Your browser doesn't support javascript.
Show: 20 | 50 | 100
Results 1 - 20 de 1.160
Filter
3.
Sci Rep ; 13(1): 8763, 2023 05 30.
Article in English | MEDLINE | ID: covidwho-20240051

ABSTRACT

As of January 2021, Australia had effectively controlled local transmission of COVID-19 despite a steady influx of imported cases and several local, but contained, outbreaks in 2020. Throughout 2020, state and territory public health responses were informed by weekly situational reports that included an ensemble forecast of daily COVID-19 cases for each jurisdiction. We present here an analysis of one forecasting model included in this ensemble across the variety of scenarios experienced by each jurisdiction from May to October 2020. We examine how successfully the forecasts characterised future case incidence, subject to variations in data timeliness and completeness, showcase how we adapted these forecasts to support decisions of public health priority in rapidly-evolving situations, evaluate the impact of key model features on forecast skill, and demonstrate how to assess forecast skill in real-time before the ground truth is known. Conditioning the model on the most recent, but incomplete, data improved the forecast skill, emphasising the importance of developing strong quantitative models of surveillance system characteristics, such as ascertainment delay distributions. Forecast skill was highest when there were at least 10 reported cases per day, the circumstances in which authorities were most in need of forecasts to aid in planning and response.


Subject(s)
COVID-19 , Pandemics , Humans , COVID-19/epidemiology , Disease Outbreaks , Public Health , Incidence , Forecasting
4.
PLoS One ; 18(6): e0286643, 2023.
Article in English | MEDLINE | ID: covidwho-20234676

ABSTRACT

The prediction of the number of infected and dead due to COVID-19 has challenged scientists and government bodies, prompting them to formulate public policies to control the virus' spread and public health emergency worldwide. In this sense, we propose a hybrid method that combines the SIRD mathematical model, whose parameters are estimated via Bayesian inference with a seasonal ARIMA model. Our approach considers that notifications of both, infections and deaths are realizations of a time series process, so that components such as non-stationarity, trend, autocorrelation and/or stochastic seasonal patterns, among others, must be taken into account in the fitting of any mathematical model. The method is applied to data from two Colombian cities, and as hypothesized, the prediction outperforms the obtained with the fit of only the SIRD model. In addition, a simulation study is presented to assess the quality of the estimators of SIRD model in the inverse problem solution.


Subject(s)
COVID-19 , Humans , Bayes Theorem , Colombia/epidemiology , Forecasting , Models, Theoretical
6.
PLoS One ; 18(3): e0283452, 2023.
Article in English | MEDLINE | ID: covidwho-2328116

ABSTRACT

In this study, we attempt to anticipate annual rice production in Bangladesh (1961-2020) using both the Autoregressive Integrated Moving Average (ARIMA) and the eXtreme Gradient Boosting (XGBoost) methods and compare their respective performances. On the basis of the lowest Corrected Akaike Information Criteria (AICc) values, a significant ARIMA (0, 1, 1) model with drift was chosen based on the findings. The drift parameter value shows that the production of rice positively trends upward. Thus, the ARIMA (0, 1, 1) model with drift was found to be significant. On the other hand, the XGBoost model for time series data was developed by changing the tunning parameters frequently with the greatest result. The four prominent error measures, such as mean absolute error (MAE), mean percentage error (MPE), root mean square error (RMSE), and mean absolute percentage error (MAPE), were used to assess the predictive performance of each model. We found that the error measures of the XGBoost model in the test set were comparatively lower than those of the ARIMA model. Comparatively, the MAPE value of the test set of the XGBoost model (5.38%) was lower than that of the ARIMA model (7.23%), indicating that XGBoost performs better than ARIMA at predicting the annual rice production in Bangladesh. Hence, the XGBoost model performs better than the ARIMA model in predicting the annual rice production in Bangladesh. Therefore, based on the better performance, the study forecasted the annual rice production for the next 10 years using the XGBoost model. According to our predictions, the annual rice production in Bangladesh will vary from 57,850,318 tons in 2021 to 82,256,944 tons in 2030. The forecast indicated that the amount of rice produced annually in Bangladesh will increase in the years to come.


