Skip to content

Instantly share code, notes, and snippets.

@idan
Created March 27, 2020 01:09
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save idan/38d312c2c347b814d4ebf6a3e1d5c9ae to your computer and use it in GitHub Desktop.
Save idan/38d312c2c347b814d4ebf6a3e1d5c9ae to your computer and use it in GitHub Desktop.
Data for Modeling COVID Things

Model and Underlying Data

Data about Hospitals

Total number of beds by state draws upon aggregates from this report from ESRI based on Definitive Healthcare.

  • It includes a listing of total actual beds, ICU beds, and licensed beds. The licensed bed number can offer a clue as to how much capacity headroom a hospital has to grow in emergency.
  • It also includes bed utilization rates.
  • The AHA also has data about this, see KFF.org. That's beds per 1000 people, and it's not as granular. The source report is available for purchase.
  • cms.gov has a report as well, which is what we used until we found the DHC report. It has one erroneous row for a hospital in NJ with an unreasonable amount of beds, but is otherwise good in that it has county/zipcode level data.

Total proportion of ICU beds

The sccm.org data about ICU facilities draws upon two sources:

  • AHA 2015: 14.3% ratio of ICU beds to normal beds
  • HCRIS 2010: 16.2% ratio of ICU beds to normal beds

The DHC dataset includes state-specfic ICU beds, but the percentages there look more like 7-8% of total beds. We're using the DHC dataset since it breaks out actual numbers.

Data about actual COVID cases

  • ECDC worldwide cases, updated daily
  • COVID Tracking, updated daily. Draws on disparate sources and grades the quality of the sources. Notes indicate confidence in the underlying source.

Other websites

These are similar modeling apps with their own sources of data and underlying models/assumptions

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment