COVID-19 Surge Planning

Purpose

Predict imaging volume based upon community data regarding COVID-19.

Tag(s)

 

Panel

Non-Interpretive, COVID-19

Define-AI ID

20160004

Originator

COVID-19 Sub-Panel
Lead Morgan McBee

Panel Chair

Ben Wandtke

Panel Reviewers

Melissa Davis, Staphanie Jo, Summer Kaplan

License

Creative Commons 4.0
Status Public Comment

 

Clinical Implementation


Value Proposition

 

Rates of COVID-19 in the population are driving government decisions regarding social distancing and patient behavior. More accurate predictions of numbers of cases will allow for more optimal staffing adjustments for both volume increases and decreases. 

Narrative(s)

 

Staffing decision makers at a local imaging facility receive data from this algorithm that forecasts the volume of imaging studies by CPT code. This informs their decisions for setting the schedule. 

For example, a second wave of COVID-19 impacts a community served by a radiology group whose volume changed significantly during the first wave  that necessitated changes in staffing levels.  The algorithm maps the trajectory of infection relative to the first wave and predicts trends in imaging demand as the wave progresses and evolves. Staffing decision makers are then able to adjust policy and scheduling to accommodate for appropriate staffing levels for the second wave.

Based on predicted volumes by modality and CPT codes, staffing suggestions can be accommodated to the subspecialist level. If the algorithm predicts a surge in neuroimaging by 50% but a decrease in screening breast examinations by 70%, the facility is able to adjust their specialist staffing and potential for redeployment to accommodate for these changes.

 

Workflow Description

 

Algorithm processes input data including prior and current local number of COVID-19 cases, expected number of cases by whichever model is deemed appropriate, and the pre-pandemic imaging volume. Inputs such as local policy, current non-imaging patient volumes, and population behavior can also optionally be included to further refine the predictions. This information is synthesized to forecast volume across different imaging modalities and is sent to policy and staffing decision makers to make appropriate adjustments. 


Data Elements


Inputs

Data Element

Data Type

Description

Notes

COVID-19 Cases/Deaths by geographic region

Tabular

Number of reported cases each day by county or city

REF 1 

Expected COVID-19 Cases/Deaths by geographic region

Tabular

Expected number of cases each day by county or city by appropriate model


Facility volume by CPT

Tabular

Pre-pandemic imaging activity for the facility by CPT code.


First wave COVID+ cases

Tabular

Volume of COVID+ cases during the first wave. Patient recovery/outcome.


Public Transportation Data (optional)

Tabular

Report on public transit - whether it's operational and any changes in frequency or access


Local business policy data (optional)

Tabular

Report on the businesses that are open to the public and the trends in traffic across different types of businesses


Community mobility reports (optional)

Tabular

Report on movement trends over time by geography, across different categories of places such as retail and recreation, groceries and pharmacies, parks, transit stations, workplaces, and residential

REF 2 

Local mask requirement (optional)

Categorical

Are masks required for public activity in the local area?


Population health data (optional)

Tabular

health metrics for the local population

REF 3 

Similar facility data (optional)

Tabular

imaging demand data for similar facilities


Local hospitalization/ ICU numbers

Tabular

number of local hospitalization/ ICU usage with respiratory symptoms

REF 4 

Referral data (optional)

Tabular

Number of office visits, operations, and inpatient census of relevant services: e.g. pulmonologist and cardiovascular office visit numbers for chest/cardiac division, orthopedic office visit numbers and orthopedic surgery cases for musculoskeletal division.



Primary Outputs


Predicted Imaging Volume by CPT Code

Definition

The predicted imaging volume grouped by CPT code for a given period of time

Data Type

Tabular

Value Set

N/A

Units

Predicted studies/ CPT code for a selected date



Predicted Imaging Volume by Modality

Definition

The predicted imaging volume grouped by modality (e.g., radiography, fluoroscopy, CT, MRI, ultrasound, nuclear medicine, interventional) for a given period of time

Data Type

Tabular

Value Set

N/A

Units

Predicted studies/ modality for a selected date




Secondary Output


Predicted Number of COVID-19 Patients

Definition

The predicted number of COVID-19 patients for a given period of time

Data Type

Tabular

Value Set

N/A

Units

Number of Patients


Predicted Number of COVID-19 Patients by Modality

Definition

The predicted number of COVID-19 patients grouped by modality (e.g., radiography, fluoroscopy, CT, MRI, ultrasound, nuclear medicine, interventional) for a given period of time

Data Type

Tabular

Value Set

N/A

Units

Predicted number of patients/modality for a selected date


Future Development Ideas


The prediction of volumes can be leveraged in potential future use cases to prioritize certain patients, indications, or study types in order to safely and efficiently reduce the backlog of rescheduled examinations.

References