Effect of predefined order sets and usability problems on efficiency of computerized medication ordering.

Computerized physician order entry (CPOE) systems are widely used in healthcare to streamline medication ordering. However, the efficiency of these systems can vary based on design and user interface characteristics. Prior research has shown that streamlined workflows and intuitive interfaces reduce task completion time and errors. That uncertainty drove this study to explore how predefined order sets impact ordering efficiency and which usability issues most hinder performance. No prior work had resolved the specific contribution of predefined order sets and the relative impact of different usability problem types on task efficiency. This gap motivated a controlled experiment to quantify these effects in a simulated clinical setting. Existing studies often focus on system adoption or error rates, but few isolate the efficiency of task execution. This paper's contribution lies in its direct measurement of mouse clicks and keystrokes as proxies for workflow efficiency. The study builds on prior work by linking usability heuristics to measurable performance metrics.

The aim of this study was to evaluate how predefined order sets affect the efficiency of computerized medication ordering. The specific problem addressed is the lack of clarity on which usability issues most significantly impact task efficiency in CPOE systems. The motivation stems from the need to optimize clinical workflows and reduce unnecessary user effort. The study sought to quantify how much predefined order sets can reduce redundant actions like mouse clicks and keystrokes. Additionally, the researchers aimed to identify which usability problems most strongly correlate with inefficiency. This work builds on prior research that suggests usability issues can hinder system adoption and performance. The study's design allowed for a direct comparison between two ordering methods in a controlled setting. The results could inform interface design improvements to enhance clinical workflow efficiency.

The study used a crossover design in a laboratory setting with a simulated CPOE system. Participants were physicians who performed medication ordering tasks using two methods: with and without predefined order sets. The number of mouse clicks and keystrokes was measured for each task. The excess number of actions was calculated as the difference between what each physician used and the minimum required. Usability problems encountered during the tasks were recorded and categorized using Zhang et al.'s heuristic principles. Statistical analysis compared the excess number of actions between the two methods. The effect of each usability problem category on efficiency was analyzed using regression models. The study controlled for individual physician variability by measuring per-participant performance.

The method with predefined order sets reduced the median excess number of mouse clicks and keystrokes by 6.2 times compared to the method without them (p<0.01). Vague and erroneous system messages increased the excess number of actions by a factor of 2.62 (95% CI 2.24–3.07). Unfamiliar language and terminology increased it by 1.28 (95% CI 1.14–1.43). Non-informative system feedback increased it by 1.15 (95% CI 1.03–1.28). Other usability problem categories had minimal impact on efficiency. The use of predefined order sets significantly improved task efficiency. The most detrimental usability issues were related to unclear or misleading system communication. The statistical significance of these findings supports the importance of addressing specific usability flaws.

The authors propose that predefined order sets can significantly improve the efficiency of computerized medication ordering by reducing redundant user actions. The study suggests that vague system messages, unfamiliar language, and non-informative feedback are the most detrimental usability issues. These findings imply that optimizing system communication can enhance workflow efficiency. The results support the idea that usability problems can have a measurable impact on clinical task performance. The study does not claim that all usability issues are equally important, only that specific categories have a significant effect. The authors do not assert that predefined order sets are essential for all systems, only that they can improve efficiency in this context. The findings may suggest that interface design should prioritize clarity and consistency in system messaging and terminology.