At UCHealth, Rx decision support nudges physicians to prescribe recommended meds
UCHealth was having troubles with its prescription process – especially a wide variation in prescribing behaviors among providers, who often opted for medications that were not first-line guideline-recommended meds for a specific indication, weren't covered by the patient's insurance, weren't on the formulary or were too expensive for the patient.
Vendor RxRevu started implementing its SwiftRX tool at UCHealth in April 2016 with the goal of engaging prescribers with best practices for disease states involving antibiotics in the emergency department. One month later, RxRevu won a Colorado state-funded grant from the Office of Economic Development and International Trade to inform prescribing for at-risk congestive heart failure patients in primary care with evidence-based guidelines.
RxRevu began its work with UCHealth by deploying SwiftRX, which presents the prescriber with medication recommendations automatically when a diagnosis is entered in the electronic health record. This allows prescribers to save clicks in their workflow versus the standard order entry process in the EHR. Click reduction greatly reduces their time searching for the most optimal medication for their patients.
In June 2018, RxRevu launched SwiftRX with real-time patient-specific pharmacy benefit management data that allows the prescriber to check the patient's out-of-pocket cost and coverage status of any medication and immediately select the order.
Today, RxRevu has expanded its footprint by supporting prescribing guidance for roughly 80 percent of UCHealth's prescription decisions, and in the form of expansion initiatives across all of primary care and virtual care.
The market for electronic prescription and Rx decision support software is relatively large. Technology vendors include AdvancedMD, DrFirst, Imprivata, MDToolbox, Practice Fusion, RxNT and Surescripts. And most of the major EHR vendors offer e-prescribing and related functionality.
MEETING THE CHALLENGE
UCHealth uses Epic for its EHR. SwiftRx is not technically within Epic but is a separate system that appears to the clinician to be integrated in such a way that the provider behaves as though it is within their EHR workflow.
"SwiftRx functions by analyzing key data points from the patient's medical record within the EHR and recommends pharmacotherapy based on the provider's impression or diagnosis," said Richard Zane, MD, chief innovation officer at UCHealth.
"For example, if I diagnose pneumonia, the tool will recognize if I am treating community-acquired pneumonia, if the patient has comorbidities that affect the treatment of pneumonia, like emphysema, if the patient has allergies, and what the antibiogram for my hospital and region are," he said.
The system then recommends antibiotics and other medications like bronchodilators, antipyretics, steroids and so forth, if indicated, he added.
"Using the example above, the physician sees the patient as she normally would, makes a diagnosis of pneumonia, and clicks the SwiftRx tab in the discharge workflow," Zane explained. "SwiftRx will recommend the medications as first-line, second-line, etc., and note any allergies. The physician then clicks on the recommendations, and the physician's orders are automatically populated."
The physician signs them as usual, and they are sent electronically to the pharmacy; the prescriptions are recorded in the EHR. It's essentially easier for the physician – fewer clicks, easier workflow, less data to gather – and there is a significantly higher chance of getting the medications right the first time, he added.
"We will pair an engineer with a provider. If something doesn't work, is clunky or not intuitive, it gets fixed immediately, and the providers get feedback."
Richard Zane, MD, UCHealth
UCHealth today supports prescribing guidance for roughly 80 percent of the health system's prescription decisions. Further, prescriber acceptance of preferred medications within the RxRevu tool has increased from 44 percent at launch to
74 percent today.
"We have five universally unalterable guiding principles when working with technology partners who wish to co-develop clinical decision support, or clinical anything for that matter," Zane said.
First, whatever is designed must function within the provider's EHR workflow; second, the tool must be easier to use and not harder; third, there must be fewer clicks and not more; fourth, there can be no "force function," meaning no hard stops or many added steps to avoid the tool; and fifth, it has to work all the time, Zane explained.
"Essentially, the tool must be the path of least resistance,” he said. “We work on the premise that providers are smart, they all have smartphones and PCs, and that using and deploying a new clinical tool has to be as intuitive as downloading a new app."
The IT team essentially tells clinicians a tool exists and what it does, and asks them to try it twice. If they don't use it or like it, the team dives deep and deliberately into optimization and iteration.
"We will pair an engineer with a provider and iterate minute to minute and day to day," Zane explained. "If something doesn't work, is clunky or not intuitive, it gets fixed immediately, and the providers get feedback."
ADVICE FOR OTHERS
Zane said it's important to have a set of core guiding principles when deploying electronic clinical decision support.
"Ours are absolute and non-negotiable," he said, referring to the aforementioned five principles. "It must be bulletproof and accurate 100 percent of the time."