Automated alerts generated using data from anesthesia information management systems (AIMS) are a promising approach to influencing the behavior of anesthesia providers, with the goal of improving care for patients undergoing surgery, according to a paper published in the September 2015 issue of Anesthesia & Analgesia.
However, developers must address a wide range of issues and concerns to ensure that alerts are reaching the right people, at the right time, to be effective in producing the desired changes, wrote Richard H. Epstein of Sidney Kimmel Medical College at Thomas Jefferson University, Philadelphia, and colleagues.
Modern AIMS routinely collect a vast amount of data on patient status and patient care. Clinicians and researchers are interested in mining these data to develop alert systems to improve key aspects of patient care and monitoring during surgery. Studies have shown that such anesthesia clinical decision support (CDS) systems can increase adherence to protocols and improve financial performance, but have yet to demonstrate improved clinical outcomes.
Designing and implementing such systems entails a “multitude of concerns” that must be recognized and addressed, to produce the desired effects. To illustrate these complexities, Epstein and coauthors shared their experience in developing two CDS interventions using AIMS data. The first case example concerns an effort to reduce fresh gas flow rates when using inhaled anesthetics. That reduction is important not only for cost control, but also to reduce the environmental impact of gases ventilated to the atmosphere.
In this case, a “post-hoc” approach was deemed an appropriate first step. An automated system was designed in which each anesthesiologist received a monthly email providing feedback on flow rates over his or her 10 most recent cases. On evaluation, fresh gas flow rates decreased significantly from before to after the alert system was implemented.
Other situations call for more immediate feedback to achieve the desired improvement. That’s illustrated in the second case example: an effort to reduce gaps in routine blood pressure (BP) during surgery. While measuring BP every five minutes is the standard, gaps of 10 minutes or longer are common.
This case required a “near real-time” approach, with alerts sent directly to the operating room workstation where the patient’s condition was being recorded. Evaluation found that the targeted improvements in BP monitoring were met, once alert intervals were decreased to six minutes. Over four years’ follow-up, the number of weekly alerts remained about the same, which suggested a “lack of long-term learning” and highlighting the need to continue the alert system.
Alert systems should not be implemented until their usefulness is confirmed, Epstein and colleagues emphasized. For example, they considered sending near real-time alerts regarding drops in blood oxygenation level, but found that most such episodes resolved within minutes, before the alert could be sent. “Lack of utility should be assumed until testing shows otherwise, even if a benefit seems apparent,” the researchers wrote.
Other concerns include possible unintended consequences of alert systems. For example, one study suggested that CDS reminders were leading to overuse use of anti-nausea drugs. Potential regulatory issues must be considered, such as the risk of running afoul of FDA rules on the use of CDS software or mobile devices. The authors also touch on the technical challenges of implementing and maintaining alert systems.
Although data collected by AIMS provide the opportunity to improve patient care in many ways, careful forethought and follow-up are needed to ensure that alerts are well-designed and effective, Epstein and coauthors said. “Our goal is to inform developers and users of CDS for AIMS about the multitude of concerns they should consider during development and implementation to increase effectiveness and mitigate potentially disruptive aspects of this technology,” they said.