Impact of a quality improvement initiative with a dedicated anesthesia team on outcomes after surgery for adult congenital heart disease

Objectives A quality improvement initiative was introduced to the adult congenital cardiac surgery program at Toronto General Hospital in January 2016. A dedicated Adult Congenital Anesthesia and intensive care unit team was introduced within the cardiac group. The use of factor concentrates was introduced. The study compares perioperative mortality, adverse events, and transfusion burden before and after this process change. Methods We performed a retrospective analysis of all adult congenital cardiac surgeries from January 2004 to July 2019. Two groups were analyzed: patients undergoing operation before and after 2016. The primary outcome was in-hospital mortality. One-year mortality and prevalence of key morbidities were analyzed as secondary outcomes. A separate analysis looked at patients who had and had not attended an anesthesia-led preassessment clinic. Results In-hospital mortality was significantly reduced in patients undergoing operation after 2016 (1.1% vs 4.3%, P = .003) despite a higher risk profile. One-year mortality (1.3% vs 5.8%, P = .003) and ventilation times (5.5 hours [3.4-13.0] vs 6.3 hours [4.2-16.2], P = .001) were also reduced. The incidence of stroke and renal failure was similar between groups. Blood product exposure was comparable, but the incidence of chest reopening decreased (1.8% vs 4.8%, P = .022), despite more patients with multiple previous chest wall incisions, on anticoagulation, and with more complex cardiac anatomy. There were no significant outcome differences between those who did or did not attend the preassessment clinic. Conclusions Both in-hospital and 1-year mortality were significantly reduced after the introduction of a quality improvement program, despite a higher risk profile. Blood product exposure remained unchanged, but there were less chest reopenings.


CENTRAL MESSAGE
In-hospital and 1-year mortality decreased after adult congenital cardiac surgery at Toronto General after the introduction of a quality improvement program despite patients with a higher risk profile.

PERSPECTIVE
Adult patients with congenital heart disease have comorbidities than can predict outcomes after surgery. Knowledge of these risk factors and management by dedicated expert teams could improve planning and perioperative and postoperative care in adult congenital cardiac surgery, and potentially lead to improved outcomes.
The population of adults living with congenital heart disease is increasing. 1 More than 85% of patients with congenital heart disease now live into adulthood. 2 Mortality after cardiac surgery in patients with adult congenital heart disease (ACHD) ranges from 0.7% to 9.4%. [2][3][4][5][6] Risk prediction tools currently applied in this population are generally those originally described for pediatric cardiac surgery. These scoring systems have been shown to poorly predict mortality and adverse outcomes in adults. 3,4,7 Patient-specific variables considered in these scores such as low weight or prematurity do not account for risk in ACHD. 8 Procedure-specific risk estimation is based on an aggregate of all age groups. 3 These scores lacked discriminative power when externally validated in an adult population. 8 The Adult Congenital Heart Surgery score, an adult congenital heart specific score derived from the Society of Thoracic Surgeons Congenital Heart Surgery Database, has performed variably, likely because it is based purely on the procedure undertaken and ignores patient-specific factors. 3,9,10 A retrospective study performed at our institution used multivariate regression to identify risk factors associated with poor outcomes after cardiac surgery in this population. 6 Mayo End-Stage Liver Disease modified score (MELD-xi) (odds ratio [OR], 1.38, 95% confidence interval [CI], 1.23-1.54, P<.001), presence of cognitive impairment (OR, 3.55, 95% CI, 1.38-9.12, P ¼ .009), more than 3 previous chest wall incisions (OR, 2.22, 95% CI, 1.02-4.85, P ¼ .045), and anatomy other than a biventricular subaortic left ventricle (other anatomy) (OR, 3.81, 95% CI, 1.79-8.11, P ¼ .001) were found to be statistically associated with poor outcomes (a composite outcome of 30-day mortality, acute kidney injury [AKI] requiring renal replacement therapy [RRT], or prolonged ventilation of more than 7 days). This has gone some way to addressing the lack of risk prediction presently available for patients with ACHD undergoing cardiac surgery.
This improved understanding of risk factors prompted the introduction of a quality improvement program with a focus on perioperative care at our institution. In January 2016, the adult congenital cardiac surgery program instituted a suite of program changes in an attempt to improve outcomes in patients identified as being at higher risk. A dedicated ACHD anesthesia team was introduced within the cardiac group. This group was responsible for the patient's entire perioperative journey through the preassessment clinic, intraoperative care, and intensive care unit stay. We hypothesized that better risk prediction at the preassessment clinic would lead to lower surgical mortality and morbidity. At the same time, the use of factor concentrates was also introduced to the surgical program, specifically prothrombin complex concentrate (PCC) and fibrinogen concentrate.
The aim of the present study was to describe mortality, adverse perioperative events, and transfusion burden before and after these program changes implemented in January 2016. We also aimed to specifically analyze the impact of attendance at the preassessment clinic on mortality and adverse surgical outcomes.

