As the growing number of coronavirus cases continues to strain limited hospital resources, college campuses with largely vacated dormitories are in a unique position to provide additional beds and facilities to hospitals. Dorms can be repurposed as alternative sites to house recovering patients and offer temporary lodging to medical personnel, coordinated through hospital systems.
At the MGGG Redistricting Lab, we usually study gerrymandering. But with the same geodata and math modeling skills, we can build a simple model that connects hospitals to nearby college campuses.
Let’s take a look at Massachusetts.
Think we should adjust which colleges and hospitals are included? Provide feedback here.
This is a mathematical model and mapping tool intended to lend quantitative support to a qualitative point: available dorm capacity can have a significant impact as the nation mobilizes infrastructure for COVID-19 treatment. (See Nation at a Glance below for a visualization showing all 50 states.)
Our model flows people from hospitals to universities in a way that accounts for bed capacity and minimizes travel time. Here are the separate populations that can be transferred from hospitals to colleges:
By freeing beds, this alliance allows hospitals to convert more beds to critical care and provides support for front-line medical workers and other first responders.
We anticipate several uses for this model in keeping with its design.
Precise model output depends critically on the quality of input data and assumptions, and the need for speed and generality makes those very approximate in this case. Therefore the applications identified above are better supported by the model than relying on detailed assignment numbers of patients and medical staff.
| College | City | Number of dorm beds | Hospital staff assigned by the model | Patients assigned by the model | Utilization (%) |
|---|
| Hospitals included in the model | City | Number of beds |
|---|
The current model only includes flows of medical personnel and non-COVID patients. This is not a predictive model. We are working to sync up with other models of projected COVID demand on hospitals and will add recovering COVID patients at that time.
Share of dorm capacity that is available for hospital use is left as a parameter in the model, considering that the three types of populations will have to be housed separately and in suitable facilities. The mathematical formulation is described here, and data sourcing is found here. See our GitHub repo for full details of model structure and parameters.
We thank Michael Apkon of Tufts Medical Center, Assaad Sayah of Cambridge Health Alliance, Dawn Brantley of the Massachusetts Emergency Management Agency, and Ozlem Ergun of Northeastern University for helpful discussions about model uses and design.
MGGG Redistricting Lab is a research team of the Tisch College of Civic Life at Tufts University. We are grateful for the major support of the National Science Foundation through the Convergence Accelerator award OIA-1937095, Network Science of Census Data.
This project is a team effort of the MGGG Redistricting Lab with critical collaboration from Parker Rule, Olivia Walch, and Austin Buchanan.
Forty-nine out of fifty states have more dormitory beds than hospital beds, based on DHS datasets. Toggle the bubbles column to sort by the dorm-to-hospital ratio.
| State | Dorm-to-Hospital Bed Ratio | Total Dorm Capacity | Total Hospital Beds |
|---|