Assessing performance of an Electronic Health Record (EHR) using Cognitive Task Analysis.
This is a review of the article titled Assessing performance of an Electronic Health Record (EHR) using Cognitive Task Analysis.
Himali Saitwal , Xuan Feng, Muhammad Walji , Vimla Patel, Jiajie Zhang
International Journal of Medical Informatics 79 (2010) 501–506
This objective of this study was to analyze the user interface of the AHLTA (Armed Forces Health Longitudinal Technology Application) EHR system used by the US Military in order to evaluate its usability, and identify any possible areas for improvement.
The AHLTA system was considered difficult to use because of the effort needed to learn and navigate the system, according to Dr Cassells, Assistant Secretary of Defense of Health Affairs, which led longer workdays for clinicians who were also then able to only see fewer patients. EHRs can have slow performance usually down to two factors; design of the GUI(Graphical User Interface and system response times. This study investigated the AHLTA GUI independent of system response time.
This study used a Cognitive Task Analysis method called GOMS (Goals, Operators, Methods and Selection rules) and a method developed by Chung et al to analyze human cognition factors across task performance with the time required to accomplish task measured using KLM (Key-stroke Level Model) to evaluate the AHLTA system. 
The KLM Method
The KLM model estimated the time to perform each of 14 tasks via the standard 8 operators, of which seven were physical operators such as point mouse to target or double click, one was a combination of physical and mental operator and one was a mental operator. The modeling formulae used are outlined in the paper.
The total number of steps for a given task ranged from 41 to 466, with a mean of 106 steps for a task. The 14 tasks were also analyzed and classified as either physical or mental. The results showed that 37% of tasks were mental which had an estimated time of 1.2 seconds.
The range of estimated time for physical tasks was between 0.57 and 0.94 seconds, while the time taken to do the 14 task varied from 35 seconds to 6.5 minutes with mental operators accounting for 50% of the time.
This study showed there was a high number of steps combined with a high degree of mental processing needed to complete a task in the AHLTA system, resulting in the clinician being required to spend 22 minutes for data entry for any given patient that required all 14 tasks.
Such high levels of mental processing for EHR tasks can lead to mental fatigue and human errors. This study indicates that reducing the overall steps involved could reduce the mental steps and consequently the mental workload on the user. This reduction in steps could be achieved with better design such drop down lists instead of free text fields and better screen layout.
This was a very interesting paper, as it demonstrated how user interface design is essential to the usability of an EHR system and how reducing the mental workload for the user could aid ease of use and mental fatigue.
While the results were quite compelling, little information was provided about the sample size. In addition, further EHRs might benefit from such a study and it would be interesting to conduct this type of study on other EHR systems currently on the market.
In addition, users seek to perform any task in an EHR system in least amount of "clicks" possible. There are physicians that measure efficiency according to the number of steps or clicks a particular task involves. It is necessary that engineers and designers take this study in consideration when developing new systems or functionalities.
- Using Human-Centered Design Theory for EHR's
- Armed Forces Health Longitudinal Technology Application (AHLTA)
- Determining differences in user performance between expert and novice primary care doctors when using an electronic health record (EHR)
- The Cognitive Complexity of a Provider Order Entry Interface
- Chung et al 2003. An Extended Hierarchical Task Analysis for Error Prediction in Medical Devices http://www.ncbi.nlm.nih.gov/pubmed/?term=An+Extended+Hierarchical+Task+Analysis+for+Error+Prediction+in+Medical+Devices