Subject(s)
Oryza , Bangladesh , Neural Networks, Computer , Incidence , Forecasting , Machine Learning , Models, Statistical
7.
Infect Dis Poverty ; 12(1): 42, 2023 Apr 21.
Article in English | MEDLINE | ID: covidwho-2327417

ABSTRACT

BACKGROUND: Global connectivity and environmental change pose continuous threats to dengue invasions from worldwide to China. However, the intrinsic relationship on introduction and outbreak risks of dengue driven by the landscape features are still unknown. This study aimed to map the patterns on source-sink relation of dengue cases and assess the driving forces for dengue invasions in China. METHODS: We identified the local and imported cases (2006-2020) and assembled the datasets on environmental conditions. The vector auto-regression model was applied to detect the cross-relations of source-sink patterns. We selected the major environmental drivers via the Boruta algorithm to assess the driving forces in dengue outbreak dynamics by applying generalized additive models. We reconstructed the internal connections among imported cases, local cases, and external environmental drivers using the structural equation modeling. RESULTS: From 2006 to 2020, 81,652 local dengue cases and 12,701 imported dengue cases in China were reported. The hotspots of dengue introductions and outbreaks were in southeast and southwest China, originating from South and Southeast Asia. Oversea-imported dengue cases, as the Granger-cause, were the initial driver of the dengue dynamic; the suitable local bio-socioecological environment is the fundamental factor for dengue epidemics. The Bio8 [odds ratio (OR) = 2.11, 95% confidence interval (CI): 1.67-2.68], Bio9 (OR = 291.62, 95% CI: 125.63-676.89), Bio15 (OR = 4.15, 95% CI: 3.30-5.24), normalized difference vegetation index in March (OR = 1.27, 95% CI: 1.06-1.51) and July (OR = 1.04, 95% CI: 1.00-1.07), and the imported cases are the major drivers of dengue local transmissions (OR = 4.79, 95% CI: 4.34-5.28). The intermediary effect of an index on population and economic development to local cases via the path of imported cases was detected in the dengue dynamic system. CONCLUSIONS: Dengue outbreaks in China are triggered by introductions of imported cases and boosted by landscape features and connectivity. Our research will contribute to developing nature-based solutions for dengue surveillance, mitigation, and control from a socio-ecological perspective based on invasion ecology theories to control and prevent future dengue invasion and localization.


Subject(s)
Dengue , Epidemics , Humans , Dengue/epidemiology , Disease Outbreaks/prevention & control , China/epidemiology , Forecasting
8.
Stud Health Technol Inform ; 302: 861-865, 2023 May 18.
Article in English | MEDLINE | ID: covidwho-2327217

ABSTRACT

BACKGROUND: Emerging Infectious Diseases (EID) are a significant threat to population health globally. We aimed to examine the relationship between internet search engine queries and social media data on COVID-19 and determine if they can predict COVID-19 cases in Canada. METHODS: We analyzed Google Trends (GT) and Twitter data from 1/1/2020 to 3/31/2020 in Canada and used various signal-processing techniques to remove noise from the data. Data on COVID-19 cases was obtained from the COVID-19 Canada Open Data Working Group. We conducted time-lagged cross-correlation analyses and developed the long short-term memory model for forecasting daily COVID-19 cases. RESULTS: Among symptom keywords, "cough," "runny nose," and "anosmia" were strong signals with high cross-correlation coefficients >0.8 ( rCough = 0.825, t - 9; rRunnyNose = 0.816, t - 11; rAnosmia = 0.812, t - 3 ), showing that searching for "cough," "runny nose," and "anosmia" on GT correlated with the incidence of COVID-19 and peaked 9, 11, and 3 days earlier than the incidence peak, respectively. For symptoms- and COVID-related Tweet counts, the cross-correlations of Tweet signals and daily cases were rTweetSymptoms = 0.868, t - 11 and tTweetCOVID = 0.840, t - 10, respectively. The LSTM forecasting model achieved the best performance (MSE = 124.78, R2 = 0.88, adjusted R2 = 0.87) using GT signals with cross-correlation coefficients >0.75. Combining GT and Tweet signals did not improve the model performance. CONCLUSION: Internet search engine queries and social media data can be used as early warning signals for creating a real-time surveillance system for COVID-19 forecasting, but challenges remain in modelling.


