- Email: [email protected]
Improving and Maintaining On-Time Start Times for Nonelective Cases in a Major Academic Medical Center
The Joint Commission Journal on Quality and Patient Safety 2019; 000:1–6
Improving and Maintaining On-Time Start Times for Nonelective Cases in a Major Academic Medical Center Dan B. Ellis, MD; Jason Santoro, RN; Dale Spracklin, RN; Vanessa Kurzweil, PhD; Stephanie Sylvia, CRNA; Peter Fagenholz, MD; Aalok Agarwala, MD, MBA
Background: As health care expenditures continue to increase, thoughtful use of perioperative resources is important. Efforts to improve operating room (OR) efficiency often focus on increasing on-time first case starts to improve OR utilization, reduce subsequent delays, and reduce adverse events. One institution, with severely limited inpatient hospital capacity and an extensive daily add-on list of surgical cases, focused efforts to improve OR efficiency by improving on-time first case starts for unscheduled, nonemergent surgeries. Methods: A multidisciplinary team was assembled to work together for this quality improvement (QI) initiative. The primary outcome measure was the percentage of cases starting on time. The team identified six key steps thought to contribute to on-time start performance. Data were collected for each of these process measures, and feedback was shared with stakeholders. Results: By measuring adherence to and giving feedback about critical steps in the preoperative process, on-time starts improved from a baseline of 65% to 85% (p = 0.041). Sustained improvement was seen even after daily measurement ceased and the QI project was completed. Conclusion: Establishing a multidisciplinary team to improve timely care of unscheduled, nonelective surgical patients; identifying key elements necessary for on-time surgical case starts; and providing feedback to clinicians were associated with a sustained improvement in OR efficiency for a traditionally difficult-to-schedule patient population.
H
ealth care expenditures in the United States continue to increase, with annual growth of 4% to 5%, and total health care expense reached $3.5 trillion in 2017.1 , 2 With a rising focus on value, thoughtful and efficient use of perioperative resources is more important than ever.3 Roughly 40% of revenue may be derived from the perioperative environment, making it one of the highest revenue-generating areas in a hospital.1 , 4 , 5 However, the operating room (OR) is also one of the costliest resources in a hospital, with OR time costing up to $60 per minute.1 , 6 OR profitability can determine a hospital’s solvency.3 , 7 To improve OR efficiency and improve patient outcomes, many institutions have focused on improving ontime starts of electively scheduled first cases and reducing turnover time between surgical cases. Cases that start later in the day tend to be associated with higher operative costs when compared to cases that start earlier in the day.8 In addition to the economic costs, which tend to be higher in patients who receive operations later in the day, patient outcomes may suffer when cases start later.8 , 9 Delays in scheduled cases may push nonelective, or add-on, cases even later into the day, requiring them to be performed after normal working hours, when surgical and anesthesia outcomes have been shown to be worse.10–16
1553-7250/$-see front matter © 2019 The Joint Commission. Published by Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.jcjq.2019.09.007
Although on-time start initiatives to improve efficiency are commonplace in hospitals, the complex adaptive perioperative workflow makes identifying and addressing issues challenging.4 , 17 , 18 Variability in surgical case types and personnel—including surgeons, anesthesia providers, circulating nurses, surgical technicians, OR assistants, and pre-op nursing staff—contribute to OR efficiency, or lack thereof. This variability can sometimes lead involved role groups (surgeons, anesthesia providers, and circulating nurses) to blame one another for lack of efficiency, as each group can become convinced that others are the cause of the delays. Many large, quaternary care hospitals such as ours often operate at or near 100% capacity, with inpatient units full almost nightly. Despite having many ORs, our institution often begins the day with more than 30 patients requiring unscheduled, add-on procedures. To accommodate this volume, several ORs are dedicated to these patients. However, many patients still wait hours for nonelective, nonemergent surgeries. These OR delays sometimes result in downstream effects by creating bottlenecks on inpatient units that subsequently result in overcrowding in the emergency department and postanesthesia care unit. To reduce unnecessary waiting for our patients, reduce wasted time for our perioperative personnel, and improve perioperative efficiency, we aimed to identify and correct issues limiting on-time starts for nonelective, nonemergent cases within one of these dedicated ORs.
