PICO

Dedicated Operating Room Teams and Clinical Outcomes in an Enhanced Recovery after

Surgery Pathway for Colorectal Surgery

Michael C Grant, MD, Andrew Hanna, MD, Andrew Benson, CRNA, Deborah Hobson, BSN, Christopher L Wu, MD, Christina T Yuan, PhD, Michael Rosen, PhD, Elizabeth C Wick, MD, FACS

BACKGROUND: Our aim was to determine whether the establishment of a dedicated operating room team leads to improved process measure compliance and clinical outcomes in an Enhanced Recovery after Surgery (ERAS) program. Enhanced Recovery after Surgery programs involve the application of bundled best practices to improve the value of perioperative care. Successful implementation and sustainment of ERAS programs has been linked to compliance with protocol elements.

STUDY DESIGN: Development of dedicated teams of anesthesia providers was a component of ERAS imple- mentation. Intraoperative provider team networks (surgeons, anesthesiologists, and certified registered nurse anesthetists) were developed for all cases before and after implementation of colorectal ERAS. Four measures of centrality were analyzed in each network based on case assignments, and these measures were correlated with both rates of process measure compliance and clinical outcomes.

RESULTS: Enhanced Recovery after Surgery provider teams led to a decrease in the closeness of anesthe- siologists (p ¼ 0.04) and significant increase in the clustering coefficient of certified registered nurse anesthetists (p ¼ 0.005) compared with the pre-ERAS network. There was no sig- nificant change in centrality among surgeons (p ¼ NS for all measures). Enhanced Recovery after Surgery designation among anesthesiologists and nurse anesthetistsdwhereby individual providers received an in-service on protocol elements and received compliance data was strongly associated with high compliance (>0.6 of measures; p < 0.001 for each group). In addition, high compliance was associated with a significant reduction in length of stay (p < 0.01), surgical site infection (p < 0.002), and morbidity (p < 0.009).

CONCLUSIONS: Dedicated operating room teams led to increased centrality among anesthesia providers, which in turn not only increased compliance, but also improved several clinical outcomes. (J Am Coll Surg 2018;226:267e276.� 2017 by the American College of Surgeons. Published by Elsevier Inc. All rights reserved.)

CME questions for this article available at http://jacscme.facs.org

Disclosure Information: Authors have nothing to disclose. Timothy J Eberlein, Editor-in-Chief, has nothing to disclose.

Drs Grant and Hanna contributed equally to this work.

Presented at the American College of Surgeons 102nd Annual Clinical Congress, Washington, DC, October 2016.

Received November 6, 2017; Revised December 1, 2017; Accepted December 1, 2017.

From the Departments of Anesthesiology and Critical Care Medicine (Grant, Benson, Wu, Yuan, Rosen) and Surgery (Hobson), the Johns Hop- kins Medical Institutions, Baltimore, MD, Department of Surgery, Univer- sity of Pennsylvania, Philadelphia, PA (Hanna), and Department of Surgery, University of California, San Francisco, CA (Wick).

Correspondence address: Elizabeth C Wick, MD, FACS, Department of Surgery, University of California San Francisco, 513 Parnassus Ave, San Francisco, CA 94143-2205. email: [email protected]

267 ª 2017 by the American College of Surgeons. Published by Elsevier Inc. All rights reserved.

Patients undergoing operations often see numerous pro- viders across a number of perioperative patient care envi- ronments. This leads to substantial variation and even introduces the potential for disparate levels of patient care. During the past several years, surgeons have begun to address variability in perioperative management through the creation of Enhanced Recovery after Surgery (ERAS) programs. Enhanced Recovery after Surgery has sparked excitement in the surgical community both domestically and abroad due to its ability to lessen the impact of surgical insults and hasten patient recovery, as adoption of ERAS programs has been associated with a reduction in rates of surgical complications,

https://doi.org/10.1016/j.jamcollsurg.2017.12.010

ISSN 1072-7515/17

Abbreviations and Acronyms

CRNA ¼ certified registered nurse anesthetist ERAS ¼ Enhanced Recovery after Surgery LOS ¼ length of stay OR ¼ operating room QAP ¼ quadratic assignment procedure SSI ¼ surgical site infection

268 Grant et al Operating Room Teamwork and Enhanced Recovery J Am Coll Surg

decreased length of hospital stay, and improvement in overall patient satisfaction.1-4 Much of the literature to date has focused on the quality of evidence associated with individual process measures included within these perioperative bundles.5,6 However, an underappreciated element of ERAS programs is the ability to promote teamwork and coordination of care, which serves to break down pre-existing silos within the various perioperative care arenas. The success of ERAS programs is inextricably linked

not only to the individual process measures, but also to the associated culture and adaptive management skills of the improvement team and clinician champions.7

