Today medical providers are charged with delivering care

The Joint Commission Journal on Quality and Patient Safety 2017; 43:524 533 CARE PROCESSES Optimizing an Enhanced Recovery Pathway Program: Development of a Postimplementation Audit Strategy Michael C. Grant, MD; Daniel J. Galante, DO; Deborah B. Hobson, BSN; Annette Lavezza, OTR/L; Michael Friedman, PT, MBA; Christopher L. Wu, MD; Elizabeth C. Wick,MD Background: Enhanced recovery pathways (ERPs) are bundled best-practice process measures associated with reduction of preventable harm, decreased length of stay (LOS), and increased overall value of care. An auditing procedure was developed to assess compliance with 18 ERP process measures and establish a system for identifying and addressing defects in measure implementation. Methods: For a one-year period, the electronic health records of 413 consecutive patients treated on a multidisciplinary ERP for colorectal surgery at an academic medical center were evaluated with the audit procedure. Patients were stratified who both met the expected LOS, as defined by LOS less than the historical (pre-erp) average LOS for the same procedure ( successes ), and exceeded the historical LOS ( outliers ). On the basis of the results of the audit process, a number of systemlevel interventions were developed. The results were then assessed for a three-month follow-up period to determine the impact on process measure compliance and LOS. Results: Detailed review of outliers identified several defects that improved following implementation of system-level changes, such as early mobility after surgery (44.4% vs. 59.5; p = 0.02). Although increased compliance through selective process measure optimization did not lead to a significant reduction in overall LOS (days; 5.2 ± 5.0 vs. 4.9 ± 3.0; p = 0.37), the audit procedure was associated with a significant reduction in outliers LOS (days; 12.2 ± 6.8 vs. 9.0 ± 2.1; p = 0.03). Conclusion: Concentrating audits in patients who fail to meet expectations on an ERP is an effective strategy to maximize identification of defects in and improve on pathway implementation. Meeting Presentation: The interim results of this study were presented at the Chesapeake Colorectal Society Meeting in Baltimore, Maryland on May 2, 2015. 1553-7250/$-see front matter 2017 Published by Elsevier Inc. on behalf of The Joint Commission. http://dx.doi.org/10.1016/j.jcjq.2017.02.011 Today medical providers are charged with delivering care that not only increases value by reducing preventable harms and improving efficiency but also must strive to optimize the patient experience. 1 This comprehensive approach to care is analogous to systems engineering where interdependent and integrated systems are aligned to optimize performance. 2 Enhanced recovery pathways (ERPs) are one of the best examples of successful adoption of this systemsbased approach in surgery. 3 6 As seen in systems engineering, the ERP team must work cohesively to engage patients and their families as partners in their own care, adhere to bestpractice patient care processes, and use robust process improvement methods to optimize implementation. Compared to previously described programs tasked with addressing individual preventable harms such as central line associated blood stream infections (CLABSIs), surgical site infections (SSIs), and catheter-associated urinary tract infections (CAUTIs), which usually have 5 to 10 process measures, most ERPs have upwards of 20 unique elements. 7 9 In addition, process measures included in ERPs are often designed to span the continuum of perioperative care from the decision to pursue surgery, through the intraoperative period, and ultimately well after the patient is discharged from the hospital. As a result, ERPs cover a number of unique geographic areas (some within the hospital walls and some in affiliated spaces) and hundreds of providers. In addition to these unique implementation challenges, there is now emerging evidence that has shown that increased compliance with pathway process measures is associated with improved patient outcomes and shorter lengths of stay. 9 11 ERP teams must therefore develop strategies to not only successfully implement complex multidisciplinary pathways over the course of the dynamic perioperative life cycle but also continuously evaluate and improve on existing systems, as well as sustain the program. To date, most groups have simply approached the development through use of ad hoc multidisciplinary groups to lead their programs, frequently coordinated by a project manager. 8,10,12 14 Many teams struggle with securing time and resources for this type of work. Although there is mounting evidence to support the technical interventions that are usually included in ERPs, 15 18 little has been reported about strategies for either process implementation or postproduction evaluation and optimization of ERP programs from the frontline provider perspective. Literature from the emerging field of implementation science has consistently identified

