304: NOVEL CRITICAL CARE OUTREACH TEAM MODEL IMPROVING RAPID RESPONSE OUTCOMES

Objective Little is understood about what contributes to perceived workload for those providing overnight coverage to hospitalized patients overnight, which limits the ability to modify these factors or to proactively identify appropriate staffing levels. The objective of this study is to understand the major contributors to perceived overnight cross-coverage workload. Methods Cross-covering advanced practice providers (APPs) in a large academic hospitalist group completed the National Aeronautics and Space Administration Task Load Index (NASA-TLX) at the end of each night shift. Other shift characteristics were collected, including patient load, assigned action items, watcher/unstable patients, newly admitted patients, number of units covered, total pages, peak pager density, rapid response team (RRT) activations, and intensive care unit (ICU) transfers. Results For 14 APP participants, who completed 271 post-shift surveys, the mean (SD) patient load was 49.9 (6.4) patients per night, and providers received a mean (SD) of 40.8 (13.7) total pages per shift. Mean (SD) NASA-TLX score was 35.1 (19.0). In multivariate modeling, total pages, action items, and any RRT or ICU transfer were associated with significant increases in the mean NASA-TLX score, with estimated effect sizes of 0.5, 0.8, and 14.3, respectively, per 1-unit increase in each shift characteristic. The greatest cumulative contributor to perceived workload was total number of pages, followed by the presence of any RRT activation/ICU transfer, with estimated effect sizes of 20.4 and 14.9, respectively. Conclusions Total number of pages was the greatest contributor to perceived workload. This study suggests that quality improvement initiatives designed to improve pager communication may considerably improve provider-perceived workload.

<h3>Background</h3> Patients with life limiting illnesses commonly present to hospital acutely and are referred to the Critical Care Outreach Team (CCOT). It is frequently concluded that escalation of care is inappropriate due to the partially or completely irreversible nature of deterioration. In such cases escalation of palliative care is fundamental. We performed a retrospective service evaluation of patient outcome following critical care outreach review in patients deemed unsuitable for critical care admission and evaluated appropriateness of referral to the Specialist Palliative Care Team (SPCT). <h3>Methods</h3> This was a two-centre study. Paper medical notes from a were reviewed. SPCT involvement was assessed for appropriateness using a conceptual framework. <h3>Results</h3> 50 patients' (evaluated by CCOT but unsuitable for a higher level of care between October and December 2018) notes were reviewed. 34 (68%) died and 16 (32%) were discharged. Of those who died, the median time from initial CCOT review to death was 4.5 days. 6 (12%) of patients were referred to the SPCT. A further 7 (14%) were reviewed by the SPCT who proactively identified patients as end of life care through the discontinuation of physiological observations. However, 45 (90%) should have been referred. <h3>Conclusions</h3> CCOTs identify patients who are inappropriate for higher-level care. Most of these patients are unlikely to survive their hospital admission and the vast majority would benefit from SPCT review for appropriate symptom management and Advanced Care Planning (ACP). The term Acute Palliative Intervention (API) could be utilised to change the language and culture of care decisions.

Rapid response team activation after major hip surgery: A case series

Continuous and wireless vital sign (VS) monitoring on hospital wards is superior to intermittent VS monitoring at detecting VS abnormalities; however, the impact on clinical outcomes remains to be confirmed. A recent propensity-matched study of primary surgical patients found decreased odds of intensive care unit (ICU) admission and mortality in patients receiving continuous monitoring. Primary surgical patients are inherently different from their medical counterparts who typically have high morbidity, including frailty. Continuous monitoring research has been limited in primary medical patients. This study aims to evaluate the clinical outcomes of primary medical patients who received either continuous or, as a contemporaneous control, intermittent vital monitoring as the standard of care using propensity matching. Propensity-matched analysis of a population-based sample of 7971 patients admitted to the medical wards between January 2018 and December 2019 at a single, tertiary United States medical center. The continuous monitoring device measures oxygen saturation, heart rate, respiratory rate, continuous noninvasive blood pressure, and either 3-lead or 5-lead electrocardiogram. Patients received either 12 hours or more of continuous and wireless VS monitoring (n=1450) or intermittent VS monitoring (n=6521). The primary outcome was the odds of a composite of in-hospital mortality or ICU transfer during hospitalization. Secondary outcomes were the odds of individual components of the primary outcome, as well as heart failure (HF), myocardial infarction (MI), acute kidney injury (AKI), and rapid response team (RRT) activations. Those who received intermittent VS monitoring had greater odds of a composite of in-hospital mortality or ICU admission (odds ratio [OR] 2.79, 95% CI 1.89-4.25; P<.001) compared with those who had continuous and wireless VS monitoring. The odds of HF (OR 1.03, 95% CI 0.83-1.28; P=.77), MI (OR 1.58, 95% CI 0.77-3.47; P=.23), AKI (OR 0.74, 95% CI 0.62-1.02; P=.06), and RRT activation (OR 0.94, 95% CI 0.75-1.19; P=.62) were similar in both groups. In this propensity-matched study, medical ward patients who received standard of care intermittent VS monitoring were at nearly 3 times greater odds of transfer to the ICU or death compared with those who received continuous VS monitoring.Our study was primarily limited by the inability to match patients on admission diagnosis due to limitations in electronic health record data. Other limitations included the number of and reasons for false alarms, which can be challenging with continuous monitoring strategies. Given the limitations of this work, these observations need to be confirmed with prospective interventional trials.

