Physician wellness is top of mind for health systems. The influx of generalized clinical alerts within EHRs, particularly medication-related warnings and built-in-house alerts, exacerbates stress among physicians.
During its EHR transition, Florida-based NCH Healthcare System optimized medication alerts frequently dismissed by physicians due to patients’ clinical status by harnessing specific patient data, including lab results and comorbidities. This resulted in targeted alerts for more at-risk patients instead of generic alerts for a broad swath of patients.
This case demonstrates the value of proactively mitigating the impact of alert fatigue on physicians and partnering closely with the pharmacy team when evaluating medication alerts. The health system found it could incrementally achieve its objectives with limited resources and in considerably less time by leveraging an off-the-shelf system rather than building from scratch, enhancing physician wellness and patient safety.
David Linz, chief medical informatics officer at NCH Healthcare System, is going to offer an in-depth look into this technological feat at HIMSS24 during an educational session entitled “Introducing More Meaningful Medication Guidance While Tackling Alert Overload.” We interviewed him to get a preview of his talk.
Q. What is the overarching focus of your session? Why is this focus important to health IT leaders at hospitals and health systems today?
A. Optimizing clinical decision support guidance to address alert fatigue and its impact on clinician wellness is not only crucial for reducing negative impacts on healthcare professionals and improving their overall efficiency and wellbeing, but it is also vital for minimizing the financial burden on healthcare organizations and enhancing patient outcomes.
The influx of generalized clinical alerts within EHRs, particularly medication-related warnings and alerts built in-house, exacerbates stress among physicians. Alert burden and the resulting alert fatigue are not solely caused by high alert volume; lack of relevance and actionability for individual patients in specific clinical settings also contribute. While alerts are crucial for patient safety, volume and lack of patient-specific context contribute to physician burnout.
Excessive alerting can ultimately harm patient care by increasing the likelihood of avoidable adverse events. Alerts that lack clarity and actionable information also pose challenges. Generalized medication warnings are too often overridden, and persistent inaccuracy due to insufficient patient specificity can lead to inappropriate overrides by providers.
In a fast-paced clinical environment, alerts that fail to concisely identify patient risks and provide actionable next steps at the right time in the physician’s workflow can increase cognitive burden and frustration, further limiting a provider’s ability to effectively utilize the alert’s message. To improve medication alert acceptance, incorporating patient-specific factors such as laboratory values, age and comorbidities consistently has been found beneficial.
Q. What is the main learning you would like session attendees to walk away with? And how is this learning vital to healthcare and/or health IT today?
A. By taking a proactive, multidisciplinary approach guided by the pharmacy team, health systems can produce more meaningful medication alerts while tackling alert overload and alert fatigue.
Accomplishing this requires leveraging specific patient data, including lab results and comorbidities to make targeted alerts more patient-specific, clearly associated with patient risk, and actionable for clinicians. An average hospital spends 60 hours per month of ongoing maintenance on custom clinical decision support.
That estimate includes only reactive adjustments, such as if a clinician reported a problem with an alert or a patient safety issue was identified. Proactively maintaining the alert content so that it leverages patient-specific information so that it is more meaningful and actionable would likely double that time commitment, if not more.
Building and maintaining customized medication CDS alerts within the EHR was far too time-consuming for our internal IT staff, especially with the tasks and duties of an EHR system switchover, which we performed in 2022.
Our health system found it could incrementally achieve its objectives with limited resources and in considerably less time by leveraging an off-the-shelf system rather than building from scratch, enhancing physician wellness and patient safety.
We discovered we could not only reduce the alert burden for our physicians, but we could also add crucial alerts to further impact patient safety. We saw a 16.6% to 37.5% reduction in the optimized alerts per week being achieved while adding more meaningful guidance for clinicians.
Another benefit for our system by offloading medication alert management is it alleviates us from having to find and add pharmacy analysts to our staff who are experienced with CDS functionality in our EHR system and medication value sets.
In addition, since alert content is continually vetted and updated by our vendor partner, our system does not have to be concerned if the information contained is timely or relevant, which saves us time and helps protect patient safety, and reduces the risk of physician frustration triggered by an irrelevant, intrusive CDS warning.
The entire solution implementation process of four new targeted medication CDS alerts, including integration with our EHR system from the largest commercial vendor in the market, consumed only one hour a week for 12 weeks. In addition, each one-hour meeting was productive in that we collaboratively built our custom alerts in real time during the session, which saved significant time for our IT staff.
Q. What is one more significant learning you would like session attendees to walk away with? And how is this learning vital to healthcare and/or health IT today?
A. Implementation of a CDS optimization system is helpful, but ongoing review and adjustments are necessary to tackle the problem of alert fatigue. A decision support committee was formed to comprehensively review medication alert data and make mitigating recommendations.
Data continuously reviewed includes the quantity of alerts, the percentage of overrides, who the alert is firing to, and feedback from clinicians. The committee is composed of pharmacists, pharmacy analysts, physicians, nurses, clinical compliance managers, risk managers and subject matter experts from our vendor.
The cadence of meetings occur bimonthly with ad-hoc workgroups in between. Alerts can be trended to show improvement or regression, compared to peer organizations, and broken down by category.
Sorting alerts by categories, such as duplicates and subcategory analgesics, helped us identify a significant amount of non-clinically relevant alerts that clinicians provided negative feedback on. By changing our duplicate allowance from 0 to 1, we were able to reduce warnings per order by 25%.
We have plans to further analyze our opportunities to change duplicate allowance levels and subcategories. We anticipate this will improve provider satisfaction of the quality of warnings and enhance attention to subsequent alerts. Combining subjective feedback with objective quantification of alerts and then improvement is helpful continuously improving the CDS environment.
The session, “Introducing More Meaningful Medication Guidance While Tackling Alert Overload,” is scheduled for March 12, 1:30-2:30 p.m. in room W311A at HIMSS24 in Orlando. Learn more and register.
Follow Bill’s HIT coverage on LinkedIn: Bill Siwicki
Email him: bsiwicki@himss.org
Healthcare IT News is a HIMSS Media publication.