Ask HN: Open Covid-19 Patient Database | Hacker News Reader

Ask HN: Open Covid-19 Patient Database

28 pointspezo19198 comments

Hello everyone,

I am wondering why don't people create a database where COVID-19 patients themselves can upload their data to help finding the answers. The dimensions I am thinking about is beyond basic "severeness" and "time to recover" but also nutrition/supplement informations or health condition informations, fitness, being too exhausted for a longer time / lack of sleep before sympthoms arise.

I know many of this is personal information and I also know it is hard to verify these pieces of information, but I think there are many people who are willing to help with their data, but there is no platform for that.

I don't know how to do this this is just the idea.

Edit: South Korea has already tested more than 140.000 people, (https://www.google.hu/amp/s/www.businessinsider.com/south-korea-coronavirus-testing-death-rate-2020-3%3famp ) it would be really interesting to see their data.

View on HN

I volunteer the next two days to help with this in any way I can. I've been working on a similar citizen science idea, related to home efficiency (beaverlodge.game), but it might have some parallels in methodology. My background is more data science and marketing, but I have a bit of coding experience as well.

I just posted an idea similar to this, without having seen yours first: https://news.ycombinator.com/item?id=22574128

add geolocation data and location based app awareness and alerting.

The problem me thinks are language barriers and cultural-socio privacy differences.

Next level technology could really help in this space, if we could only get our global sh*t together...

Do not allow PII! Be very careful about what you gather as publicly sourced data.

Think about what can happen if false data is entered. Consider both malicious and ignorant cases.

Yes. I think in times like this, one of the things we need to guard against is misinformation.

Transparency is important but it must exist alongside trust -- if the latter is compromised, the implementation of correct actions may be compromised.

Crowdsourced systems have an implicit assumption that the downside of acting on incorrect information is small. I'm not sure if this assumption holds here. Imagine someone entering large swaths of data that are false but plausible, and then less scrupulous news reporters pick it up, arrive at the wrong conclusion and then fast feedback loops inadvertently propagate it without fact checks (you know how the news cycle works).

With Wikipedia, there's at least the mechanism of curation (flawed as it may be) and citations. With an open patient database, there's no means of verification.

There's also the assumption that people are good at entering data without constraints correctly. I see this assumption violated all the time.

Locate the servers outside the USA/EU them all for whatever you want. Hippa is outdated and designed for a pre Internet Age. We could have our whole health history running in Apple health yet visible to researchers if allowed. This is technically possible but the present healthcare industry/government does not allow this.

Please do not give bad legal advice in a public forum. This could have grave consequences for someone who naively follows such guidance.

Whether you agree with a law is not relevant to whether you reside or operate within a jurisdiction it applies to.

You can put lives at risk by sharing bad or the wrong type of information. People who are scared often make bad decisions.

In the wrong environment, sharing someone's identity can be equivalent to swatting them and their community.

There are deep ethics involved with the decisions that structured many of these laws. Anytime you are dealing with personally identifiable information, you have to consider the consequences of trivializing privacy.

Always consider and know the laws for your domain and also take seriously the ethics.

If you aren't ready to do this, choose to contribute by a different way. There are many ways to develop resources that can reduce harm, and currently large online communities brainstorming those ways.

I'm down to help. You could skip a UI and go straight jupyter w csv file - get some nice visualizations out there quick. Google Forms.

Reply on news.ycombinator.com