Introduction: Identification of patients in crisis and early activation of the Rapid Response Team (RRT) can improve patient outcomes. However, if knowledge gaps exist related to activation triggers or recognizing subtle physiological changes in patients, nurses may not activate the RRT reliably, even if they are worried. We sought to understand, if a proactive, relationship-based approach could increase the RRT activation rate. Methods: This practice-focused project implemented RRT+, a proactive safety practice, designed to support the psychological safety of bedside nurses, and increase the number of adult RRT activations. RRT nurses proactively sought out patients at risk for deterioration, provided education to nurses about signs of deterioration, created a safety plan, then followed up at two and four-hour intervals. Combined, these four elements were predicted to raise the intentions and change behavior of nurses to activate the RRT. Multiple process and outcome measures were recorded and evaluated at 3 and 6-month intervals using a mixed-methods approach. RRT activations were evaluated for magnitude of effect and statistical significance, while qualitative survey and focus group data were evaluated for clinical significance. This project was monitored continuously for unintended adverse events. Results: A statistically significant increase in RRT activations was observed. For the initial three-months, RRT activations expected were (x ̅) ± (SD) x ̅=14.92 ± (7.83) and observed x ̅=21.80 ± (11.82); a 46.1% increase. A large and statistically significant effect was observed at the three-month interval d=0.943; t (df) = -4.18 (4), p = 0.013 (two tailed), and six-month d=.885, -2.86 (4), p=0.045 intervals compared to baseline. Testimony from stakeholders was positive, RRT responders had a favorable view of RRT+ “I like it, nurses are asking us questions because we’re approaching them… it’s having an impact on RRT activation.” Conclusions: This project shifted a reactive RRT model into an effective proactive model, aimed at improving patient care. Point-of-care education and safety planning helped bedside nurses better respond to patient deterioration and seek help. RRT+ elements to address worry, focused education, and proactive safety planning are sustainable practices and transferrable to most settings.

Rapid response teams (RRTs) are groups of health-care providers, implemented by hospitals to respond to distressed hospitalized patients on the hospital wards. Patients assessed by the RRT for deterioration may be admitted to the intensive care unit (ICU) or may be triaged to remain on the wards, putting them at risk of recurrent deterioration and repeat RRT activation. Previous studies evaluating outcomes of patients with recurrent deterioration and multiple RRT activations have produced conflicting results. We used a prospectively collected multicenter registry from 2 hospitals within a single tertiary-level hospital system between 2012 and 2016. Comparisons were made between patients with a single RRT activation and those with multiple RRT activations over the course of their admission. Primary outcome was in-hospital mortality, which was analyzed using multivariable logistic regression. A total of 5995 patients who had any RRT activation were analyzed. Of that, 1183 (19.7%) patients had recurrent deterioration and multiple RRT activations during their admission. Risk factors for recurrent deterioration included admission from a home setting (as opposed to a long-term care facility), RRT activation during nighttime hours, and delay (>1 hour) to RRT activation. Recurrent deterioration was associated with increased odds of mortality (adjusted odds ratio [OR]: 1.44 [1.28-1.64], P = <.001). Increasing number of RRT activations were associated with increasing risk of mortality. Patients with recurrent deterioration had prolonged median hospital length of stay (21.0 days vs 12.0 days, P < .001), while patients with only a single activation were more likely to be admitted to the ICU (adjusted OR: 2.30 [1.96-2.70], P < .001). Recurrent deteriorations leading to RRT activations among hospitalized patients are associated with increased odds of mortality and prolonged hospital length of stay. This work identifies a group of patients who warrant closer attention to help reduce adverse outcomes.