One way to foster the development of these adaptive skills is through the creation of operating room (OR) teams, which involves the engagement and support of surgical leadership, incorporation of select perioperative anesthesia and nursing providers, and concerted “cohorting” of patient care. Although there is limited evidence to suggest that dedicated OR teams can increase productivity and enhance teamwork and safety climate,8 the impact of this approach has been difficult to study in a systematic fashion. In other industries and areas of medicine, the degree of teamwork has been characterized through the application of social network analysis, which allows for the quantification of inter-provider communication and organization through various metrics of social interaction. At least one group found that improved teamworkdas captured by a network measure of provider overlap with patientsdled to better outcomes after coronary artery bypass grafting.9 In this light, our group sought to better understand the potential impact of ERAS programs on teamwork through the use of social network analysis on a recently implemented ERAS for colorectal surgery program. Our hypothesis was that creating OR teams through the identification of surgical, anesthesia, and nursing leadership coupled with recruitment of desig- nated ERAS providers across the perioperative environ- ment would lead to both improvements in metrics of teamwork, as well as surgical outcomes.

METHODS

Enhanced Recovery after Surgery program description

The Enhanced Recovery after Surgery for colorectal surgery program

After IRB approval, the ERAS program was implemented in February 2014 at Johns Hopkins Hospital, a 1,056-bed quaternary care hospital, in all patients scheduled for colorectal surgery by 1 of 5 surgeons. Patients were educated and individual consent was obtained for partic- ipation in the ERAS program. The pathway included bundled preoperative, intraoperative and postoperative process measures encompassing the core principles of ERAS (ie enhanced patient engagement, reduced opioid consumption, avoidance of prolonged fasting, early ambulation, goal-directed IV fluids, and best evidence to reduce preventable harms). As reported previously, implementation resulted in a significant reduction in length of stay, surgical site infections, and variable direct costs, as well as improvement in the patient experience.9-11

Designation of operating room teams

Before the implementation of the ERAS program, anes- thesia providers (attending anesthesiologist and anesthesia resident or certified registered nurse anesthetist [CRNA]) were assigned ad hoc to colorectal surgical cases by a cen- tral anesthesiologist scheduler. Providers were selected from the entire pool of clinical providers assigned to a particular clinical day. To maximize staffing resources to all OR locations, anesthesia providers were not preferen- tially scheduled to specific procedures or surgeons. As part of the official ERAS program implementation, a dedicated team of colorectal ERAS anesthesia providers was identified and a pilot program developed to ensure only core ERAS providers were assigned to these cases. Af- ter development of a formal ERAS protocol, volunteers from the faculty and CRNA pool were identified who were interested in joining the core ERAS provider team. The training associated with certification to be part of the ERAS provider cohort involved the following: self- identification of providers (CRNAs and anesthesiologists) who would be open to review the available consensus guidelines literature, were comfortable with practicing the anticipated anesthesia protocol elements and willing to standardize their approach to colorectal anesthesia de- livery; orientation to the protocol during one of the monthly quality meeting for the cohort of CRNAs and attending anesthesiologists who would be involved in the initiative; inclusion of providers among a specific ERAS listserv to receive ongoing updates (ie new literature

Figure 1. Graphic representation of (A) pre-Enhanced Recovery af- ter Surgery (ERAS) program and (B) ERAS networks. Blue node, surgeon; red node, anesthesiologist; yellow node, certified regis- tered nurse anesthetist (CRNA). Size of each node is based on relative degree of centrality. Internode line thickness corresponds to number of shared operations. Node shape is based on ERAS designation: diamond, ERAS; circle, non-ERAS.

Vol. 226, No. 3, March 2018 Grant et al Operating Room Teamwork and Enhanced Recovery 269

and protocol developments). Colorectal ERAS procedures were marked on the OR schedule ahead of time by the surgeons’ staff and the anesthesiologist scheduler began to assign only members of the core ERAS provider team to these cases. The goal was to provide at least an ERAS core anesthesiologist or a core ERAS CRNAdand ideally bothdto all colorectal ERAS procedures. Anesthesiology residents were not included as part of the core ERAS pro- vider team because of frequent rotation and inability to ensure consistent coverage. Furthermore, it was believed that an additional provider type might undermine suc- cessful evaluation of the impact of a dedicated team of anesthesia providers on clinical outcomes.