2
Dan B. Ellis, MD, et al.
METHODS
This project was undertaken as a quality improvement (QI) initiative at Massachusetts General Hospital. As such, it was not formally supervised by the Institutional Review Board per hospital policy. Massachusetts General Hospital is a large quaternary-care hospital in Boston with 1,018 licensed hospital beds, 58 ORs, and more than 20 out-ofOR anesthetizing locations. Nonelective, nonemergent surgical patients are often cared for in one of four dedicated ORs: one each for neurosurgery and orthopedic surgery, and two for general surgery patients. One of these two ORs is allocated for nonscheduled patients from the emergency department or inpatient units requiring acute care surgery, and the second OR provides care for a combination of elective scheduled and nonscheduled acute care surgery patients. Emergency surgical cases are also preferentially placed into these rooms when necessary. Six attending surgeons and a cohort of surgical residents manage this surgical service. To create buy-in and include necessary stakeholders, a multidisciplinary team of trauma and acute care surgeons, anesthesiologists, nurse anesthetists, circulating nurses, and hospital administrators was assembled. Using standard process improvement priniciples,19 the team identified the goal (improving efficient care for nonelective, nonemergent general surgical patients), determined metrics to define improvement (improved percentage of first cases beginning by 8:00 am), and identified several changes that could be implemented to meet the goal after first mapping the existing and ideal processes. The following critical process steps necessary to start a case on time were mapped and identified: 1. A surgical case must be booked at least two hours before the scheduled start time. 2. A perioperative huddle to discuss surgical and anesthesia issues associated with the case must occur by 7:15 am daily. 3. The surgical consent, signed by provider and patient, must be in the patient’s chart by 7:15 am. 4. The anesthesia consent, signed by provider and patient, must be in the patient’s chart by 7:15 am. 5. The surgical history and physical (H&P) must be completed by 7:15 am. 6. The 24-hour update to the surgical H&P if the patient was seen by the surgical service more than 24 hours before the intended operation must be completed by 7:15 am. Compliance with each of these six process steps would serve as the project’s process measures. A case would qualify for an on-time start only if the patient was wheeled into the OR by 8:00 am. In addition, if an emergency surgery was in progress at 6:00 am, the room would be excluded from data collection. The prospective QI initiative was divided into three stages. During the first stage, baseline data for on-time starts
Improving and Maintaining On-Time Start Times
for patients who presented for nonelective, nonemergent general surgeries were collected. Between January 1, 2017, and April 21, 2017, six attending general surgeons; several surgical residents; a variety of anesthesia residents, nurse anesthetists, and anesthesia attendings; and two circulating nurses from the Trauma and Acute Care Surgery service cared for patients. Circulating nurses used a checklist to record whether each of the six process measures had been met for each case in addition to whether the case started on time. From April 21 through April 28, 2017, data were reviewed by the multidisciplinary leadership team. Over the course of one week, results were reported to individual surgeons and anesthesia providers. The data included personal performance in the setting of performance of peer providers. For example, if one surgeon was available for a 7:15 am huddle only 50% of the time, this surgeon would see his or her performance in comparison to peers. On April 28, 2017, the second stage of the QI initiative began. During this stage, data were again prospectively gathered over a three-month period. Similar to the first stage of the initiative, a member of the circulating nursing staff recorded whether each of the six process measures had been met and whether the case started on time. These new data were then presented to surgeons and anesthesiologists. Five attending surgeons who participated in the first stage of the QI initiative participated in the second stage of the initiative. One attending surgeon who had participated in the initial phase left the practice, and an additional surgeon joined the practice. A variety of surgical residents and anesthesia providers participated in the second stage. The same two circulating nurses who were assigned to the Trauma and Acute Care Surgery service participated in this stage as well. In August 2017 the postintervention data were again reported to the surgeons and anesthesiologists who were involved in the initiative. The final phase of the project consisted of a retrospective audit of case start data from August 1, 2017, through December 31, 2017. Case start times were extracted from the electronic medical record (EMR) and, as with the previous two phases, if an emergency surgery was in progress at 6:00 am, the room was excluded from analysis. Statistical analysis was performed using R, version 3.4.3 (R Foundation for Statistical Computing, Vienna). Analysis of case types in each period was performed using the 3-sample test for equality of proportions without continuity correction. Overall on-time start rate was compared preand postintervention using a 2-sample test for equality of proportions without continuity correction. Post hoc tests were then performed for each of the six process measures comparing pre- and postintervention. P values for the process measures were corrected for multiple testing using the method of Benjamini and Hochberg,20 and these corrected values are displayed in Figure 1. A statistical process control
Volume 000, No. , November 2019
3
Figure 1: Percentage of cases starting on time and associated process measures before (gray bars) and after (black bars) intervention. P values shown for process measures are corrected for multiple post hoc hypothesis testing. N.S., not significant; H&P, history and physical.