To improve understanding of afferent limb behaviour in acute hospital ward settings, to define and specify who needs to do what differently and to report what afferent limb behaviours should be targeted in a subsequent multi-phase, theory-based, intervention development process. Focused ethnography was used including direct observation of nursing staff enacting afferent limb behaviours and review of vital signs charts. An observation guide focused observation on "key moments" of the afferent limb. Descriptions of observations from between 7 January 2019-18 December 2019 were recorded in a field journal alongside reflexive notes. Vital signs and early warning scores from charts were reviewed and recorded. Field notes were analysed using structured content analysis. Observed behaviour was compared with expected (policy-specified) behaviour. Observation was conducted for 300hr. Four hundred and ninety-nine items of data (e.g., an episode of observation or a set of vital signs) were collected. Two hundred and eighty-nine (58%) items of data were associated with expected (i.e. policy-specified) afferent limb behaviour; 210 (42%) items of data were associated with unexpected afferent limb behaviour (i.e. alternative behaviour or no behaviour). Ten specific behaviours were identified where the behaviour observed deviated (negatively) from policy or where no action was taken when it should have been. One further behaviour was seen to expedite the assessment of a deteriorating patient by an appropriate responder and was therefore considered a positive deviance. Afferent limb failure has been described as a problem of inconsistent staff behaviour. Eleven potential target behaviours for change are reported and specified using a published framework. Clear specification of target behaviour will allow further enquiry into the determinants of these behaviours and the development of a theory-based intervention that is more likely to result in behaviour change and can be tested empirically in future research.

Rapid response teams (RRT) were first proposed as a means of reducing inpatient morbidity and mortality. Modifying RRT activation triggers poses a potential risk for delayed recognition of a deteriorating patient. Trigger modifications have not been validated for safety. To determine if RRT trigger modifications are associated with: increased frequency of recurrent RRT activation; increased length of stay (LOS); increased intensive care admission; and increased in-hospital mortality. A retrospective audit of all RRT activations occurring at the Launceston General Hospital (LGH) over an 18-month period was performed. Rate of recurrent RRT activations did not decrease with the use of trigger modifications around the time of RRT activation, and for patients with two modifications, the frequency increased (1.98 vs 1.32, P = 0.007). LGH LOS increased for patients with any trigger modifications compared to those with none (11 vs 9, P = 0.0002), and for patients with two modifications (11.5 vs 9, P = 0.010). Total hospital LOS increased for patients with any modifications compared to patients with none (12 vs 10, P = 0.002). There was no significant association between trigger modifications and frequency of intensive care unit admission. The relative risk of in-hospital death increased with increasing numbers of trigger modifications (relative risk 1.38-4.89). Trigger modifications are associated with increased hospital LOS and increased rate of in-hospital death and do not reduce the number of recurrent events. For patients in whom escalation of care is not appropriate, the presence of multiple trigger modifications at the time of an RRT activation may be a useful trigger for conversations around goals of care.

Forming a Partnership to Save Lives

Early identification and treatment of clinical deterioration is crucial for improving outcomes among hospital patients. A high-acuity response team (HART) program can integrate early warning systems and proactive rounding by critical care nurses to prevent unplanned escalations in care. During the COVID-19 pandemic, a HART program was inconsistently implemented because of intensive care unit staffing shortages. Barriers to optimizing the HART nurse role included inconsistent practices, lack of clear role expectations, and frequent reassignment of HART nurses to compensate for staffing shortages. Postpilot implementation of the HART program began in October 2019. Critical care nurses were designated as HART nurses, responsible for monitoring the Rothman Index, and assisted bedside nurses with high-acuity patients. Data were collected from 2019 to 2023 and were analyzed using IBM SPSS Statistics, version 29, with statistical significance defined as P ≤ .05. The HART program significantly reduced 24-hour intensive care unit readmissions by 33.9% and 72-hour readmissions by 32.7%. HART nurse consultations increased by 35.7%. There were clinically significant decreases in code blue emergencies outside the intensive care unit (22.2%) and overall (16.7%), although no statistically significant differences were found for rapid response team activations or unplanned intensive care unit transfers. The HART nurse program effectively integrates early warning systems and rapid response teams, significantly reducing intensive care unit readmissions and improving patient care. Clear role expectations and dedicated staffing are needed, and continuous stakeholder engagement and resource allocation are essential for sustaining the program's success.