Data variable extraction

Operating room staffing

The intraoperative anesthesia record (Metavision; iMdSoft) was used to identify the surgeon, anesthesiolo- gist, and CRNA of record for each individual procedure. In the event that more than 1 person was entered in either role, the person who was present for the longest period of time was determined to be the primary provider for the procedure. This information was transcribed into a pre- scribed proforma.

Clinical outcomes

The primary end point of interest was overall index hos- pitalization length of stay (LOS). Secondary outcomes included surgical site infection (SSI) and overall postoper- ative morbidity, which were obtained from the hospital American College of Surgeons NSQIP database, as well as rate of return to the OR, which was obtained from the hospital administrative data records.

Enhanced Recovery after Surgery program process measures

The charting location of all preoperative, intraoperative and postoperative anesthesia process measures, as outlined in the institutional anesthesia protocol, were identified in the electronic medical record (Sunrise Clinical Manager; Eclipsys Inc) and individual queries were developed to facilitate their automated extraction to a centralized ERAS database. This permitted updated information about both individual and process measure compliance throughout the duration of the program.

Network development and analysis of centrality

Case-specific assignments for 3 separate provider typesdattending surgeons, attending anesthesiologists, and CRNAsdare recorded for each colorectal surgical procedure during the study period. Two social networks were first constructed: one describing the Pre-ERAS

network from January 1, 2013 to January 31, 2014, and the second describing the ERAS network from February 1, 2014 to February 15, 2016. This network was created with the following parameters: individual providers are represented as “nodes,” any providers that were assigned to the same procedure were connected by “ties”dphysical lines adjoining 2 nodes, the size of each node correlates to the number of times that individ- ual provider was assigned to a surgical procedure during the study period, and the thickness of each tie correlates with the number of times those 2 providers shared a pro- cedure. Provider role is indicated by color, with sur- geons, anesthesiologists, and CRNAs as blue, red, and yellow, respectively. Within the ERAS network only, node shape corresponds to ERAS certification, with diamond-shaped nodes indicating certification and cir- cular nodes indicating no certification. Figure 1 displays both the pre-ERAS and ERAS networks using these conventions. For the sake of this study, we were interested in the

concept of node centrality because it provides quantitative measures to either identify the most influential

270 Grant et al Operating Room Teamwork and Enhanced Recovery J Am Coll Surg

nodesdones that most dictates the overall data flowdor the nodes that promote the greatest level of network cohe- siveness. Our group evaluated the pre-ERAS and ERAS networks using 4 measures of centrality, which include:

1. Degree centrality: Perhaps the simplest to conceptu- alize, this is defined as the total number of ties on a given node. It best estimates the likelihood the node will transfer or receive data or information within the network.

2. Closeness centrality: Focuses on how close a node is to all the other nodes in the network. Central nodes are characterized by having “minimum steps” to all other nodes; meaning that the paths (ie geodesics) linking a central node to the other nodes are as short as possible.12 Greater closeness is associated with greater influence, as a node does not need to rely on the other nodes for its interactions because it is tied to all others.

3. Eigenvector centrality: Can be considered the most direct measure of the influence of a node on all other nodes in the network. In simple terms, it assigns weight to a node based on the strength of all the nodes with which it forms ties. If a node is connected to high-ranking node, its own ranking is strengthened.

4. Clustering coefficient: At the whole network level, this is a measure of the overall connectedness of a set of nodes. Ranging from 0 to 1, it is a ratio of the number of actual connections between a set of nodes over the maximum number of possible connections between

Figure 2. Social network analysis terms, definitions, an anesthetist; OR, operating room.

those nodes. When referring to a single node, the clus- tering coefficient measures the extent to which neigh- boring nodes are connected.

Figure 2 describes several social network terms and pro- vides a correlate to the current study.

Statistical analysis

All relevant data for each surgical encounter were com- bined into a single database. Data were processed and analyzed with the software programs Excel, version 14.0 (Microsoft Inc) and STATA, version 14.2 (Stata Corp) statistical package. Network analysis was per- formed using UCINET 6 (Analytic Technologies). Cen- trality measures calculated from any social network are, by design and definition, non-independent observations, meaning that using ordinary least squares or logistic regression, which assume observational independence, would greatly overstate the statistical significance of any effect size. To circumvent this problem, social network analysis uses a method called “quadratic assign- ment procedure” (QAP) to generate a sampling distribu- tion for effect size comparison. This is done by permutating the dependent variable across the indepen- dent variables 10,000 times and creating 10,000 completely new networks, each with its own effect esti- mate. The actual effect size seen in the original network is then compared with this empirically derived sampling distribution to generate a p value. Comparisons of means and multivariate logistic regressions where any