p-chart was generated and analyzed using QI Macros 2017 (KnowWare International Inc., Denver). Special cause variation was identified using Western Electric rules for control chart analysis.
RESULTS
The Trauma and Acute Care Surgery service cares for general surgery patients and performs a robust variety of cases. No significant differences were found in case types between observation periods (Table 1). Prior to the intervention, 65% of nonelective, nonemergent cases started on time (Figure 1). During the threemonth postintervention period, the overall rate of on-time starts increased to 85%, and this increase was statistically significant (95% confidence interval, 0.01–0.39; p = .041). After identifying a significant improvement in the outcome measure of on-time starts, we performed post hoc significance tests on each of the six process measures (Figure 1): 1. Case booked at least two hours before the scheduled start time 2. Perioperative huddle at 7:15 am 3. Signed surgical consent in the patient’s chart by 7:15 am 4. Signed anesthesia consent in the patient’s chart by 7:15 am 5. Complete surgical H&P by 7:15 am 6. 24-hour update to the surgical history and physical by 7:15 am (if applicable)
All the process measures showed significant improvement from the preintervention to postintervention period except the “Signed anesthesia consent in the patient’s chart” measure (Figure 1; p values displayed are corrected for multiple hypothesis testing). There was no consistent pattern of anesthesia staffing. In fact, a different anesthesia team (anesthesiologist and nurse anesthetist or anesthesiologist and anesthesia resident) cared for patients on each day of the observation. Therefore, data for anesthesia teams are reported in aggregate as the large quantity of variables prevents meaningful interpretation of data. However, granular anesthesia data were reported to anesthesia leadership so they could be referenced in discussions with individual providers. To evaluate whether the improvement in on-time starts was maintained after the postintervention period of daily monitoring had concluded, we conducted a retrospective audit of case starts using EMR data from the five months following the postintervention monitoring period. This revealed that the monthly percentage of on-time starts during this period follows a similar trend to the postintervention monitoring period (Figure 2). When graphed on a statistical process control chart using preintervention data as a baseline, the postintervention and retrospective audit data together meet criteria for special cause variation using Western Electric rules. This suggests that the improvement in on-time starts cannot be accounted for simply by the variation established by the baseline data and indicates that the underlying process has changed sufficiently to establish a new baseline.
4
Dan B. Ellis, MD, et al.