Subject(s)
COVID-19 , Communicable Diseases, Emerging , Social Media , Humans , COVID-19/epidemiology , Communicable Diseases, Emerging/diagnosis , Communicable Diseases, Emerging/epidemiology , Cough , Search Engine , Internet , Forecasting
9.
BMC Med Educ ; 23(1): 288, 2023 Apr 27.
Article in English | MEDLINE | ID: covidwho-2325553

ABSTRACT

BACKGROUND: Early- and mid-career academics in medicine, dentistry and health sciences are integral to research, education and advancement of clinical professions, yet experience significant illbeing, high attrition and limited advancement opportunities. OBJECTIVES: Identify and synthesise published research investigating challenges and opportunities related to diversity and inclusion, as experienced by early and mid-career academics employed in medicine, dentistry and health sciences disciplines. DESIGN: Rapid review. DATA SOURCES: OVID Medline, Embase, APA PsycInfo, CINAHL and Scopus. METHODS: We systematically searched for peer reviewed published articles within the last five years, investigating challenges and opportunities related to diversity and inclusion, as experienced by early and mid-career academics employed in medicine, dentistry and health sciences. We screened and appraised articles, then extracted and synthesised data. RESULTS: Database searches identified 1162 articles, 11 met inclusion criteria. Studies varied in quality, primarily reporting concepts encompassed by professional identity. There were limited findings relating to social identity, with sexual orientation and disability being a particularly notable absence, and few findings relating to inclusion. Job insecurity, limited opportunities for advancement or professional development, and a sense of being undervalued in the workplace were evident for these academics. CONCLUSIONS: Our review identified overlap between academic models of wellbeing and key opportunities to foster inclusion. Challenges to professional identity such as job insecurity can contribute to development of illbeing. Future interventions to improve wellbeing in academia for early- and mid-career academics in these fields should consider addressing their social and professional identity, and foster their inclusion within the academic community. REGISTRATION: Open Science Framework ( https://doi.org/10.17605/OSF.IO/SA4HX ).


Subject(s)
Cultural Diversity , Workplace , Humans , Female , Male , Forecasting , Dentistry
10.
BMJ Open ; 13(5): e071241, 2023 05 05.
Article in English | MEDLINE | ID: covidwho-2316043

ABSTRACT

OBJECTIVES: The quest to measure and improve diagnosis has proven challenging; new approaches are needed to better understand and measure key elements of the diagnostic process in clinical encounters. The aim of this study was to develop a tool assessing key elements of the diagnostic assessment process and apply it to a series of diagnostic encounters examining clinical notes and encounters' recorded transcripts. Additionally, we aimed to correlate and contextualise these findings with measures of encounter time and physician burnout. DESIGN: We audio-recorded encounters, reviewed their transcripts and associated them with their clinical notes and findings were correlated with concurrent Mini Z Worklife measures and physician burnout. SETTING: Three primary urgent-care settings. PARTICIPANTS: We conducted in-depth evaluations of 28 clinical encounters delivered by seven physicians. RESULTS: Comparing encounter transcripts with clinical notes, in 24 of 28 (86%) there was high note/transcript concordance for the diagnostic elements on our tool. Reliably included elements were red flags (92% of notes/encounters), aetiologies (88%), likelihood/uncertainties (71%) and follow-up contingencies (71%), whereas psychosocial/contextual information (35%) and mentioning common pitfalls (7%) were often missing. In 22% of encounters, follow-up contingencies were in the note, but absent from the recorded encounter. There was a trend for higher burnout scores being associated with physicians less likely to address key diagnosis items, such as psychosocial history/context. CONCLUSIONS: A new tool shows promise as a means of assessing key elements of diagnostic quality in clinical encounters. Work conditions and physician reactions appear to correlate with diagnostic behaviours. Future research should continue to assess relationships between time pressure and diagnostic quality.


