What websites do medical researchers use?

[FallenApple]

Google is pretty much useless. Most items entered are just responses by non MDs or PhDs with no rigor whatsoever.

I'm looking for a website that is scientific with an abundant source of numerical figures/values. (p-vals, methodology, source code for their data analysis etc.)

 

[DrClaude]

PubMed. But you won't find any data in it, you'll have to do the work yourself of going back to the published articles.

 

[Ygggdrasil]

Usually, one would consult the scientific literature to find data, methodology, and source code for particular studies (though more specialized subfields may have specific depositories for data). As DrClaude mentioned, PubMed is the main portal biomedical researchers use to search for papers. Google Scholar and Web of Science are other good search engines.

 

[jim mcnamara]

If you prefix your search with

nih: [search pattern]

Or pubmed:, etc., you filter out most useless references. You will also get some few publications meant for non-professionals. These are quality content.
google scholar gets me too many references behind a paywall, NIH (pubmed) does that a lot less.

 

[Auto-Didact]

I predominantly use Pubmed, Embase, Cochrane and Uptodate.

 

[onoturtle]

source code for their data analysis

Ha, I wish such a resource for this existed. Would help with reproducibility. But a lot of science is also done with expensive proprietary software, so even if you had the code or saved project, you may not necessarily be able to run it.

The reality is you read a paper, download their data (at least in genomics we got things like NCBI GEO website for that!), then try to follow the written methodology from the paper/supplements using whatever tools you have (not necessarily what was used in the paper, but even if you do, they may not spell out all the parameters) and hope for the best.

Reproducible science is a problem.

 

[Ygggdrasil]

Ha, I wish such a resource for this existed. Would help with reproducibility. But a lot of science is also done with expensive proprietary software, so even if you had the code or saved project, you may not necessarily be able to run it.

The reality is you read a paper, download their data (at least in genomics we got things like NCBI GEO website for that!), then try to follow the written methodology from the paper/supplements using whatever tools you have (not necessarily what was used in the paper, but even if you do, they may not spell out all the parameters) and hope for the best.

Reproducible science is a problem.

For analysis done in python, some researchers will put their analysis into an interactive python notebook (like a jupyter notebook) and provide a link to the notebook in their paper.

 

[onoturtle]

I've never encountered that. I did attend a workshop for R that used jupyter notebooks. I suppose one could use that, using the script to download the data and run a simple analysis. I just don't see this happening with the area I'm in with terabytes of omics data, processing that takes hours on relatively big machines, and tools probably not even installed/accessible on the machine running jupyter (e.g. proprietary IPA for pathway analysis is pretty popular). I suppose a jupyter notebook for a microarray experiment could work where the script downloads data from GEO or whatever and use a bunch of BioConductor R libraries to do the analysis. That usually isn't very intensive.