Volume 43, No., October 2017 525 both frontline provider engagement and regular review of performance using valid data as an essential step to successful and sustainable quality improvement programs. 2,19 21 We hypothesized that we could address both of these elements through development of an audit strategy that is sensitive to the burden of data collection and taps frontline providers wisdom to continually improve our multidisciplinary ERP program. In this article, we report on the audit strategy development and implementation, results of preliminary data analysis in our program, and examples of process improvement initiatives implemented to improve compliance with certain measures. METHODS Setting and Sample We have previously reported on the development and results of the multidisciplinary ERP for colorectal surgery at Johns Hopkins Hospital, a 1,059-bed tertiary care academic medical center. 22,23 Starting on February 1, 2014, all patients undergoing elective outpatient colorectal surgery by one of five colorectal surgeons were placed on the pathway. The surgeries included not just traditional colorectal procedures but also small bowel procedures such as ileostomy reversal procedures. The program was implemented as an extension of the work of the Johns Hopkins Hospital colorectal surgery Comprehensive Unit-based Safety Program (CUSP) that was established in 2010 to improve teamwork and safety culture and to reduce preventable harms such as SSIs and venous thromboembolism (VTE) 24,25 and to improve the value of care delivered. Enhanced Recovery Pathway Development and Implementation The ERP included a total of 18 process measures, which were divided into three categories: 1. Prior to Surgery the time frame from initial patientsurgeon encounter to the immediate preoperative period 2. Date of Surgery 3. After Surgery postanesthesia care unit until initial 30- day outpatient surgical follow-up. In recognition of the importance of auditing performance, as part of the implementation of the pathway, the ERP processes were reviewed with frontline providers, and consensus was reached on how to define compliance vs. noncompliance and identify where the information was currently being documented in the electronic health record (EHR). In some instances, there was no clear designated area within the electronic document to report, and the initial agreement was to include a nursing note. A comprehensive list of included process measures and documentation strategy is outlined in Table 1. It was agreed that a patient was considered compliant with a process measure only if the task was documented as completed (for example, ordering a medication alone was not sufficient; documentation stating that the drug was administered was necessary). The Johns Hopkins University Institutional Review Board deemed the chart review of pre-erp patients, as well as the prospective enrollment of ERP patients and associated process measure auditing strategy, to fall under the auspices of quality improvement, and therefore it was deemed exempt from convened review. Patients were excluded from this study if their procedures were deemed urgent or emergent (nonscheduled), outpatient in nature (admit date = discharge date), or occurred greater than one day after hospital admission. Development of Enhanced Recovery Pathway Audit Extraction Process To assess for compliance with key process measures, an ERP form was developed (Figure 1) to provide structure for data abstraction from the EHR and the team s discussions about areas for improvement. The form included patient identifiers, surgeon, and procedure and inpatient date ranges, as well as definitions for key process measures organized into the three respective procedure phases as previously outlined. After review of the patient s chart, individual process measures were scored on either a dichotomous (Yes [ Y ] or No [ N ]) or continuous (that is, volume) fashion where appropriate. Space was made available for free-text descriptions of relevant perioperative complications or comments that were not otherwise represented elsewhere in the document. Study Design Initially, the year prior to ERP implementation (pre-erp: February 2014 January 2015) was evaluated to establish historical lengths of stay (LOS) of each individual procedure subtype. Thereafter, following program implementation, all patients undergoing ERP for colorectal surgery were reviewed with the aid of the audit form to establish accurate and consistent documentation practices. Patients were then divided into respective groups based on individual LOS. Patients with LOS that exceeded the pre-erp average LOS for the given procedure were identified and labeled as outliers, while those patients who had LOS equal to or less than the historical average for the given procedure represented the successes. This auditing period was conducted for a total of 12 months. Initially, all data were abstracted by a nurse coordinator, surgical resident, and certified registered nurse anesthetist, but as the process evolved, frontline providers from each area (preoperative holding [nurses], operating room [certified registered nurse anesthetists] and inpatient unit [nurses]) took ownership of some of the processes relevant to their work area and reviewed the patient charts of those identified as outliers and determined compliance with their relevant processes. After the year-long auditing period, a number of improvement measures were employed on the basis of the results of the interim audit, and subsequent compliance was evaluated thereafter for a period of three months to determine if those improvements led to subsequent improvements in

