A doctor pulls up an adolescent’s electronic health record and the EHR triggers a recommendation to screen for depression. Positive findings from the screening are entered into the EHR and a tool for creating a shared care plan pops up along with an option to refer the patient to a mental health provider.
Welcome to the era of clinical decision support software. CDS tools, as in this scenario, can recommend things the doctor might been done anyway. Or they might suggest “best practices”.
In a PowerPoint designed to convince the C-suite of the benefits to be gained from clinical decision support, HIMSS (Healthcare Information and Management Systems) shows how a CDS tool designed to encourage the recommendation of H2 blockers for gastroesophageal reflux disease can change staff behavior.
In the eight weeks leading up to implementation of a CDS tool for alerting physicians to order this blocker, H2 prescriptions never comprised more than 20% of the total ordered. In the first week after implementation, H2 orders broke 80%. They rose to nearly 100% in week 7 before settling back to 95% in week 8.
In the case of this blocker, as in that of the pop-up advice regarding the adolescent patient, CDS is a strategic tool, one that can be wielded to achieve priority care delivery objectives.
Those objectives may be driven by the need to boost or maintain revenues by maximizing reimbursement; meet regulations set by Medicare; or improve care and safety. Any or all three may be involved. It depends on administrators and what they see as institutional priorities. One thing is certain: the involvement of CDS in patient care will increase.
More and more reimbursement is being tied to positive patient outcomes. CDS is being leveraged to improve the processes to achieve those outcomes. Meanwhile, under the guise of “meaningful use,” CDS has become the centerpiece of Medicare and Medicaid EHR Incentive programs.
From office visits to the emergency room, CDS is impacting American medical care. In one scenario, lab results and imaging studies are posting on the 30 patients in the ER. A multi-patient monitor updates patient status, highlighting which patients have new orders to process; which have abnormal results; which have been there for two hours…three hours…four hours; which of those have inpatient beds ready.
In myriad ways, CDS tools are not meant – nor can they – replace clinical judgment. They are intended to help staff make good decisions and make them quickly.
Plenty can go wrong. CDS tools are only as good as the administrators who decide what will be pop up; the staff who act on them; and the IT staff who implement them.
An intervention, nobly intended but poorly designed, can distract caregivers and disrupt workflow. Too many alerts desensitize staff, frustrating them with disruptions, causing them to ignore or bypass recommendations – the good as well as the bad.
Beset by “alert fatigue,” a provider will begin “to override or ignore further alerts without attending to them, which can decrease the care improvements expected from the tools.”
Considering what’s at stake it’s easy to appreciate the need for adequate planning and the allocation of adequate resources. This applies not just when implementing CDS tools but keeping them up to date and functioning well. CDS tools are not “something that can be implemented and forgotten.”
These considerations lead to the three credos of CDS:
- Speed is essential. Efficiency is king. Too many mouse clicks will kill a CDS tool.
- Be selective. Too many alerts will disrupt workflow and lead to “alert fatigue.”
- Find a CDS champion – or two. Leadership is needed in the C-suite as well as the trenches. Administrators have to support CDS, just as users have to buy into its use.
To achieve its potential, CDS will have to be effective and efficient. Its tools must make work easier.
As we enter this transition to CDS, alerts and recommendations have to be trimmed to no more than will annoy the medical staff. They must become part of accepted work patterns, injecting improved effectiveness and efficiency through the use of best practices.
Ultimately, they must become transparent, transformed into automated tasks that streamline workflow.
The trick will be creating this kind of automation without diminishing human control over the processes – or at least doing so in a way that does not upset human sensibilities.