Incorporation of near-real-time hospital occupancy data to improve hospitalization forecast accuracy during the COVID-19 pandemic

Preiss, Alexander; Hadley, Emily; Jones, Kasey; Stoner, Marie; Kery, Caroline; Baumgartner, Peter; Bobashev, Georgiy; Tenenbaum, Jessica D.; Carter, Charles W.; Clément, K.; Rhea, Sarah

doi:10.1016/j.idm.2022.01.003

Cited by 7 publications

(7 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To evaluate the accuracy of forecasting models, an error evaluation method was widely used in previous literature and selected as an evaluation metric, MAPE. The smaller the MAPE value, the higher the forecasting accuracy [27], [28]…”

Section: E Measures For the Comparison Of Methods To Evaluate The For...mentioning

confidence: 99%

Predicting the Welfare Cost of Premature Deaths Based on Unsafe Sanitation Risk using SutteARIMA and Comparison with Neural Network Time Series and Holt-Winters

Annas

Ahmar

Hidayat

2023

JOIV : Int. J. Inform. Visualization

View full text Add to dashboard Cite

Unhealthy and unsafe sanitation will make it easier for various diseases to attack the body. In addition, unsafe sanitation will also affect a country's economy, including declining welfare, tourism losses, and environmental losses due to the loss of productive land. The research aimed to estimate the welfare cost of premature deaths based on unsafe sanitation risks using the SutteARIMA, Neural Network Time Series, and Holt-Winters. The study analyzed estimates and projections of the welfare cost of premature deaths based on the risks of unsafe sanitation of BRICS countries (Brazil, Russia, Indonesia, China, and South Africa). The data in this research used secondary data. Secondary time series data was taken from the Environment Database of the OECD. Stat. (Mortality and welfare cost from exposure to environmental risks). The data on the study was based on variables: welfare cost of premature deaths, % GDP equivalent, risk: unsafe sanitation, age: all, sex: both, unit: percentage, and data from 2005 to 2019. The three forecasting methods (SutteARIMA, Neural Network Time Series, and Holt-Winters) were juxtaposed in fitting data to see the forecasting methods' reliability and accuracy. The accuracy of forecasting results was compared based on MAPE and MSE values. The results of the research showed that the SutteARIMA and NNAR(1,1) methods were best used to predict the welfare cost of premature deaths in view of unsafe sanitation risks for BRICS countries.

show abstract

Section: E Measures For the Comparison Of Methods To Evaluate The For...mentioning

confidence: 99%

Predicting the Welfare Cost of Premature Deaths Based on Unsafe Sanitation Risk using SutteARIMA and Comparison with Neural Network Time Series and Holt-Winters

Annas

Ahmar

Hidayat

2023

JOIV : Int. J. Inform. Visualization

View full text Add to dashboard Cite

show abstract

“…Nonetheless, the course of a pandemic is a non-stationary situation, in which hospitalization parameters may vary between different waves and places, and evolve over time. Integration of near-real-time hospital occupancy data into the model can have a large impact on improving forecast accuracy [ 23 ].…”

Section: Introductionmentioning

confidence: 99%

Estimation of patient flow in hospitals using up-to-date data. Application to bed demand prediction during pandemic waves

et al. 2023

View full text Add to dashboard Cite

Hospital bed demand forecast is a first-order concern for public health action to avoid healthcare systems to be overwhelmed. Predictions are usually performed by estimating patients flow, that is, lengths of stay and branching probabilities. In most approaches in the literature, estimations rely on not updated published information or historical data. This may lead to unreliable estimates and biased forecasts during new or non-stationary situations. In this paper, we introduce a flexible adaptive procedure using only near-real-time information. Such method requires handling censored information from patients still in hospital. This approach allows the efficient estimation of the distributions of lengths of stay and probabilities used to represent the patient pathways. This is very relevant at the first stages of a pandemic, when there is much uncertainty and too few patients have completely observed pathways. Furthermore, the performance of the proposed method is assessed in an extensive simulation study in which the patient flow in a hospital during a pandemic wave is modelled. We further discuss the advantages and limitations of the method, as well as potential extensions.

show abstract

“…Models have used a variety of data inputs to predict hospitalizations due to COVID-19. Some of these data inputs include previous hospitalization data [9, 10], population mobility data [11, 9], realtime hospital occupancy data [12], trends in genomic variants [13], and internet search queries and chats from a public-facing Health Bot [14]. Though forecasts of hospitalizations are important, these forecasts were sometimes inaccurate at different stages of the pandemic due to characteristics of the data inputs.…”

Section: Introductionmentioning

confidence: 99%

“…In addition to these data sources, one important data input often used to forecast hospitalizations is the number of reported cases in a geographic location [16, 11, 12]. Epidemiological intuition suggests that there should be a lagged relationship between diagnosis of a COVID-19 case and subsequent hospitalization.…”

Section: Introductionmentioning

confidence: 99%

“…For example, the utility of mobility data as a predictor of transmission and subsequent hospitalizations declined as the pandemic evolved and human behavior and response changed over time [15]. The utility of these data sources in informing disease transmission varies with public health precautionary policies [11], the capacity of healthcare systems, and missing data [12].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Assessing the utility of COVID-19 case reports as a leading indicator for hospitalization forecasting in the United States

Reich

Wang

Burns

et al. 2023

Preprint

View full text Add to dashboard Cite

Identifying data streams that can consistently improve the accuracy of epidemiological forecasting models is challenging. Using models designed to predict daily state-level hospital admissions due to COVID-19 in California and Massachusetts, we investigated whether incorporating COVID-19 case data systematically improved forecast accuracy. Additionally, we considered whether using case data aggregated by date of test or by date of report from a surveillance system made a difference to the forecast accuracy. Evaluating forecast accuracy in a test period, after first having selected the best-performing methods in a validation period, we found that overall the difference in accuracy between approaches was small, especially at forecast horizons of less than two weeks. However, forecasts from models using cases aggregated by test date showed lower accuracy at longer horizons and at key moments in the pandemic, such as the peak of the Omicron wave in January 2022. Overall, these results highlight the challenge of finding a modeling approach that can generate accurate forecasts of outbreak trends both during periods of relative stability and during periods that show rapid growth or decay of transmission rates. While COVID-19 case counts seem to be a natural choice to help predict COVID-19 hospitalizations, in practice any benefits we observed were small and inconsistent.

show abstract

Incorporation of near-real-time hospital occupancy data to improve hospitalization forecast accuracy during the COVID-19 pandemic

Cited by 7 publications

References 19 publications

Predicting the Welfare Cost of Premature Deaths Based on Unsafe Sanitation Risk using SutteARIMA and Comparison with Neural Network Time Series and Holt-Winters

Predicting the Welfare Cost of Premature Deaths Based on Unsafe Sanitation Risk using SutteARIMA and Comparison with Neural Network Time Series and Holt-Winters

Estimation of patient flow in hospitals using up-to-date data. Application to bed demand prediction during pandemic waves

Assessing the utility of COVID-19 case reports as a leading indicator for hospitalization forecasting in the United States

Contact Info

Product

Resources

About