526 Michael C. Grant, MD, et al Audit Tool for Enhanced Recovery Implementation Table 1. Enhanced Recovery Pathway Process Measures and Documentation Methodology Process Measure Prior to Surgery Preoperative counseling regarding surgery, anesthesia, pain management and recovery Preoperative visit with enterostomal therapist if ostomy is planned for procedure Mechanical bowel preparation with oral antibiotics Chlorhexidine gluconate washes Carbohydrate drink (20 ounces) 2 hours prior to surgical procedure Within Periprocedural Area Preprocedure multimodal analgesia and PONV medications (oral acetaminophen, gabapentin, and celecoxib, and scopolamine patch) Maintenance of normothermia by use of forced-air warming device Preoperative venous thromboembolism (VTE) prophylaxis administered prior to incision Prophylactic intravenous antibiotic administration before incision Avoidance of urinary catheter placement for procedures less than 2 hours Documentation Methodology Outpatient preoperative surgical clinic note, preoperative anesthesia consultation note Ostomy nursing note, preoperative anesthesia consultation note Preoperative nursing note on day of surgery, preoperative electronic checklist Preoperative nursing note on day of surgery, preoperative electronic checklist Preoperative nursing note on day of surgery, preoperative electronic checklist Preoperative nursing note on day of surgery, intraoperative anesthesia record, electronic record Preoperative nursing note on day of surgery, intraoperative anesthesia record Preoperative nursing note on day of surgery, intraoperative anesthesia record, electronic record Intraoperative anesthesia record Operating room nursing record, intraoperative anesthesia record Mobilization to chair in recovery area (PACU) PACU nursing note on day of surgery, electronic medical record Use of regional anesthesia (epidural or transversus abdominis plane [TAP] block) Preoperative nursing note on day of surgery, intraoperative anesthesia record, acute pain service note Postoperative Recovery/After Surgery Multimodal pain control to minimize opioids Postoperative nursing, daily nursing record, electronic medication record Early and progressive mobilization Postoperative nursing, daily nursing record, electronic medical record Early resumption of oral intake Postoperative nursing, daily nursing record, electronic medical record Minimal or no use of intravenous fluids Postoperative nursing, daily nursing record, electronic medication record Early removal of Foley catheter (if placed) Postoperative nursing, daily nursing record, electronic medical record PONV, postoperative nausea and vomiting; PACU, postanesthesia care unit. process measure compliance or further reductions in LOS. Therefore, our study had three distinct phases: 1. Determination of pre-erp LOS (four three-month quarters [Q] Q1, Q2, Q3, Q4) 2. ERP successes and outliers auditing phase (four threemonth quarters Q1, Q2, Q3, Q4) 3. Audit-based process measure improvement and postimplementation review (one quarter QPost) Statistical Analysis All data were processed and analyzed with Excel 14.1.0 (Microsoft Corporation, Redmond, Washington) and JMP 9.0.0 (SAS Institute Inc., Cary, North Carolina). Comparisons were made between means using the paired or unpaired Student s t-test or ANOVA for repeated measure, where appropriate. Medians and interquartile ranges were compared by the Mann-Whitney U test for unpaired data. Categorical variables were assessed by the chi-square or Fisher s exact test, where appropriate. Data are given as mean ± standard deviation; median (interquartile range); or number (percentage), as appropriate. The nature of the hypothesis testing was two-tailed, with p < 0.05 defining significance. Statistical analyses were performed with the assistance of an inhouse statistician. RESULTS Historical Length of Hospital Stay In the year prior to ERP implementation (pre-erp), a total of 397 patients underwent surgery within the traditional colorectal surgical treatment paradigm. The pre-erp average LOS was determined for each procedure subtype. The overall average LOS for the pre-erp colorectal cohort was 6.9 days. The average LOS for individual pre-erp procedures was 7.2 days for open colon, 5.1 days for laparoscopic colon, 7.4 days for open rectal, 5.7 days for laparoscopic rectal, 3.8 days for open small bowel, 3.2 days for laparoscopic small bowel, and 4.9 days for other procedures.

