Issue link: https://viewer.e-digitaledition.com/i/489205
"There are some really interesting, untapped opportunities in crowdsourcing, not only for enhancing effi ciency in work but for creating new knowledge." — Marc Halterman, MD, PhD, Associate Professor in the Department of Neurology at the University of Rochester and Co-Founder of Collaborative Informatics, LLC Altruism, competition, education and intel- lectual stimulation are important factors, he says. The crowd has helped bring more than 50 percent of CrowdMed patients — more than 500 individuals — closer to a correct diagnosis or cure since the service launched in April 2013. CrowdMed has an obvious limitation: In-person patient examinations aren't possible, which can make physicians skepti- cal about the service's accuracy. But the groundwork laid by patients' own physicians is vital, Heyman says. "Most patients have seen several physi- cians by the time they consult CrowdMed, and we like it that way," he says. "CrowdMed doesn't replace seeing physicians; it doesn't even work very well unless patients have seen one or two physicians fi rst. CrowdMed's role is to provide a short, insightful list of suggestions that patients can discuss with their physicians." Clinician-to-crowd Applications Marc Halterman, MD, PhD, Associate Professor in the Department of Neurology at the University of Rochester and Co-Founder of Collaborative Informatics, LLC, and his colleagues have investigated a different form of medical crowdsourcing: a downloadable mobile application that allows physicians to glean information from their peers. Dr. Halterman and his team described their work with the app they created, called DocCHIRP, in papers published in 2014 in Health Informatics Journal and Journal of Hospital Medicine. When physicians used DocCHIRP to pose questions to the crowd, members of the crowd received notifi cations on their smartphones and had the option to answer immediately; the average fi rst response occurred within 19 minutes. Eighty-fi ve providers — 91 percent of whom were physicians — participated in DocCHIRP's 244-day pilot period. "We expected a lot of providers' questions to revolve around diagnostic dilemmas, but, interestingly, the majority of inquiries were medication-related," Dr. Halterman says. "Providers wanted to know about dosing, side effects and off-label uses. Another interesting discovery is that providers were excited about being able to communicate with other specialists and allied health workers to ask, 'Is this the optimal referral or therapy?' Those kinds of questions might otherwise end up on a to-do list that the provider might not handle until the end of the day or the next day." The follow-up study Dr. Halterman and his team conducted after the trial revealed some user concerns about notifi - cation-driven crowdsourcing, including its potential to contribute to distracted doctor- ing, the safety of making recommendations based on digital curbside consults, data security and providers' reluctance to expose personal knowledge gaps. "Physicians want to help their colleagues but in a way that makes the best use of time," Dr. Halterman says. "Thirty percent of users accessed their accounts after hours on a desktop computer, so in addition to using their mobile devices, users liked checking in after hours to review the answers to their posts. This program may be most useful for simple questions that are barriers to workfl ow, such as, 'I have the answer to this question on the tip of my tongue, but I can't think of it — can you all help?' "Crowdsourcing systems are at their most robust with large numbers of partici- pants. This type of scaling is particularly important in developing 'smart' systems that can evenly distribute consults to the crowd so that providers don't experience notifi cation fatigue." ■ M D N E W S . CO M ■ MD NEWS Long Island | 2 5