MATERIAL AND METHODS
The Research Ethics Board of University Health Network approved the study protocol and publication of data (Research Ethics Board 15-9178, July 3, 2015, renewed July 3, 2021). Patient written consent for the publication of the study data was waived by the Research Ethics Board due to the retrospective and anonymous nature of the study. After obtaining this approval, a retrospective analysis was conducted of all cardiac surgeries performed in a single center in patients with congenital heart disease more than 16 years from January 1, 2004, to July 31, 2019. The following operations were excluded from analysis: heart transplant, lung transplant, transvenous or epicardial pacemaker implantation or revision, chest reexploration for bleeding or tamponade, and reexploration for a revision of the initial repair during the same admission. The setting was a quaternary referral ACHD cardiac surgery program located within a quaternary adult cardiac surgery program, where ACHD cardiac surgeries are performed by experienced congenital heart disease surgeons with an exclusive congenital heart disease practice. Data were analyzed in 2 cohorts, before and after January 1, 2016.

Adult Congenital Heart Disease Cardiac Surgery Quality Improvement Initiatives
Initiative 1 was the creation of a dedicated subspecialty ACHD cardiac anesthesia and intensive care unit team: Before January 1, 2016, care for patients with ACHD undergoing cardiac surgery could be provided by any member of a large group of cardiac anesthesiologists (n ¼ 40) and intensivists (n ¼ 10) within the wider cardiac surgery program. Creation of a dedicated subspeciality ACHD cardiac anesthesia team was introduced into clinical practice on January 1, 2016. After that date, members of the ACHD cardiac anesthesia team (n ¼ 3) coordinated care for all ACHD patients undergoing cardiac surgery, from preoperative assessment through intraoperative care, until stable postoperative management was established in the cardiac surgical intensive care unit.
Initiative 2 was the introduction of a preoperative ACHD anesthesia assessment clinic: After January 1, 2016, the ACHD cardiac anesthesia To view the AATS Annual Meeting Webcast, see the URL next to the webcast thumbnail.

Abbreviations and Acronyms
AHCD ¼ adult congenital heart disease AKI ¼ acute kidney injury replacement therapy  team aimed to see all patients with ACHD before cardiac surgery in a new,  dedicated ACHD preoperative anesthesia clinic. At the clinic, individual risk  assessment is performed using previously identified risk factors (cognitive  impairment, MELD-xi score, more than 3 previous chest wall incisions,  and anatomy other than a subaortic systemic left ventricle) 6 and advance care planning conversations take place to discuss potential adverse outcomes of surgery, identify substitute decision makers, and clarify goals of care. These clinic assessments were relayed to the ACHD team at multidisciplinary case conference discussions to inform and tailor the operative plan. Initiative 3 was the optimization of perioperative coagulation. During the same time frame use of factor concentrates, PCC and fibrinogen concentrate were introduced to the institution's cardiac surgical program targeting normalization of clotting profile in the operating room.

