Researching Efficient Approaches to Delirium Identification-Sustaining Effective Translation

This study will integrate daily delirium screening using the UB-CAM into routine care for every older adult admitted to 6 acute medical/surgical units at 3 hospitals in 2 states. The investigators will conduct a stepped-wedge implementation design, where UB-CAM screening will be launched at a new unit every 6 months over a 42-month study. The investigators will utilize evidence-based implementation strategies and test fidelity, accuracy, durability, and impact using a combination of: 1) electronically captured outcomes measured in all older patients admitted to the study units (administrative sample), 2) research RSDAs (Investigators will utilize masking by not making these RSDA delirium assessment determinations available to the clinical nurses) coupled with patient and care partner reported outcomes measured in a consented nested sample of older adults on the study units (patient-oriented sample), and 3) qualitative research using focus groups and 1:1 interviews with hospital staff, patients, and family care partners. The investigators propose the following Specific Aims, which will also be assessed in subgroups: older adults with AD/ADRD, those aged 80 and older, and their care partners.

Study Samples and Settings: The study population is hospitalized older adults admitted to 6 general medicine/surgery units at 3 hospitals in Pennsylvania and Massachusetts. There are 2 patient study samples:

Patient-Oriented Sample (N=1050) (Aims 1, 2): A non-randomized sample of patients admitted to the 6 study units will provide informed consent and undergo up to 3 reference standard delirium assessments (RSDAs), which will be used in the Aim 1 accuracy analyses. These patients and their family care partners will also be interviewed by telephone 1 month after hospital discharge to measure outcomes for Aim 2. An average of 1 patient per week per unit will be enrolled, or 25 patients over each 6-month study period. There are 6 units and 7 study periods, yielding a total of 1050 patients. Each patient will contribute at least 1 and up to 3 RSDAs. Based on READI, the investigators anticipate a 25% discharge rate each day, thus day 1 RSDAs will be available on 100% of the sample, day 2 on 75%, and day 3 on 56%, yielding 2362 paired nurse screen-RSDAs for Aim 1. The investigators will take advantage of the large eligible: enroll ratio (at least 10:1) to enrich this population for subgroups of interest. Specifically, The investigators will enroll 300 patients with AD/ADRD, and 250 patients from subgroups, which will enable subset analyses by sex, age 80 and older, and the presence of AD/ADRD. The investigators do this by preferentially approaching persons with AD/ADRD and those 80 and older among our eligible patient list. The MPIs used this strategy in READI, which achieved its targets of 35% AD/ADRD.

Administrative Sample (N=20,000) (Aim 3): This consists of all patients aged 70 and older admitted to the 6 study units throughout the study period. Outcomes for Aim 3 will be obtained from pharmacy data (for antipsychotic use), ICD discharge codes (for aspiration and pressure injury), incident reports (for falls), and hospital orders (for restraint use) via a waiver of informed consent. Based on information provided by our study hospitals (Section 3.3.3), each unit has 25-40 patients (average 30), of which approximately 1/3, or 10, will be aged 70 or older. With average length of stay of 3.5 days, there will be approximately 20 patients aged 70 or older admitted to each study unit each week. This results in 520 patients per unit every 6-month study period, or 21,840 patients over 6 units and 42 months. The investigators conservatively round this number to 20,000 patients. This will provide sufficient sample size to examine rarer outcomes related to both complications and management of delirium. These data will be de-identified and the investigators will not attempt to link it back to the patient-oriented sample.

1. To test the fidelity, accuracy, and durability of implementing daily nurse screening for delirium using the UB-CAM in routine care. Hypothesis: Screening will be performed in over 90% of eligible days after launch and remain over 90% throughout the study period. Accuracy will be improved relative to the pre-screening period. Focus groups with hospital staff will identify impediments and facilitators to inform fidelity, durability, and future implementation efforts. (Patient-oriented Sample, Staff sample) 2. To evaluate the impact of UB-CAM delirium screening on patient and care partner centered outcomes assessed at one month follow-up interviews. Hypothesis: UB-CAM screening will increase Healthy Days at Home (HDAH), decrease delirium-related distress, and improve perceptions of communication and collaboration relative to the pre-screening period. These same trends will be seen in subgroups of persons with AD/ADRD. (Patient-oriented sample). 3. To evaluate the impact of UB-CAM screening on rates of delirium complications (falls, pressure injury, aspiration) and adverse delirium management (psychoactive medication use, restraint use). Hypothesis: UB-CAM screening will reduce delirium-related complications and adverse delirium management, relative to the pre-screening period (Administrative sample).

• Aim 1: Fidelity: To Find the Percent of Eligible Days With At Least One Completed Nurse Delirium Screen [Time frame: From hospital admission through discharge, an average of 1 week]
• Aim 2: Patient/Family Care Partner Distress [Time frame: One month after hospital discharge]
• Aim 3: Rate of Use of Physical Restraints or Use of Antipsychotic Medications [Time frame: From hospital admission through discharge, an average of 1 week]