"A conclusion of my work is that mobile apps work. If you have something, make a mobile app and it will work," Dr. Paul Fontelo, director of the Lister Hill National Center for Biomedical Communications at the National Library of Medicine, said Tuesday while presenting a paper at the American Medical Informatics Association (AMIA) annual symposium in Washington.

That may be an oversimplification, but NLM has had quite a bit of success with its PubMed for Handhelds (PubMed4Hh) app.

NLM, which has had a website for PubMed since 2002, released an iPhone app in late July 2012. "We immediately noticed a surge in searches that progressively amplified over the succeeding weeks," read the paper, which appeared in the conference proceedings.

In the six months following the app's introduction, total usage of three search tools — PICO (Patient/Problem, Intervention, Compared to, and Outcome, a method used for structuring clinical questions), askMedline and the BabelMeSH cross-language search — jumped by 280 percent compared to the previous six months, Fontelo reported.

What's more, 97.8 percent of queries through the app were clinical in nature, the majority of those related to therapy.

"When an app is listed in iTunes, its visibility increases. People write about it in social media and reviews are written, so it is possible that some of the growth could have come from the increasing of the popularity for PubMedHh, leading to new users," the paper said.

"Mobile apps are behavior-changing," Fontelo explained. But only if they are useful, he added.

Fontelo noted that iTunes now lists more than 900,000 apps for Apple iOS. "That's ridiculous," he said. "Not all of them are good apps." Users have to be "judicious" in choosing apps, particularly when dealing with clinical knowledge, Fontelo said.

NLM has tried to optimize PubMed4Hh by focusing on what Fontelo called "TBL," namely the bottom line, with short summaries of research abstracts that can fit within a 160-character text message. He said this helps people find things quickly and easily share medical research with colleagues. "The correlation was quite good" between TBL summaries and actual abstracts, Fontelo said.

"We're always modifying the algorithm to make it more accurate," he continued.

Upcoming work will focus on improving an Android version of PubMed4Hh, according to Fontelo. An app has been available for Android devices since late last year, "but it's really not very robust," Fontelo admitted.