Data Collection and Analysis
Patients with ACHD cardiac surgery were identified from the prospectively maintained cardiac surgery database. Patient's electronic healthcare records were accessed for data collection. The patients were divided into 2 groups for analysis: all those who underwent surgery from January 1, 2004, to December 31, 2015, inclusive (the "before" group) and those who underwent surgery from January 1, 2016, to July 31, 2019 (the "after" group). Twenty-three preoperative variables were collected for each patient (Table 1). Anatomy was defined as a 2 ventricle, subaortic left ventricle or other. Chest wall incisions included thoracotomy or sternotomy. The primary outcome was in-hospital mortality in each group. The secondary outcomes were 1-year mortality, duration of postoperative mechanical ventilation, incidence of AKI requiring institution of RRT, incidence of stroke (defined as a new-onset neurological insult as diagnosed by a neurologist or stroke physician), incidence of intraoperative blood product exposure (BPE) (a dichotomous outcome defined as the exposure to any of packed red blood cells, plasma, platelets, or cryopreciptitate), and incidence of chest reopening for bleeding or tamponade. A composite outcome of in-hospital mortality, prolonged mechanical ventilation for greater than 7 days, or AKI requiring RRT was also compared. The same preoperative variables and outcomes were also compared between those who received blood products and those who did not.
Further analysis of the "after" group was also performed to attempt to delineate any association of the introduction of the specialist ACHD anesthesia preoperative clinic on the composite outcome of in-hospital mortality, prolonged ventilation for more than 7 days, or AKI requiring RRT. Heart transplant recipients were included in this analysis because the ACHD surgical team assumed care of ACHD transplants toward the end of 2015. Propensity score matching (1:1 matching) was used to allow the effects of treatments to be estimated on the observations by reducing the potential treatment selection bias. 11 Preoperative clinic attenders (intervention group) were matched to nonattenders using similar preoperative variables (Tables E1 and E2). The McNemar test for matched pairs was used to compare groups.
Statistical methods. Clinical characteristics were summarized using descriptive statistics. Continuous variables were characterized using median and interquartile range; dichotomous or polytomous variables were characterized using frequencies. Between-group comparisons were evaluated using Wilcoxon rank-sum tests for continuous variables and Fisher exact tests for dichotomous and polytomous variables.
The monthly morbidity rate/mortality time series surrounding the introduction of the program changes was analyzed using an interrupted time series model. Each year was divided into 3 equal parts: January to April, May to August, and September to December. Within each time interval, outcome rates were computed. Percentages were subjected to empirical logistic transformation.

