11

Is there anywhere a database that includes case-level information on COVID-19 cases?

Such data would typically include patient demographics and/or health information, location, time of onset/testing, and case outcomes.

Even data from one (or several) specific locations would be most helpful.

This question refers to case-level data. For aggregate data, see this question (COVID19 - Corona Virus data).

  • 3
    Given the sensitivity of such PII and PHI under HIPAA and other patient/respondent privacy laws, it is unlikely to be available at case-level. If it ever does, it will more than likely be masked to some degree. – Kotebiya Mar 18 at 13:47
  • 2
    Also looking for such data. I‘m interested in looking at mortality contingent on demographics and individual patient information. – Peter Mar 20 at 21:17
  • 4
    @Kotebiya sure, but even masked data (coarsened age, location, etc.) could be useful. – dzeltzer Mar 24 at 15:30
5

The Anthem XPRIZE Pandemic Alliance has pledged the availability of such de-identified data including data on previous outbreaks (i.e. swine-flu, influenza, SARS etc.) It is unclear, however, how soon the datasets will be released to researchers as it was just announced.

| improve this answer | |
  • 2
    any news so far? – cccnrc Apr 3 at 14:44
4

I've also found a fairly detailed dataset on github

https://github.com/beoutbreakprepared/nCoV2019/tree/master/latest_data

This is quite exciting as it holds the following data

  • Demography: age,sex
  • Geographic: city, province, country, wuhan(0)_not_wuhan(1), latitude, longitude, geo_resolution, lives_in_Wuhan, location, country_new
  • Dates: date_onset_symptoms, date_admission_hospital, date_confirmation, date_death_or_discharge, travel_history_dates
  • Symptoms: symptoms
  • Other disease: chronic_disease_binary, chronic_disease
  • Exposure: travel_history_location, reported_market_exposure, travel_history_binary
  • Outcome: outcome,
  • Other: additional_information, source, sequence_available, notes_for_discussion, admin3, admin2, admin1, admin_id, data_moderator_initials

The downside, is that the data is of mixed quality, but it is critiqued by medical and data specialists. Some fields are completely left blank, but as a research source it could be a great place to start.

(This answer was moved from a more general question)

| improve this answer | |
3

South Korea has opened patient data. From what I glean from these tweets:

https://twitter.com/sang_woo_park/status/1247222889222483969

South Korea released medical history for all #COVID19 patients based on their insurance claims for the past five year:

https://hira-covid19.net/

https://twitter.com/sang_woo_park/status/1247386655549526017

The website is a bit cryptic. The sample data can be found in the "notice" menu under the "community" tab after logging in, but the data schema is in Korean. If you're serious about analyzing the data, I might be able to help with the translation...

https://twitter.com/FabiKrauer/status/1247467016006651905

„Raw data will not be available for download nor transmission. Researchers will use the sample data on the website (https://covid19data.hira.or.kr) to produce an analysis code and submit the code to HIRA through this website. In return, HIRA will deliver the analysis result values“

PDF user manual: https://rtrod-assets.s3.ap-northeast-2.amazonaws.com/static/tools/manual/COVID-19+website+manual_v2.1.pdf


It requires registration, and I think you have to do the analysis inside the platform.

enter image description here

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.