Volume 43, No., October 2017 527 Audit Form for the Enhanced Recovery Pathway (ERP) Program for Colorectal Surgery at Johns Hopkins Hospital Figure 1: This audit form was used by the multidisciplinary team to investigate all of the patients with an extended length of stay on the enhanced recovery pathway. The form provided structure to identify defects that needed to be addressed in order to optimize the implementation. Importantly, the focused approach of just using this form with patients with a prolonged length of stay balanced the need to audit our practices against the burden of manual data collection. LOS, length of stay; POD, postoperative day; IV, intravenous; PCA, patient-controlled analgesia; NSAID, nonsteroidal antiinflammatory drug. Study Population In the study period (February 2014 January 2015), a total of 413 patients were treated using the ERP for colorectal surgery. The most common types of procedures were colon (n = 160 patients; 38.7%), followed by rectal (n = 115; 27.8%), small bowel (n = 113; 27.4%), and other colorectal procedures (n = 25; 6.1%). Seventy-eight (18.9%) procedures were performed laparoscopically. As compared to pre- ERP, the ERP for colorectal surgery resulted in a significant reduction in overall average LOS (days; 7.0 ± 5.9 vs. 5.4 ± 5.1; p < 0.001). Defining the Outlier Population On the basis of our definition, a total of 102 patients were defined as outliers (24.7%), and 311 patients were considered successes (75.3%). Table 2 outlines and compares the

528 Michael C. Grant, MD, et al Audit Tool for Enhanced Recovery Implementation Table 2. Enhanced Recovery Pathway Program Demographic and Selected Process Measure Data Total N = 413 Successes n = 311 Outliers n = 102 P Value Age (years; average) 52.9 52.7 56.9 0.02 Male Sex, n (%) 207 (50.1) 154 (50.0) 53 (52.0) 0.73 Caucasian Race, n (%) 319 (77.2) 243 (78.1) 76 (74.5) 0.50 Surgery Length (minutes) 222 ± 136 206 ± 124 276 ± 161 < 0.001 Estimated Blood Loss (ml) 200 ± 360 150 ± 250 370 ± 560 < 0.001 Mean Length of Stay (days) 5.4 ± 5.0 3.6 ± 1.76 12.2 ± 6.8 < 0.001 Procedure Type, n (%) Open Colon 127 (30.8) 95 (30.5) 32 (31.4) 0.90 Laparoscopic Colon 33 (8.0) 29 (9.3) 4 (3.9) 0.09 Open Rectal 106 (25.7) 71 (22.8) 35 (34.3) 0.03 Laparoscopic Rectal 9 (2.2) 6 (1.9) 3 (2.9) 0.70 Open Small Bowel 77 (18.6) 61 (19.6) 16 (15.7) 0.46 Laparoscopic Small Bowel 36 (8.7) 30 (9.6) 6 (5.9) 0.31 Other Procedure 25 (6.1) 19 (6.1) 6 (5.9) 1.00 Stoma Creation 103 (24.9) 74 (23.8) 29 (28.4) 0.35 Overall Open 335 (81.1) 246 (79.1) 89 (87.3) 0.08 Overall Laparoscopic 78 (18.9) 65 (20.9) 13 (12.7) < 0.001 Anesthesia/Analgesia, n (%) Pre-Op Multimodal Meds 254 (61.5) 214 (68.8) 40 (39.2) < 0.001 Total Opioid Requirement 159 ± 311 140 ± 254 231 ± 459 0.03 (morphine equivalents; mg) Postoperative IV Fluid (ml) POD1 2,572 ± 1,671 2,445 ± 1,307 3,034 ± 2,566 < 0.001 POD2 1,731 ± 1,454 1,534 ± 1,069 2,457 ± 2,249 < 0.001 POD3 1,220 ± 1,362 906 ± 1,125 2,375 ± 1,530 < 0.001 Diet, n (%) POD1 PO Diet 386 (93.5) 301 (96.8) 85 (83.3) < 0.001 POD2 PO Diet 364 (90.5)* 289 (96.3) 75 (73.5) < 0.001 POD 3 PO Diet 258 (82.4) 203 (96.2) 55 (53.9) < 0.001 Mobility, n (%) 182 (44.1) 149 (47.9) 33 (32.4) 0.006 Achieved Mobility Goals POD1 Achieved Mobility Goals POD2 187 (46.5)* 150 (50.0) 37 (36.3) 0.02 Achieved Mobility Goals POD3 139 (44.4) 90 (42.7) 49 (48.0) 0.37 Alterations in n due to discharges from hospital on POD1 & POD2: *n = 402 (all patients). n = 313 (all patients). n = 300 (ERP success group). n = 211 (ERP success group). IV, intravenous; POD, postoperative day; PO, postoperative. Italicized values are statistically significant. relevant baseline demographics and surgical subtypes for each group. The outliers average LOS was significantly longer than that of the successes (days; 12.2 vs. 3.6; p < 0.001). On the basis of the included covariates, outliers were likely to be elderly as compared to the successes group. In addition, patients who underwent open rectal procedures were more likely to be outliers than were other procedure subtypes. Results of ERP Audit Process: Select Process Measure Compliance Prior to Surgery. All included patients received preoperative surgical counseling, including information relevant to the surgical procedure, anesthesia, pain management, and surgical recovery. Similar numbers of patients from the outliers and successes groups were instructed to complete a mechanical bowel preparation (87.3% vs. 84.9%, respectively; p = 0.63). There was no significant difference between outliers and successes in compliance with mechanical bowel preparation (82.4% vs. 82.3%; p = 0.99), use of preoperative chlorhexidine gluconate washes (75.5% vs. 76.2%; p = 0.87), or consumption of carbohydrate drink (14.7% vs. 15.4%; p = 0.87). Analgesia. Outliers were significantly less likely to receive immediate preoperative multimodal pain and postoperative nausea and vomiting medications (39.2% vs. 68.8%; p < 0.001) compared to successes. In addition, outliers received a significantly larger amount of total opioids (morphine equivalents [milligrams; mg]; 231 ± 459 vs. 140 ± 254; p < 0.03) compared to successes.