RESULTS
A total of 1064 operations performed in 1032 patients were included in the study. There were 784 patients in the "before" group and 280 patients in the "after" group. The baseline characteristics of the groups are displayed in Table 1. There was a significantly higher incidence of preoperative risk factors in the "after" group. The number of patients with other anatomy, with more than 3 previous chest wall incisions, and on preoperative anticoagulation, as well as bilirubin and creatinine levels, were all significantly higher. Sixty patients (22.1%) in the "after" group had other anatomy compared with 76 patients (9.7%) in the "before" group (P <.001). Table 2 describes expanded diagnostic categories for both groups. Eighty-five patients (31.9%) in the "after" group had 3 or more prior chest wall incisions compared with 82 (10.4%) in the "before" group (P<.001). Sixty-three patients (23.3%) in the "after" group were anticoagulated before surgery compared with 123 (15.7%) in the "before" group (P ¼ .006). The median bilirubin level in the "after" group was 0.70 mg/dL (0.53-1.08) compared with 0.64 mg/dL (0.41-0.88) in the "before" group (P<.001), and the median creatinine level in the "after" group was 0.87 mg/dL (0.78-1.00) compared with 0.84 mg/dL (0.74-0.97) in the "before" group (P <.003).
In-hospital mortality was 1.1% (3/274) in the "after" group compared with 4.3% in the "before" group (34/ 784) (P ¼ .012) ( Table 3). The interrupted time series model (Figures 1 and 2) shows that before the program changes the in-hospital mortality rate was decreasing at a yearly rate of 0.075% from January 2004 to December 2015. From January 2016 to July 2019, this decreased at a yearly rate of 0.36%.
One-year mortality in the "after" group was 1.3% (3/ 238) compared with 5.8% (38/657) in the "before" group (P ¼ .003). Analysis of the causes of death shows a heterogenous set of diagnoses with no discernible pattern. Procedures such as Ebstein surgery and Fontan revision feature highly in both groups, as might be expected (Table  E3). Other secondary outcomes are displayed in Table 3. The median postoperative ventilation time was shorter in the "after" group (5.5 [3.42-13.00] vs 6.38 hours [4.17-16.17], P ¼ .001). There was no difference in incidence of AKI requiring RRT (2.6% vs 3.3%, P ¼ .69) or stroke (1.5% vs 1.8%, P ¼ .79). There was also no difference in the composite outcome between the groups (5.6% vs 6.9%, P ¼ .57). At 1-year follow up, 6 of 7 patients with AKI in the "after" group had recovered their renal function and were free of dialysis and all were alive.
