NUR 6310 Module 3 Usability Design Paper: A Sepsis Screening Flowsheet Built on Usability Principles Example

Seven Fields Instead of Twenty-Two: Redesigning a Nursing Sepsis Screen at an Englewood Hospital for Usability

Student Name

M.S. in Nursing, Nursing Informatics, Thomas Edison State University

NUR-6310: Nursing Informatics: Systems Life Cycle

Mentor Name

Month Day, Year

Seven Fields Instead of Twenty-Two: Redesigning a Nursing Sepsis Screen at an Englewood Hospital for Usability

The Problem

The hospital and every figure in this paper are invented. Two years ago a 360-bed hospital in Englewood added a sepsis screen to the nursing admission and shift assessment. The screen had 22 fields, including vital signs that were already charted elsewhere, a list of infection sources and a free-text comment. When certain combinations were entered, a pop-up alert told the nurse to notify the provider. Nurses described the screen as slow, and an audit found that 31 percent of screens were left incomplete. The alert fired on about one in five patients, and nurses said most of those patients did not have sepsis. In the past year, two patients with sepsis were recognized late on units where the screen had been completed but the alert dismissed.

The informatics nurse leading the redesign was asked to make the screen faster without making it less safe.

What Usability Means

Zhang and Walji (2011) proposed TURF as a unified framework for usability of electronic health records. In their definition, usability has three parts: the system must be useful for the work, usable in practice and satisfying to the people who rely on it. Useful means the system supports the work that needs to be done. Usable means it is easy to learn, efficient to use and error tolerant. Satisfying means users have a good subjective impression of it. TURF analyzes four parts of a system: the tasks it supports, its users, how information is represented on the screen and the functions it offers. The framework separates the work the domain requires from the extra work a poorly designed interface adds, which the authors called overhead.

Applied to the sepsis screen, the TURF view showed a large overhead. Re-entering vital signs that were already charted added work without adding information. The infection source list asked nurses to make a judgment better made by a provider. The free-text comment was rarely read.

Why Usability Is a Safety Issue

Howe et al. (2018) reviewed patient safety event reports from three health care systems over several years and identified reports that mentioned an electronic health record product and described a usability problem. Most involved medications, and a share of those reports described events that may have reached the patient and caused harm. The study showed that usability is not a matter of convenience alone: confusing screens, poor layout and hard-to-find information contribute to events that hurt patients. The authors argued for better usability testing by vendors and by the hospitals that configure their systems.

Heuristic Evaluation

Nielsen (1993) described heuristic evaluation, in which a few evaluators independently examine an interface against a short list of recognized usability principles, such as consistency, error prevention, keeping users informed of system status, minimizing what users must remember and avoiding unnecessary information. He reported that a small group of three to five evaluators finds most usability problems, making the method inexpensive enough to use early and often.

The informatics nurse, an informatics pharmacist and a usability analyst from the vendor each reviewed the current screen. They found 17 problems. The most serious were the duplicated vital signs, an alert with no explanation of why it fired, a dismiss button placed where the acknowledge button sits on other alerts and no way to see the patient's trend in heart rate or temperature on the same screen.

The Redesigned Screen

The new screen has seven fields. Temperature, heart rate, respiratory rate, systolic pressure and mental status are pulled from the flowsheet, with the last three values shown beside each so the nurse sees trends. The nurse answers two questions only: whether infection is suspected and whether the patient looks worse than at the last assessment. If criteria are met, the alert explains which ones and offers two choices, notify the provider now or document why not. The dismiss option was replaced by a required reason, chosen from a short list.

The informatics nurse tested the new screen with eight nurses from three units using ten written patient scenarios, five with sepsis and five without. Nurses completed the screen in a median of 40 seconds, compared with more than two minutes for the old one, and correctly escalated all sepsis scenarios. Two nurses found the reason list confusing, and it was reworded before go-live.

Working With the Vendor

Not every change was within the hospital's control. Pulling vital signs automatically required a configuration change that the vendor supported, but standardizing alert button layout across the whole record required a request to the vendor's product team, which will take a later release. In the meantime, the hospital changed the layout of its own locally built alerts and documented the inconsistency in its training materials. The informatics nurse also shared the heuristic findings with the vendor's usability team, since other hospitals using the same product likely face the same problems.

Design Rules for Future Screens

The project produced five rules for future nursing screens at the hospital: never ask a nurse to re-enter data the record already holds; show trends beside values; explain every alert; use the same button layout on every alert; and test with at least five nurses before go-live.

Limits

Scenario testing is not real practice, and nurses may behave differently at the bedside under pressure. The safety report study described possible harm but could not measure how often usability problems cause it. Fewer fields may miss rare presentations of sepsis.

Conclusion

The old sepsis screen carried a heavy overhead of unnecessary work and an alert that explained nothing. TURF showed where the overhead lay, Nielsen's heuristics found specific problems quickly and evidence from safety reports showed why usability matters for patients. The redesigned screen asks less of nurses and tells them more, and the design rules carry the lessons into future work.

References

Howe, J. L., Adams, K. T., Hettinger, A. Z., & Ratwani, R. M. (2018). Electronic health record usability issues and potential contribution to patient harm. JAMA, 319(12), 1276-1278. https://doi.org/10.1001/jama.2018.1171

Nielsen, J. (1993). Usability engineering. Academic Press.

Zhang, J., & Walji, M. F. (2011). TURF: Toward a unified framework of EHR usability. Journal of Biomedical Informatics, 44(6), 1056-1067. https://doi.org/10.1016/j.jbi.2011.08.005