Volume 43, No., October 2017 529 Table 3. Post Process Measure Improvement for Selected Interventions Pre Process Improvement N = 413 Post Process Improvement N = 100 P Value Carbohydrate Consumption 63 (15.3) 67 (67.0) < 0.001 n (%) Pre-Op Multimodal Medications 254 (61.5) 76 (76.0) 0.01 n (%) Total Opioid Requirement 159 ± 311 83 ± 140 0.03 (morphine equivalents; mg) Mobility, n (%) Achieved Mobility Goals POD1 182 (44.1) 43 (43.0) 0.82 Achieved Mobility Goals POD2 187 (46.5)* 48 (49.5) 0.65 Achieved Mobility Goals POD3 139 (44.4) 44 (59.5) 0.02 Alterations in n due to discharges from hospital on POD1 & POD2: *n = 402. n = 313. n = 97 (all patients). n = 74. POD, postoperative day. Italicized values are statistically significant. Postoperative Metrics. Outliers received significantly more fluid on postoperative day (POD) 1, POD2 and POD3 compared to successes (p < 0.001). In addition, outliers were significantly less likely to have appropriate oral intake on POD1, POD2, and POD3 compared to successes. Finally, outliers were significantly less likely than the successes to achieve predetermined mobility goals (Table 2). Multidisciplinary Assessment of Clinical Course. A multidisciplinary team (composed of surgeons, anesthesiologists, medical doctors, and certified registered nurse anesthetists) reviewed the individual completed audit forms, and the cases were further discussed to identify opportunities to optimize the existing pathway to both improve compliance and focus efforts on outlier-specific variables. Given the time limitations of the frontline providers, the discussion focused on the outlier populations. In addition, all outliers received a formal clinical assessment designating the cause for their prolonged LOS. The categories were as follows: 1. Ileus 2. Postoperative complications (such as SSI [superficial, deep, organ/space], VTE, urinary tract infection, pneumonia, postoperative hemorrhage, reoperation, or cardiovascular event), as defined by the American College of Surgeons National Surgical Quality Improvement Program (NSQIP) User Guide 26 3. Intraoperative complications 4. Exacerbation of underlying comorbidities 5. Multifactorial Of the patients who had a longer than expected LOS, the majority (71; 69.6%) were categorized as having postoperative ileus. Other reasons were postoperative complications (10; 9.8%), intraoperative complications (8; 7.8%), exacerbation of underlying comorbidities (3; 2.9%), and multifactorial reasons (10; 9.8%). Optimizing Enhanced Recovery Pathway On the basis of the audit data, the multidisciplinary team targeted several distinct areas to improve process measure compliance. This targeting included interventions to improve compliance with the following: 1. Administration of the preoperative carbohydrate drink (because of poor overall study compliance across both groups) 2. Administration of preoperative multimodal analgesic and postoperative nausea and vomiting medications (to potentially improve rates of ileus secondary to narcotic administration) 3. Improvement in early and progressive postoperative mobility in the early postoperative phase (POD0, POD1, and POD2) On the basis of discussions in the multidisciplinary forum and review of the data with key stakeholders, system-level fixes were identified to improve compliance (see Discussion). Data regarding postprocess measure improvement are shown in Table 3. Most notably, improvements led to an increase in the compliance with preoperative carbohydrate beverage consumption (p < 0.001) and administration of the preoperative medication bundle (p = 0.01), which may have led to a significant reduction in total opioid requirements (p = 0.03). Interventions designed to help improve mobility rates did not lead to a significant increase in achievement of mobility goals on POD1 (p = 0.82) or POD2 (p = 0.65). However, post process measure improvement patients were significantly more likely to reach mobility goals by POD3 (p = 0.02). Program Sustainability and Reducing Outlier Length of Stay LOS data associated with individual quarters (threemonth segments) of each individual study phase can be found

530 Michael C. Grant, MD, et al Audit Tool for Enhanced Recovery Implementation A 10 Overall LOS over Duration of Study Period LOS (days) 8 6 4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 QPost pre-erp ERP B 15 Outliers and Successes LOS over ERP Study Period * LOS (days) 10 5 Outliers Successes Average *p < 0.03 0 Q1 Q2 Q3 Time Frame Figure 2: Through continuous review and optimization of our enhanced recovery pathway (ERP) implementation, the initial reduction in LOS for the duration of the study period was sustained (Figure 2a). The specific focus on intensive evaluation of the outlier population led to a reduction in LOS in this high-utilization group (Figure 2b). Teams should consider that there is opportunity for ongoing improvement in both the traditional and nontraditional ERP patients and should develop a culture of ongoing process improvement in order to achieve the optimal benefits from this work. Figure 2a depicts mean ± standard error measurement. LOS, length of stay; Q, quarter (three-month interval); Post, post process measure improvement. Q4 QPost in Figure 2a. Of note, LOS was significantly reduced in the initial quarter following ERP implementation, compared to both the prior quarter (pre-erp Q4; 6.7 ± 5.3 vs. 5.3 ± 5.2 days; p = 0.02) and the prior year (pre-erp; 6.9 ± 5.8 vs. 5.4 ± 5.1 days; p < 0.001), an effect that was maintained each quarter throughout the remainder of the study period. Optimization of the ERP (QPost; Figure 2b) did not lead to a significant change in LOS for either the overall cohort (average; 5.2 ± 5.0 vs. 4.9 ± 3.0 days; p = 0.37) or the successes (3.2 ± 1.4 vs. 3.3 ± 1.5 days; p = 0.92). However, following optimization of the ERP, the average LOS for outliers was significantly reduced compared to the prior ERP quarter (Q4 ERP; 12.5 ± 6.6 vs. 9.0 ± 2.1 days; p = 0.03) as well as compared to all outliers over the prior year (ERP; 12.2 ± 6.8 vs. 9.0 ± 2.1 days; p = 0.03). DISCUSSION Our study shows that, through the implementation of a formal audit process, multidisciplinary performance review, and intervention strategy, patients who absorb the greatest perioperative resources the outlier group might receive increased levels of process measures compliance and subsequent reduction in index hospitalization LOS. We noted that many of the included process measures depend on a highfunctioning and coordinated team of providers who