d study correlates. CRNA, certified registered nurse

Vol. 226, No. 3, March 2018 Grant et al Operating Room Teamwork and Enhanced Recovery 271

network measure was involved was therefore performed use QAP. Categorical variables not involving network metrics were assessed by the chi-square or Fisher’s exact test, where appropriate. Univariable regression using QAP was performed for outcomes variables of interest based on provider type, ERAS designation, and central- ity measures. Multivariable logistic regression using QAP was conducted for each of the outcomes variables of in- terest incorporating all results of univariable analysis with p < 0.1. For the sake of this study, p < 0.05 was considered statistically significant.

RESULTS

Pre-Enhanced Recovery after Surgery vs Enhanced Recovery after Surgery global network comparison

The ERAS network was associated with a statistically sig- nificant increase in the overall clustering coefficient (0.81 vs 0.51; p ¼ 0.009) compared with the pre-ERAS pro- vider network. This finding was not associated with a sig- nificant difference in overall degree (10.0 ties vs 9.1 ties; p ¼ 0.53) between ERAS and pre-ERAS networks, despite an increase in both numbers of operations (763 vs 401) and number of involved provider personnel (204 vs 169). Visual representation of the network was constructed so that strongly tied nodes are closer together than weakly tied nodes. Figure 1 graphically depicts both the pre-ERAS and ERAS networks. On qualitative inspec- tion, the ERAS network is more densely populateddwith more nodes and tiesdthan the pre-ERAS network. The ERAS network also appears to display a large clustering

Table 1. Comparing Normalized Centrality Measures between P after Surgery

Characteristic Whole network

Surgeon

Degree 0.418

Closeness 0.604

Eigenvector 0.343

Clustering coefficient 0.145

Anesthesiologist

Degree 0.041

Closeness 0.453

Eigenvector 0.070

Clustering coefficient 4.09

Certified registered nurse anesthetist

Degree 0.041

Closeness 0.450

Eigenvector 0.067

Clustering coefficient 3.83

*Statistically significant. ERAS, Enhanced Recovery after Surgery program.

effect between a select group of individuals, much more so than the pre-ERAS network. In examining the mem- bers involved in the dense clustering of the ERAS network, there is noted to be a redistribution of cases to select personnel. In illustration of this, node sizesdwhich correlate with node degreedshow that attending anesthe- siologists (red nodes) and CRNAs (yellow nodes) are relatively more central to the ERAS network than in the pre-ERAS network, where the central members were over- whelmingly surgeons (blue nodes), as indicated by their relatively larger node sizes.

Pre-Enhanced Recovery after Surgery vs Enhanced Recovery after Surgery role comparisons

As shown in Table 1, the centrality measures of surgeons differed significantly from both Anesthesiologists and CRNAs within the whole network. These differences were all statistically significant (p < 0.001). There was no statistical difference among any centrality measure when comparing anesthesiologists with CRNAs. There is an increase in surgeon clustering coefficient, albeit a nonsignificant one (p ¼ 0.085), from pre-ERAS to ERAS, which suggests a more strongly connected neigh- borhood of anesthesiologists and CRNAs. There is a sta- tistically significant increase in CRNA clustering coefficient from pre-ERAS to ERAS (p ¼ 0.005), indi- cating that the surgeon-anesthesiologists network became more strongly connected as well. Anesthesiologists exhibited a decrease in average closeness from pre-ERAS to ERAS network (p ¼ 0.039).

re-Enhanced Recovery after Surgery and Enhanced Recovery

Pre-ERAS ERAS % Change p Value

0.400 0.437 9 0.725

0.595 0.613 3 0.759

0.337 0.348 3 0.877

0.111 0.178 60 0.082

0.043 0.040 �7 0.536 0.471 0.435 �8 0.0386* 0.078 0.063 �20 0.111 3.94 4.22 7 0.742

0.044 0.040 �10 0.458 0.456 0.444 �7 0.084 0.071 0.063 �12 0.287 2.81 4.61 64 0.005*

Table 2. Multivariate Regression of Role and Enhanced Recovery after Surgery Certification on Centrality Measures

Centrality measure Effect of CRNA vs anesthesiologist p Value Effect of ERAS (vs non-ERAS) p Value

Degree 2.39 0.032* 7.956 <0.0001*

Closeness 0.0114 0.083 0.0372 <0.0001*

Eigenvector 0.0113 0.03* 0.037 <0.0001*

Clustering coefficient 0.092 0.475 0.915 0.037*

Predictor variables included role (CRNA vs anesthesiologist) and ERAS certification (ERAS vs non-ERAS). *Statistically significant. CRNA, certified registered nurse anesthetist; ERAS, Enhanced Recovery after Surgery program.

