The Beta Angel Project’s PubMed Search Strategy

By: Tallie Casucci & Taylor Reed

china-clouds-cold-1586205
The search can be cloudy, but beautiful.  Photo by Sam Lim

UPDATE (1/21/2019): Added “Post-publication feedback and revision” to the end of this article.  Check it out.

The Beta Angel Project uses a PubMed search strategy to identify research relevant to inclusion within the Beta Angel Project’s Research Inventory of performance-related climbing research. The search strategy consists of three parts: (1) mountaineering and ice climbing, (2) technical face climbing (bouldering, lead, speed, competition), and (3) the exclusion of irrelevant concepts, such as specific methodologies, animals, and plants.

Each section is noted by a color: mountaineering (BLUE); technical face climbing (GREEN); removal search terms (RED).

“Mountaineering”[Mesh] OR alpinis*[tw] OR mountaineer*[tw] OR (climb*[tw] AND (alpine[tw] OR ice[tw] OR icicle[tw] OR mixed[tw] OR mountain*[tw] OR snow[tw])) OR bouldering*[tw] OR boulderer*[tw] OR buildering[tw] OR climber*[tw] OR psicobloc[tw] OR (“deep water”[tw] AND (solo[tw] OR soloing[tw])) OR (climb*[tw] AND (athlete[tw] OR athletes[tw] AND boulder*[tw] OR competiti*[tw] OR crack[tw] OR cracks[tw] OR face[tw] OR free[tw] OR gym[tw] OR gyms[tw] OR gymnasium*[tw] OR indoor[tw] OR “on sight”[tw] OR onsight[tw] OR “red point”[tw] OR redpoint[tw] OR rock[tw] OR solo[tw] OR soloing[tw] OR speed[tw] OR sport[tw] OR “top rope”[tw] OR trad[tw] OR wall[tw] OR walls[tw])) OR “campus board”[tw] OR “crimp grip”[tw] OR “half crimp”[tw] OR hangboard[tw] OR “hang board”[tw] OR “open crimp”[tw] OR “slope grip”[tw] NOT (“climbazole”[Supplementary Concept] OR “Drosophila melanogaster”[Mesh] OR “In Vitro Techniques”[Mesh] OR “Mice”[Mesh] OR “Mice, Knockout”[Mesh] OR “Stair Climbing”[Mesh] OR stair[tw] OR hill[tw] OR robot[ti] OR amphibi*[ti] OR animal*[ti] OR ant[ti] OR ants[ti] OR beaver*[ti] OR bat[ti] OR bats[ti] OR beef[ti] OR bovine[ti] OR breeding[ti] OR bull[ti] OR climbazole[tiab] OR canine[ti] OR castoris[ti] OR cat[ti] OR caterpillar[ti] OR cattle[ti] OR cats[ti] OR chicken*[ti] OR chimp*[ti] OR cockroach*[tw] OR cow[ti] OR dog[ti] OR dogs[ti] OR drosophila[ti] OR equine[ti] OR fish[ti] OR foal[ti] OR foals[ti] OR frog[ti] OR geckos[ti] OR hedgehog*[ti] OR honeybee*[ti] OR horse[ti] OR horses[ti] OR hummingbird*[ti] OR insect*[ti] OR livestock[ti] OR lizard*[ti] OR mice[ti] OR mosquito*[ti] OR monkey*[ti] OR mouse[ti] OR murine[ti] OR plant[ti] OR plants[ti] OR pork[ti] OR porcine[ti] OR protozoa*[ti] OR purebred[ti] OR rabbit[ti] OR rabbits[ti] OR rat[ti] OR rats[ti] OR reptile*[ti] OR rodent*[ti] OR sheep[ti] OR spider[ti] OR spiders[ti] OR thoroughbred[ti] OR veterinary*[ti] OR vulture*[ti])

For the Beta Angel Project’s Research Inventory, we primarily use search strategy parts 2-3 (green and red sections), since we are interested in technical face climbing performance.

We chose PubMed, the premier biomedical literature database, because anyone can search it freely around the world. There is a major caveat with PubMed for climbing research; it does not contain everything. For example, there may be relevant engineering literature about climbing that will not be in PubMed. Additionally, “grey literature,” such as conference proceedings, dissertations/theses, and reports, are vital for the Beta Angel Project’s Research Inventory.

