Dear Oana,
Thanks for trying EPPI-Reviewer and for contacting us on the forum.
The problem you are facing is probably a sum of separate issues (internal and external) and is likely to be the result of having to deal with too many PubMed results.
If your query retrieves more than 3000 items from PubMed, and you are trying to import them all in one go, a few things may go wrong:
1) for big numbers, PubMed returns an unreliable number of hits. In particular, it seems that it usually states that there are N results but only sends N minus X results when asked to send them all. In my tests, whenever I tried to retrieve all results of a query that returned thousands of results, I always got less than the original number. This happened during local "ad hoc" tests that were not affected from any other known problem. I guess this is a common issue: even Google usually overestimates the number of results. In any case, if you have this problem, the only solution is to search on the PubMed webpage, export the results as a text file and then import them into EPPI-Reviewer.
2) The PubMed service may be busy and send the results to our server too slowly. A timeout would occur, but I'm pretty sure that nothing would be imported if this is the case. For this and other reasons, I'm guessing this problem can be excluded (at least for now).
3) The timeout you get may be triggered when EPPI-Reviewer refreshes the current list of sources. This may happen when a large import was completed, EPPI-Reviewer gets busy digesting the new data and a timeout is fired. I may be able to tell if that's the case by looking at the complete error message. You may want to send a screenshot to EPPISupport@ioe.ac.uk.
A few hints that may help:
First and foremost. If possible, try not to import more that 2-3000 items at a time. The PubMed import tab allows to select how many items to import in one go, this is done via the boxes on the bottom left of the window. In real-life I have been able to successfully import up to 8000 items from PubMed in one go, but I wouldn't recommend to! Please note that because of issue 1) PubMed may actually return fewer items than the expected. For example, if you ask for items from 2001 to 4000 you may receive 1993 items instead of the expected 2000. This is a known issue, but since it happens on the PubMed side I was never able to get to the bottom of it.
Second: to avoid polluting your list of items, you may delete old, possibly incomplete, sources. Safest way would be to delete them "forever". This is done in two steps: on the main window, select the sources tab on the right column. Click the bin icon to delete the source(s). Open the "manage/add sources" window and navigate to the "Manage Sources" tab. In here, you can delete forever sources that are already marked as deleted: select a deleted source from the right column, click the red "X" button and confirm.
Third: for some reason, the PubMed web-page does not seem to be affected by issue 1). If you are worried that the direct import may leave out some results, you will need to go through the long route. This implies searching on the web page, saving all results in PubMed text format and then importing them through the "import new" tab. Please note that the size issue is even more limiting when importing text files: you will have to divide your files in batches of 2-3000 items, otherwise each import may fail completely.
I realise the picture I'm drawing is not ideal: we are painfully aware of the limitations on how many items can be imported in one go, but we were unable to find a reliable solution (keeping in mind that we can't assume anything about our users' connection speed and other limitations).
I hope this will help, please do not hesitate to get in touch again if you'll need further assistance.
Best wishes,
Sergio