272 Grant et al Operating Room Teamwork and Enhanced Recovery J Am Coll Surg

Role and Enhanced Recovery after Surgery desig- nation comparisons within the Enhanced Recovery after Surgery network

Within the ERAS network, there are minimal differences in centrality measures noted between the overall role of anesthesiologist and CRNA. However, when analyzing these end points based on ERAS designation, all 4 central- ity measures were shown to have a statistically significant increase in ERAS compared with non-ERAS personnel. The results of bivariable regression analyses using role and ERAS designation to predict centrality measures are shown in Table 2. The provider role of CRNA (irrespec- tive of ERAS designation) was associated with a signifi- cant increase in both degree and eigenvector within the ERAS network compared with the role of anesthesiolo- gist. In addition, ERAS designation (irrespective of pro- vider role) was associated with a significant increase in all centrality measures compared with non-ERAS designation.

Effect of provider role and Enhanced Recovery after Surgery designation on clinical outcomes

Overall rates of individual clinical outcomes within the ERAS network based on both provider role and ERAS designation are reported in Table 3. As shown, within the CRNA provider role, ERAS designation was associ- ated with a statistically significant increase in the rate of high compliance (defined as >0.6 compliance with ERAS process measures; p < 0.001), as well as a

Table 3. Effect of Team Member Designations on Primary Out

Outcomes

Certified registered nurse anesth

ERAS (n ¼ 334)

Non-ERAS (n ¼ 73)

n % n %

Compliance score >0.6 242 72.5 17 23.3

Length of stay <6 d 164 49.1 33 45.2

Surgical site infection 34 10.2 12 16.4

All morbidity 66 19.8 23 31.5

Return to operating room 9 2.7 6 8.2

*Statistically significant. ERAS, Enhanced Recovery after Surgery program.

significant decrease in the rate of postoperative morbidity (p ¼ 0.031), and a decrease in the rate of return to the operating room (p ¼ 0.035). Within the anesthesiologist provider role, ERAS designation was associated with a sta- tistically significant increase in the rate of high compliance (p < 0.001). Table 4 reports the association between the overall

anesthesia team ERAS designation and clinical out- comes. When at least one member of the team is ERAS designated, there is a significant increase in the number of high compliance cases (p < 0.001), as well as cases with LOS <6 days (p ¼ 0.048) and a significant decrease in the rate of SSI (p ¼ 0.007) and return to the OR (p ¼ 0.029) compared with procedures when there are no ERAS designated anesthesia team members. When both members of the anesthesia team are ERAS designated, there is a significant increase in the number of high compliance cases compared with teams where at least one member is not ERAS designated (p < 0.001). Interestingly, there are no significant differences in the rest of the clinical outcomes between these 2 groups.

Effect of provider role, Enhanced Recovery after Surgery designation, and individual centrality mea- sures on clinical outcomes

A multivariable regression model using the previously described QAP method was developed to determine in- dependent predictors of various clinical outcomes. The

comes within the Enhanced Recovery after Surgery Network

etist Anesthesiologist

p Value

ERAS (n ¼ 342)

Non-ERAS (n ¼ 65)

p Valuen % n %

<0.001* 235 68.7 24 36.9 <0.001*

0.606 179 52.3 31 47.7 0.502

0.126 38 11.1 8 12.3 0.780

0.031* 73 21.3 15 23.1 0.756

0.035* 11 3.2 3 4.6 0.570

Table 4. Effect of Anesthesia Team Designation on Primary Outcomes within the Enhanced Recovery after Surgery Network

Outcomes

Both ERAS (n ¼ 297)

At least 1 non-ERAS (n ¼ 110)

p Value

At least 1 ERAS

(n ¼ 379)

Both non- ERAS

(n ¼ 28) p Valuen % n % n % n %

Compliance score >0.6 219 73.7 40 36.4 <0.001* 258 68.1 1 3.6 <0.001*

Length of stay <6 d 157 52.9 51 46.4 0.316 188 49.6 8 28.6 0.048*

Surgical site infection 32 10.8 14 12.7 0.581 37 9.8 8 28.6 0.007*

All morbidity 57 19.3 31 27.7 0.080 80 21.1 9 32.1 0.233

Return to operating room 9 3.2 5 4.6 0.570 11 2.9 3 10.7 0.029*

*Statistically significant. ERAS, Enhanced Recovery after Surgery program.