Subject(s)
Physicians , Working Conditions , Humans , Prospective Studies , Forecasting , Primary Health Care
11.
Health Psychol ; 42(5): 335-342, 2023 May.
Article in English | MEDLINE | ID: covidwho-2320457

ABSTRACT

OBJECTIVE: The term "long-COVID" refers to the persistence of neurological symptoms after being ill with COVID-19 (e.g., headaches, fatigue, and attentional impairment). Providing information about long-COVID (i.e., "diagnosis threat") increased subjective cognitive complaints among recovered COVID-19 patients compared with those exposed to neutral information (Winter & Braw, 2022). Notably, this effect was particularly prominent among more suggestible participants. Our aim in the current study was to validate these initial findings and to explore the impact of additional variables (e.g., suggestibility). METHOD: Recovered patients (n = 270) and controls (n = 290) reported daily cognitive failures after being randomly assigned to either a diagnosis threat (exposure to an article providing information regarding long-COVID) or a control condition. RESULTS: Recovered patients, but not controls, reported more cognitive failures in the diagnosis threat condition compared with the control condition. Diagnosis threat added significantly to the prediction of cognitive complaints based on relevant demographic variables and suggestibility. Diagnosis threat and suggestibility interacted (i.e., suggestible individuals were particularly vulnerable to the impact of a diagnosis threat). CONCLUSIONS: Diagnosis threat may contribute to the persistence of complaints regarding cognitive impairment among recovered COVID-19 patients. Suggestibility may be an underlying mechanism that increases the impact of diagnosis threat. Other factors, such as vaccination status, may be at play though we are only at the initial stages of research concerning their impact. These may be the focus of future research, aiding in identifying risk factors for experiencing COVID-19 symptoms past the resolution of its acute phase. (PsycInfo Database Record (c) 2023 APA, all rights reserved).


Subject(s)
COVID-19 , Cognitive Dysfunction , Humans , Cognitive Dysfunction/diagnosis , Cognitive Dysfunction/etiology , Forecasting , Cognition , Fatigue/diagnosis , Fatigue/epidemiology , Fatigue/etiology , COVID-19 Testing
12.
PLoS One ; 18(5): e0285407, 2023.
Article in English | MEDLINE | ID: covidwho-2320283

ABSTRACT

Improving forecasting particularly time series forecasting accuracy, efficiency and precisely become crucial for the authorities to forecast, monitor, and prevent the COVID-19 cases so that its spread can be controlled more effectively. However, the results obtained from prediction models are inaccurate, imprecise as well as inefficient due to linear and non-linear patterns exist in the data set, respectively. Therefore, to produce more accurate and efficient COVID-19 prediction value that is closer to the true COVID-19 value, a hybrid approach has been implemented. Thus, aims of this study is (1) to propose a hybrid ARIMA-SVM model to produce better forecasting results. (2) to investigate in terms of the performance of the proposed models and percentage improvement against ARIMA and SVM models. statistical measurements such as MSE, RMSE, MAE, and MAPE then conducted to verify that the proposed models are better than ARIMA and SVM models. Empirical results with three real datasets of well-known cases of COVID-19 in Malaysia show that, compared to the ARIMA and SVM models, the proposed model generates the smallest MSE, RMSE, MAE and MAPE values for the training and testing datasets, means that the predicted value from the proposed model is closer to the actual value. These results prove that the proposed model can generate estimated values more accurately and efficiently. As compared to ARIMA and SVM, our proposed models perform much better in terms of error reduction percentages for all datasets. This is demonstrated by the maximum scores of 73.12%, 74.6%, 90.38%, and 68.99% in the MAE, MAPE, MSE, and RMSE, respectively. Therefore, the proposed model can be the best and effective way to improve prediction performance with a higher level of accuracy and efficiency in predicting cases of COVID-19.


Subject(s)
COVID-19 , Humans , COVID-19/epidemiology , Models, Statistical , Pandemics , Forecasting , Malaysia
13.
J Am Med Inform Assoc ; 29(12): 2089-2095, 2022 11 14.
Article in English | MEDLINE | ID: covidwho-2319255

ABSTRACT

The coronavirus disease 2019 (COVID-19) pandemic has caused millions of deaths around the world and revealed the need for data-driven models of pandemic spread. Accurate pandemic caseload forecasting allows informed policy decisions on the adoption of non-pharmaceutical interventions (NPIs) to reduce disease transmission. Using COVID-19 as an example, we present Pandemic conditional Ordinary Differential Equation (PAN-cODE), a deep learning method to forecast daily increases in pandemic infections and deaths. By using a deep conditional latent variable model, PAN-cODE can generate alternative caseload trajectories based on alternate adoptions of NPIs, allowing stakeholders to make policy decisions in an informed manner. PAN-cODE also allows caseload estimation for regions that are unseen during model training. We demonstrate that, despite using less detailed data and having fully automated training, PAN-cODE's performance is comparable to state-of-the-art methods on 4-week-ahead and 6-week-ahead forecasting. Finally, we highlight the ability of PAN-cODE to generate realistic alternative outcome trajectories on select US regions.


