SUMSearch

SUMSearch is a free medical, meta-search engine sponsored by University of Texas Health Science Center at San Antonio. It has an easy-to-use interface and is currently available in English, Spanish, and Italian, with a French module coming soon. SUMSearch was created by Robert Badgett, an internal medicine doctor and professor at the University of Texas Health Science Center at San Antonio. His instruction includes classes teaching medical students how to perform effective research in medical sources. His background and knowledge of the problems faced while looking for quality medical evidence led to this user-friendly meta-search engine.

The main SUMSearch search screen is simple and clean. It has been designed for medical students to perform evidence-based research [1], so it can easily be used by a beginning searcher. SUMSearch performs a meta-search and adds contingency searching to help narrow down results when too many are found or can expand results when too few are found. The results can be further researched, making this a good tool for the expert searcher as well.

When a search is run, SUMSearch performs a meta-search. It only searches databases with recognized qualified medical information, such as textbooks and medical journals. By default, search terms are applied to Wikipedia, PubMed, the National Guidelines Clearinghouse, and the Cochrane Library's Database of Abstract of Reviews of Effectiveness (DARE).

While the search terms are added in the search field on the left, the right-hand top panel changes to offer a link to check Medical Subject Headings (MeSH). When the link is clicked, a new window opens displaying potential terms from the MeSH database. Warnings can also appear for suggestions on how to reword or change the search. The Check My Strategy button also opens that portion of the screen. It is a good tool for novice searchers to be able to check their terminology and make changes.

The results come back divided into sections. The top section (salmon colored on the screen) gives a summary of the searches performed. The bottom section (the blue background color of the screen) gives specific sources found. There is a link from the summary section to the detail section.

The first portion of the summary has header text stating the results are easy to read but may not be the most up-to-date information. This portion of the summary screen focuses on textbooks, reviews, general articles, and guidelines. Links from the summary section jump to the matching portion of the detail section below.

The second portion of the summary has header text stating that the information is more up to date but not as easy reading as the first group. This portion covers systematic reviews and original research in DARE and PubMed. Other systematic review sources can also be examined, such as the national association's site that applies to the search.

The PubMed search results include a link showing the number of searches run. That link will bring up the different wording and conditions of the searches run in PubMed. The expert searcher can work with those conditions in PubMed to narrow down the results even more.

After the results are listed, SUMSearch adds extra information that is helpful for the novice searcher. There is assistance for narrowing down a search, such as pictured in Figure 1. Notice the warning for too many original studies found in the initial search. SUMSearch offers links to automatically refocus the search to a subsection of the first result. The researcher also can return to the original screen and use the focus portion of the screen to narrow the search.

Figure 1.

SUMSEarch search results

The bottom part of the screen in the blue area lists part of the results from the search. They can be reached through the links in the top salmon-colored section or the user can scroll down to see a portion of the results. There sections are color coded by a vertical bar to the left of the results. For example, the systematic reviews searched in DARE and the Cochrane abstracts have a red box where the summary shows and a red line down the left side of the results from that search. A color code key appears to the left of the list of results in the blue area. The key floats down the screen, so it is always visible to the user even if the header of the list is off the top of the screen.

If there are more citations available than are listed on the result screen, a link to the list can be found in each search. Each of the citations that are shown links to the online source. Many of the sources are full text. Others, especially the ones found in the PubMed searches, lead to the citation and then may require a subscription to obtain the article or source.

The home search screen in SUMSearch has a Focus section. A request can be immediately restricted to different predetermined filters: adverse treatment affects, diagnosis, etiology/causation, intervention, physical findings, prognosis, and screening/prevention. A few automatic limits can be set: age, human or all, and language. Unfortunately, SUMSearch does not have any limits by date. A close examination of some of the PubMed search strategies shows that they were limited to the past ten years. If someone only wants to examine the newest findings within the past three years or so, the list has to be visually checked.

Firefox and Internet Explorer 7 plug-ins are available to add SUMSearch to the search bar.

I ran a search in PubMed and SUMSearch that included the term “evidence-based.” In SUMSearch, the search only found results in PubMed, the same ones I found in my PubMed search. Knowing that the search engine is programmed to search only quality sites, I removed the descriptor “evidence-based” from my search. Then I got numerous recommended resources. Choosing the different filters in the original search positively impacted my results for the different focuses.

Different word spellings will change the SUMSearch results. I ran a search that included the word “pediatric.” After checking the results, I ran the search again using “pediatrics,” the proper MeSH term. My results were fewer than without the “s.” A researcher may want to run searches using different spellings or terms for greater results.

As a bonus, the searcher can choose to have a separate window open showing current medical news while waiting for the search to run. The main screen warns that a search can take up to forty-five seconds. I found that by the time I had opened the extra window and started glancing at the first headline, the original search was usually completed. The news feature is not one I would use. I found a prominently featured link to another site that was broken. I reported it through the contact information. I received a response within a couple days thanking me for the notice and informing me that the link was fixed.

I found SUMSearch an easy-to-use helpful meta-search engine. The screen is plain and not cluttered. As I stated before, a new searcher can get good results from it. Educational links throughout the site help the new searcher understand the results being obtained. An expert searcher can run an initial search, then dig deeper by refining strategies already created by the SUMSearch programming. The biggest drawback I found was the inability to limit by date, but I can work with that. SUMSearch appears to list its found sources in reverse chronological order, so I can stop reading a list once I have reached my target year in the citation. I like the ease of SUMSearch and the assistance available on the screens. I now recommend SUMSearch to my research clients as part of their writing tools.