| Issue Number | 430 |
|---|---|
| Summary | [Literature] Citation Statistics for EIC Meeting |
| Created | 2017-04-14 14:49:43 |
| Issue Type | Improvement |
| Submitted By | Juthe, Robin (NIH/NCI) [E] |
| Assigned To | Kline, Bob (NIH/NCI) [C] |
| Status | Closed |
| Resolved | 2017-05-16 09:25:51 |
| Resolution | Fixed |
| Path | /home/bkline/backups/jira/oceebms/issue.206552 |
Margaret is pulling together some citation numbers to share with the Editors-in-Chief in our meeting coming up on May 4. Cynthia was able to get most of the numbers Margaret is interested in, but we wondered if you might be able to help us with an ad-hoc query in the EBMS to come up with a few additional numbers.
The attached spreadsheet shows the number of citations imported by the medical librarians between Jan 1-Dec 31, 2016 and then the number for each subsequent step in the review process down to the number of citations cited in PDQ summaries (the last number is coming from a CDR report and we recognize that it’s an approximation since some citations are likely in markup or added and subsequently removed from the summaries.)
We’d like to get a sense of how many articles we assign to Board members for review that end up getting a “yes” or a “no” response, and how many received “no responses”. We define a “yes” response as ANY positive response (i.e., deserves citation, merits discussion, or merits revision of the text) and a “no” response as ONLY “warrants no changes” (no “yes” responses were received for the article for any topic associated with that Board).
We’d like to determine this number on a Board level. In other words, if the same citation is assigned for review for multiple topics on the same Board, we only want that citation counted once. I tried using our existing reports – Literature Reviews, Articles Without Responses, Search – to get these numbers but I wasn’t able to collapse them by Board (each citation displayed once for each topic).
Lastly, since Margaret plans to present these data as a percentage of the number of full-text articles sent to the Boards for review, we wondered if you would be able to determine this number from the system as well. The numbers Cynthia has on the spreadsheet (rows 34-40) should be fairly close, as these represent all articles that received an approval at the full-text review state. Theoretically, the numbers you come up with for citations that received a “yes”, “no”, or “no response” should add up the total number of articles sent to each Board.
Please let us know if you have any questions or need any more info. As I mentioned, the EIC meeting is on May 4, so we'd like to have these numbers sometime before then. Thank you!
So to summarize, you want a report with six rows, one for each board. Each row should contain:
the name of the board
the number of articles imported in 2016 and subsequently assigned for review
the number of those articles which received at least one response which was not "warrants no changes"
the number of those articles which received only "warrants no changes" responses
the number of those articles which received no responses at all
Questions:
We understand that the report could change over time (though it's pretty unlikely that the changes would be significant) as late assignments/responses are recorded, right?
If there is any import event for an article (for this board) in 2016, does it get counted (assuming it got assigned for review)? Or does the first (or the last) import event for the article have to have occurred (for this board) in 2016?
Can I assume that we want to consider an article as "assigned for review" if either of the following conditions is satisfied?
It has at least one reviewer assigned for one of this board's packets for the article
It has at least one response for one of this board's packets for the article
That way we'd (usually) avoid articles that went out without any assignments (just for information) but pick up responses which were made before the board member was dropped from the reviewer list for the packet.
Robin and I conferred via telephone and agreed that the summary of the report structure was correct and that the report will be mutable over time (question #1). For question #2, we decided to include an article if there was a row added in 2016 to the ebms_import_action table for it. We realize that this will mean that if we were to run a report next year for 2017, it would include some of the same articles, and Robin said that if we get to the point where we want to run the report more than this one time we may well revisit the logic for it.
We didn't get to the third question, so perhaps you can respond to that one in the ticket, ~juther. :-)
Re #3, your assumptions look good to me. Thanks!
Looks like we may need to revisit the decision we made earlier about which articles to include as I had roughly four times as many articles as Cynthia did. I believe the bulk of that discrepancy comes from the rows in the ebms_import_event table which reflect replacement imports for records updated by NLM. I suggest we filter my query using the import disposition column. Here are the valid values for that column:
import dispositions:
imported
reviewReady
notListed
duplicate
topicAdded
replaced
error
We don't need to worry about the rows with a disposition of error because the article ID column is NULL for those, but we'll probably want to skip the ones that have notListed, duplicate or replaced. What do you think?
I went ahead with the filtering I suggested in the previous comment (so my numbers would match Cynthia's). Results are attached to the ticket. Boy, that Peds board is rockin' it! :-)
As a side note, the second condition for question #3 above never had to be applied, because for none of the articles were all of the reviewers dropped from all of the board's packets for the article after a review had been snuck in. Of course, we always expected that to be a pretty rare edge case, and we were right. :-)
Let me know if this looks like what you need.
Please close this ticket if everything looks OK.
This was very helpful. Thank you!
| File Name | Posted | User |
|---|---|---|
| EBMS statistics 2016.xlsx | 2017-04-14 14:49:41 | Juthe, Robin (NIH/NCI) [E] |
| oceebms-430-prod.xlsx | 2017-04-20 09:50:06 | Kline, Bob (NIH/NCI) [C] |
Elapsed: 0:00:00.000692