Subject(s)
COVID-19 , Pandemics , Humans , Forecasting , Models, Theoretical
14.
Rev Esp Quimioter ; 35(5): 444-454, 2022 Oct.
Article in English | MEDLINE | ID: covidwho-2308158

ABSTRACT

A "Pandemic/Disaster Law" is needed to condense and organize the current dispersed and multiple legislation. The State must exercise a single power and command appropriate to each situation, with national validity. The production of plans for the use of land and real estate as potential centers for health care, shelter or refuge is recommended. There should be specific disaster plans at least for Primary Health Care, Hospitals and Socio-sanitary Centers. The guarantee of the maintenance of communication and supply routes is essential, as well as the guarantee of the autochthonous production of basic goods. The pandemic has highlighted the need to redefine the training plans for physicians who, in their different specialties, have to undertake reforms that allow a more versatile and transversal training. National research must have plans to be able to respond quickly to questions posed by the various crises, using all the nation's resources and in particular, all the data and capabilities of the health sector. Contingency plans must consider ethical aspects, and meet the needs of patients and families with a humanized approach. In circumstances of catastrophe, conflicts increase and require a bioethical response that allows the best decisions to be made, with the utmost respect for people's values. Rapid, efficient and truthful communication systems must be contained in a special project for this sector in critic circumstances. Finally, we believe that the creation of National Coordination Centers for major disasters and Public Health can contribute to better face the crises of the future.


Subject(s)
COVID-19 , Disasters , Forecasting , Humans , Pandemics , Public Health
15.
Soc Sci Med ; 326: 115950, 2023 06.
Article in English | MEDLINE | ID: covidwho-2307108

ABSTRACT

Life expectancy in the United States is decreasing. Health disparities are widening. Growing evidence for and integration of social and structural determinants into theory and practice has not yet improved outcomes. The COVID-19 pandemic reinforced the fact. In this paper, we argue that the biomedical model and its underlying scientific paradigm of causal determinism, which currently dominate population health, cannot meet population health needs. While criticism of the biomedical model is not new, this paper advances the field by going beyond criticism to recognize the need for a paradigm shift. In the first half of the paper, we present a critical analysis of the biomedical model and the paradigm of causal determinism. In the second half, we outline the agentic paradigm and present a structural model of health based on generalizable, group-level processes. We use the experience of the COVID-19 pandemic to illustrate the practical applications of our model. It will be important for future work to investigate the empirical and pragmatic applications of our structural model of population health.


Subject(s)
COVID-19 , Pandemics , Humans , United States , COVID-19/epidemiology , Life Expectancy , Forecasting
16.
Lancet ; 401(10385): 1340, 2023 04 22.
Article in English | MEDLINE | ID: covidwho-2293745
17.
Clin Oncol (R Coll Radiol) ; 35(5): 314-317, 2023 05.
Article in English | MEDLINE | ID: covidwho-2293173
18.
BMC Public Health ; 23(1): 782, 2023 04 28.
Article in English | MEDLINE | ID: covidwho-2305654