Chest reopening for bleeding or tamponade was reduced in the "after" group (1.8% vs 4.8%, P ¼ .022). Incidence of BPE was also not significantly different between the groups, with 65.7% (184/280) in the "after" group and 62.2% (488/ 784) in the "before" group being exposed (P ¼ .31). There were differences in exposure to individual blood products however. Plasma exposure was reduced in the "after" group (35.4% vs 43%, P ¼ .028), whereas platelet and cryoprecipitate exposure were significantly increased (52.1% vs   Figure 3) shows that before the program changes the rate of BPE was decreasing at a yearly rate of 0.41%. At the time of the program changes BPE increased by 9.03%. After the program changes, from January 2016 to July 2019, BPE increased at a yearly rate of 0.024%. Despite the introduction of PCC and fibrinogen, they were used in only 19.6% (55/280) and 21.8% (61/280) of patients in the "after" group respectively (Table 1). Sixtysix percent (184/280) of these patients received blood products, and 34% (96/280) did not. PCC was administered to  29.9% (55/184) of patients who received blood products compared with 0% of those who did not receive any blood products (P < .001). Thirty percent (56/184) of patients receiving blood products were administered fibrinogen compared with 5.2% (5/96) of those who did not receive any blood products (P <.001) ( Table 4).

Impact of Blood Product Exposure on Morbidity and Mortality
A total of 672 patients (63%) received blood products ( , and more postoperative AKI requiring RRT (4.9% vs 0%, P < .001). The inhospital mortality was higher in patients who received blood products (5.4% vs 0.3%, P < .001), and a higher number of them reached the composite outcome (10.0% vs 0.5%, P <.001).

Secondary Analysis
The secondary analysis included 285 surgeries on patients for a congenital lesion after January 1, 2016. A total of 234 patients attended the preoperative clinic (the intervention group), and 51 did not (the control group). Patient demographics are presented in 14.0% -12 -11 -10 -9 -8 -7 -6 -5 -4 Years since intervention In-hospital mortality Interrupted time series model for in-hospital mortality. The monthly in-hospital mortality rate surrounding the program changes was analyzed using an interrupted time series model. Each year was divided into 3 equal parts: January to April, May to August, and September to December. Within each time interval, outcome rates were computed. Percentages were subjected to empirical logistic transformation. Before the program changes, in-hospital mortality rates were decreasing at a yearly rate of À0.075% (95% CI, -0.179 to 0.028, P ¼ .154). At the time of the intervention, there is a decrease in in-hospital mortality rate of À0.091% (95% CI, -4.764 to 4.583, P ¼ .97). The sustained effect is a decrease in the yearly rate of change of the in-hospital mortality rate of À0.285% (95% CI, -0.908 to 0.339, P ¼ .37) after the intervention. The analysis suggests that after the program changes, the in-hospital mortality rate decreased at a yearly rate of 0.36%.
(5.6%) in the intervention group and 6 patients (11.8%) of the control group reached the composite outcome (P ¼ .12). There were 51 matched pairs. After matching the standardized mean difference (raw bias) of 33 of 37 variables decreased to less than 10% (Table E2). In the matched sample, the composite adverse outcome occurred in 9.8% of the treatment group versus 11.8% of controls (P ¼ 1).

