**Check the result count before exporting.** Note the total hits shown for the search. Expected: if the count is in the tens of thousands, a single export is the likely timeout cause.

**Export a small test set.** Apply one narrow filter, such as a single year, and export that subset in the summary or csv format. Success check: a file downloads quickly, proving export itself works.

**Batch the real export.** Export by date slices or by journal subsets that together cover the search, keeping each batch near the size of the successful test export. Expected: each batch downloads without a timeout.

**Prefer a lighter format.** If the summary format times out, try the csv format for the same batch, which carries less text per record. Success check: the lighter file completes where the heavier one did not.

**Save the search and retry off peak.** If even a modest batch times out, save the exact search, wait, and retry the same batch later. Expected: the same batch succeeds without changes, confirming a transient service slowdown.

**Merge and dedupe the batches.** Combine the downloaded files and dedupe by the PubMed identifier column. Success check: the merged count matches the expected unique records for the slices exported.