ABSTRACT

BACKGROUND: The COVID-19 pandemic has highlighted the role of infectious disease forecasting in informing public policy. However, significant barriers remain for effectively linking infectious disease forecasts to public health decision making, including a lack of model validation. Forecasting model performance and accuracy should be evaluated retrospectively to understand under which conditions models were reliable and could be improved in the future. METHODS: Using archived forecasts from the California Department of Public Health's California COVID Assessment Tool ( https://calcat.covid19.ca.gov/cacovidmodels/ ), we compared how well different forecasting models predicted COVID-19 hospitalization census across California counties and regions during periods of Alpha, Delta, and Omicron variant predominance. RESULTS: Based on mean absolute error estimates, forecasting models had variable performance across counties and through time. When accounting for model availability across counties and dates, some individual models performed consistently better than the ensemble model, but model rankings still differed across counties. Local transmission trends, variant prevalence, and county population size were informative predictors for determining which model performed best for a given county based on a random forest classification analysis. Overall, the ensemble model performed worse in less populous counties, in part because of fewer model contributors in these locations. CONCLUSIONS: Ensemble model predictions could be improved by incorporating geographic heterogeneity in model coverage and performance. Consistency in model reporting and improved model validation can strengthen the role of infectious disease forecasting in real-time public health decision making.


Subject(s)
COVID-19 , Communicable Diseases , Humans , Pandemics , Retrospective Studies , COVID-19/epidemiology , SARS-CoV-2 , Communicable Diseases/epidemiology , California/epidemiology , Public Policy , Decision Making , Hospitalization , Forecasting
19.
Elife ; 122023 04 21.
Article in English | MEDLINE | ID: covidwho-2303644

ABSTRACT

Background: Short-term forecasts of infectious disease burden can contribute to situational awareness and aid capacity planning. Based on best practice in other fields and recent insights in infectious disease epidemiology, one can maximise the predictive performance of such forecasts if multiple models are combined into an ensemble. Here, we report on the performance of ensembles in predicting COVID-19 cases and deaths across Europe between 08 March 2021 and 07 March 2022. Methods: We used open-source tools to develop a public European COVID-19 Forecast Hub. We invited groups globally to contribute weekly forecasts for COVID-19 cases and deaths reported by a standardised source for 32 countries over the next 1-4 weeks. Teams submitted forecasts from March 2021 using standardised quantiles of the predictive distribution. Each week we created an ensemble forecast, where each predictive quantile was calculated as the equally-weighted average (initially the mean and then from 26th July the median) of all individual models' predictive quantiles. We measured the performance of each model using the relative Weighted Interval Score (WIS), comparing models' forecast accuracy relative to all other models. We retrospectively explored alternative methods for ensemble forecasts, including weighted averages based on models' past predictive performance. Results: Over 52 weeks, we collected forecasts from 48 unique models. We evaluated 29 models' forecast scores in comparison to the ensemble model. We found a weekly ensemble had a consistently strong performance across countries over time. Across all horizons and locations, the ensemble performed better on relative WIS than 83% of participating models' forecasts of incident cases (with a total N=886 predictions from 23 unique models), and 91% of participating models' forecasts of deaths (N=763 predictions from 20 models). Across a 1-4 week time horizon, ensemble performance declined with longer forecast periods when forecasting cases, but remained stable over 4 weeks for incident death forecasts. In every forecast across 32 countries, the ensemble outperformed most contributing models when forecasting either cases or deaths, frequently outperforming all of its individual component models. Among several choices of ensemble methods we found that the most influential and best choice was to use a median average of models instead of using the mean, regardless of methods of weighting component forecast models. Conclusions: Our results support the use of combining forecasts from individual models into an ensemble in order to improve predictive performance across epidemiological targets and populations during infectious disease epidemics. Our findings further suggest that median ensemble methods yield better predictive performance more than ones based on means. Our findings also highlight that forecast consumers should place more weight on incident death forecasts than incident case forecasts at forecast horizons greater than 2 weeks. Funding: AA, BH, BL, LWa, MMa, PP, SV funded by National Institutes of Health (NIH) Grant 1R01GM109718, NSF BIG DATA Grant IIS-1633028, NSF Grant No.: OAC-1916805, NSF Expeditions in Computing Grant CCF-1918656, CCF-1917819, NSF RAPID CNS-2028004, NSF RAPID OAC-2027541, US Centers for Disease Control and Prevention 75D30119C05935, a grant from Google, University of Virginia Strategic Investment Fund award number SIF160, Defense Threat Reduction Agency (DTRA) under Contract No. HDTRA1-19-D-0007, and respectively Virginia Dept of Health Grant VDH-21-501-0141, VDH-21-501-0143, VDH-21-501-0147, VDH-21-501-0145, VDH-21-501-0146, VDH-21-501-0142, VDH-21-501-0148. AF, AMa, GL funded by SMIGE - Modelli statistici inferenziali per governare l'epidemia, FISR 2020-Covid-19 I Fase, FISR2020IP-00156, Codice Progetto: PRJ-0695. AM, BK, FD, FR, JK, JN, JZ, KN, MG, MR, MS, RB funded by Ministry of Science and Higher Education of Poland with grant 28/WFSN/2021 to the University of Warsaw. BRe, CPe, JLAz funded by Ministerio de Sanidad/ISCIII. BT, PG funded by PERISCOPE European H2020 project, contract number 101016233. CP, DL, EA, MC, SA funded by European Commission - Directorate-General for Communications Networks, Content and Technology through the contract LC-01485746, and Ministerio de Ciencia, Innovacion y Universidades and FEDER, with the project PGC2018-095456-B-I00. DE., MGu funded by Spanish Ministry of Health / REACT-UE (FEDER). DO, GF, IMi, LC funded by Laboratory Directed Research and Development program of Los Alamos National Laboratory (LANL) under project number 20200700ER. DS, ELR, GG, NGR, NW, YW funded by National Institutes of General Medical Sciences (R35GM119582; the content is solely the responsibility of the authors and does not necessarily represent the official views of NIGMS or the National Institutes of Health). FB, FP funded by InPresa, Lombardy Region, Italy. HG, KS funded by European Centre for Disease Prevention and Control. IV funded by Agencia de Qualitat i Avaluacio Sanitaries de Catalunya (AQuAS) through contract 2021-021OE. JDe, SMo, VP funded by Netzwerk Universitatsmedizin (NUM) project egePan (01KX2021). JPB, SH, TH funded by Federal Ministry of Education and Research (BMBF; grant 05M18SIA). KH, MSc, YKh funded by Project SaxoCOV, funded by the German Free State of Saxony. Presentation of data, model results and simulations also funded by the NFDI4Health Task Force COVID-19 (https://www.nfdi4health.de/task-force-covid-19-2) within the framework of a DFG-project (LO-342/17-1). LP, VE funded by Mathematical and Statistical modelling project (MUNI/A/1615/2020), Online platform for real-time monitoring, analysis and management of epidemic situations (MUNI/11/02202001/2020); VE also supported by RECETOX research infrastructure (Ministry of Education, Youth and Sports of the Czech Republic: LM2018121), the CETOCOEN EXCELLENCE (CZ.02.1.01/0.0/0.0/17-043/0009632), RECETOX RI project (CZ.02.1.01/0.0/0.0/16-013/0001761). NIB funded by Health Protection Research Unit (grant code NIHR200908). SAb, SF funded by Wellcome Trust (210758/Z/18/Z).


