A new CDS era
Enterprise EHR vendors are striving to improve their CDS performance monitoring capabilities, while new clinical knowledge management vendor solutions offer powerful tools to monitor, maintain and update CDS artifacts. And third-party CDS platforms are taking advantage of improvements in interoperability solutions, such as FHIR, SMART on FHIR, and CDS Hooks, to operate in real time on EHR data and return analytics-driven CDS recommendations directly into provider workflows.
The revolution is coming just in time. Clinicians are about to encounter a massive influx of new data types and volumes, including new kinds of patient-generated health data (PGHD), social determinants of health (SDH) data, and genomic data. Patients, payers, regulators, and clinician colleagues will soon expect providers to incorporate these data types into clinical decision making. It's not an exaggeration to predict that this tsunami of data will alter the standard of care across the terrain of medical practice.
To use these data effectively, clinicians will need in-workflow, pertinent, patient-specific information and recommendations at the point of care. Such information will have to be generated by intelligent systems using predictive and prescriptive analytics to interpret the data and devise guidance using combinations of AI techniques. In addition to effective data aggregation techniques and data governance, many organizations will need to reexamine their approach to CDS program management if they are to avoid the pitfalls of past CDS efforts.
The Achilles' heel of CDS
Historically, the Achilles' heel of CDS has been ineffective decision support and the resulting alert fatigue. Factors contributing to alert fatigue included EHR and drug database limitations, poor
CDS design, rushed implementations, ineffective governance and demand management, and lack of clinical knowledge management—especially CDS performance monitoring.
A robust CDS program can help providers avoid these pitfalls and utilize CDS mechanisms at their full potential. All too often, CDS focuses on fulfilling regulatory and administrative requirements rather than providing pertinent, patient-specific care advice. Too many "drug safety" alerts lack the specificity to make them effective, resulting in inappropriate alerting. Badly designed CDS interrupts the wrong provider—or the right provider at the wrong time and with the wrong mechanism. And far too many organizations have no data on the performance of their existing CDS: Are alerts being overridden? Acted on? By whom, under what circumstances? If an alert is not being acted on or an order set not utilized, why not? Lack of CDS performance monitoring seems to be the norm rather than the exception in the industry.
Stanford Health Care's improved CDS approach
Leading organizations have developed clinical and administrative governance mechanisms that help them select strategic objectives, prioritize effectively, make decisions, and adhere to the fundamentals of clinical knowledge management to build and maintain effective CDS programs. Stanford Health Care, for example, has implemented CDS performance monitoring that permits managers to review decision support adherence at the level of provider type, years in practice, diagnosis, patient insurance status, and more. Stanford has a mandatory three-month monitoring period for every new CDS element, after which stakeholders review CDS performance with the target clinical audience and determine how to improve alert performance. Stanford utilizes a CDS proposal evaluation form and process to ensure that only pertinent, appropriate CDS is built and added to their systems.
Approaches like this can help organizations both to improve their current CDS performance and to lay the foundation for the coming new generation of CDS demands and opportunities. You can learn more about effective CDS program management from our Best Practice Guide.
Download our cheat sheets
Get up to speed on the new and emerging technologies for today's digital health systems.