Improving and Maintaining On-Time Start Times
Table 1. Number of Cases by Type by Observation Period Case Type
Preintervention n (%)
Postintervention n (%)
Active Surveillance n (%)
Wound debridement PEG Lap cholecystectomy I&D Lap appendectomy Lymph node biopsy Exploratory laparoscopy Amputation Exam under anesthesia Groin cyst excision Lap colectomy Muscle biopsy Temporal artery biopsy Fixation of rib fracture Low anterior resection Skin graft Pancreatic debridement Exploratory laparotomy Hernia repair Cholecystectomy (open) Tracheostomy Excision of mass Creation of colostomy Diagnostic laparoscopy
7 (18.9) 7 (18.9) 5 (13.5) 4 (10.8) 3 (8.1) 2 (5.4) 2 (5.4) 2 (5.4) 1 (2.7) 1 (2.7) 1 (2.7) 1 (2.7) 1 (2.7) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
9 (22.5) 1 (2.5) 12 (30.0) 3 (7.5) 6 (15.0) 0 (0) 3 (7.5) 0 (0) 0 (0) 0 (0) 0 (0) 2 (5.0) 0 (0) 1 (2.5) 1 (2.5) 1 (2.5) 1 (2.5) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
17 (18.3) 7 (7.5) 17 (18.3) 10 (10.8) 9 (9.7) 1 (1.1) 5 (5.4) 0 (0) 3 (3.2) 0 (0) 1 (1.1) 1 (1.1) 1 (1.1) 1 (1.1) 1 (1.1) 1 (1.1) 2 (2.2) 7 (7.5) 3 (3.2) 1 (1.1) 2 (2.2) 1 (1.1) 1 (1.1) 1 (1.1)
PEG, percutaneous endoscopic gastronomy; I&D, incision and drainage.
DISCUSSION
To improve on-time start times for nonelective, nonemergent cases, we assembled a multidisciplinary team to address the different elements associated with first-case delays. Our process (identifying metrics, setting expectations, and feeding data back to individual providers) was associated with a 20 percentage point improvement in on-time starts. Although many groups have identified and addressed perioperative delays,21–23 our initiative is unique in terms of the patient population we studied. To date, there has not been a published example of improving on-time starts in nonelective, nonemergent surgical patients. This may be in part due to unpredictability and the heterogeneous nature of these types of cases as compared to scheduled, elective cases. However, the success of our initiative could potentially be applied to other services. We believe the success of our initiative is due to a number of factors. First, we established a multidisciplinary team to develop mutually agreed-upon metrics and group norms. After the metrics were established with buy-in from relevant stakeholders, positive and negative feedback could confidently be relayed to the involved clinicians and staff. Second, we agreed as a team to challenge the previously held belief that the Trauma and Acute Care Surgery service was inherently different from any other service line and unchangeable. Believing in our team and our process was key to this project’s success.
The third important factor was that key team members had recent outside experience. Having worked at a variety of other institutions prior to joining our hospital, team members could identify alternative approaches to the established patterns of care in our OR. Often, members of an institution can become entrenched in existing processes and have difficulty finding innovative solutions to long-standing problems. Recent experience with other systems and different processes allowed for a more critical appraisal of our current system. Given the paucity of literature surrounding this challenging patient population, this experience proved invaluable. A fourth factor contributing to our success was that we used the data for individual feedback in the setting of full-group performance. Although neither rewards nor punishments were tied to individual performance in this initiative, providers repeatedly commented on how well they performed relative to their peers. Finally, we were able to deploy a consistent set of providers to collect data. Over the 12-month project period, seven surgeons (one attending surgeon left the institution, and another surgeon joined the practice during the initiative), two new surgical residents per month, and a rotating cast of anesthesiologists, anesthesia residents, and certified nurse anesthetists were involved in these cases. Given the size and complexity of our teaching institution, it is nearly impossible to have a stable group of providers. However, two circulating nurses were consistently assigned to
Volume 000, No. , November 2019
5
Figure 2: This p-chart displays the percentage of cases starting on time by month throughout 2017. † Points that meet one of the Western Electric rules for identifying special cause variation.
the Trauma and Acute Care Surgery service and remained on service throughout the entire initiative. Limitations
There are several limitations to our study. First, our QI initiative was hampered by a limited sample size. It is possible that we would observe a more statistically significant signal if the observation period was increased and a greater number of cases was included. However, our preintervention data are consistent with our institutional performance— 51% to 58% of cases at our hospital started on time during the time frame in question. Second, it is possible that a general culture shift toward QI contributed to our success, as a variety of other perioperative process improvement projects were taking place concurrently with our intervention. Third, the second phase of this QI initiative took place during the summer, when new trainees were joining our hospital. Perhaps having fresh trainees that were willing to try a new QI initiative on the surgical and anesthesia services during the intervention phase of this study affected our results. Finally, it is possible that our results were influenced by the Hawthorne effect, as the postintervention
period was actively monitored using checklists to track processes and outcomes. However, this is less likely in light of our findings that the increase in on-time starts was maintained in the five months after the circulating nursing staff ceased tracking process measures.