Subject(s)
COVID-19 , Communicable Diseases , Epidemics , Humans , COVID-19/diagnosis , COVID-19/epidemiology , Forecasting , Models, Statistical , Retrospective Studies
20.
Sci Rep ; 13(1): 6750, 2023 04 25.
Article in English | MEDLINE | ID: covidwho-2296393

ABSTRACT

In this work the applicability of an ensemble of population and machine learning models to predict the evolution of the COVID-19 pandemic in Spain is evaluated, relying solely on public datasets. Firstly, using only incidence data, we trained machine learning models and adjusted classical ODE-based population models, especially suited to capture long term trends. As a novel approach, we then made an ensemble of these two families of models in order to obtain a more robust and accurate prediction. We then proceed to improve machine learning models by adding more input features: vaccination, human mobility and weather conditions. However, these improvements did not translate to the overall ensemble, as the different model families had also different prediction patterns. Additionally, machine learning models degraded when new COVID variants appeared after training. We finally used Shapley Additive Explanation values to discern the relative importance of the different input features for the machine learning models' predictions. The conclusion of this work is that the ensemble of machine learning models and population models can be a promising alternative to SEIR-like compartmental models, especially given that the former do not need data from recovered patients, which are hard to collect and generally unavailable.


Subject(s)
COVID-19 , Pandemics , Humans , Spain/epidemiology , COVID-19/epidemiology , SARS-CoV-2 , Machine Learning , Forecasting
SELECTION OF CITATIONS
SEARCH DETAIL