We welcome feedback about this search strategy! Did we miss a term? Did you find an article that this strategy does not retrieve?

tommy-lisbin-644940-unsplash
Everyone needs a search strategy.  Jon Sui climbing in Joshua Tree National Park.  Photo by Tommy Lisbin

The Nitty Gritty on Search Strategy Development

Over the past few months the Beta Angel Project has worked with Tallie Casucci, a librarian, to create and refine this search strategy. The topic of climbing is challenging because there are other non-sport climbing topics, such as stair climbing exercises, climbing plants, and robots. For a true, comprehensive, evidence-synthesis search (i.e. systematic review or scoping review methodologies), you would have to search climb* alone. Climb* retrieves thousands of false hits (the asterisk finds word ending variants: climb, climbs, climber, climbing, etc.). For the purposes of the Beta Angel Project’s Research inventory, it is impractical to review thousands of articles, so we narrowed the search by combining climb* with other concepts to better describe our population (rock, walls, gyms, etc.).

We tested the search concepts and strategy numerous times. For example, the concept cliff retrieves biology and anthropology articles, so cliff was excluded from the search. Cave was another concept we tested early in the strategy development stages, but it also retrieved biology, anthropology, and caver literature (i.e. not climbing in beautiful limestone caves). The Beta Angel Project also provided a list of article titles within the current Research Inventory that did not contain the climb* terms in the title. The librarian used this list to further develop the strategy to capture articles that do not use the familiar climbing concepts.

Additionally, the librarian worked with the U.S. National Library of Medicine (i.e. PubMed support) to index specific phrases, such as “crimp grip,” “slope grip,” and “hang board.” Now that these terms are in the PubMed phrase index,[1] the search will retrieve these exact phrases in the title, abstract, and author keywords. These phrases are necessary because some articles investigate grip strength, but do not specify the population as climbers. As more studies are published about grip positions and climbing-specific training apparatuses (i.e. Moon Board, Lattice Board, etc.), we will work with the U.S. National Library of Medicine’s support to add these to the phrase index. They require at least three articles using the phrase in the title, abstract, or author keywords.

Our most important test was the exclusion portion of the search strategy (part 3). On October 30, 2018, we searched for only the 2018 technical face climbing literature (note: this was a different search strategy than the one above). We identified 41 relevant technical face climbing performance articles published between January 1, 2018, and October 30, 2018. Without part 3, our success rate was 26.6% (41/154). With part 3, our success rate increased to 41.4% (41/99). This retrieval difference highlights the real challenge with searching for climbing literature due to the number of false hits which discuss other forms of climbing. The librarian double-checked the 55 excluded articles and confirmed them as false hits. It is worth noting that if articles are published concerning a bat biting a climber or comparing climber and chimpanzee grip strength, the search strategy with part 3 will exclude those articles. We will occasionally test and compare the search results in the future to ensure we do not miss these potential articles.

Finally, after peer-review from friends and colleagues, we tested additional concepts. We tested the climbing scales (Yosemite Decimal, French Rating, Ewbank) and found our search already captured the relevant articles. Other terms – route finding, traditional, lead, artificial, rope, outdoor, and pitch – connected to climb* retrieved mostly false hits and again our search strategy already captured the relevant articles. Of the new terms added in November, we only retrieved four new results. This gave us confidence that our search strategy is pretty comprehensive and focused. There are still false hits, due to the nature of mountain sports, PubMed’s capabilities, and author’s abstracts.

action-adult-adventure-946337
Photo by Rodrigo

What can you do?

First, send us feedback on the search strategy! Did we miss a term? Did you find an article that this strategy does not retrieve? Even though the strategy was peer-reviewed, we may be missing concepts. As the field of climbing research expands, we expect to make revisions.

Second, use this search strategy to write summaries for the Beta Angel Research Inventory.

Third, stay up-to-date on the most recent climbing literature. Simply create a free My NCBI account on PubMed. Copy and paste the above search strategy. Run the search and use the “create alert” feature.

Fourth, consider meaningful author keywords when submitting your manuscript for publication. “Climbing” alone is not a helpful term. Additionally, clearly define your population in your abstract.

