Effect of Implementing Discharge Readiness Assessment in Adult Medical-Surgical Units on 30-Day Return to Hospital

Key Points Question What is the effect of adding structured nurse assessment of patient readiness for discharge to standard medical-surgical unit discharge practices on 30-day return to hospital? Findings In this multisite cluster randomized clinical trial, when patient self-assessments were combined with readiness assessment by nurses, high-readmission units showed a reduction in 30-day hospital returns. Mixed results were observed for nurse assessments only and for low-readmission units. Meaning Adding a structured discharge readiness assessment by the discharging nurse that includes patient self-assessment to standard practice for hospital discharge may reduce readmissions and emergency department or observation visits.

version for use in clinical practice. The 8-item RN-RHDS was more strongly associated with readmission than the 8-item PT-RHDS in 2 adult samples in the Midwest and Eastern US. [23][24] Patients with low readiness by nurse 49 assessment experienced more than a 6-fold increase in the likelihood of being readmitted. 24

50
Currently there is no standard approach or tool available for routine use in clinical practice. The body of 51 evidence from prior studies by the research team lays the foundation for the recommendation that pre-discharge 52 readiness assessment be implemented as a standard nursing practice for hospital discharge. 3 53 Aims

54
We propose to conduct a multi-site study to determine the impact on post-discharge utilization 55 (readmission and ED visits) and costs of implementing discharge readiness assessment as a standard nursing practice 56 for adult medical-surgical patients discharged to home, using a reliable, valid, and clinically meaningful tool (RHDS 57 8-item short form). The value of implementing discharge readiness assessment as a standard practice lies in timely, 58 rapid, and systematic determination of nurse and patient perspectives on 4 aspects of discharge readiness (personal 59 status, knowledge, coping ability, expected support). The assessment results enable the nurse to initiate pre-60 discharge risk-mitigating actions that enhance patient readiness and avert adverse post-discharge outcomes that may 61 result in ED visits and readmissions.

62
The proposed research extends the research team's prior observational studies to an implementation study.

63
The study tests, in a stepped approach, the impact of implementing discharge readiness assessment by the 64 discharging nurse as standard nursing practice, and the incremental value of informing the nurse assessment with the 65 patient's perspective, and of requiring that the nurse initiates and documents risk-mitigating actions for patients with 66 low readiness scores.

76
A prospective, parallel cohort, stepped implementation study design 25 with difference-in-difference 77 analysis will be used to evaluate outcomes (likelihood of readmissions and ED visits within 30 days post-discharge) 78 for 4 sequential cohorts of patients on units where usual care (baseline cohort 0) is followed by a stepwise examined between each sequential implementation cohort (first difference) while also controlling for any changes in 83 institutional practices or seasonal trends using usual care cohorts at each concurrent phase (second difference). The 84 implementation will occur at the unit-level, incorporating the implementation protocols into unit operational 85 processes for hospital discharge. The decision to implement at the unit-level, with comparable units serving as 86 control units, was made because spill-over effects of nurse training and cross assignment of nurses to 87 implementation and control patients on the same unit would likely contaminate the control patient samples.

88
We will use the following timeline for data collection on each unit.

100
We will include two units per participating hospital. One will be randomly selected as the implementation 101 unit and the other as the control unit. Where possible, these units will be of s similar type (either medical, surgical,

102
or medical-surgical). Critical care units and oncology specialty units will be excluded. We will work with hospitals 103 to achieve a balance of medical, surgical, and mixed medical-surgical unit types across the hospitals. be properly adjusted for confounding due to any unrelated system-wide changes in patient care practices or seasonal 115 trends that may occur during the study period, the start and end date of each implementation and data collection 116 phase will be same for paired implementation and control units; frequency weighting will be used to adjust for 117 sample size differences among units.

118
Because the study design includes a time trend, patients will be enrolled throughout the entire study period as 119 described in the study design. Therefore, the number of patients enrolled per hospital will vary based on volume and 120 will exceed the minimum sample estimates.

122
The study will evaluate implementation of assessment of readiness for discharge for adult medical-surgical 123 patients being discharged to home. We use a stepped implementation approach with sequential introduction of 124 structured tools for discharge readiness assessment and documentation of nursing actions to improve discharge 125 transition outcomes for patients with low readiness. We will evaluate the incremental benefits of 3 protocols: (1) For the READI1 protocol, the discharging nurse was instructed to complete the form on the day of hospital 137 discharge normally within 4 hours prior to discharge, normally after completion of discharge preparation. readiness risk both in anticipation of discharge and at the time of discharge readiness assessment and that the use of the NDAG can serve as a trigger for action by the nurse. We will analyze total nurse actions as well as differentiate 180 between anticipatory and reactive nurse actions.

