A real-life example is the best way to show the value of a good order set within a locally programmable electronic medical record (EMR). Of course, any EMR that is not locally programmable would seem categorically untenable to me – no two organizations are alike on a host of internal and external measures that require a flexible EMR solution.
Here’s a great story of how one large multi-hospital Allscripts client—as part of its continuous efforts to improve care—helped clinicians and the organization help stroke patients leveraging local programmability. They had successfully addressed a host of clinical challenges one at a time, and next turned specifically to better serving patients with new-onset stroke.
Using best practices more consistently
First, the 13 generally accepted, evidence-based best practices for treating stroke patients provided the goal. For example, completing a lipid profile within 24 hours, or using the Barthel Index and modified Rankin Scale for assessing severity and thereafter improvements or downturns.
Next, the organization extracted retrospective data from Sunrise, using commonplace SQL queries, to learn how often clinicians completed each practice. They were focused on improvement, so they were not surprised to find several areas for improvement.
For example, the data showed that Barthel and Rankin evaluations occurred less than 10% of the time, to which one physician legitimately complained, “That is ridiculous; we always do those evaluations as part of our routine.” Upon further investigation, they found that low frequency was due to lapses in documentation (e.g., spelling) which resulted in missing coding, not subpar care.
Then we determined to re-design ideal workflow and documentation processes, and the organization actively included the full continuum of team members in the discussion: physicians, nurses, ER clinicians, admission staff and medical records clerks. They re-mapped the documentation process so that information would be added when and by whomever would ensure reduced possibility of missing anything, and avoid leaving it all up to the physician to manage.
As part of this process, the organization developed the stroke admission order set which ensured best data capture leading to ideal clinical decision support (CDS) and best percentages on the 13 criteria.
Clinical decision support improves results by 40% in two months
Within two calendar months the organization saw vast improvement:
Figure 1: The organization virtually doubled clinical results for stroke patients.
Beyond compliance measures, outcomes improvements for stroke patients also included:
- 35% fewer 31-day readmissions
- 9% more patients discharged to home, rather than rehab or skilled nursing facility (SNF)
- 7% reduced length-of-stay (LOS)
- $250,000 to $500,000 annual savings for stroke-related discharges
What’s good for the organization is also good for the patient
The organization was getting patients out of the hospital and better completing documentation requirements. But we wondered whether we improved outcomes for patients after they left the hospital or just discharged them more efficiently?
Deeper post-discharge studies showed that readmissions were significantly reduced (p<0.01) for home-discharges and SNF patients as well, and other cost-related utilization for ED, clinics and primary care physicians also fell significantly. In fact one third as many readmissions from home or SNFs. SNF impacts were so favorable that the managing director of one SNF estimated she could admit an additional 130 stroke cases annually without changing staffing or bed counts.
Users of the Sunrise EMR can improve organizational processes and CDS capabilities leveraging the free Stroke Outcomes Toolkit. They can either adopt approaches proven elsewhere, customize their own approach reflecting Outcomes Toolkit and similar strategies, or bring in Allscripts experts if needed for a customized result.
It’s time to stop using financial data to inform clinical decisions
Of course, this example depends on using Sunrise as the EMR. These achievements happened because the organization could program the EMR to match specific needs and continuous improvement imperatives. These successes are only genuinely achievable with programmable EMRs, with data extraction capabilities to evaluate improved efficacy and care.
Otherwise organizations are stuck with old-school 30- to-90-day data lags for analyzing primarily financial data to determine what needs to get better, or whether it did. Old-school approaches don’t improve care “now” for real patients and their clinicians.
CDS must be a real part of the way we care for and monitor patients to achieve a real, substantial and statistically significant difference in quality and outcomes. Programmed order sets help physicians do the right thing, consistently, the first time – true CDS. Capturing needed documentation elements sooner with minimized burden to physicians, leads to best outcomes and ideal organizational use of the real power of good, programmable, adaptable EMRs.