Hi Edwin,
Thanks for your response. I appreciate your patience, and I'm sorry I
didn't get back to you earlier.
In response to your questions, there are fewer citations from the API than
exist in PubMed for the author's name for a given time period. The
precision of search is also very poor -- it's capturing only a handful of
articles with authors from our organization, despite the affiliation terms
in the strategy. When we tested the strategy for the API, the precision is
very high. The automatic mapping in both versions shouldn't be an issue for
this search because all of our terms are tagged, but I was concerned that
either the differences between the interfaces or the ATM was causing issue.
I tested the strategy in both versions and they retrieve the same set and
the details show there isn't any automatic mapping being applied. I did
contact NLM about an API with new PubMed and unfortunately, they don't know
when a new version will be available for the new PubMed.
Elizabeth Huggins
MALIS '11
www.linkedin.com/pub/elizabeth-huggins/43/417/172
On Fri, Mar 27, 2020 at 8:10 AM EDWIN VINCENT SPERR <[log in to unmask]> wrote:
> Elizabeth --
>
> What kind of discrepancies are you seeing? Is it when you are checking
> your results against the New PubMed interface?
>
> Besides a tweaked UI, the new iteration of PubMed (
> https://pubmed.ncbi.nlm.nih.gov/) features a different default sort order
> ("Best Match") that can show slightly different results from the
> traditional Most Recent sort. Note, this happens not only with general
> searching, but even when retrieving known items.
>
> I note this because the PubMed API returns items in Most Recent order by
> default and indeed can not show a Best Match view (I think it also uses the
> older iteration of Automatic Term Mapping). There has been communication
> from NLM/NCBI that a new version of the API is coming at some point, but I
> don't know about a schedule...
>
> *Edwin V. Sperr, MLIS*
>
> AU/UGA Medical Partnership
>
> Office of Graduate Medical Education
>
> Clinical Information Librarian
>
>
>
> St. Mary’s Hospital
>
> 1230 Baxter Street
>
> Athens, GA 30606
>
>
>
> p: 706-389-3864
>
> e: [log in to unmask] | [log in to unmask]
>
> w: medicalpartnership.usg.edu <http://www.medicalpartnership.usg.edu/>
>
>
>
> ------------------------------
>
> Date: Thu, 26 Mar 2020 16:35:07 -0500
> From: Elizabeth Huggins <[log in to unmask]>
> Subject: PubMed API
>
> Hi all,
>
> I hope everyone is doing well during these ever-changing times. I work at a
> health sciences library and we're using an NCBI's API (E-utilities)
> <https://www.ncbi.nlm.nih.gov/home/develop/api/> to search PubMed. We work
> with IT to use the API to retrieve our faculty's citations monthly. I've
> noticed some discrepancies in the results of the API over the last few
> months and I would be grateful for insight from anyone else using this API.
>
> Here are a few of specific questions I have:
>
> - Does anyone else use NCBI's API in PubMed on a regular basis to
> retrieve faculty publications?
> - If so, what's your experience with it and how does your library
> implement?
> - Has anyone encountered any challenges with the results of a regular
> retrieval?
>
> I'd love to hear from you and learn from your experience. Please email me
> at [log in to unmask], my university email.
>
> Thank you, and stay safe and well!
> Elizabeth
>
> Elizabeth Huggins
> MALIS '11
> www.linkedin.com/pub/elizabeth-huggins/43/417/172
>
> ------------------------------
>
>
>