182
For this study, the NDAG will be completed following discharge readiness assessment (RN-RHDS+PT-RHDS) for 183 all patients on the implementation units. Nurses will be instructed that initiation and documentation of an action on  %BSN). These data will be collected 199 monthly from nursing administrative databases and specifications will be consistent with NQF/NDNQI.

201
Education regarding the implementation protocols and training in study procedures will occur at multiple

213
Site PI training will occur via webinar/videoconferencing using Go-To-Meeting®. All training documents 214 and all study materials will be available on the study website. Start-up training for unit nurses will be conducted by procedures, including human subjects' protection. All training materials will be available to sites on the designated 220 website. PowerPoint slides will contain voice-over for online training purposes for staff not able to attend in-person 221 training. Training for phases 2 and 3 will be via in-person or online training, as selected by site PIs.

222
Because this is a unit-based implementation approach, logs will be kept of participation in training to assure 223 all staff nurses have been trained. For staff not attending, site PIs will assure that training has occurred by 224 documentation of review of study materials posted on the study website. Staff will login to the site to confirm that 225 they have met the training requirement.

HUMAN SUBJECTS PROTECTION
Hospital IRBs: The study protocol will be reviewed and approved by the researchers' university IRBs. Each local 228 IRB will also review and make a determination.

229
Human Subjects Protection: The study evaluates the implementation of a process of care that augments existing 230 discharge processes. The contribution of the acute care clinical nurse in the complex processes of preparing patients 231 for discharge is a study focus. Implementation of a new protocol for nurse assessment of a patient's readiness for 232 discharge will occur in 3 sequential phases to evaluate each of 3 components of the discharge readiness assessment

239
The READI study received expedited approval from Marquette University's IRB on September 9, 2013. Marquette

240
University IRB will serve as the primary IRB for the study. All participating hospital IRBs will review and make a 241 determination of the most appropriate mechanism for IRB approval. We encourage the execution of an Institutional

242
Authorization Agreement with Marquette University IRB to facilitate continuing review during the 3-year study 243 period.

244
We provide the following information to consider in making a determination of the appropriate mechanism for IRB 245 approval at hospital sites:

246
We are requesting a waiver of documentation of consent based on the following factors:     6. The data will be de-identified by the site PI before release to the research team.

261
All data will be de-identified at the study site by the site PI. Nurse RHDS Assessment forms will contain a section 262 at the bottom of the patient study ID and the patients label containing identifying information. After completion by 263 the discharging nurse, these forms will be collected by the site PI, the identification section will be detached from 264 the form. The de-identified RHDS forms will be sent to the study team; the detached identifiers with link to study ID 265 will be used by the site PI to create a study log on an excel spreadsheet which will be retained at the study site. The 266 study log will be password protected by the site PI. This log will be given to IT services to generate the electronic 267 data files for the study. The site PI will verify completeness of the electronic extract and then remove all identifiers 268 from the file. Following verification of transmission to the study team via secure FTP and screening of the 269 transmitted file, the site PI will destroy all on-site paper and electronic identifiers. The site PI will retain a copy of 270 the de-identified electronic file.

271
Data Protection: All data will be de-identified at the study site. Nurse RHDS Assessment forms will contain a 272 section at the bottom of the patient study ID and the patients label containing identifying information. After 273 completion by the discharging nurse, these forms will be collected by the site PI, the identification section will be 274 detached from the form. The de-identified RHDS forms will be sent to the study team; the detached identifiers with 275 link to study ID will be used to create a study log on an excel spreadsheet which will be retained at the study site.

276
The study log will be password protected by the site PI. This log will be given to IT services to generate the 277 electronic data files for the study.

278
The site PI will verify completeness of the electronic extract and then remove all identifiers from the file.

279
Following verification of transmission to the study team via secure FTP and screening of the transmitted file, the site

280
PI will destroy all on-site paper and electronic identifiers. The site PI will retain a copy of the de-identified 281 electronic file.

282
Data will be submitted to the primary study site at Marquette University. Data will be uploaded by study 283 sites to a secure Office 365 password protected database. Site will have access only to their own secure space and 284 will not be able to access data submitted by other participating hospitals.