Your mission, should you choose to accept it (queue music): take responsibility for the orchestration of over a thousand healthcare experiments running simultaneously. To do this, your team has been granted access to robust technology that runs experimental trials 24 hours per day, 7 days a week, 365 days a year. These assessments are carried out across every medical specialty, engage all members of the care delivery team, and touch virtually every process in healthcare.
Intrigued? Good. These experiments are already deployed live today within your health system, in a production environment, and involve real clinicians and patients at the point of care. And optimizing processes based on findings is an opportunity you don’t want to miss.
The experiments referenced here come in the form of electronic clinical decision support, or CDS, and interventions span a broad range of tools within the electronic health record system. For example, the intervention may be an alert that interrupts a physician with information about a patient’s risk factor for a blood clot, an order set that encourages application of all the appropriate evidence-based orders for a stroke patient, a documentation template that pulls in lab results that can change a patient’s care plan, to name a few.
I consider these experiments because there is abundant evidence that the way users interact with these tools is highly variable and can fail spectacularly. In other words, the teams that create these support tools aren’t actually sure what will happen after they design and implement them. Will end users misinterpret the text? Bypass the planned workflow altogether? Get increasingly frustrated with IT based on poor design or workflow integration?
To better understand why these experiments are a core element in creating better outcomes, I want to compare them to the practice of prescribing a drug to a sick patient. Imagine the practice of two otherwise identical physicians. The first clinician prescribes an antibiotic and sends the patient home without any follow up, while the second clinician prescribes an antibiotic and checks in a week later to assess whether symptoms have resolved. The latter exemplifies the iterative nature of driving improvement in the imperfect and complex environment of medicine where an outcome isn’t guaranteed. In essence, this reflects a plan-do-study-act (PDSA) cycle in the delivery of healthcare at the level of clinician and patient. While the aforementioned first physician’s approach has a chance of working out just fine, it is not hard to recognize that the lack of follow up may delay an appropriate adjustment in management and lead to patient dissatisfaction. Which clinician would you want to work with?
Drugs and devices undergo a rigid approval and monitoring process prior to being authorized for patient use. By the time they are leveraged to a live healthcare setting, their likely impact is well known. In contrast, the ongoing monitoring and evaluation of CDS intervention post launch is imperative because, although they have been experimentally validated in several settings, their almost limitless permutations of workflows and available tools render their efficacy highly contextual. There is no question that ongoing monitoring is harder work than just building the intervention and moving on. But it is not an impossible mission! Platforms like Phrase Health exist to make this evaluation more accessible across stakeholders within an organization.
Health systems risk poor quality outcomes, frustrated clinicians, and increased costs due to variation when they underinvest in actionable insights around CDS performance and the clinical informatics teams that orchestrate it. Those who prioritize ongoing monitoring and iterative improvement based on learnings enjoy higher quality outcomes, reduced clinician burnout, and improved ROI. It doesn’t take Secret Service agent instincts to see the value of investment there.