Designing Effective EHR Alerts | SMART Toolkit

Most Electronic Health Records (EHRs) have the ability to display best practice alerts (BPAs) and clinical reminders to end users. Numerous studies have evaluated the use of EHR Alerts and how they impact the way clinical care is delivered. The result is clear: When EHR Alerts are well designed, they work.

Clinical Decision Support can help or hurt clinical practice.

EHR Alerts are a form of Clinical Decision Support (CDS) and include technology-mediated reminders, advisories and alerts. While EHR Alerts have been used to reduce the incidence of HAIs, poorly designed alerts have also been identified as contributing to alert fatigue and overload, which can lead to clinician frustration with the EHR.

This module offers a standardized language to allow individuals impacted by EHR Alerts to articulate issues that can be resolved through a robust design process. Effective EHR Alerts require collaborative efforts before, during, and after their design. High-quality EHR Alerts involve leadership and end users throughout the organization to ensure that the design process weighs the costs and benefits of the choices required to develop an effective EHR Alert.

Quiz: Which type of system has been found to have the lowest number of errors? Select one of the images below.

When humans and machines work together, each using their strengths toward a common goal, they create the most resilient systems. Researchers who specialize in human-computer interactions have noted that systems where humans or computers make decisions without the benefit of the other can result in outcomes that are less than ideal.

Human-led Systems

Humans are imperfect decision-makers, particularly during complex or stressful situations. Biases of all sorts have been documented by researchers, with negative effects on the quality of decision making. For example, confirmation bias, the dynamic that makes one see patterns in the data that reinforce a currently held belief, is just one of the factors that can reduce the quality of decision making. Biases, limitations in memory, the demand to multi-task - these all can contribute to poor decision-making. On top of that, the particularly challenging environment of healthcare, including factors such as high workload, high stakes, long hours, and fragmentation of care, also contributes to difficulty making decisions.

Machine-led Systems

Machines are imperfect decision-makers, particularly because they are unable to gather the nuance of the human experience. As tools, machines are exceptionally capable of storing and retrieving data from multiple sources, at an instant. Further, machines can be built to leverage multiple simultaneous sources of data. However, machines can be data rich and information poor. They can be home to vast quantities of data, but fail to manage the subjective well.

Human-machine Systems

Human-machine systems seek to leverage the strengths of both people and machines, while providing a buffer against the weaknesses. Well-designed human-machine systems can leverage the machine to cue the human to consider important information when making a clinical decision. The human takes information provided by the machine and, combined with their clinical expertise, makes an informed decision. Human-machine systems in the infection prevention context can include tools such as clinical decision support and EHR alerts.

EHR Alerts are one way in which humans and machines can work together to reduce HAIs. EHR Alerts should not be expected to replace clinical decision making, rather they are tools that support clinicians in making evidence-informed decisions in the care of their patients. The important thing to note is that a well-designed Alert can result in improved outcomes, and a poorly designed one can not only yield poor outcomes, but can also impact the outcomes of the well-designed ones. One framework for looking at the design of an EHR alert is the Five Rights of Clinical Decision Support.

Designing Effective EHR Alerts

The most powerful consideration in developing and improving an EHR Alert is attention to design. One of the best frameworks for thinking about the design of EHR Alerts is the "Five Rights model for CDS". This framework was developed in 2009 and focused on improving medication use through the use of CDS systems and has been refined since its initial design.

THE FIVE RIGHTS ARE:

  1. The Right Information
  2. The Right Person
  3. The Right Format
  4. The Right Channel
  5. The Right Time

The Right Information

Decorative information icon

In accordance with CDC recommendations, one hospital set an alert every 24 hours after the insertion of a Foley as a reminder for nursing staff to ensure the catheter is only in place as long as needed.

An EHR Alert should be evidence-based, derived from a set of recognized guidelines, or based on a national performance measure. As hospitals start their EHR Alert deployment, they should focus on alerts that have established best practice and are based on high-level clinical evidence. When EHR Alerts derive from evidence that is of lower quality, their implementation can be contentious and may undermine trust in both the particular alert as well as Alerts, more generally.

The Right Person

To create an environment where individuals are working at the highest level of their licensure, EHR Alerts should be directed to individuals with the minimum licensure required to address the alert. The use of team-based alerts can seem like a good idea, but they can contribute to alert fatigue.

Decorative silhouettes of people

The Right Channel

Decorative icon of megaphone with arrows pointing outward to people icons

EHR Alerts are often associated with other systems. For instance, patient portals and EHR-tethered smartphone apps for clinical and patient care, can be used to convey messages to patients about their need to practice good hygiene. While EHR Alerts are generally targeted to members of a patient's care team, effective Alerts need not be limited to the EHR. The goal of an alert is to raise awareness on the right technology to make a difference.

