Case Fatality Ratio Estimates for the 2013–2016 West African Ebola Epidemic: Application of Boosted Regression Trees for Imputation

Forna, Alpha; Nouvellet, Pierre; Dorigatti, Ilaria; Donnelly, Christl A.

doi:10.1093/cid/ciz678

Cited by 24 publications

(20 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…18 22 23 36 In Liberia, the 2014-16 Ebola outbreak The perceived likelihood of contracting COVID-19 was relatively lower than reported during the Ebola outbreak when 83.7% felt at risk of contracting Ebola in low incidence counties. 14 The case fatality rate for COVID-19 in Liberia is around 4.7%, 9 much lower compared with that for EVD of around 79.2%, 37 which partly explains the lowrisk perception. Individuals who perceived themselves to be in poorer health, however, had a higher self-reported risk of acquiring COVID-19, among our survey sample.…”

Section: Discussionmentioning

confidence: 97%

COVID-19-related knowledge, attitudes and practices: a mixed-mode cross-sectional survey in Liberia

et al. 2021

View full text Add to dashboard Cite

ObjectivesTo examine the knowledge, attitudes and practices (KAP) of COVID-19 of rural and urban residents in Liberia to inform the development of local social and behaviour change communication strategies.DesignCross-sectional, mixed-mode (online and telephone) survey using non-probability sampling.SettingAll 15 counties in Liberia with a focus on Maryland County.ParticipantsFrom 28 May to 28 June 2020, data were collected from a total of 431 adults aged 18 years and older (telephone 288 (66.8%); online 143 (33.2%)) out of a total of 741 contacts.Main outcome measuresKAP scores. Frequencies and proportions were calculated, followed by univariate and multivariable analyses to examine the association between KAP scores and the sociodemographic variables.ResultsAround 69% of the online survey respondents were younger than 35 years of age, compared with 56% in the telephone interviews. The majority (87%) of online respondents had completed tertiary education, compared with 77% of the telephone respondents. Male participants, on average, achieved higher knowledge (52%) and attitude scores (72%), in contrast to females (49% and 67%, respectively). Radio (71%) was the most cited source for COVID-19 information, followed by social media (63%). After controlling for sociodemographic variables, adaptive regression modelling revealed that survey mode achieved 100% importance for predicting knowledge and practice levels with regard to COVID-19.ConclusionsThe survey population demonstrated moderate COVID-19 knowledge, with significant differences between survey mode and educational level. Correct knowledge of COVID-19 was associated with appropriate practices in Maryland County. Generalisation of survey findings must be drawn carefully owing to the limitations of the sampling methods. Yet, given the differences in knowledge gaps between survey modes, sex, education, occupation and place of residence, it is recommended that information is tailored to different audiences.

show abstract

Section: Discussionmentioning

confidence: 97%

COVID-19-related knowledge, attitudes and practices: a mixed-mode cross-sectional survey in Liberia

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Here, we report CFR estimates based on all reported cases (confirmed, probable, and suspected). Case classification (confirmed, probable, and suspected) was an important predictor (explained 10.1% of observed CFR variation) of EVD CFR for the West African epidemic and was accounted for in the BRT models used in this study (Forna et al, 2019).…”

Section: Data Sourcesmentioning

confidence: 90%

“…estimates that account for the missingness in survival outcomes unlike the crude estimates that do not account for missingness in survival outcomes) at national and district-level. Briefly, BRT models were calibrated to predict the survival outcomes using the 24 most important predictors of CFR (Forna et al, 2019). Hyperparameters (tree complexity (tc = 27), learning rate (lr = 0.001), bag faction (bf = 0.75) and data partitioning = 0.65) were optimised to maximise out of sample accuracy, allowing us generate district-level and quarterly CFR estimates, along with the relative uncertainty, using a nonparametric bootstrap approach (Forna et al, 2019).…”

Section: Spatiotemporal Analysismentioning

confidence: 99%

“…Briefly, BRT models were calibrated to predict the survival outcomes using the 24 most important predictors of CFR (Forna et al, 2019). Hyperparameters (tree complexity (tc = 27), learning rate (lr = 0.001), bag faction (bf = 0.75) and data partitioning = 0.65) were optimised to maximise out of sample accuracy, allowing us generate district-level and quarterly CFR estimates, along with the relative uncertainty, using a nonparametric bootstrap approach (Forna et al, 2019). Residuals from the adjusted BRT model were also calculated.…”

