138 - Improving Hospitalized Patient Sleep by Minimizing Unnecessary Overnight Vital Sign Measurements

Poster Number: 138
Publication Number: 138.1482

Background: Despite its importance in illness recovery, the sleep of hospitalized children is frequently interrupted. Vital sign (VS) measurements are the most common caregiver-reported disruption.

Objective: This quality improvement intervention aimed to increase the proportion of patient-nights with fewer than two overnight VS measurements by 20% within six months.

Design/Methods: This study included all patients admitted to hospital medicine services at a university-affiliated children’s hospital. Overnight was defined as 10PM to 6AM. Preparation for and informal discussion of the project began 10/25/22 with a project go-live 2/16/23. The intervention included provider education, use of a rounding checklist, VS order changes, and admission and disease-specific orderset changes. Outcome measures were overnight VS and blood pressure (BP) measurements. Process measures were non-every 4-hour (non-Q4H) VS and BP orders, and the balancing measure was rapid responses. We used statistical process control charts and defined special cause variation as 8 points above/below the mean for our outcome and process measures. We used Chi-square or Fischer’s exact tests to compare our balancing measure and demographic characteristics. We assessed for inequities based on language of care and race/ethnicity using a generalized linear mixed effects model.

Results: Our pre-intervention period included 4401 patients and 16132 patient-nights between 1/1/22 and 2/15/23 and our post-intervention period included 1732 patients and 6991 patient-nights between 2/16/23 and 7/31/23. Table 1 includes patient demographics pre- and post-intervention. Patient-nights with fewer than two VS measurements increased from 13% to 23%, while those with fewer than two BP measurements increased from 39% to 58% (Fig 1). Non-Q4H VS orders increased from 13% to 24% (Fig 2). Post-intervention, 54% of BP orders were non-Q4H. There were fewer rapid responses post-intervention [pre-intervention, n=98 (0.61%) vs post-intervention, n=7 (0.10%), p < 0.001]. Pre-intervention, there were no differences in VS measurements based on language of care or race/ethnicity; however, post-intervention, non-Hispanic white patients had a higher odds of fewer than two overnight VS measurements (OR 1.54, 95% CI 1.23-1.92).

Conclusion(s): Following implementation of a multi-pronged intervention, there were fewer overnight VS and BP measurements without an associated increase in patient deterioration events. The potential of this intervention to introduce disparities requires attention.