Forecasting ward-level bed requirements to aid pandemic resource planning: Lessons learned and future directions

Abstract

During the COVID-19 pandemic, there has been considerable research on how regional and country-level forecasting can be used to anticipate required hospital resources. We add to and build on this work by focusing on ward-level forecasting and planning tools for hospital staff during the pandemic. We present an assessment, validation, and deployment of a working prototype forecasting tool used within a modified Traffic Control Bundling (TCB) protocol for resource planning during the pandemic. We compare statistical and machine learning forecasting methods and their accuracy at one of the largest hospitals (Vancouver General Hospital) in Canada against a medium-sized hospital (St. Paul's Hospital) in Vancouver, Canada through the first three waves of the COVID-19 pandemic in the province of British Columbia. Our results confirm that traditional statistical and machine learning (ML) forecasting methods can provide valuable ward-level forecasting to aid in decision-making for pandemic resource planning. Using point forecasts with upper 95% prediction intervals, such forecasting methods would have provided better accuracy in anticipating required beds on COVID-19 hospital units than ward-level capacity decisions made by hospital staff. We have integrated our methodology into a publicly available online tool that operationalizes ward-level forecasting to aid with capacity planning decisions. Importantly, hospital staff can use this tool to translate forecasts into better patient care, less burnout, and improved planning for all hospital resources during pandemics.

Keywords: COVID-19; Forecasting; Machine learning; Pandemic resource planning; Traffic Control Bundling; Ward-level forecasting.

Fig. 1

Number of COVID-19 patients at St. Paul’s Hospital from April 2020 until June 2021. The red line represents the total number of COVID-19 positive patients assigned to the Red ward. The yellow line represents the total number of patients assigned to both the Red and Yellow wards. The black dotted line represents the set capacity level of the beds allocated to both the Red and Yellow ward

Fig. 2

Number of COVID-19 patients at Vancouver General Hospital from March 2020 until June 2021. The red line represents the total number of COVID-19 positive patients assigned to the Red ward. The yellow line represents the total number of patients assigned to both the Red and Yellow wards. The black dotted line represents the set capacity level of the beds in the Red ward only

Fig. 3

Positive correlation (r =  + 0.72 for Red and r =  + 0.54 for Red + Yellow) between the total number of patients in the COVID-19 wards at Vancouver General Hospital (VGH) (on the left vertical axis) and a 14-day lag of the COVID-19 positivity rate within the Vancouver Coastal Health Authority (VCHA) (on the right vertical axis)

Fig. 4

Positive correlation (r =  + 0.89 for Red and r =  + 0.90 for Red + Yellow) between the total number of patients in the COVID-19 wards at Vancouver General Hospital (VGH) (on the left vertical axis) and an 18-day lag of the daily reported confirmed cases within the Vancouver Coastal Health Authority (VCHA) (on the right vertical axis)

Fig. 5

Positive correlation (r =  + 0.88 and r =  + 0.87) between the total number of patients in the COVID-19 wards at Vancouver General Hospital (VGH) (on the left vertical axis) and a 12-day lag of the positivity rate within the Vancouver Coastal Health Authority (VCHA) (on the right vertical axis)

Fig. 6

Expanding window cross-validation procedure

Fig. 7

Cross-validation of the ARIMAX forecasting method at SPH and VGH during the greatest demand volatility of the second and third waves for each hospital. Top row: expanding window forecasts 1, 3, 5, and 7 days ahead in the combined COVID-19 red and yellow wards at SPH (left) and the combined COVID-19 red and yellow wards at VGH (right). Middle row: expanding window forecasts for 5 days ahead in the combined COVID-19 red and yellow wards at SPH with total planned capacities for the COVID-19 wards set by hospital staff (left) and the combined COVID-19 red and yellow wards at VGH (right), along with the 95% prediction interval. Bottom row: same as middle row except using expanding window forecasts for 7 days ahead

Fig. 8

SPH ward-level forecasts, time horizon = 7 days, 95% prediction intervals, with black squares representing actual values

Fig. 9

VGH ward-level forecasts, time horizon = 7 days, 95% prediction intervals, with black squares representing actual values

Fig. 10

VGH – Lag Analysis of the Daily Reported Cases in VCHA and total COVID-19 patients (Red and Yellow wards). VGH – Lag Analysis of the VCHA Positivity Rate and total COVID-19 patients (Red and Yellow wards). SPG – Lag Analysis of the VCHA Positivity Rate and total COVID-19 patients (Red and Yellow wards)

Fig. 11

Capacity planning module integrated within a ward-level forecasting tool