CONCLUSION
By bringing together a multidisciplinary group of champions, including surgery, anesthesia, and nursing, we were able to change long-standing institutional practice and culture. The success of the initiative depended on a commitment from the circulating nursing staff to collect data, and support from anesthesiologists, nurse anesthetists, and attending surgeons. By working together for a common goal, the entire team benefited from better communication, teamwork, and sustained improvement. Improvement efforts centered on provider feedback may help hospitals improve the rate of on-time starts for nonelective, nonemergent surgeries, reduce patient wait time, and improve OR capacity.
6
Dan B. Ellis, MD, et al.
Improving and Maintaining On-Time Start Times
9. Asfour L, et al. In surgeons performing cardiothoracic surgery is sleep deprivation significant in its impact on morbidity or mortality? Interact Cardiovasc Thorac Surg. 2014;19:479–487. 10. Taffinder NJ, et al. Effect of sleep deprivation on surDan B. Ellis, MD, is Assistant Division Chief, Division of General Surgery Anesthesia, Department of Anesthesia, Critical Care and Pain geons’ dexterity on laparoscopy simulator. Lancet. 1998 Oct Medicine, Massachusetts General Hospital (MGH), Boston. Jason San10;352:1191. toro, RN, is Operating Room Nurse, Department of Perioperative 11. Grantcharov TP, et al. Laparoscopic performance after one Nursing, MGH. Dale Spracklin, RN, is Registered Nurse, Department night on call in a surgical department: prospective study. of Perioperative Nursing, MGH. Vanessa Kurzweil, PhD, is Project BMJ. 2001 Nov 24;323:1222–1223. Manager, Department of Anesthesia, Critical Care and Pain Medicine, 12. Komen N, et al. After-hours colorectal surgery: a risk MGH. Stephanie Sylvia, CRNA, is Certified Registered Nurse Anesthetist, Department of Anesthesia, Critical Care and Pain Medicine, factor for anastomotic leakage. Int J Colorectal Dis. MGH. Peter Fagenholz, MD, is Surgeon, Department of Surgery, MGH. 2009;24:789–795. Aalok Agarwala, MD, MBA, is Chief Medical Officer, Massachusetts 13. Ricci WM, et al. Is after-hours orthopaedic surgery associEye and Ear, Boston. Please address correspondence to Dan B. Ellis, ated with adverse outcomes? A prospective comparative [email protected]. study. J Bone Joint Surg Am. 2009;91:2067–2072. 14. Chacko AT, et al. Does late night hip surgery affect outcome? J Trauma. 2011;71:447–453. 15. Fechner G, et al. Kidney’s nightshift, kidney’s nightmare? REFERENCES Comparison of daylight and nighttime kidney transplanta1. Clark A, et al. Early endocrine attending surgeon prestion: impact on complications and graft survival. Transplant ence increases operating room efficiency. J Surg Res. Proc. 2008;40:1341–1344. 2016;205:272–278. 16. Wright MC, et al. Time of day effects on the inci2. Centers for Medicare & Medicaid Services. Nadence of anesthetic adverse events. Qual Saf Health Care. tional Health Expenditures 2017 Highlights. Ac2006;15:258–263. cessed Oct 17, 2019. https://www.cms.gov/Research- 17. Wright JG, Roche A, Khoury AE. Improving on-time surgical Statistics- Data- and- Systems/Statistics- Trends- and- Reports/ starts in an operating room. Can J Surg. 2010;53:167–170. NationalHealthExpendData/downloads/highlights.pdf. 18. Vardaman JM, Cornell PT, Clancy TR. Complexity and 3. Cima RR, et al. Use of Lean and Six Sigma methodolchange in nurse workflows. J Nurs Adm. 2012;42:78–82. ogy to improve operating room efficiency in a high-vol- 19. Associates in Process Improvement. Home page. Accessed ume tertiary-care academic medical center. J Am Coll Surg. Oct 17, 2019. https://www.apiweb.org. 2011;213:83–92. 