Observation chart design features affect the detection of patient deterioration: a systematic experimental evaluation

• Published in: Journal of advanced nursing Vol. 72; no. 1; pp. 158 - 172
• Main Authors: Christofidis, Melany J., Hill, Andrew, Horswill, Mark S., Watson, Marcus O.
• Format: Journal Article
• Language: English
• Published: England Blackwell Publishing Ltd 01.01.2016; Wiley Subscription Services, Inc
• Subjects: Accuracy; Adult; Australia; Charts; design; deterioration; Disease Progression; Early Diagnosis; Equipment Design; Factorial experiments; Female; human factors; Human factors research; Humans; Male; Medical Records - standards; Nursing; Observation - methods; observation chart; Response time; subjective judgements; systematic; Vital Signs; Australia; observation chart; subjective judgements; deterioration; human factors; design; nursing; systematic
• ISSN: 0309-2402; 1365-2648
• DOI: 10.1111/jan.12824

Aim To systematically evaluate the impact of several design features on chart‐users’ detection of patient deterioration on observation charts with early‐warning scoring‐systems. Background Research has shown that observation chart design affects the speed and accuracy with which abnormal observations are detected. However, little is known about the contribution of individual design features to these effects. Design A 2 × 2 × 2 × 2 mixed factorial design, with data‐recording format (drawn dots vs. written numbers), scoring‐system integration (integrated colour‐based system vs. non‐integrated tabular system) and scoring‐row placement (grouped vs. separate) varied within‐participants and scores (present vs. absent) varied between‐participants by random assignment. Methods 205 novice chart‐users, tested between March 2011–March 2014, completed 64 trials where they saw real patient data presented on an observation chart. Each participant saw eight cases (four containing abnormal observations) on each of eight designs (which represented a factorial combination of the within‐participants variables). On each trial, they assessed whether any of the observations were physiologically abnormal, or whether all observations were normal. Response times and error rates were recorded for each design. Results Participants responded faster (scores present and absent) and made fewer errors (scores absent) using drawn‐dot (vs. written‐number) observations and an integrated colour‐based (vs. non‐integrated tabular) scoring‐system. Participants responded faster using grouped (vs. separate) scoring‐rows when scores were absent, but separate scoring‐rows when scores were present. Conclusion Our findings suggest that several individual design features can affect novice chart‐users’ ability to detect patient deterioration. More broadly, the study further demonstrates the need to evaluate chart designs empirically.