The Right Intervention Format

EHR Alerts can be delivered in a variety of formats: break the glass, hard stop alerts, slowdown alerts, pop-up messages, reminders, and many others. Each of these formats offer different behavioral nudges. For instance, hard stop alerts that require entry to move beyond have been shown to be very effective at shaping behavior but are also unpopular because of their impact on workflow. Effective EHR Alert design requires that implementers understand what formats are in their arsenal, the potential impact on clinical workflow, and take these into account when selecting the right tool for the problem.

Decorative icon of pencil and paper

An analysis of view times on click through alerts (e.g. a pop-up message that requires acknowledgement) has shown that clinicians can spend less than 0.1 seconds on the notice - suggesting that they may be less effective.

The Right Time in the Workflow

Decorative diagram of a gear and generic shapes

One of the hardest parts of building effective EHR Alerts is mapping them into the workflow. Further, it can be more difficult because the act of building an Alert can change the workflow itself. Designers have to be sensitive to the timing in the workflow to ensure that the alert is presented in a context-aware form. This principle has implications because timing can mean time in the process or time in the day.

In both of these cases, timing is everything:

Alerting too late in the process

A physician is trying to prescribe a medication that interacts with the patient's existing medications. The alert only appears after entering the pertinent information for the prescription, when the physician attempts to send the script to the pharmacy. An appropriately timed alert would present as soon as the medication is selected. Later alerts mean wasted effort.

Alerting too late in the day

A reminder to remove a Foley might be triggered for only one shift to isolate responsibility and ensure that the question is asked at a time when the appropriate clinical team is available to make that decision.

A common tension with health technology is between the desire to introduce new technology and its impact on current clinical processes. When considering the implementation of a CDS system, clinical workflows must be both understood and documented to ensure that the right information is delivered to the right person at the right time. As EHR Alerts are designed, getting them as close to the point where the least amount of effort is wasted should be the goal.

EHR Alerts should be built around specific goals and aims. Individuals designing alerts should do so with emotional intelligence, the input of the planned recipients of the alert and with particular attention to ensuring that the clinician recognizes their responsibility to make the decision. Flow charts and swimlane diagrams can be used to illustrate how the alert can fit into the clinical workflow.

EHR Alerts should be monitored to assess their impact and value, both quantitatively (measured against desired outcomes) and qualitatively (in terms of clinician adoption and utilization). If an alert triggers too many irrelevant interventions, the designer may need to go back to the drawing board and redesign the scope of the alert in the context of new information. The designer must also be cognizant of the unintended consequences of Alert design. Finally, successful EHR Alerts should be celebrated - they are hard to build, hard to get right, and are the results of significant efforts to help patients and clinicians receive and deliver the best care possible.

Roles and Responsibilities

Frontline Managers play a crucial role in ensuring that an alert is well received. They serve as clinical practice champion, educator, and clinical advocate to the designers. An EHR Alert designer wants their Alert to function well. Frontline Managers are responsible for providing feedback about the way an alert is implemented by using the Five Rights to express how they are or are not met in practice.

"The right BPA at the right time can save a life … The wrong BPA at the wrong time will be ignored by frontline staff within a week."

Frontline Managers are in a unique position to provide 360 degree feedback on making alerts effective. In some cases, they may need to provide coaching to the clinical staff on the rationale and use of the alert. In others, a Frontline Manager may need to reach out to the Infection Preventionist to discuss concerns related to the implementation of the alert recommendations or in response to questions on best practice. The Frontline Manager is the voice of the clinical team in ensuring that the alerts are their most effective.

Frontline Managers have first hand knowledge of what works - and what doesn't - in the real world, they play a critical role in the success of high quality alerts. Frontline Managers can help EHR Alerts gain the staff buy-in needed to be effective by creating a process where alerts are validated with the frontline clinical care team and incorporating staff input. Without buy-in of both design and implementation, the effect (and intent) of the alert may be reduced. Frontline staff need a champion through which they can channel their issues and concerns because when individuals believe their concerns are heard, systems are more successfully adopted.

The identification of best practices that are well established, particularly in the context of HAIs, is an ideal role for Infection Preventionists. As clinical professionals with specialized knowledge about HAIs, as well as site specific knowledge about the hospital itself, Infection Preventionists play a critical role in the identification, design and evaluation of effective EHR Alerts. Infection Preventionists keep up on the latest research and clinical practice guidelines for the organization and can provide the recommendations necessary to outline the proactive steps that a healthcare organization should be taking to reduce CAUTIs and CLABSIs.

