OHDSI Home | Forums | Wiki | Github

Patient-level data for covid-19 from diverse sources, how can we standardize those?

(Seng Chan You) #1

I found four patient-level datasets from Korea, Singapore, Hongkong, and US.
Can we standardize these datasets?

I posted this issue to the github, too.

(Selva) #2

Hi @SCYou,

I am from Singapore and would be interested to participate in this effort.

Currently we have standardized our questionnaire and put it into cdm.observation table.

Is that what you are asking?

(Seng Chan You) #3

@SELVA_MUTHU_KUMARAN , do you mean results of questionnaires for suspected covid-19 patients in Singapore? Then, yes, it would be what I’ve been looking for!

(Selva) #4

Sorry, no. I didn’t mean that. What I meant was we usually standardize our other survey questionnaires from hospital and put it into observation table.

btw @SCYou - May I know whether you want to just convert the questionnaires (questions) which are hand out to the covid-19 patients?

Or if you wish to have access to their (covid-19 patients) responses, then I guess it might take a bit of time

I was told that it is sensitive and not open as of now

(Hainiwen) #5

looks great! The hongkong database needs a bit patient information collection and formatting. I can offer some time to convert if we fix the strings that need to be collected. Let me know where I can join.

(Selva) #6

Count me in as well. Happy to be involved

(Alexander Sivura) #7

I’m happy to get involved

(Alexander Sivura) #8

I’m going to spend some time today to download and convert into structured fromat (e.g. csv or json) from html
this source https://wars.vote4.hk/en/cases/

(Alexander Sivura) #9

Hongkong data is already available as spreadsheet: https://docs.google.com/spreadsheets/d/e/2PACX-1vSr2xYotDgnAq6bqm5Nkjq9voHBKzKNWH2zvTRx5LU0jnpccWykvEF8iB_0g7Tzo2pwzkTuM3ETlr_h/pub?gid=0&range=A2:ZZ

(Akilah Kamaria) #10

Kudos! This is great information.

(Vojtech Huser) #11

On date of confirmed diagnosis, we do:


condition_cid=<covid19> (from snomed 840539006 , see https://www.wikidata.org/wiki/Q84263196)

(Jan 2020 snomed seems not to be in Athena web version) No results when doing


Symptom onset

would go to OBSERVATION

(Alexander Sivura) #12

There are cases from the Hong Kong source
I created GitHub PR too.

(Keesvanbochove) #13

These datasets are very useful for research, thanks for sharing! I will try to hunt down if there is any open data in The Netherlands from RIVM or NIVEL.