Fifth, visit your local library. If your library does not contain the climbing material that you want, you can either suggest a purchase or request it through interlibrary loan. Most public university libraries in the US will allow you to use their computers for free, which means you can access full-text articles!

Finally, consider including a “social media friendly” image in your publication. Many authors, trainers, and coaches, like Beta Angel, will post research summarizes and/or their impressions on social media. Beta Angel has found that images with participants and/or meaningful graphs tend to receive more likes, comments, and interactions. This is another way to promote your research.

[1] “half crimp” is in the queue to be added.  Request received 12/7/2018

Post-publication feedback and revisions

After posting our article, we received important feedback from Professor and Dr. Volker Schöffl, Adjunct Professor of Trauma and Orthopedic Surgery, FAU Erlangen-Nuremberg, FRG; Adjoint Assistant Professor of Emergency Medicine at the University of Colorado School of Medicine, USA; and Visiting Professor at the School of Clinical and Applied Sciences, Leeds Becket University, UK.  Dr. Schöffl suggested that while the search strategy works for sports science, it doesn’t necessarily work for sports medicine.  This is a little bit of a problem for us because nearly one-sixth of the current Beta Angel research inventory (BARI) is made up of climbing injury-related research.  As a result, we spent some time diving into his constructive recommendations.

First, Dr. Schöffl noted that there were no sports medicine keywords associated with specific injuries (i.e. pulley injury) or areas of the anatomy prioritized in climbers (i.e. fingers). We searched for pulley and finger injuries within PubMed and found that other athletes and populations also have hand injuries. While these non-climbing hand injuries are interesting and valuable, the search retrieves too many results. Fortunately, most of the literature indicates that the population were climbers or the authors provide a climbing-related keyword. Since the Beta Angel Research Inventory does not focus on specific injuries, the general climbing-specific search terms are likely enough, but we are open to evidence that we are missing climbing-specific sports medicine articles. This is why we included “half crimp,” “slope grip,” and similar keywords, since some of those articles do not specify the participants (or cadavers) as climbers, but this research is very relevant to rock climbers.

As a prominent sports medicine researcher, we decided to look at Dr. Schöffl’s own list of publications. With our current strategy, we do not retrieve 25 of his articles. While some are not relevant to the BARI, others are. In those cases, Dr. Schöffl and his colleagues used many climbing terms in the full-text article, climbing as an author keyword, and “climbing athlete” in the title & abstract. Unfortunately, PubMed does not search the full-text article, so we are limited to searching for words in the title, abstract, and keywords. As we discussed above, climbing alone retrieves way too many false hits for stair climbing, climbing plants, and other false hits, so we will miss one of Dr. Schöffl’s articles. After evaluating Dr. Schöffl’s work, we added climb*[tw] AND (athlete[tw] OR athletes[tw]) to the search terms above. This addition to our search has the added bonus of returning some parkour articles, which may have a climbing-related application.  We will consider including these from time to time.

Second, Dr. Schöffl suggested some performance diagnostic terms, such as cpet and spiroergometry.  A quick and dirty PubMed search: (cpet OR “Cardiopulmonary Exercise Test” OR Spiroergometry) AND (climb* OR mountaineer*) returns 719 searches, of which most are false hits. Fortunately, the relevant articles already include our other search terms, such as rock climbing, indoor climbing, mountain climbing, etc.  However, it’s important to note that these two diagnostic tools are performance-related and incredibly important from the perspective of pushing evidence-based practice into the climbing community. As a result, we would like to offer a potential solution – we suggest beginning work with the sports medicine community to create an “injury- and diagnostic-specific” search strategy (similar to the color sections above). This potential additional search strategy component would provide value to researchers. We are open to begin collecting search terms, assuming this would be of value?

Finally, Dr. Schöffl was concerned that some terms, such as “red point,” are not considered science-specific and as a result would probably not be used in the science community. Red point and similar terms are actually used in science abstracts because some authors will include their participants’ red point abilities. We initially included red point because we expected more sport psychology articles, but actually it retrieves sports science-related climbing articles.

It’s important to remember that the context for this search strategy is for the Beta Angel website, which has the purpose of bridging science and practice. As a result, we have the burden of providing a “manageable volume of science” to the climbing community. We would like to be comprehensive, but comprehensive within the context of climbing.