Pubmed Entrez Data

Question

[d.caliano]

Hi, I have to download from pubmed all articles inserted into the database within a certain date. I do it, by inserting the filter, [edat] in the esearch request.

When i make an efetch request to recover all the information about the article, i don't know how to find the edat. I thought it was contained here:

<PubMedPubDate PubStatus="entrez">
<Year>2012</Year>
<Month>1</Month>
<Day>3</Day>
<Hour>6</Hour>
<Minute>0</Minute>
</PubMedPubDate>

But some articles have an entrez data, out of the range I request in the esearch. So this is not the real entrez data.

Some ideas?

[Neilfws]

Can you edit this question to include an example of a search query, showing the date format you used?

Answer 1

[Hayssam]

Hi,

You should have a look at the <PubMedPubDate PubStatus="pubmed"> field. This usually correspond to the [EDAT] field of the record in MEDLINE format.

As an example, search for articles of a specific author with EDAT == the 26th of October 2012, e.g. with

http://www.ncbi.nlm.nih.gov/pubmed/?term=(%222012%2F10%2F26%22%5BEDAT%5D+%3A+%222012%2F10%2F26%22%5BEDAT%5D)+McIntyre

Open the 7th entry (about the SHIELD project), both in MEDLINE and XML format. The XML node matching the EDAT MEDLINE field is the <PubMedPubDate PubStatus="pubmed">. I usually found the MEDLINE fields more documented and more matching manual queries than the XML. Note also that there might be some voodoo processing on the server side to interpret your query!

[d.caliano]

It seems you're right. But I'm not sure that there's that data for every article in pubmed

[Hayssam]

I never made this comparison (I usually go for the MEDLINE files). If you're right and that this occur in a significant amount of entries, I'm sure it could be interesting to your analysis results to PUBMED admins. It's their job to keep to database consistent ;)

[d.caliano]

I've just made some tests and seems that the pubmed data there is always. But for the newest article (edat from 2013/09/01 to 2013/11/05) is completely wrong. So the problem remains

[Hayssam]

I think these articles are indicated as being under processing. Can you give some PUBMEDID examples of these?

[d.caliano]

yes this is the efetch xml: http://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&id=24016960,24016983,24017236&retmode=text&rettype=xml

[Hayssam]

I don't see the discrepancies about this articles. If you look at the MEDLINE records : http://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&id=24016960,24016983,24017236&retmode=MEDLINE&rettype=MEDLINE

The EDAT corresponds to the pubstatus="pubmed" node. Did I miss something?

[d.caliano]

yes, but my search request was with the range 2013/09/01 - 2013/11/05 so why i have articles with edat = 2005 ? It seems edat i request in the esearch is actually CRDT date in the xml

[Neilfws]

I do not see any articles with EDAT year = 2005 in that example XML.

EDAT = Entrez date which is <pubstatus="entrez"> NOT <pubstatus="pubmed">. Year is 2013 in all cases.

[Neilfws]

Yes: what you missed is that EDAT does not correspond to "pubmed", but to "entrez".