Background: Hospital-acquired delirium is a common cause of increased morbidity and mortality. In an earlier pilot study, we identified through surveys that adult medicine inpatients slept an average of 4 hours. Also, multiple factors including environmental noise, vital sign checks, and lab draws were recognized as barriers to sleep. Based on this information and prior research, we developed and encouraged the use of a multifaceted delirium prevention protocol that can be optionally ordered by providers. The protocol includes clinical decision support and order options for providers and promotes nurse-driven delirium prevention interventions. We intend to measure the effectiveness of this new protocol in sleep improvement and reducing delirium-related complications such as falls, restraint use, and longer length of stay among elderly patients.

Methods:

This ongoing IRB-approved study takes place on a single adult general medicine floor. Inclusion criteria defined elderly patients as 60 years and older. A survey inquiring about the quantity and quality of sleep was administered to a convenience sample of 100 patients post-protocol and later compared to previously collected survey results of 100 patients pre-protocol. Aggregate outcomes including use of sleep aids, pharmacologic and physical restraints, and length of stay were obtained through EHR data queries, which 9 months of pre-protocol data was compared to 23 months of post-protocol data. We used chi-squared tests for differences in proportions and t-tests for continuous variables.  Data was analyzed by SAS software.

Results:

On average, 28% of patients have received the delirium prevention protocol since its activation in January 2015. Protocol use has increased monthly with provider education, most recently with 40% of eligible patients receiving the protocol during hospitalization. Compared to pre-protocol results, implementation of the delirium prevention protocol was associated with a trend toward increased amount of sleep: from 4.1 (95% CI 3.7 to 4.4) to 4.6 hours (95% CI 4.1 to 5.1), p=0.08. Quality of sleep was rated as an average of 3 on a 0 to 5 scale, which was unchanged in post-protocol data. Only 13% of the surveyed post-protocol patients were initiated on the protocol. Median length of stay, use of physical or chemical restraints, and sitter use was unchanged between pre-protocol and post-protocol patients. While benzodiazepine and zolpidem use was unchanged, the proportion of patients on melatonin increased from 8% to 42% (p<0.01).

Conclusions:

This delirium prevention protocol was partially adopted into admission workflow with provider education. It was associated with a trend toward increased amount of sleep and significant increase in the use of melatonin, a more appropriate sleep aid than benzodiazepines or zolpidem for the elderly. The “opt-in” nature of the protocol limited its utilization. The lack of significant improvement in sleep quality or quantity may have been due to the small proportion of surveyed patients who actually received the protocol. An additional limitation of this study was the lack of Confusion Assessment Method (CAM) scores to assess the change in incidence of delirium. The next phase of this study will involve tracking delirium incidence with q shift nurse-driven CAM scoring and reassess sleep and aggregate outcomes over time during which the delirium prevention protocol order will be system generated for 100% of our elderly patients.