Vol. 226, No. 3, March 2018 Grant et al Operating Room Teamwork and Enhanced Recovery 273

results of this analysis are shown in Table 5. Based on multivariable logistic regression using QAP, CRNA closeness, and use of an anesthesiologist or CRNA with ERAS designation were independently associated with high compliance with ERAS process measures. When high compliance was added into the multivariable regression model, it was independently associated with cases that met the goal LOS (<6 days), as well as reduced SSI and overall morbidity. In addition, central- ity measure of CRNA closeness was independently asso- ciated with meeting goal LOS (p ¼ 0.021) and reduction in returns to the OR (p ¼ 0.031). The cen- trality measure of CRNA clustering coefficient was inde- pendently associated with reduction in overall morbidity (p ¼ 0.038) and reduction in returns to the OR (p ¼ 0.019).

Table 5. Significant Results of Multivariable Logistic Regressi

Outcomes and predictor variable Odds ratio

Compliance score > 0.6*

Anesthesiologist, ERAS designation 3.09

CRNA, ERAS designation 5.83

CRNA closeness 1.21

Length of stay <6 dz

CRNA closeness 1.13

Compliance score >0.6 1.92

Surgical site infection, compliance score >0.6z 0.49

All morbidityz

CRNA clustering coefficient 0.76

Compliance score >0.6 0.51

Return to the operating roomz

CRNA closeness 0.67

CRNA clustering coefficient 0.33

*Covariables included ERAS designation, centrality measures for anesthesiologis in p < 0.1). yStatistically significant. zCompliance score >0.6 was also included among covariables for the length of s room multivariable regression analyses. CRNA, certified registered nurse anesthetist; ERAS, Enhanced Recovery after S

DISCUSSION Analysis of the social networks associated with the recruit- ment of designated anesthesia providers as part of the implementation strategy of an ERAS program for colo- rectal surgery resulted in a significant increase in the over- all centrality of the cohort. Not only was the ERAS network shown to have a significant increase in the overall clustering coefficient, but there were a number of provider type-specific changes in several other measures of central- ity as well, with a significant increase in clustering among CRNAs and significant decrease in closeness among anes- thesiologists. Enhanced Recovery after Surgeryedesig- nated anesthesia providersdboth anesthesiologists and CRNAsdwere noted to have significant increases in all 4 measures of centrality as well. Enhanced Recovery after Surgery designation was shown to correlate positively with

on for Clinical Outcomes

Lower 95% Upper 95% p Value

1.47 6.51 <0.001y

2.60 13.11 <0.001y

1.05 1.39 0.006y

1.01 1.28 0.021y

2.94 1.23 0.010y

0.26 0.95 0.020y

0.60 0.97 0.038y

0.31 0.84 0.009y

0.46 0.96 0.031y

0.13 0.83 0.019y

t, CRNA, and surgeon from univariable analysis (any measure that resulted

tay <6 d, surgical site infections, all morbidity, and return to the operating

urgery.

274 Grant et al Operating Room Teamwork and Enhanced Recovery J Am Coll Surg

process measure compliance, an effect that ultimately led to downstream effects on several major clinical outcomes. High compliance was independently associated with reduced LOS, SSIs, and morbidity. In short, each of the clinical outcomes evaluated were positively influenced through the recruitment of dedicated ERAS-certified providers. Several more granular conclusions can be made from

these network data. Importantly, despite a relative increase in the potential size of the network on implemen- tation of ERASdas reflected by an increase in the number of cases involved in our analysisdthere was not a signif- icant increase in the overall degree of the network. At first glance, this might suggest that centrality is shifted from one provider type (surgeons) to another (anesthesia pro- viders). This plausible explanation might stem from the fact that our intervention specifically designated anes- thesia providers to ERAS cases, and surgeons’ roles largely continued unchanged. However, this explanation is un- likely, as there was no significant difference in degree among all provider groups with ERAS implementation. Far more plausible an explanation lies in the interpreta- tion of the other measures of centrality. Although degree is a measure of volume (ie the number of ties per node), clustering coefficient is a measure of interaction (ie how strongly connected the network is). Our study notes a sig- nificant increase in clustering coefficient among CRNAs with a concomitant increase (albeit nonsignificant one) among surgeons. Put simply, the ERAS network is more strongly connected than the pre-ERAS network independent of its larger size. In addition, our data show a decrease in closeness among anesthesiologists, which is most likely due to an inherent consolidation of their role associated with the introduction of the ERAS program. The goal of this study is to determine whether incorpo-