20. Benjamini Y, Hochberg Y. Controlling the false discovery 4. Franklin J, Franklin T. Improving preoperative throughput. rate: a practical and powerful approach to multiple testing. J Perianesth Nurs. 2017;32:38–44. J R Stat Soc Series B Stat Methodol. 1995;57:289–300. 5. Fowler PH, et al. Perioperative workflow: barriers to effi- 21. Overdyk FJ, et al. Successful strategies for improving opciency, risks, and satisfaction. AORN J. 2008;87:187–208. erating room efficiency at academic institutions. Anesth 6. Garner P. Complexities in the operating room. In: Lim G, Analg. 1998;86:896–906. Herrmann JW, editors. 62nd Annual Conference and Expo 22. Sarin P, et al. Specialized ambulatory anesthesia teams conof the Institute of Industrial Engineers 2012. Red Hook, NY: tribute to decreased ambulatory surgery recovery room Curran. p. 807–814 https://pdfs.semanticscholar.org/a256/ length of stay. Ochsner J. 2012;12:94–100. 269cee9d586a7bd324c83e47efb3b899d98e.pdf . 23. Wong J, et al. Delays in the operating room: signs of an 7. Deldar R, et al. Improving first case start times using Lean in imperfect system. Can J Surg. 2010;53:189–195. an academic medical center. Am J Surg. 2017;213:991–995. 8. Yount KW, et al. Late operating room start times impact mortality and cost for nonemergent cardiac surgery. Ann Thorac Surg. 2015;100:1653–1659. Conflicts of Interest. All authors report no conflicts of interest.
Improving and Maintaining On-Time Start Times for Nonelective Cases in a Major Academic Medical Center Dan B. Ellis, MD; Jason Santoro, RN; Dale Spracklin, RN; Vanessa Kurzweil, PhD; Stephanie Sylvia, CRNA; Peter Fagenholz, MD; Aalok Agarwala, MD, MBA
Background: As health care expenditures continue to increase, thoughtful use of perioperative resources is important. Efforts to improve operating room (OR) efficiency often focus on increasing on-time first case starts to improve OR utilization, reduce subsequent delays, and reduce adverse events. One institution, with severely limited inpatient hospital capacity and an extensive daily add-on list of surgical cases, focused efforts to improve OR efficiency by improving on-time first case starts for unscheduled, nonemergent surgeries. Methods: A multidisciplinary team was assembled to work together for this quality improvement (QI) initiative. The primary outcome measure was the percentage of cases starting on time. The team identified six key steps thought to contribute to on-time start performance. Data were collected for each of these process measures, and feedback was shared with stakeholders. Results: By measuring adherence to and giving feedback about critical steps in the preoperative process, on-time starts improved from a baseline of 65% to 85% (p = 0.041). Sustained improvement was seen even after daily measurement ceased and the QI project was completed. Conclusion: Establishing a multidisciplinary team to improve timely care of unscheduled, nonelective surgical patients; identifying key elements necessary for on-time surgical case starts; and providing feedback to clinicians were associated with a sustained improvement in OR efficiency for a traditionally difficult-to-schedule patient population.