DISCUSSION
Mortality after cardiac surgery in adults with congenital heart disease has decreased in our center after January 2016 when compared with the preceding 12 years. This decrease is coincident with the aforementioned changes in the ACHD cardiac surgical program. Both in-hospital and 1-year mortality were reduced. Not only was the overall in-hospital mortality rate reduced in the "after" group, but its rate of decrease was greater than in the "before" group, at a rate of 0.36% per year compared with 0.075%. Ventilation times were slightly reduced, and there was no significant difference in incidence of AKI requiring RRT or stroke. There was also no difference in the composite outcome.
This was all despite more complex patients with more complex anatomy, a larger number of patients with 3 or more previous chest wall incisions, and more anticoagulated patients. Although the "after" group had statistically higher creatinine and bilirubin levels, these differences were not clinically significant.
Our results compare favorably with the published literature where mortality ranges from 0.7% to 9.4%. [2][3][4][5] Our contemporary cohort also compares well to reported morbidity rates with a recent similar single-center study reporting rates of AKI of 2.1%, stroke of 1.4%, and prolonged ventilation of 11.4%. 2  In-hospital and one-year mortality after adult congenital cardiac surgery were reduced after the introduction of a quality improvement programme including a dedicated anesthesia team Mortality after adult congenital cardiac surgery; before and after the introduction of a quality improvement programme FIGURE 2. Mortality after adult congenital cardiac surgery, before and after the introduction of a quality improvement program, was assessed. A dedicated ACHD anesthesia team was introduced within the cardiac group. Use of factor concentrates was introduced contemporaneously. In-hospital and 1-year mortality rates were reduced. Incidence of AKI, stroke, and BPE was unchanged. Chest reexplorations were reduced. This was despite more complex patients. AKI, Acute kidney injury; ICU, intensive care unit; ACHD, adult congenital heart disease. ventilation greater than 7 days was 3.9% in the "after" group. Although overall BPE was unchanged between the 2 groups, the interrupted time series analysis shows that at the time of the program changes there was a sudden increase in BPE and that this increase was sustained. This was unexpected; however, the patients in the "after" group were more likely to be anticoagulated and have 3 or more previous chest wall incisions. Furthermore, point-of-care coagulation testing with a treatment algorithm was introduced for all cardiac surgery patients in 2015. These factors, along with the introduction of the specialist ACHD group and accompanying proactive coagulation profile management, could account for the increase. These patients received less plasma but more platelets and cryoprecipitate. This was despite the introduction of factor concentrates. Multiple small retrospective studies have shown reduced blood loss, lower red cell transfusion burden, and no increase in thromboembolic complications when using PCC after cardiopulmonary bypass compared with fresh frozen plasma. [12][13][14] A meta-analysis including 14 randomized controlled trials and 1035 patients has shown that fibrinogen use in cardiac surgery reduces mortality, blood loss, and numbers of units of red blood cells when compared with placebo or other comparators. 15 Use of factor concentrates in our study was limited. Of patients who received any blood product, only 29.9% of them were administered PCC and 30.4% fibrinogen. Of those patients who did not receive a transfusion, none of them were administered PCC and only 5.2% fibrinogen. This seems to suggest that factor concentrates were used sparingly in bleeding patients and not at all in those not being transfused. Perhaps incorporating factor concentrates into the point-of-care treatment algorithm, as reported previously, would increase use. 16,17 Those who received blood products had a higher inhospital mortality and incidence of AKI, and longer ventilation times. This is in line with previously published work showing increase in mortality and morbidity after transfusion in cardiac surgery. 18,19 Whether BPE itself is the cause of increased mortality or simply a marker of more complex patients is debatable. Patients who received blood products were significantly more comorbid than those who did not.
In any case, despite overall BPE not being reduced, a proactive approach to transfusion, the coagulation profile, and intravascular volume has resulted in exposure remaining at reference levels despite more anticoagulated patients who  had 3 or more previous sternotomies. Furthermore, the rate of chest reopening for bleeding or tamponade was also significantly reduced in the "after" group. Of all the heterogenous program changes, it was thought that the introduction of the specialist ACHD team within the cardiac anesthesia group was the change most likely to impact perioperative outcomes. Although there is evidence that the impact of the anesthesiologist on perioperative mortality after cardiac surgery is negligible, the same extensive study confirms that patient risk is the single most important predictor of outcome (accounting for 95.7% of mortality variation in 10 UK centers). 20 This group has previously established risk factors for morbidity and mortality after cardiac surgery in patients with ACHD. 6 It was hypothesized that a preoperative assessment by the ACHD anesthesia team could be associated with a reduction in mortality due to better patient selection and decision making. Preassessment clinics have been shown to reduce duration of hospital stay in a recent systematic review. 21 Previously used risk assessment tools in ACHD were developed for use in the pediatric population. They poorly predict mortality and morbidity in adults. 3,4,7 To that end, a secondary analysis of the "after" group was performed. Attendance at the preoperative clinic was the intervention assessed. Although the composite outcome was achieved by only 5.6% in the intervention group compared with 11.8% in the control group, this was not significant (P ¼ .12). In an effort to control for confounders, after 1:1 matching, there was no difference between the groups. There were only 51 matched pairs because only 51 patients did not attend the preassessment clinic. These were more likely to be urgent inpatients rather than elective cases. Although we cannot show that the preoperative assessment clinic alone had an impact on outcome, we believe that it does influence perioperative and postoperative management. Some benefits of the clinic, such as appreciation of previously unrecognized risk factors and delaying or postponing of surgery were difficult to measure. Furthermore, although we believe it to be an important aspect of the program changes, it is only one part of a larger heterogenous quality improvement initiative that these nonattenders still benefited from. Thus, it is difficult to demonstrate its exclusive effect.

