Alerting to Action: Intelligent Clinical Decision Support Lowers Costs, Improves Outcomes

The movement from fee-for-service to pay for value means many things for health system administrators, not least of all the need to provide affordable, high-quality care beyond the four walls of the hospital. It also drove, thanks in part to the HITECH Act and the need to meet Meaningful Use criteria, the implementation of expensive electronic health record (EHR) systems. The latter was intended to make the former possible – by investing in EHR systems, health system leaders hoped to have the data necessary to make value improvement actionable.

Shifting focus to overall value spurred hospital administrators to reduce unnecessary variation in care, part of the estimated 30 percent of all Medicare clinical care spending that is considered wasteful. In the ambulatory setting – now part of the overall value equation for cross-continuum care – health system leaders also needed a way to assist clinicians in making real-time decisions based on scientific evidence. But the EHR alone, as it was soon realized, did not offer health systems the analytics or actionable insights necessary to succeed in a value-based environment.

Many EHRs fall short in providing guidance to clinicians within the workflow – at the point of decision-making. A meta-analysis of clinical decision support (CDS) systems concluded that by providing context-specific information at the point of care, the odds of providers adopting guideline recommendations are 112 times higher.

Putting CDS to the Test at Cedars-Sinai

That’s where Stanson Health, a Premier company, comes in.

Cedars-Sinai Medical Center (CSMC) of Los Angeles, California, is an 886-bed tertiary academic medical center with both inpatient and ambulatory settings and services. Citing challenges with driving down costs, length of stay and variability in practice, Scott Weingarten, MD, MPH, Senior Vice President and Chief Clinical Transformation Officer of CSMC, believed real-time, intelligent CDS was the key to clinical improvement.

“EHRs make it possible to alert or nudge a physician during the actual decision-making process, as part of their normal workflow. That’s critical for change management. Education or retrospective data alone has a significantly lower chance of impacting clinical decision making, and may be a less successful strategy to measurably improve quality and safety of care and reduce costs. Altogether, education, data and real-time feedback have been shown to improve the quality and value of patient care,” said Dr. Weingarten, who is also the Co-Founder and Chairman of the Board for Stanson Health.

“Coupled with progress in artificial intelligence and machine learning, we are now at a point where intelligent decision support – that learns based on what alerts physicians actually follow and find useful – can be embedded within the EHR workflow. That gives us a tremendous opportunity to drive significant quality improvement and reduce undesirable variation in care.”

Stanson’s clinical team reviews available evidence – including Choosing Wisely guidelines along with many other sources of information – and makes it machine-readable and houses it in a content management system where it can be mapped to all the different data elements in an organization’s EHR (the major EHRs currently integrate with Stanson). The technology currently uses the native rules engine of the EHR to link the guidelines to the patients to whom it may apply, but guidelines are also being deployed into a cloud-based rules engine that pings physicians when new evidence is available.

Similar to the “You might be interested in this item” function of a retail e-commerce site, Stanson’s guidelines appear in the physician workflow as they input tests and treatment paths for a patient. If a physician prescribes an adult insomnia patient an antipsychotic, for instance, as a first-line intervention, the system will prompt the physician to evaluate an alternative based on available guidelines.

The physician then has the option to accept, override or ignore the prompt (see Figure 1).

CSMC implemented 92 alert-based clinical decision support interventions from ABIM Foundations’ Choosing Wisely campaign of more than 400 medical society recommendations into their inpatient and ambulatory EHR. In an observational study of 26,424 patient encounters and 18 of the highest-volume alerts, CSMC analyzed providers who were either always adherent to the alerts in the treatment group or never adherent in the control group.

Figure 1:

Encounters in the treatment group – in which providers followed all of the CDS guidelines – demonstrated significantly better outcomes compared to the control group, including:

• Lower readmissions rates: Patient encounters in which the provider was non-adherent to alerts were more likely to be readmitted within 30 days.
• Shorter lengths of stay: Non-adherent encounters had a 6.2 percent increase in length of stay compared to adherent.
• Lower complication rates: The odds ratio for complication increased by 1.29 for non-adherent versus adherent encounters.
• Lower costs: The total encounter cost increased 7.3 percent for non-adherent encounters versus adherent encounters.

Patient encounters that received all of the guidelines were associated with a nearly $1,000 savings for a median encounter.

“The results of our study show a strong association between Stanson clinical decision support and better patient care at a lower cost,” Weingarten noted.

And Dr. Weingarten sees Stanson’s clinical decision support as a happy marriage with Premier’s Performance Services business.

Working Together

Quality improvement analytics in PremierConnect^® complement Premier’s newly-acquired Stanson technology. PremierConnect analyzes outcomes based on physician, facility and procedure to show where improvement opportunities – in readmission rates, mortality, length of stay and utilization, for example – exist among an organization’s staff and service line. That data can be used to inform which physicians need coaching or where evidence-based standards need to be inserted into the provider workflow.

Data tracking what action the physician took has now been added to PremierConnect, creating a continuous, scalable and sustainable approach to outcomes improvement.

“Premier has done a spectacular job building robust data and analytics capabilities that can pinpoint where there is opportunity in a health system’s operational, clinical or financial processes. Its collaboratives, consulting and improvement technology, like Stanson’s, give providers and administrators the steps to drive those improvements home in practice.”