Health Data

Alert Fatigue: An Unintended Consequence

This blog explores health informatics—a collaborative activity connecting people, process, and technologies to produce trusted data for better decision-making.


By Barbara A. Ryznar, RPh, MSHI, RHIA, CHDA, CPHI, CPHIMS, CAPM

 

I ignore many of the routine alert messages embedded in the software applications I use regularly for my work, automatically clicking right on by. The purpose of these prompts is to guide the user through the application process and support processing decisions. Some alerts are for business tasks as we bill in real-time; some are sparked from triggers within the clinical decision support functions. A single transaction can elicit numerous messages intended to grab the user’s attention and produce an appropriate response. The term used to describe this desensitization to these messages is alert fatigue, and I admit that I have it. And in reality, some of these alerts should be tuned out as they are not significant and disrupt the user’s thought process. However, it is important that the critical ones be addressed and not become lost in the sea of messaging.

I have used a multitude of pharmacy software systems throughout my career. These merge business functions and clinical functions with prompts, messaging, and alerts that seem to occur at every step. A variety of messages are baked into the software design to ensure accurate data is captured, for billing compliance, and to promote clinical decision-making for patient safety. But there are many that are false alarms that break the user’s cognitive flow and slow the process with multiple or redundant questions that require a response to proceed. Those that are business-dominant serve to support quality data at the point of entry and compliance with billing requirements, which are both important to satisfy the functional requirements of the software. For the novice user, the benefit of this step-by-step guidance can be helpful; but for the experienced user, they are a bit of a nuisance more likely to be tuned out and clicked through to get to the next screen.

When electronic prescribing came into the picture, a new message appeared at the bottom of the screen to alert the user that new e-prescriptions were present with a message, “Do you want to process them now? (Y/N).” A response is needed before any activity can continue. This is my pet peeve with the system because it is as though it wants to change my workflow to address the new arrivals when I am attempting to complete the tasks that are of higher priority—namely, the waiting customers. I wish that I could turn this alert off; it’s a task that I will get to when I have time available, not necessarily in the immediate moment. In other words, the system should work WITH me and not try to dictate the direction of my main concerns. On the flip side, there are steps that could use a prompt to smooth workflow and to prevent the need for reprocessing.

When it comes to clinical decision support, there are also alerts that are both significant and insignificant. The system has a built-in safety feature that produces a “hard halt” that prevents further processing until the high alert flag has been addressed and resolved appropriately. I like this feature as it guarantees that the most important drug-drug interactions or allergies must be addressed. Other systems that I have worked with use color coding with “red alerts” being the most serious and mandatory for response. The value of this aspect of the software relies on the accuracy of the patient-specific data available. For example, if the patient’s allergy record is not up to date, the trigger is not present. This circles back to the importance of the accuracy at data entry and the importance of having current information resources in the clinical support system.

So, all the prompts, alerts, and messages are necessary to ensure that quality data is entered at the source to meet the purposes of use and reuse downstream. Some software applications are just better designed to deliver a user experience that promotes both accuracy and efficiency in the process. A major factor that has made one system I’ve had experience with so laborious to use is that it is “old” regarding technology and has been upgraded several times to meet new requirements. While this is good for business, each upgrade has brought new screen, prompts, and cognitive load. Applications with built-in alert features and quality of the clinical decision support functions can significantly affect productivity and the user experience. Informatics plays a vital role in designing a system that can prevent disruption of the user’s cognitive flow while still ensuring that data collected at the source of entry supports both speed and accuracy. Perhaps new systems in the near future will achieve a better balance.

 

Barbara A. Ryznar is a community pharmacist at a Medicine Shoppe Pharmacy and a part-time adjunct instructor in the MSHI program at University of Cincinnati.