Titin.net

Link to the original article.

Authors conducted task analysis and focus groups to study a Longitudinal Medical Record (LMR). They studied the usability of one of these LMR’s modules. This module is the results manager. This module allows physicians to follow up on their patients’ laboratory results. Physicians would assess these results and prepare letters for their patients. Based on the inputs from these two qualitative studies, the authors proposed modifications to the existing results manager module.
It was interesting for me to realise that the criticism I have to our totally different Electronic Medical Record were similar to the ones raised by physicians in this study. This means that systems’ developers are similar and though make similar systems. Here is a summarized list of findings with some commentary:

1. For a clinician to accomplish a simple task s/he needs to go through many screens and access different modules. This can lead to “loose of overview”.

2. Physicians do care about response time. The slowness lead one physician to open multiple screens. This can lead to confusion as confirmed by this physician. The slowness of contemporary EMRs could be related to the reliance on relational databases. Relational databases do provide concurrency. But, concurrency is not as important in health care. Patients are at one place at one time and only one health care professional is handling their clinical data at a given moment. So, the use of different database models as XML files to handle patients clinical data can be a better solution.

3. Balance of having more information on the screen and yet not looking too crowded. This is related to the first point. The need to go through different screens to gather necessary information can lead to loss of overview. The other issue, screens being too busy can lead to increase cognitive load. We need to decrease our cognitive load to have more attention resources to place on the task at hand to reduce the likelihood of mistakes. Cognitive load theory defines the amount of “mental energy” required to process a given amount of information. When screen elements are too close, we require more time and attentional cognitive resources. And according to the authors these two elements are lacking in the clinical user’s multi-tasking workflow and interrupt-driven environments. Solutions to this problem could be to have scrollable screens and more white space on the screen. Having white space around elements can draw physicians’ attention. The authors also suggest the use of visual hierarchy, where related information is grouped to gather and using intellengent information retrieval to display only the information related to the physician current task.

4. One interesting observation is that physicians did not use the available letters’ templates as they felt these templates were not enough for their needs. This is a settle but an important point. There are big projects as openEHR that rely on temples. Templates are seen as a method to standardise electronic medical data. This standardization will make interoperability between different medical systems easily developed and once developed much more reliable. But, we should be careful were to use and where not to use templates.

5. A valuable advice by the authors is to discuss with users during training how will the new system be integrated in the workflow

6. The modules used while patients are with physicians were not tested. Testing these would be more challenging and would raise ethical and patient confidentiality concerns. Focus groups could be a way. Yet this study does shed light on a vital area of health informatics.