rating an intraoperative ERAS team into a hospital’s pro- gram improves care processes and patient outcomes. The notion that teamwork impacts clinical processes and out- comes is not new.13 In general, team performance is described in terms of inputs (ie relatively stable character- istics of the team, its members, task, and setting), mediators or processes (ie their interactions with one another prepar- ing for, completing, and reflecting on their work), and out- comes (ie performance outcomes, learning).14 Most interventions (eg team training, surgical team check- lists)15,16 for improving teamwork in surgical services have focused on the team’s interactions or processes. This work is unique in that it focuses on the composition of the team (ie attributes of the team members). Although our study did not assess changes in team interaction, there are 2 logical mechanisms from earlier team research that

could explain the observed effect: team familiarity and shared mental models. First, teams whose members have a history of working together perform better than those that do not.17,18 The scheduling component of this inter- vention increased the regularity with which people on the ERAS team shared patient cases and had the opportunity to become familiar with one another. Second, teams that hold shared mental models (eg similar or complementary ideas about which protocols are appropriate for a given pa- tient and what those protocols entail) do better than teams that do not.19 By volunteering to become a part of the ERAS team, providers were given detailed information about ERAS protocols and ERAS patients were clearly designated to ensure all knew when those protocols were appropriate. Although all 4 measures of centrality provide unique insight as to the predominance of any specific node (or provider type), the sum effect of the available data sug- gest that the introduction of the ERAS programdid, in fact, lead to more teamwork between providers. In fact, the centrality measures describing improved

teamwork were positively associated with several noted clinical outcomes. Enhanced Recovery after Surgery designation and CRNA closeness were associated with increased process measure compliance, which was, in turn, associated with reductions in LOS, SSI, and post- operative morbidity. Several measures of centrality among CRNAs were also associated with these end points, including reduction in rates of reoperation. This large and widespread effect of the CRNA pro- viders, both in terms of their ERAS designation and their centrality measures, is quite interesting. It is logical to conclude that regular assignment to familiar cases (and protocols) would lead to improved rates of compliance, but it is more compelling to show that it further correlates with improved clinical outcomes. Others have previously shown that increased compli- ance with ERAS protocols is associated with improved outcomes,4 but obtaining strong compliance remains a difficult aspect of clinical care. Within the analyzed OR network, CRNAs are the decision makers of many of the variables that the ERAS protocol describes in the preoperative and intraoperative setting, such as fluid management, sedation, and narcotic use. It is clear that successful compliance of ERAS and its effect on improved outcomes in this cohort were largely affected by the ERAS certification of CRNAs. This suggests that at least one approach to improving compliance, and therefore outcomes, is to build designated teams of pro- viders, particularly those at the forefront of adminis- tering care. To initiate this program, our strategy included identifi-

cation of a core group of attending anesthesiologists and

Vol. 226, No. 3, March 2018 Grant et al Operating Room Teamwork and Enhanced Recovery 275

CRNAs. This was done at the exclusion of another front- line provider group, the anesthesiology residents. This was done in an effort to further reduce the potential for vari- ation in practice, as residents rotated frequently and the volume of the residency (more than 120 residents during the study period) posed a particular challenge in providing adequate initial education and consistent pro- gram updates. This was done because the idea of following a protocol for anesthesia in the general ORs was a new concept and we wanted to control the number of people who were involved in the pilot. Although there were certainly isolated examples where inexperienced/ uncohorted providers overlapped with “uncertified ERAS providers,” this was very uncommon. More recently, as the ERAS program expanded to other services (liver resection, urology, gynecologic oncology), we began to incorporate the residents as rotating providers as well, provided they undergo training through a similar didactic as the CRNA cohort. There are several important limitations to this study.

The results of our study are based on the selection of a pri- mary provider for each individual procedure. Although our selection accounts for identification of providers with the greatest contact time in any given case, our anal- ysis does not capture the potential for alterations in anes- thetic plans or process measure compliance associated with intraoperative provider relief. In addition, it should be recognized that before-and-after analysis, as performed in this study, is potentially confounded by uncontrolled variables, parallel patient care initiatives, and even simple regression to the mean that cannot be entirely accounted for in multivariable analysis. That said, our model does incorporate patient-related variables, including demo- graphics, American Society of Anesthesiology score, and procedure-specific variables, most notably open vs laparo- scopic approach. Finally, our study was conducted in the setting of a large, quaternary academic institution. As such, staffing models, resource allocation and scheduling should all be accounted for when attempting to reproduce similar results in an alternative setting.