Study Limitations
Although mortality has decreased significantly after this initiative, we have not succeeded in pinpointing why. The retrospective nature of our study prevented us from using established quality improvement methodology. Ideally, data should be collected prospectively to document the status quo, an intervention implemented and refined, and then further prospectively collected data would document the new paradigm. Each intervention should also be studied in isolation. We believed that the most significant intervention was the establishment of a specialist ACHD anesthesia team. The preoperative clinic was an effort to better risk stratify our patients and customize their management plans. With only 51 matched pairs, we were unable to show a benefit despite the overall composite outcome being lower in the intervention group. Our study did not account for the surgical procedure. Our center is a large quaternary referral center with a high proportion of complex procedures. Procedure complexity is associated with outcome, with mortality after atrial septal defect repair being reported as low as 0.3% but that after a Fontan conversion as high as 9%. 3,4 Cardiac transplant recipients were also excluded because the ACHD surgical team only assumed care of ACHD transplants in 2015.

CONCLUSIONS
We performed this study to assess the impact of a suite of heterogenous quality improvement interventions on outcomes after cardiac surgery in patients with ACHD. Both in-hospital and 1-year mortality have decreased. Ventilation times have been reduced. Incidence of AKI and stroke remain unchanged. This is despite a more complex patient cohort with more complex anatomy, more chest wall incisions, and more anticoagulation. However, we have not succeeded in proving that the outcome changes were specifically related to the intervention set, thus "association" needs to be emphasized over "causation." Our next step is to externally validate our identified patient risk factors and to build a risk prediction tool as has been done recently by a UK-based group. 9

Webcast
You can watch a Webcast of this AATS meeting presentation by going to: https://www.aats.org/resources/1474.

Conflict of Interest Statement
The authors reported no conflicts of interest. The Journal policy requires editors and reviewers to disclose conflicts of interest and to decline handling or reviewing manuscripts for which they may have a conflict of interest. The editors and reviewers of this article have no conflicts of interest.
Key Words: adult congenital heart disease, blood transfusion, cardiac surgery, morbidity, mortality, quality improvement, risk prediction
The authors studied a change in the protocol after the induction of a PCC and fibrinogen concentrate on 3 outcomes: mortality, adverse postoperative events, and transfusion burden. They studied 2 periods, one preceding 2016 before the institution of these changes and the 3 years after 2016 over 1000 patients. Congratulations on terrific outcomes with your complex patients. Their mortality rate decreased commensurately to approximately 1.1% from 4.3%, and this was also not just a short-term mortality improvement but one that went all the way to 1 year after surgery. I have 3 questions for the authors. The Society of Thoracic Surgeons ACHD working group, an effort that's been led by Dr Nelson, has been working to develop a risk model to improve the ascertainment of patients with ACHD and to appropriately risk adjust those patients whose risk previously has been extrapolated from pediatrics or adult-acquired patients' data. How did you risk adjust these patients? You showed a commensurate increase in more complex anatomy, but did you actually use any validated risk adjustment tool?
Dr Bill Walsh (Dublin, Ireland). Our institution has a risk predictor model that has been presented, and it's based on MELD-XI scores, BMI, presence of cognitive impairment, and the number of previous incisions and anatomy. They would be the factors we would use to attempt to predict postoperative course in our patients.
Dr Karamlou. Okay. The impact of dedicated teams has been studied by Mike Gaies and colleagues in the PHN trial with PC4. In that study, the authors demonstrated something tangible, important to our community, and that is the engagement of a multidisciplinary group, as well as traction with your center and follow-up, will result in a sustainable improvement after a change in protocol. Gaies and colleagues demonstrated that in early extubation protocols after the end of the trial, centers that adopted it and embraced it widely actually had a sustained improvement. Although you didn't present it, is how did you develop sustainability of these changes in protocols and how did you disseminate this throughout your heart center?
Dr Walsh. From January 2016, every patient who presented for ACHD surgery was managed by a small group of 3 of the staff cardiac anesthetists, and they would be seen in the preoperative assessment clinic only managed by those patients intraoperatively and managed by those in the postoperative intensive care unit also. That program has been sustained since inception.
Dr Karamlou. You mentioned an interrupted time series model. Did you use a washout period to avoid contamination of data from one period to the next?
Dr Walsh. Not that I'm aware of. Unidentified Speaker 1. This risk stratification has as one of the objectives that of averting the occurrence of nonsurgical bleeding postoperatively. You're just ready with more blood products and you stratify the patient. So, I haven't seen your incidence of reexploration from bleeding. Was it higher? Was it lower? Higher because of the higher complexity or did you notice a change?
Dr Walsh. That wasn't a variable we recorded, but anecdotally, with the increased complexity and blood product use, there possibly was a slight increased reexploration, but I don't know that for certain. We didn't collect those data.
Unidentified Speaker 1. What is your takeaway from this?
Dr Walsh. My takeaway is that narrowing the team and being responsible for looking after these complex patients with higher throughput and better individual skill sets will hopefully lead to reduced mortality and the ability to better risk stratify and deal with more complex patients.