Forecasting COVID-19 Hospital Census: A Multivariate Time-Series Model Based on Local Infection Incidence.

BACKGROUND: COVID-19 has been one of the most serious global health crises in world history. During the pandemic, health care systems require accurate forecasts for key resources to guide preparation for patient surges. Forecasting the COVID-19 hospital census is among the most important planning decisions to ensure adequate staffing, number of beds, intensive care units, and vital equipment. OBJECTIVE: The goal of this study was to explore the potential utility of local COVID-19 infection incidence data in developing a forecasting model for the COVID-19 hospital census. METHODS: The study data comprised aggregated daily COVID-19 hospital census data across 11 Atrium Health hospitals plus a virtual hospital in the greater Charlotte metropolitan area of North Carolina, as well as the total daily infection incidence across the same region during the May 15 to December 5, 2020, period. Cross-correlations between hospital census and local infection incidence lagging up to 21 days were computed. A multivariate time-series framework, called the vector error correction model (VECM), was used to simultaneously incorporate both time series and account for their possible long-run relationship. Hypothesis tests and model diagnostics were performed to test for the long-run relationship and examine model goodness of fit. The 7-days-ahead forecast performance was measured by mean absolute percentage error (MAPE), with time-series cross-validation. The forecast performance was also compared with an autoregressive integrated moving average (ARIMA) model in the same cross-validation time frame. Based on different scenarios of the pandemic, the fitted model was leveraged to produce 60-days-ahead forecasts. RESULTS: The cross-correlations were uniformly high, falling between 0.7 and 0.8. There was sufficient evidence that the two time series have a stable long-run relationship at the .01 significance level. The model had very good fit to the data. The out-of-sample MAPE had a median of 5.9% and a 95th percentile of 13.4%. In comparison, the MAPE of the ARIMA had a median of 6.6% and a 95th percentile of 14.3%. Scenario-based 60-days-ahead forecasts exhibited concave trajectories with peaks lagging 2 to 3 weeks later than the peak infection incidence. In the worst-case scenario, the COVID-19 hospital census can reach a peak over 3 times greater than the peak observed during the second wave. CONCLUSIONS: When used in the VECM framework, the local COVID-19 infection incidence can be an effective leading indicator to predict the COVID-19 hospital census. The VECM model had a very good 7-days-ahead forecast performance and outperformed the traditional ARIMA model. Leveraging the relationship between the two time series, the model can produce realistic 60-days-ahead scenario-based projections, which can inform health care systems about the peak timing and volume of the hospital census for long-term planning purposes.

A predictive internet-based model for COVID-19 hospitalization census.

The COVID-19 pandemic has strained hospital resources and necessitated the need for predictive models to forecast patient care demands in order to allow for adequate staffing and resource allocation. Recently, other studies have looked at associations between Google Trends data and the number of COVID-19 cases. Expanding on this approach, we propose a vector error correction model (VECM) for the number of COVID-19 patients in a healthcare system (Census) that incorporates Google search term activity and healthcare chatbot scores. The VECM provided a good fit to Census and very good forecasting performance as assessed by hypothesis tests and mean absolute percentage prediction error. Although our study and model have limitations, we have conducted a broad and insightful search for candidate Internet variables and employed rigorous statistical methods. We have demonstrated the VECM can potentially be a valuable component to a COVID-19 surveillance program in a healthcare system.

Modeling COVID-19 latent prevalence to assess a public health intervention at a state and regional scale

BACKGROUND: Emergence of COVID-19 caught the world off-guard and unprepared, initiating a global pandemic. In the absence of evidence, individual communities had to take timely action to reduce the rate of disease spread and avoid overburdening their healthcare systems. Although a few predictive models have been published to guide these decisions, most have not taken into account spatial differences and have included assumptions that do not match the local realities. Access to reliable information that is adapted to local context is critical for policymakers to make informed decisions during a rapidly evolving pandemic. OBJECTIVE: The goal of this study was to develop an adapted susceptible-infected-removed (SIR) model to predict the trajectory of the COVID-19 pandemic in North Carolina (NC) and the Charlotte metropolitan region and to incorporate the effect of a public health intervention to reduce disease spread, while accounting for unique regional features and imperfect detection. METHODS: Using the software package R, three SIR models were fit to infection prevalence data from the state and the greater Charlotte region and then rigorously compared. One of these models (SIR-Int) accounted for a stay-at-home intervention and imperfect detection of COVID-19 cases. We computed longitudinal total estimates of the susceptible, infected, and removed compartments of both populations, along with other pandemic characteristics (e.g., basic reproduction number). RESULTS: Prior to March 26, disease spread was rapid at the pandemic onset with the Charlotte region doubling time of 2.56 days (95% CI: (2.11, 3.25)) and in NC 2.94 days (95% CI: (2.33, 4.00)). Subsequently, disease spread significantly slowed with doubling times increased in the Charlotte region to 4.70 days (95% CI: (3.77, 6.22)) and in NC to 4.01 days (95% CI: (3.43, 4.83)). Reflecting spatial differences, this deceleration favored the greater Charlotte region compared to NC as a whole. A comparison of the efficacy of intervention, defined as 1 - the hazard ratio of infection, gave 0.25 for NC and 0.43 for the Charlotte region. Also, early in the pandemic, the initial basic SIR model had good fit to the data;however, as the pandemic and local conditions evolved, the SIR-Int model emerged as the model with better fit. CONCLUSIONS: Using local data and continuous attention to model adaptation, our findings have enabled policymakers, public health officials and health systems to proactively plan capacity and evaluate the impact of a public health intervention. Our SIR-Int model for estimated latent prevalence was reasonably flexible, highly accurate, and demonstrated the efficacy of a stay-at-home order at both the state and regional level. Our results highlight the importance of incorporating local context into pandemic forecast modeling, as well as the need to remain vigilant and informed by the data as we enter into a critical period of the outbreak.