H
ealth care expenditures in the United States continue to increase, with annual growth of 4% to 5%, and total health care expense reached $3.5 trillion in 2017.1 , 2 With a rising focus on value, thoughtful and efficient use of perioperative resources is more important than ever.3 Roughly 40% of revenue may be derived from the perioperative environment, making it one of the highest revenue-generating areas in a hospital.1 , 4 , 5 However, the operating room (OR) is also one of the costliest resources in a hospital, with OR time costing up to $60 per minute.1 , 6 OR profitability can determine a hospital’s solvency.3 , 7 To improve OR efficiency and improve patient outcomes, many institutions have focused on improving ontime starts of electively scheduled first cases and reducing turnover time between surgical cases. Cases that start later in the day tend to be associated with higher operative costs when compared to cases that start earlier in the day.8 In addition to the economic costs, which tend to be higher in patients who receive operations later in the day, patient outcomes may suffer when cases start later.8 , 9 Delays in scheduled cases may push nonelective, or add-on, cases even later into the day, requiring them to be performed after normal working hours, when surgical and anesthesia outcomes have been shown to be worse.10–16
1553-7250/$-see front matter © 2019 The Joint Commission. Published by Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.jcjq.2019.09.007
Although on-time start initiatives to improve efficiency are commonplace in hospitals, the complex adaptive perioperative workflow makes identifying and addressing issues challenging.4 , 17 , 18 Variability in surgical case types and personnel—including surgeons, anesthesia providers, circulating nurses, surgical technicians, OR assistants, and pre-op nursing staff—contribute to OR efficiency, or lack thereof. This variability can sometimes lead involved role groups (surgeons, anesthesia providers, and circulating nurses) to blame one another for lack of efficiency, as each group can become convinced that others are the cause of the delays. Many large, quaternary care hospitals such as ours often operate at or near 100% capacity, with inpatient units full almost nightly. Despite having many ORs, our institution often begins the day with more than 30 patients requiring unscheduled, add-on procedures. To accommodate this volume, several ORs are dedicated to these patients. However, many patients still wait hours for nonelective, nonemergent surgeries. These OR delays sometimes result in downstream effects by creating bottlenecks on inpatient units that subsequently result in overcrowding in the emergency department and postanesthesia care unit. To reduce unnecessary waiting for our patients, reduce wasted time for our perioperative personnel, and improve perioperative efficiency, we aimed to identify and correct issues limiting on-time starts for nonelective, nonemergent cases within one of these dedicated ORs.
2
Dan B. Ellis, MD, et al.
METHODS
This project was undertaken as a quality improvement (QI) initiative at Massachusetts General Hospital. As such, it was not formally supervised by the Institutional Review Board per hospital policy. Massachusetts General Hospital is a large quaternary-care hospital in Boston with 1,018 licensed hospital beds, 58 ORs, and more than 20 out-ofOR anesthetizing locations. Nonelective, nonemergent surgical patients are often cared for in one of four dedicated ORs: one each for neurosurgery and orthopedic surgery, and two for general surgery patients. One of these two ORs is allocated for nonscheduled patients from the emergency department or inpatient units requiring acute care surgery, and the second OR provides care for a combination of elective scheduled and nonscheduled acute care surgery patients. Emergency surgical cases are also preferentially placed into these rooms when necessary. Six attending surgeons and a cohort of surgical residents manage this surgical service. To create buy-in and include necessary stakeholders, a multidisciplinary team of trauma and acute care surgeons, anesthesiologists, nurse anesthetists, circulating nurses, and hospital administrators was assembled. Using standard process improvement priniciples,19 the team identified the goal (improving efficient care for nonelective, nonemergent general surgical patients), determined metrics to define improvement (improved percentage of first cases beginning by 8:00 am), and identified several changes that could be implemented to meet the goal after first mapping the existing and ideal processes. The following critical process steps necessary to start a case on time were mapped and identified: 1. A surgical case must be booked at least two hours before the scheduled start time. 2. A perioperative huddle to discuss surgical and anesthesia issues associated with the case must occur by 7:15 am daily. 3. The surgical consent, signed by provider and patient, must be in the patient’s chart by 7:15 am. 4. The anesthesia consent, signed by provider and patient, must be in the patient’s chart by 7:15 am. 5. The surgical history and physical (H&P) must be completed by 7:15 am. 6. The 24-hour update to the surgical H&P if the patient was seen by the surgical service more than 24 hours before the intended operation must be completed by 7:15 am. Compliance with each of these six process steps would serve as the project’s process measures. A case would qualify for an on-time start only if the patient was wheeled into the OR by 8:00 am. In addition, if an emergency surgery was in progress at 6:00 am, the room would be excluded from data collection. The prospective QI initiative was divided into three stages. During the first stage, baseline data for on-time starts
