Every time you enter a PubMed search, the friendly computers at the NCBI work to translate your terms(s) into a carefully codified search, often combining the words that you typed with a series of Medical Subject Headings (MeSH). For example, "death" becomes:
"death"[MeSH Terms] OR "death"[All Fields]"
...while "heart attacks and aspirin" turns into the slightly more wordy
("myocardial infarction"[MeSH Terms] OR ("myocardial"[All Fields] AND "infarction"[All Fields]) OR "myocardial infarction"[All Fields] OR ("heart"[All Fields] AND "attacks"[All Fields]) OR "heart attacks"[All Fields]) AND ("aspirin"[MeSH Terms] OR "aspirin"[All Fields])"
Search Workbench presents exactly the same translation to you as if you had scrolled down to the "Search details" box in PubMed. You may find that this interpolation of your original terms is not exactly what you are looking for — experienced searchers can modify it by hand using the "Edit this search" option.
If you choose the icon next to your search, you'll open it in a new PubMed window. Just copy the address from this window and use that URL to run the search again.
If you want to construct a PubMed search of any complexity, you have to think about how you want to combine your terms using Boolean logic. Search Workbench helps to facilitate this process by showing you how the results for your search terms in PubMed compare to one another in total size as well as how they overlap.
These diagrams are not only displayed for individual searches (so long as you have more than one term), but they are also displayed for comparisons as well. Comparing searches directly is useful for seeing exactly how variants of a search do or do not overlap. Note: The display logic underlying the Venn diagram will almost always give a good sense of set proportions, but plotting intersections onto a two-dimensional plane can be somewhat inexact for any diagram involving more than two sets.
For more fun with Venn diagrams, check out PubVenn.
As you might have guessed, the "Proportion by year" chart shows you how the results for a particular search compare with PubMed as a whole over time. Remember, these are not raw totals, but proportions; it is entirely possible for the proportions for a search to remain steady (or decrease) in the face of increasing citations as the total database increases dramatically each year.
Note: with some searches, you may see the graph tail off dramatically in the last year or two. This is likely an artifact of citations being included in PubMed before they are formally indexed; any search that relies upon MeSH headings will underperform in relative terms until the indexing catches up.
You can see these diagrams in a larger format by visiting PubMed by Year.
A hedge is a pre-built part of a search that helps you focus your retrieval on a particular topic.