Reducing HAIs to zero is the mission of every Infection Preventionist. BPAs offer one way for Infection Preventionists' to share their knowledge at the point of care.

High-quality EHR Alerts can augment clinical decision making with helpful, timely and directed guidance concerning specific policies and procedures that contribute to improved patient care. They aim to promote and facilitate positive change or recognize excellence rather than merely to codify current accepted practice. The Infection Preventionists have a particular responsibility to identify and follow the evidence to guide clinical practice.

Poorly designed advisories provide little to no clinical value. As the impact of specific EHR Alerts are evaluated and subsequently adapted, clinicians can feel like they are trying to hit a moving target. Clinical Leadership plays a critical role in ensuring the message is heard - that the organization places a premium on reducing HAIs and that EHR Alert adaptation is simply part of the process. EHR Alert designers need to work with leadership to identify goals, understand the current state of the world, define what success in this area looks like, create solutions that bridge this gap, and test these solutions with actual users of the system. In this context, leaders can heighten awareness, work to facilitate acceptance and reinforce that the design process is informed by those who experience the Alerts.

Hospital Administrators provide material and organizational support, delegate authority and remove obstacles when building a culture that accepts the use of EHR Alerts. Administrative Leaders are in a position to act as vocal champions for well-designed alerts, and notably can act as champions for data informed approaches to adoption. Making a well-designed BPA takes a lot of people working together and requires support at the leadership level to become successful.

Administrative Leaders can become BPA design champions

Support from Administrative Leaders can come in the form of hiring specialists tasked with working with clinicians, Frontline Managers, Infection Preventionists and Clinical Leadership to build a culture that accepts alerts as something that adds to the quality of care delivery. One role present in hospitals is the Clinical Informaticist or EHR Build Specialist. These individuals are specialists with technical knowledge about the EHR and domain knowledge about clinical care. Frequently, these individuals play a central role in advancing the use of alerts across the care system. Further, they can signal that it's ok for Clinical Leaders, Frontline Managers, Infection Preventionists to invest their time and talent in figuring out what is going to work. Administrative leaders also have to work to create partnerships with the organization's Information Technology department and leadership to execute effective alerts.

Remember, in a hospital setting, there is nothing that happens, big or small, good or bad, without the support of the administrators who oversee their implementation. Administrative leaders oversee the implementation for all hospital policies, and can lead the charge for meaningful, thoughtful and respectful alerts across the organization. In some cases, all an Administrative Leader has to do to support the change is remove the barriers to change.

EHR Alert Guides

EHR Alert Facilitator Guide

Portable Document Format (PDF)

Use the Facilitator Guide Script to support the presentation of the Facilitator Slides.

Download EHR Alert Facilitator Guide
View EHR Alert Facilitator Guide in new tab

EHR Alert Presentation

PowerPoint Document (PPTX)

The slide deck contains summary point images for the presentations and can be copied and pasted onto organizational templates. When transferring the deck to organizational PowerPoint templates, organizations are asked to ensure they incorporate the use of the SMART logo and acknowledgment in their presentation. The slide deck has been designed to incorporate notes from the website to provide ease of access.

Download EHR Alert Presentation

EHR Alert Case Study

Portable Document Format (PDF)

The Case Study is intended as a group exercise grounded in discussion. Individuals participating in a group where the Case Study is being used should be given a copy of the materials.

Download EHR Alert Case Study
View EHR Alert Case Study in new tab

EHR Alert Case Study slides

PowerPoint Document (PPTX)

The Case Study slide deck presents the questions from the case study for use during meetings.

Download EHR Alert Case Study slides

Works Cited & References

  1. Hollnagel, Erik. (2005). Designing for joint cognitive systems. 47 - 51. 10.1049/ic:20050450. Cognitive systems engineering (CSE).
  2. Woods, D. D. (1985). Cognitive Technologies: The Design of Joint Human-Machine Cognitive Systems. AI Magazine, 6(4), 86.
  3. Merkert, Johannes; Mueller, Marcus; and Hubl, Marvin, "A Survey of the Application of Machine Learning in Decision Support Systems" (2015). ECIS 2015 Completed Research Papers. Paper 133.
  4. Zerilli, J., Knott, A., Maclaurin, J. et al. Algorithmic Decision-Making and the Control Problem. Minds & Machines 29, 555-578 (2019). https://doi.org/10.1007/s11023-019-09513-7
  5. Tversky A, Kahneman D. Judgment under uncertainty: Heuristics and biases. Science. 1974 Sep 27;185(4157):1124-31.