Outcomes of pediatric rapid response teams at a hematology/oncology center in Riyadh, Saudi Arabia

We thank Overdyk and Broens1 for their educational comments on our systematic review and meta-analysis. The prospective randomized controlled trial by Ochroch et al2 found that continuous pulse oximetry (CPOX) did not reduce intensive care unit (ICU) transfers on a postcardiothoracic surgery floor. However, in a subgroup analysis, they found that CPOX significantly decreased ICU transfers due to pulmonary complications. It is important to note that the study by Ochroch et al2 was underpowered to detect real subgroup effects, thereby increasing the risk of false-positive results. The other prospective controlled study by Taenzer et al3 included the cohort of elderly patients undergoing joint replacement surgeries in a patient surveillance system unit. The comparison unit included the cohort of patients from urology, gynecology, vascular surgery, and general surgery. They found that CPOX monitoring reduced the need for ICU transfers; however, no subgroup analysis was performed to detect whether ICU transfers were due to respiratory complications.3 Overdyk and Broens1 suggested that ICU transfers in the study by Taenzer et al3 are less likely to be skewed by hemodynamic causes. This hypothesis is not supported in the growing literature which shows that there is an increased incidence of major complications like pulmonary embolism, sepsis, nonischemic cardiac complications, and pneumonia after joint replacement surgeries.4 Thus, ICU transfers were not exclusively due to pulmonary causes. Having extracted data from the prospective controlled studies,2,3,5,6 we decided that it was not appropriate to pool results from trials comparing outcomes that are dissimilar. In other words, it would be inappropriate to pool the data on pulmonary causes of ICU transfer from Ochroch et al2 and data on all-cause ICU transfer from Taenzer et al.3 Overdyk and Broens1 suggested combining the 2 data sets from Taenzer et al3 and Ochroch et al.2 This may increase the risk of bias, raising questions on the validity of the results. To avoid bias in our meta-analysis, we pooled the data on all-cause ICU transfer from both studies.2,3 Based on these findings, we concluded that all-cause ICU transfer had a risk reduction of 34% with CPOX when compared to standard monitoring. Due to lack of data in the literature, our meta-analysis failed to differentiate ICU transfers with respect to pulmonary versus other causes. We agree with Overdyk and Broens1 that pulse oximetry is not a measure of circulatory insufficiency. Pulse oximetry device is designed to measure specific physiologic outcome (oxygen saturation), and as such, it measures respiratory adverse events better than cardiac adverse events.7 Our systematic review makes it clear that there are not enough outcome studies using oxygen saturation to evaluate its utility for improving patient outcomes with respect to pulmonary causes of adverse events. Mahesh Nagappa, MDDepartment of Anesthesia and Perioperative MedicineLondon Health Sciences Centre and St. Joseph Health CareWestern UniversityLondon, Ontario, Canada Thach Lam, MDFrances Chung, MBBSDepartment of Anesthesiology and Pain MedicineToronto Western HospitalUniversity Health NetworkUniversity of TorontoToronto, Ontario, Canada[email protected]

Postoperative Bedside Critical Care of Thoracic Surgery Patients.

Evidence-Based Review and Discussion Points

Technology-driven intensive care unit (ICU) environments can be dehumanizing for critically ill patients, their families, and team members who care for them. Prior educational efforts to humanize the ICU have focused on training clinicians to provide human-centered care for patients and families, with relatively little attention on promoting human-centered environments for team members. We used the framework of team coaching to develop an interprofessional educational program focused on humanizing the ICU. We sought to evaluate whether this program improves participants' perceived capability to promote a human-centered environment for the ICU team. We conducted a 3-year, mixed methods evaluation of a novel interprofessional ICU team coaching program in a medical/surgical/trauma ICU within an academic medical center. The 3-session program is led by a certified coach and annual cohorts are a subset of the ICU team. Session topics are: role identification, end-of-life care, and practices of dignity and respect. We conducted pre- and post-coaching participant surveys and qualitative content analysis of coaching reflections, open-ended survey responses, and an audio-recorded debriefing session. Over 3 years, 43 ICU team members representing 7 professional roles (advanced practice providers, attending physicians, chaplains, fellow physicians, nurses, respiratory therapists, social workers) participated. The overall session attendance rate was 70%. Before participation, most reported a need for training to promote a human-centered environment for patients (78.1%), families (81.3%) and other members of the team (81.3%). After coaching, this unmet need decreased to 28.6% for patients (p < 0.01), 42.9% for families (p < 0.01), and 33.3% for team members (p < 0.01). Participants' self-reported ability to provide optimal human-centered care was similar pre-(84.4%) and post-coaching (81.8%; p = 0.80). Participants' scores on the interprofessional Consultation and Relational Empathy (CARE) measure improved from 37.1 (95% CI 34.3-39.9) pre-coaching to 40.7 post-coaching (95% CI 38.8-42.6;p = 0.047). Qualitative analysis revealed commitment to growth in interprofessional relationships among participants and a desire to expand the program's reach. Our findings confirm an unmet need among ICU team members for training in how to promote human-centered ICU environments and suggest that interprofessional team coaching is a promising tool to help meet this need.