Section: Spatiotemporal Analysismentioning

confidence: 99%

“…In Forna et al, a methodological framework built around Boosted Regression Trees (BRT) with adjustments for imperfect model sensitivity and specificity, generated more representative CFR estimates for the West African Ebola epidemic than the crude CFR estimates by imputing survival outcomes for the 44% of the individual cases with missing survival outcomes (Forna et al, 2019). Here, we use a similar BRT imputation framework to investigate the spatiotemporal heterogeneities in CFR for the 2013-2016 West African Ebola epidemic.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Spatiotemporal variability in case fatality ratios for the 2013–2016 Ebola epidemic in West Africa

Forna

Dorigatti

Nouvellet

et al. 2020

International Journal of Infectious Diseases

Self Cite

View full text Add to dashboard Cite

Background: For the 2013-2016 Ebola epidemic in West Africa, the largest Ebola virus disease (EVD) epidemic to date, we aim to analyse the patient mix in detail to characterise key sources of spatiotemporal heterogeneity in the case fatality ratios (CFR). Methods: We applied a non-parametric Boosted Regression Trees (BRT) imputation approach for patients with missing survival outcomes and adjusted for model imperfection. Semivariogram analysis and kriging were used to investigate spatiotemporal heterogeneities. Results: CFR estimates varied significantly between districts and over time over the course of the epidemic. BRT modelling accounted for most of the spatiotemporal variation and interactions in CFR, but moderate spatial autocorrelation remained for distances up to approximately 90 km. Combining districtlevel CFR estimates and kriged district-level residuals provided the best linear unbiased predicted map of CFR accounting for the both explained and unexplained spatial variation. Temporal autocorrelation was not observed in the district-level residuals from the BRT estimates. Conclusions: This study provides new insight into the epidemiology of the 2013-2016 West African Ebola epidemic with a view of informing future public health contingency planning, resource allocation and impact assessment. The analytical framework developed in this analysis, coupled with key domain knowledge, could be deployed in real time to support the response to ongoing and future outbreaks.

show abstract

Implementation of stacking based ARIMA model for prediction of Covid-19 cases in India

Swaraj

Verma

Kaur

et al. 2021

Journal of Biomedical Informatics

View full text Add to dashboard Cite

Background Time-series forecasting has a critical role during pandemics as it provides essential information that can lead to abstaining from the spread of the disease. The novel coronavirus disease, COVID-19, is spreading rapidly all over the world. The countries with dense populations, in particular, such as India, await imminent risk in tackling the epidemic. Different forecasting models are being used to predict future cases of COVID-19. The predicament for most of them is that they are not able to capture both the linear and nonlinear features of the data solely. Methods We propose an ensemble model integrating an autoregressive integrated moving average model (ARIMA) and a nonlinear autoregressive neural network (NAR). ARIMA models are used to extract the linear correlations and the NAR neural network for modeling the residuals of ARIMA containing nonlinear components of the data. Comparison: Single ARIMA model, ARIMA-NAR model and few other existing models which have been applied on the COVID-19 data in different countries are compared based on performance evaluation parameters. Result The hybrid combination displayed significant reduction in RMSE(16.23%), MAE(37.89%) and MAPE (39.53%) values when compared with single ARIMA model for daily observed cases. Similar results with reduced error percentages were found for daily reported deaths and cases of recovery as well. RMSE value of our hybrid model was lesser in comparison to other models used for forecasting COVID-19 in different countries. Conclusion Results suggested the effectiveness of the new hybrid model over a single ARIMA model in capturing the linear as well as nonlinear patterns of the COVID-19 data.

show abstract

Case Fatality Ratio Estimates for the 2013–2016 West African Ebola Epidemic: Application of Boosted Regression Trees for Imputation

Cited by 24 publications

References 33 publications

COVID-19-related knowledge, attitudes and practices: a mixed-mode cross-sectional survey in Liberia

COVID-19-related knowledge, attitudes and practices: a mixed-mode cross-sectional survey in Liberia

Spatiotemporal variability in case fatality ratios for the 2013–2016 Ebola epidemic in West Africa

Implementation of stacking based ARIMA model for prediction of Covid-19 cases in India

Contact Info

Product

Resources

About