Volume 43, No., October 2017 531 understand the rationale of the pathway, believe in the individual components, and feel empowered to promote adherence. Although some of the process measures (daily mobility goals, opioid requirements, and intravenous fluid use) could be described as outcomes in isolation, in an effort to keep the team focused on the overarching goal of reducing LOS and reducing preventable harms such as SSIs, we considered them process measures. To improve some process measures, a system-level fix was sufficient, but for others a more complex cultural change in the way postoperative care is provided was necessary. For example, through the audit process we identified that patient compliance with drinking a carbohydrate-rich beverage prior to surgery was low. Patients were given the drink during their preoperative visit and instructed to drink it on the way to the hospital. They were reminded of this at the time of their preoperative anesthesia visit and during a preoperative phone call the night before surgery. Although the preoperative nurses were instructed to add the compliance question to their intake assessment, a dedicated item was not originally added to the EHR because resources were limited for modifications to the current EHR, as the entire health system was scheduled to change electronic vendors within the next 18 months. Therefore, the group had intended to document the patient s response as a free-text item at the bottom of the checklist. On review of the audit data, it was established that the low compliance was likely a result of documentation failure. Therefore, with senior level executive support, we were able to justify adding this question as a structured field in the checklist, and compliance improved dramatically. On further review, although the documentation accounted for a portion of the noncompliance, there were instances in which the patient did not consume the beverage either because of lack of education or confusion. To address this gap, the preoperative area nurses started to stock the beverage on the unit and, if the omission was noted more than two hours before surgery, the patient was provided the drink in the preoperative area. This example illustrates the power of focusing on quality improvement at the unit level and empowering frontline providers to identify defects, barriers, and solutions. In addition to increasing the likelihood that the fix is feasible, it also supports local ownership and accountability for performance. Initial compliance was similarly low for administration of the preoperative medication bundle (including acetaminophen, celecoxib, and gabapentin by mouth and a scopolamine patch applied topically). To facilitate improved compliance, a dedicated preoperative order set was developed, and formal instruction was provided to anesthesia providers on how to locate and use the function. The medications were also listed on a hard-copy checklist attached to the patient s chart in order to cue the preoperative nursing staff to ensure that patients received the medication bundle along with other relevant preoperative process measures. The order set also cued providers to carry medication administration into the postoperative phase, which reinforced multimodal analgesia administration and likely reduced the reliance on opioid-based medications. Through our audit process, low compliance with early mobility on the inpatient units was also identified. On further evaluation, this process required a more in-depth examination with key stakeholders, including patients and family members, to improve compliance. In collaboration with the Department of Physical Medicine and Rehabilitation, we conducted a survey of nurses on the inpatient unit to establish the barriers and knowledge, attitudes, and beliefs related to early mobility after surgery. 27 Twenty-three staff nurses from the unit responded to the electronic survey. The results indicated that the nursing staff on the unit believed that there was, in fact, adequate leadership support from the unit nurse manager for mobility, as well as appropriate equipment and proper training. The main challenge, however, was related to the cumbersome documentation of mobility in the EHR and the ambiguous nature of the mobility goals themselves. Nursing believed more direct, objective communication to both the patient and bedside nurse regarding mobility goals from the care team would be helpful. To address these barriers, the daily mobility goals in the ERP were changed from designations such as walk in hall twice by postoperative Day 2 to a formal alignment with the hospital s standard communication tool (Highest Level of Mobility) to describe and communicate patient mobility goals. In addition, the unit posted a calendar on patients doors and empowered patients and their families to document daily activity on this sheet. The clinical technicians then used this to document in the EHR. Finally, mobility was incorporated into the daily unit-based care coordination rounds, and the prior day s performance as well as currentday s goals were reviewed. Strides have been made in developing a culture of early ambulation, but this process measure in particular has been challenging, reinforcing how complex some of the ERP processes are to implement. Another important association noted in our results was that one of the principal reasons for prolonged LOS in the outlier group was related to ileus. Not surprisingly, two of the selected process measures are likely to have provided a positive impact on rates of postoperative ileus. Through application of the preoperative medication bundle, which included multimodal analgesics such as acetaminophen, celecoxib, and gabapentin, patients might rely less heavily on opioid-based analgesics, which have repeatedly been shown to carry a side effect profile of nausea and constipation. In addition, greater adherence to mobility goals might have improved rates of ileus, as well promoting earlier return of bowel function. Although our study was not designed to specifically comment on causation, it is nonetheless compelling that measures associated with improving ileus the chief nidus for prolonged LOS might also reduce outliers LOS. The development and application of a focused auditing process for a colorectal ERP program has proven to be a