CONCLUSIONS In an era where hospitals are increasingly seeking to increase the value of health care, ERAS programs have been successful not only in reducing inter-provider vari- ability, but in improving postoperative outcomes. To ensure evidence-based process measures are used, pro- viders are tasked with developing mechanisms to improve overall compliance with program elements. Through the designation of dedicated ERAS provider teams, our study can shed light on at least one key aspect to not only

improve compliance, but extract increased value from the perioperative care environment.

Author Contributions

Study conception and design: Hanna, Rosen, Wick Acquisition of data: Grant, Hanna, Benson, Hobson, Wu, Wick

Analysis and interpretation of data: Grant, Hanna, Benson, Hobson, Wu, Yuan, Rosen, Wick

Drafting of manuscript: Grant, Hanna, Yuan, Rosen, Wick

Critical revision: Grant, Hanna, Benson, Hobson, Wu, Yuan, Rosen, Wick

REFERENCES

1. Wind J, Polle SW, Fung Kon Jin PH, et al. Systematic review of enhanced recovery programmes in colonic surgery. Br J Surg 2006;93:800e809.

2. Grant MC, Yang D, Wu CL, et al. Impact of Enhanced Recovery after Surgery and fast track surgery pathways on healthcare-associated infections: results from a systematic re- view of meta-analysis. Ann Surg 2017;265:68e79.

3. Cakir H, van Stijn MF, Lopes Cardozo AM, et al. Adherence to enhanced recovery after surgery and length of stay after colonic resection. Colorectal Dis 2013;15:1019e1025.

4. ERAS Compliance Group. The impact of enhanced recovery protocol compliance on elective colorectal cancer resection: re- sults from an international registry. Ann Surg 2015;261: 1153e1159.

5. Larson DW, Lovely JK, Cima RR, et al. Outcomes after implementation of a multimodal standard care pathway for laparoscopic colorectal surgery. Br J Surg 2014;101: 1023e1030.

6. Bakker N, Cakir H, Doodeman HJ, et al. Eight years of expe- rience with enhanced recovery after surgery in patients with colon cancer: impact of measures to improve adherence. Sur- gery 2015;157:1130e1136.

7. Wick EC, Galante DJ, Hobson DB, et al. Organizational culture changes result in improvement in patient-centered outcomes: implementation of an integrated recovery pathway for surgical patients. J Am Coll Surg 2015;221: 669e677.

8. Proczko M, Kaska L, Twardowski P, et al. Implementing enhanced recovery after bariatric surgery protocol: a retrospec- tive study. J Anesth 2016;30:170e173.

9. Hollingsworth JM, Funk RJ, Garrison SA, et al. Association between physician teamwork and health system outcomes after coronary artery bypass grafting. Circ Cardiovasc Qual Out- comes 2016;9:641e648.

10. Stone AB, Grant MC, Pio Roda C, et al. Implementation costs of an Enhanced Recovery after Surgery program in the United States: a financial model and sensitivity analysis based on expe- riences at a quaternary academic medical center. J Am Coll Surg 2016;222:219e225.

11. Fabrizio AC, Grant MC, Siddiqui Z, et al. Is enhanced recov- ery enough for reducing 30-d readmissions after surgery? J Surg Res 2017;217:45e53.

276 Grant et al Operating Room Teamwork and Enhanced Recovery J Am Coll Surg

12. Wasserman S, Faust K. Social Network Analysis: Methods and Applications. Cambridge, UK: Cambridge University Press; 1994.

13. Schmutz J, Manser TD. Do team processes really have an ef- fect on clinical performance? A systematic literature review. Br J Anaesth 2013;110:529e544.

14. Ilgen DR, Hollenbeck JR, Johnson M, et al. Teams in organi- zations: from input-process-output models to IMOI models. Annu Rev Psychol 2005;56:517e543.

15. Neily J, Mills PD, Young-Xu Y, et al. Association between implementation of a medical team training program and sur- gical mortality. JAMA 2010;304:1693e1700.

16. Levy SM, Senter CE, Hawkins RB, et al. Implementing a sur- gical checklist: more than checking a box. Surgery 2012;152: 331e336.

17. Xu R, Carty MJ, Orgill DP, et al. The teaming curve: a lon- gitudinal study of the influence of surgical team familiarity on operative time. Ann Surg 2013;258:953e957.

18. Kurmann A, Keller S, Tschan-Semmer F, et al. Impact of team familiarity in the operating room on surgical complications. World J Surg 2014;38:3047e3052.

19. DeChurch LA, Mesmer-Magnus JR. The cognitive underpin- nings of effective teamwork: a meta-analysis. J Appl Psychol 2010;95:32e53.