Andrew Kelly

Interprofessional Teamwork Innovation Model (ITIM) to promote communication and patient-centred, coordinated care

Background Despite recommendations and the need to accelerate redesign of delivery models to be team-based and patient-centred, professional silos and cultural and structural barriers that inhibit working together and communicating effectively still predominate in the hospital setting. Aiming to improve team-based rounding, we developed, implemented and evaluated the Interprofessional Teamwork Innovation Model (ITIM). Methods This quality improvement (QI) study was conducted at an academic medical centre. We followed the system’s QI framework, FOCUS-PDSA, with Lean as guiding principles. Primary outcomes included 30-day all-cause same-hospital readmissions and 30-day emergency department (ED) visits. The intervention group consisted of patients receiving care on two hospitalist ITIM teams, and patients receiving care from other hospitalist teams were matched with a control group. Outcomes were assessed using difference-in-difference analysis. Results Team members reported enhanced communication and overall time savings. In multivariate modelling, patients discharged from hospitalist teams using the ITIM approach were associated with reduced 30-day same-hospital readmissions with an estimated point OR of 0.56 (95% CI 0.34 to 0.92), but there was no impact on 30-day same-hospital ED visits. Difference-in-difference analysis showed that ITIM was not associated with changes in average total direct costs nor average cost per patient day, after adjusting for all other covariates in the models, despite the addition of staff resources in the ITIM model. Conclusion The ITIM approach facilitates a collaborative environment in which patients and their family caregivers, physicians, nurses, pharmacists, case managers and others work and share in the process of care.

1454: ASSOCIATION OF THIAMINE ADMINISTRATION WITH OUTCOMES IN SEPTIC SHOCK PATIENTS

Critical Care Medicine • Volume 46 • Number 1 (Supplement) www.ccmjournal.org Learning Objectives: Thiamine is an essential component of aerobic metabolism in humans. Deficiency in this co-factor disrupts Kreb’s cycle functionality, decreases adenosine triphosphate production, and may lead to harmful effects including lactic acidosis, hypotension, and death. Mounting evidence has shown that critically ill patients are commonly thiamine deficient. Thus, we hypothesized that critically ill patients with septic shock exposed to thiamine would demonstrate improved lactate clearance and other positive clinical outcomes compared to those not receiving thiamine. Methods: A retrospective, single-center, matched-cohort study was conducted in adult patients that were admitted with an ICD-9/10 diagnosis code of septic shock to either the medicine or surgery ICU between 1/1/2013-1/1/2017. Patients that received intravenous thiamine supplementation within 24 hours of hospital admission were identified, and compared to a matched-cohort of controls not receiving thiamine. Patients were excluded if they did not meet the Sepsis-3 criteria for septic shock on chart review, developed septic shock after admission to the hospital, or if missing data. Mahalanobis distance matching was used to match in a 1:2 fashion. The primary objective was to determine if thiamine supplementation in septic shock patients reduces time to lactate clearance, defined as achieving a serum lactate level of < 2mmol/L, using a competing-risk regression model with death as a competing event. Secondary objectives included incidence of acute kidney injury (AKI) and renal replacement therapy (RRT) initiation. A Cox proportional hazards model was constructed to assess ICU mortality. Results: We screened 2,272 patients, of whom 1,049 were eligible. The study consisted of 123 thiamine patients matched with 246 control patients. Baseline demographics were balanced. Receipt of thiamine was associated with a greater likelihood of lactate clearance over time (SHR 1.33 [1.03–1.71], p = 0.028) and reduced ICU mortality (HR 0.67 [0.49–0.91], p = 0.01) compared to controls. There were no differences in AKI or RRT initiation. Conclusions: Thiamine administration in septic shock patients was associated with improved lactate clearance and reduced ICU mortality. A randomized, prospective study is needed to confirm these results.

1406: ASSESSING THE IMPACT OF ALBUMIN USE ON VASOPRESSOR DURATION AND MORTALITY IN SEPTIC SHOCK

Critical Care Medicine • Volume 46 • Number 1 (Supplement) www.ccmjournal.org Learning Objectives: Albumin use in septic shock is controversial, with large randomized controlled trials generally indicating no difference in outcomes when compared to crystalloids. However, a subgroup analysis from the ALBIOS trial generated the hypothesis that mortality benefit with albumin may be reserved to patients with septic shock, and possibly be associated with a reduction in the duration of vasopressor use. Methods: Retrospective cohort study using ICD9/10 codes to identify septic shock (and confirmed with vasopressor use) in adult patients admitted to the ICU over a 4 year period. Baseline demographics were collected, including SOFA scores, peak lactate, maximum number of vasopressors, and liver disease, among other variables. A patient was classified as receiving albumin if they received 25% albumin while concomitantly receiving vasopressor therapy. A multivariable competing-risk regression model was used to assess time on vasopressor therapy (with death as a competing event) and a multivariable logistic regression model used to assess mortality. Results: Of 2,489 patients screened, 1,168 patients were included. In bivariate analyses, receipt of albumin (n = 493) was generally associated with increased severity of illness, liver disease, worse shock, longer vasopressor duration, worse acute kidney injury, and worse in-hospital mortality compared to no albumin (n = 675). In competing-risk regression adjusting for imbalances in baseline characteristics, use of 25% albumin was not associated with a reduced time on vasopressors (SHR 0.91 95% CI 0.78–1.06; p = 0.24). The receipt of 25% albumin was not associated with hospital mortality in logistic regression modeling (OR 1.13 95% CI 0.83–1.57; p = 0.43). Conclusions: In this retrospective analysis using multivariable modeling and competing-risk analyses, 25% albumin was not associated with a reduction in time on vasopressors or in-hospital mortality in patients with septic shock, however, selection bias may still remain despite attempts to adjust in multivariable modeling.