532 Michael C. Grant, MD, et al Audit Tool for Enhanced Recovery Implementation successful approach that balanced the need for providing data to drive improvement with the burden of manual data collection. By transitioning the auditing procedure from a nurse coordinator, resident, and nurse anesthetist collaborative to local frontline providers, much of the burden of documentation was absorbed into the daily work flow. Given that nursing was documenting these metrics into the electronic record, these efforts were not repeated by study staff. This also further promoted local champions and personal accountability for relevant pathway elements and prompted further conversations about system-level barriers that may have affected performance. As we stated earlier, this led to substantial gains in compliance with selected measures. Much focus with regard to ERP is on reducing the overall LOS, and we have been successful in maintaining a consistent reduction in LOS throughout the duration of this program. Although a comprehensive audit and optimization procedure did not necessarily improve on those LOS values, we do show a significant reduction in the outliers LOS. It is important for ERP implementation teams to recognize the opportunity for ongoing improvement in both the traditional and nontraditional ERP patients and to develop a culture of ongoing process improvement to achieve the optimal benefits of ERPs. Although we do not know the mechanism behind this improvement, through our audit process much of our data discussion and performance review with frontline providers focused on this outlier group. Therefore, a portion of the effect may be attributable to improved compliance overall, but another portion may be the result of heightened awareness of these patients, their complications, and reasons for prolonged hospitalization. This is significant because this outlier group, while not frequently cited in ERP publications, does appear to constitute a subset of the population of patients that uses a large amount of hospital resources. This suggests, at the very least, that further study is warranted as to how to use both ERP and focused data-driven performance review to specifically affect this group. In addition, given our initial auditing results suggesting that failures in compliance are concentrated among the outliers, it may be appropriate to focus future auditing efforts exclusively on this group of patients. This may be a beneficial approach for two reasons. First, it promotes frontline provider ownership of performance (through the auditing procedure itself), and second, it reduces some of the documentation burden to these providers and allows them to use that additional time to diagnose the actual causes of noncompliance with ERP process measures. Our study also provides context for the impact of patientrelated factors that can be categorized as modifiable vs. nonmodifiable in nature. Clearly, our results suggest that patients represented as outliers may represent either more complex procedures (for example, open rectal) or potentially sicker patients (for example, more complex or progressive disease, coexisting organ dysfunction), which despite best efforts may not have many modifiable factors. But it is important to remember that even in this complex subpopulation, there is likely always opportunity to deliver better care. The strategy employed in this article serves to represent a framework for systematically and successfully addressing variables associated with disparate process measure application, systemlevel deficiencies, or compliance issues. Success in this case was measured not only by improved compliance but by a reduction in the overall range of the outlier population LOS. Limitations The study has important limitations. Our audit strategy has been used only at a single institution and is reflective of our locally adapted ERP. Therefore, although the concepts and processes are likely widely applicable, the tool itself requires local adaptation. To reduce confusion and streamline processes, we placed all elective colorectal surgery patients undergoing abdominal surgery (including ileostomy reversals, hernia repairs, and small bowel resections) on the pathway with the agreement that, if clinically indicated, the treatment can be altered at any time. Thus, this represents a more pragmatic approach, as opposed to one that is strictly defined by common procedural terminology (CPT) codes, which are only attainable postsurgery. This approach is part of the reason that not all patients included in our study had associated NSQIP data captured. We believe that, ultimately, ERP should be considered the new norm for all general surgery patients. This is in contrast to many programs in which the pathway is reserved for patients whom the surgeon believes are best suited. If the latter approach is taken, there may be fewer overall patients on the pathway, as well as fewer outlier patients. In this case, consideration should be made to audit all patients and not focus exclusively on outliers. We also recognize that clinical documentation varies greatly from hospital to hospital. It is therefore vital for teams to define ERP process measures and identify both the location and quality of documentation associated with those measures in the electronic or paper health record. In this manner, our experiences will differ greatly from those of other institutions. CONCLUSION Despite an inability to achieve universal compliance with individual ERP process measures, our group has observed a significant reduction in average LOS across all included patients. Furthermore, the use of a novel auditing strategy has helped to not only identify the outliers among those undergoing an ERP program but aid in the interventions that led to a significant reduction in subsequent LOS among a resource-intensive outlier population. This suggests that ensuring the delivery of evidence-based care by using approaches that lead to improved process measure compliance remains a vital aspect of ERP. The audit process described in this article outlines one such strategy the continual involvement of postimplementation evaluation, which fosters teamwork, focuses on patient engagement, and even

