When used correctly, data and analytics foster great potential to help improve healthcare systems and clinician performance. However, choosing which data set to decipher and how to best explain data outcomes to physicians can be a daunting task.
For hospitals looking to make the switch from fee-for-service into the realm of value-based care, applying data effectively can be critical to making decisions. Panelists at Becker's Hospital Review 7th Annual CEO + CFO Roundtable in Chicago, Nov. 12-14, discussed some of these dilemmas and shared insights on the most important steps to take when applying data to improve healthcare outcomes.
1. Prioritize data to specific situations. Health systems maintain copious amounts of data, so choosing which data set to even begin with can be difficult. Kayla Pelegrin, MD, chief clinical officer of data informatics at 3M, advises starting with data you know has been standardized.
"Start with data that's much smaller and a more trusted set, instead of trying to go out and gather as much data as you possibly can. Because the challenge is how do you drive insight and knowledge out of the data you already have? So, start with smaller data sets that are more standardized, more well-known, and then expand that clinical knowledge from there as opposed to trying to start with your largest data set and not being able to pull anything out of it."
2. Meet with physicians to present data. Once you have extracted a data set, the new challenge becomes communicating that information with your physician. Jim Gera, president of Fusion5, part of the healthcare logistics company Owens and Minor family, stresses meeting with the physician a few times to walk him or her through the data in a way that explains what the numbers represent.
"What I've done before is shown [the physician] all [of the system's physicians] ranked in a simple spreadsheet but blinded because you want to show the cost and outcomes. You want the physicians to all start figuring out which one of those blinded names is them, and who's performing poorly and who's performing better. Once they start to get engaged, unblind the data. Now everybody can see transparently who's performing. And be very transparent and open with the data. As soon as you start to establish that process tell them … 'hey, here's our interpretation of what's happening with the data, and here's where we think there may be some solutions. Let's have a follow-up discussion to see if that's where we can make an impact.'"
3. Physician leaders can help communicate data outcomes. While Dr. Pelegrin agrees interacting with physicians in person is an effective way to explain data results, she also points out the person delivering the information should ultimately be someone the physician can relate to.
"I think physicians tend to respond better when the message is delivered by another clinician. It doesn't necessarily have to be a physician or a nurse, but somebody who has lived in their role and actually understands it. Having a data analyst come in and review [data performance] with a physician… [the physician's] going to throw that person out of the room. Or, they might be polite and just ignore the information that's provided. So, I think part of on the front end I would make sure that you find one or two clinicians, physicians, nurses, whoever, that have strong reputations and relationships within the organization. It will really, really go a long way in those physicians adopting the change and even believing in the results."