing awareness on the risk posed by the COVID-19 pandemic, stimulates the exploration and creation of preventive tools that are accessible to the population. OBJECTIVE: To develop a predictive model that allows evaluating the risk of mortality in the event of SARS-CoV-2 virus infection. METHODS: Exploration of public data from 16,000 COVID-19-positive patients to generate an efficient discriminant model, evaluated with a score function and expressed by a self-rated preventive interest questionnaire. RESULTS: A useful linear function was obtained with a discriminant capacity of 0.845; internal validation with bootstrap and external validation, with 25 % of tested patients showing marginal differences. CONCLUSION: The predictive model with statistical support, based on 15 accessible questions, can become a structured prevention tool.


Subject(s)
COVID-19/prevention & control , Models, Statistical , Adolescent , Adult , Aged , COVID-19/mortality , Child , Child, Preschool , Discriminant Analysis , Female , Humans , Infant , Linear Models , Male , Middle Aged , Risk , Young Adult
19.
Zhonghua Liu Xing Bing Xue Za Zhi ; 42(3): 421-426, 2021 Mar 10.
Article in Chinese | MEDLINE | ID: covidwho-1534264

ABSTRACT

Objective: To compare the performances of different time series models in predicting COVID-19 in different countries. Methods: We collected the daily confirmed case numbers of COVID-19 in the USA, India, and Brazil from April 1 to September 30, 2020, and then constructed an autoregressive integrated moving average (ARIMA) model and a recurrent neural network (RNN) model, respectively. We applied the mean absolute percentage error (MAPE) and root mean square error (RMSE) to compare the performances of the two models in predicting the case numbers from September 21 to September 30, 2020. Results: For the ARIMA models applied in the USA, India, and Brazil, the MAPEs were 13.18%, 9.18%, and 17.30%, respectively, and the RMSEs were 6 542.32, 8 069.50, and 3 954.59, respectively. For the RNN models applied in the USA, India, and Brazil, the MAPEs were 15.27%, 7.23% and 26.02%, respectively, and the RMSEs were 6 877.71, 6 457.07, and 5 950.88, respectively. Conclusions: The performance of the prediction models varied with country. The ARIMA model had a better prediction performance for COVID-19 in the USA and Brazil, while the RNN model was more suitable in India.


Subject(s)
COVID-19 , Forecasting , Humans , Models, Statistical , Neural Networks, Computer , SARS-CoV-2
20.
BMC Public Health ; 21(1): 2132, 2021 11 20.
Article in English | MEDLINE | ID: covidwho-1526611

ABSTRACT

BACKGROUND: The global spread of COVID-19 has shown that reliable forecasting of public health related outcomes is important but lacking. METHODS: We report the results of the first large-scale, long-term experiment in crowd-forecasting of infectious-disease outbreaks, where a total of 562 volunteer participants competed over 15 months to make forecasts on 61 questions with a total of 217 possible answers regarding 19 diseases. RESULTS: Consistent with the "wisdom of crowds" phenomenon, we found that crowd forecasts aggregated using best-practice adaptive algorithms are well-calibrated, accurate, timely, and outperform all individual forecasters. CONCLUSIONS: Crowd forecasting efforts in public health may be a useful addition to traditional disease surveillance, modeling, and other approaches to evidence-based decision making for infectious disease outbreaks.


Subject(s)
COVID-19 , Disease Outbreaks , Forecasting , Humans , Intelligence , Models, Statistical , SARS-CoV-2
SELECTION OF CITATIONS
SEARCH DETAIL
...