Volume 43, No., October 2017 533 promotes local and organizationwide cultural shifts. In addition to a strong implementation scheme, successful ERP programs require a comprehensive process for ongoing improvement. Conflicts of Interest. All authors report no conflicts of interest. Michael C. Grant, MD, is Assistant Professor of Anesthesiology and Critical Care Medicine, Johns Hopkins University School of Medicine, Baltimore. Daniel J. Galante, DO, is Chief Resident in Surgery, Sinai Hospital, Baltimore. Deborah B. Hobson, BSN, is Patient Safety Nurse, Johns Hopkins Hospital, Baltimore, and Annette Lavezza, OTR/L, is Manager, Acute Care Rehabilitation Therapy Services. Michael Friedman, PT, MBA, is Director, Rehabilitation Therapy Services, Johns Hopkins Hospital, and Co- Director, Johns Hopkins Center for Activity and Mobility Promotion, Johns Hopkins Hospital. Christopher L. Wu, MD, is Professor of Anesthesiology and Critical Care Medicine, Johns Hopkins University School of Medicine. Elizabeth C. Wick, MD, formerly Associate Professor of Surgery and Vice Chair for Quality and Safety, Johns Hopkins University School of Medicine, is Associate Professor of Surgery, University of California, San Francisco, and Core Faculty Member, Armstrong Institute for Patient Safety and Quality, Johns Hopkins Medicine, Baltimore. Please address correspondence to Elizabeth C. Wick, elizabeth.wick@ucsf.edu. REFERENCES 1. Institute of Medicine. Crossing the Quality Chasm: A New Health System for the 21st Century. Washington, DC: National Academy Press, 2001. 2. Booth A, Carroll C. How to build up the actionable knowledge base: the role of best fit framework synthesis for studies of improvement in healthcare. BMJ Qual Saf. 2015;24:700 708. 3. Fearon KC, et al. Enhanced recovery after surgery: a consensus review of clinical care for patients undergoing colonic resection. Clin Nutr. 2005;24:466 477. 4. Wind J, et al. Systematic review of enhanced recovery programmes in colonic surgery. Br J Surg. 2006;93:800 809. 5. Kahokehr A, et al. Implementation of ERAS and how to overcome the barriers. Int J Surg. 2009;7:16 19. 6. Teeuwen PH, et al. Enhanced Recovery After Surgery (ERAS) versus conventional postoperative care in colorectal surgery. J Gastrointest Surg. 2010;14:88 95. 7. Hechenbleikner EM, et al. Implementation of surgical quality improvement: auditing tool for surgical site infection prevention practices. Dis Colon Rectum. 2015;58:83 90. 8. Ahmed J, et al. Enhanced recovery after surgery protocols compliance and variations in practice during routine colorectal surgery. Colorectal Dis. 2012;14:1045 1051. 9. Smart NJ, et al. Deviation and failure of enhanced recovery after surgery following laparoscopic colorectal surgery: early prediction model. Colorectal Dis. 2012;14:e727 734. 10. Ahmed J, et al. Compliance with enhanced recovery programmes in elective colorectal surgery. Br J Surg. 2010;97:754 758. 11. Kahokehr AA, et al. Influences on length of stay in an enhanced recovery programme after colonic surgery. Colorectal Dis. 2011;13:594 599. 12. Ahmed J, et al. Predictors of length of stay in patients having elective colorectal surgery within an enhanced recovery protocol. Int J Surg. 2010;8:628 632. 13. Lassen K, et al. Consensus review of optimal perioperative care in colorectal surgery: Enhanced Recovery After Surgery (ERAS) Group recommendations. Arch Surg. 2009;144:961 969. 14. Cakir H, et al. Adherence to Enhanced Recovery After Surgery and length of stay after colonic resection. Colorectal Dis. 2013;15:1019 1025. 15. Thiele RH, Bartels K, Gan TJ. Inter-device differences in monitoring for goal-directed fluid therapy. Can J Anaesth. 2015;62:169 181. 16. Keller DS, et al. The effect of transversus abdominis plane blocks on postoperative pain in laparoscopic colorectal surgery: a prospective, randomized, double-blind trial. Dis Colon Rectum. 2014;57:1290 1297. 17. Keller DS, Ermlich BO, Delaney CP. Demonstrating the benefits of transversus abdominis plane blocks on patient outcomes in laparoscopic colorectal surgery: review of 200 consecutive cases. J Am Coll Surg. 2014;219:1143 1148. 18. Aarts MA, et al. Adoption of Enhanced Recovery After Surgery (ERAS) strategies for colorectal surgery at academic teaching hospitals and impact on total length of hospital stay. Surg Endosc. 2012;26:442 450. 19. Bauer MS, et al. An introduction to implementation science for the non-specialist. BMC Psychol. 2015 Sep 16;3:32. 20. Brewster AL, et al. Integrating new practices: a qualitative study of how hospital innovations become routine. Implement Sci. 2015 Dec 5;10:168. 21. Pannick S, Sevdalis N, Athanasiou T. Beyond clinical engagement: a pragmatic model for quality improvement interventions, aligning clinical and managerial priorities. BMJ Qual Saf. 2016;25:716 725. 22. Wick EC, 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;221:669 677. 23. Wu CL, et al. Initiating an enhanced recovery pathway program: an anesthesiology department s perspective. Jt Comm J Qual Patient Saf. 2015;41:447 456. 24. Wick EC, et al. Implementation of a surgical Comprehensive Unit-based Safety Program to reduce surgical site infections. J Am Coll Surg. 2012;215:193 200. 25. Timmel J, et al. Impact of the Comprehensive Unit-based Safety Program (CUSP) on safety culture in a surgical inpatient unit. Jt Comm J Qual Patient Saf. 2010;36:252 260. 26. American College of Surgeons (ACS). User Guide for the 2014 ACS NSQIP Participant Use Data File (PUF). Chicago: ACS, 2015. Accessed Apr 2, 2017. https://www.facs.org/ ~/media/files/quality%20programs/nsqip/nsqip_puf _userguide_2014.ashx. 27. Hoyer EH, et al. Barriers to early mobility of hospitalized general medicine patients: survey development and results. Am J Phys Med Rehabil. 2015;94:304 312.