Safety Action Feedback and Engagement (SAFE) Loop

This cluster randomized controlled trial will test whether a novel intervention, the Safety Action Feedback and Engagement (SAFE) Loop, enhances hospital incident reporting practices, improves nurses' perceptions of incident reporting, and lowers rates of high-priority medication events, as compared with Cedars-Sinai Medical Center's (CSMC's) existing incident reporting system.

The specific aims are to determine whether SAFE Loop:

• Aim 1: improves incident reporting practices by increasing the rate at which nurses report high-priority medication incidents (H1.1) and the number of contributing factors described per report (H1.2).
• Aim 2: improves nurses' attitudes toward incident reporting, including perceptions of feedback and communication about error (H2.1) and of the frequency with which events are reported (H2.2).
• Aim 3: reduces rates of high-priority medication events (H3.1) and high-priority medication events involving harm (H3.2).

Setting and Subject Population: 20 acute care nursing units at Cedars-Sinai Medical Center in Los Angeles, where about 1980 nurses provide over 294,000 patient-days of care per year (about 115,000 hospital visits).

Quality Improvement (QI) Activities: (1) Safe Loop Activities, including obtaining input from nurses, educational activities training them in how to focus on target events and write high-quality reports, stimulated time-limited reporting, integration of information from reports and other QI data sources from within the hospital and other information, designing changes to improve medication safety, and providing feedback to nurses. (2) Randomization of nursing units to intervention or control and early vs. later implementation. The will assure that the implementation occurs in a uniform and manageable fashion. (3) Implementation activities: planning, engaging, executing, reflecting/evaluating.

Primary Research Procedures:

1. Nurse surveys (N=1,980): We will attempt to survey about 1980 nurses on study nursing units. Survey recruitment will be done through email, with e-informed consent via Redcap.
2. Nurse interviews on implementation (N=37): We will interview 32 nurses (10 unit leaders and 22 registered nurses) to characterize implementation. We will recruit nurses and unit leaders will by email and word-of-mouth. Prelim work: We will conduct cognitive tests of the interview script with up to 5 nurses to make refinements (2 unit leaders, 3 floor nurses)
3. Analysis of Incident Reports (N=1,963): We will examine incident reports related to medication events during the study period. We anticipate that 1,663 patients will have eligible incident reports. Prelim work: We will conduct practice analyses on up to 300 patients with incident reports to refine study instruments and train analysts. We may seek to publish findings based on prelim work.
4. Analysis of Medication Events (N=1,520): We will randomly sample 1,520 hospitalizations during the study period. We will review medical records and access pharmacy iVents reports to identify events. We will also obtain variables from the electronic health record on all these patients as well (for case mix adjustment). Prelim work: We will conduct practice analyses on up to 100 medical records to refine instruments and train nurse and physician analysts.
5. Nurse interviews for DSM (N=18): As part of our Data Safety and Monitoring Plan, we will interview 15 nurses on the intervention nursing units that first implement the intervention to assess and monitor for adverse consequences to the nurses. We will recruit nurses by having a research nurse go to the nursing units and ask to speak with nurses privately. Preliminary work: We will pilot test the interview guide with 3-5 nurses.

Total number of hospital visits: Approximately 172,500 over 2.5 years. Length of follow up: start of hospital visit (e.g., admission) to discharge from hospital or end of study, whichever is earlier.

How Participation Differs from Standard of Care: In current systems of hospital care, voluntary incident reporting is a technique through which frontline personnel -usually nurses-submit reports that describe events that posed risks to safety, particularly incidents that were intercepted or happened not to cause harm. In contrast, participation in the study involves the Safety Action Feedback and Engagement (SAFE) Loop. The SAFE Loop has five key attributes designed to transform hospital incident reporting systems into much more effective tools for improving patient safety: obtaining nurses' input about which medication safety problems to address; focusing on selected high-priority medication events; prompting nurses to report high-priority medication events during a designated period and training them to write more informative reports; integrating information about medication safety problems on hospital nursing units from the incident reports, investigations, and other internal and external sources; and providing feedback to nurses on the problems identified on the nursing units and mitigation plans.

Any parts of the protocol not conducted at CSMC? Not applicable. All parts will be conducted at CSMC.

Advancing knowledge or clinical practice: This study will be a highly rigorous and comprehensive evaluation of an incident reporting intervention in healthcare. This appears to be the first RCT of an intervention designed to improve incident reporting in healthcare, and the first study to examine the effect of an incident reporting intervention on an important patient safety outcome, medication event rates. Prior studies have only used observational designs and examined effects on the timeliness of reporting, the numbers of reports, and the types of incidents reported.

Lastly, this study will transform how incident reporting systems function in hospitals: (A) Engaging Nurses in Decisions about Which Patient Safety Problems to Address. Prior strategies for improving safety include top-down initiatives. However, clinician engagement is essential because safety is an emergent property that results from millions of small actions they take each day. (B) Prompting Nurses to Report High-priority Events and Write More Informative Reports. Emphasizing high-priority events increases the likelihood of making meaningful improvements in safety. This project emphasizes the value of incident reporting as a tool for gaining insights into the causes of critical events, based on successes of incident reporting at improving safety in other high-risk industries. (C) Creating Standardized Procedures for Investigating High-priority Events. Procedures for following up on reports are variable and not well standardized; therefore, this will be a major advance in the field. (D) Providing Feedback to Nurses about How Reports Were Used to Improve Safety. Interventions that improve incident reporting are more likely to succeed and be sustained if frontline clinicians receive feedback that their reports were useful and observe changes in systems of care and event rates-experts call this the 'safety action feedback loop.' Few interventions have attempted to strengthen this loop in healthcare.