Improving and Maintaining On-Time Start Times
for patients who presented for nonelective, nonemergent general surgeries were collected. Between January 1, 2017, and April 21, 2017, six attending general surgeons; several surgical residents; a variety of anesthesia residents, nurse anesthetists, and anesthesia attendings; and two circulating nurses from the Trauma and Acute Care Surgery service cared for patients. Circulating nurses used a checklist to record whether each of the six process measures had been met for each case in addition to whether the case started on time. From April 21 through April 28, 2017, data were reviewed by the multidisciplinary leadership team. Over the course of one week, results were reported to individual surgeons and anesthesia providers. The data included personal performance in the setting of performance of peer providers. For example, if one surgeon was available for a 7:15 am huddle only 50% of the time, this surgeon would see his or her performance in comparison to peers. On April 28, 2017, the second stage of the QI initiative began. During this stage, data were again prospectively gathered over a three-month period. Similar to the first stage of the initiative, a member of the circulating nursing staff recorded whether each of the six process measures had been met and whether the case started on time. These new data were then presented to surgeons and anesthesiologists. Five attending surgeons who participated in the first stage of the QI initiative participated in the second stage of the initiative. One attending surgeon who had participated in the initial phase left the practice, and an additional surgeon joined the practice. A variety of surgical residents and anesthesia providers participated in the second stage. The same two circulating nurses who were assigned to the Trauma and Acute Care Surgery service participated in this stage as well. In August 2017 the postintervention data were again reported to the surgeons and anesthesiologists who were involved in the initiative. The final phase of the project consisted of a retrospective audit of case start data from August 1, 2017, through December 31, 2017. Case start times were extracted from the electronic medical record (EMR) and, as with the previous two phases, if an emergency surgery was in progress at 6:00 am, the room was excluded from analysis. Statistical analysis was performed using R, version 3.4.3 (R Foundation for Statistical Computing, Vienna). Analysis of case types in each period was performed using the 3-sample test for equality of proportions without continuity correction. Overall on-time start rate was compared preand postintervention using a 2-sample test for equality of proportions without continuity correction. Post hoc tests were then performed for each of the six process measures comparing pre- and postintervention. P values for the process measures were corrected for multiple testing using the method of Benjamini and Hochberg,20 and these corrected values are displayed in Figure 1. A statistical process control
Volume 000, No. , November 2019
3
Figure 1: Percentage of cases starting on time and associated process measures before (gray bars) and after (black bars) intervention. P values shown for process measures are corrected for multiple post hoc hypothesis testing. N.S., not significant; H&P, history and physical.
p-chart was generated and analyzed using QI Macros 2017 (KnowWare International Inc., Denver). Special cause variation was identified using Western Electric rules for control chart analysis.
RESULTS
The Trauma and Acute Care Surgery service cares for general surgery patients and performs a robust variety of cases. No significant differences were found in case types between observation periods (Table 1). Prior to the intervention, 65% of nonelective, nonemergent cases started on time (Figure 1). During the threemonth postintervention period, the overall rate of on-time starts increased to 85%, and this increase was statistically significant (95% confidence interval, 0.01–0.39; p = .041). After identifying a significant improvement in the outcome measure of on-time starts, we performed post hoc significance tests on each of the six process measures (Figure 1): 1. Case booked at least two hours before the scheduled start time 2. Perioperative huddle at 7:15 am 3. Signed surgical consent in the patient’s chart by 7:15 am 4. Signed anesthesia consent in the patient’s chart by 7:15 am 5. Complete surgical H&P by 7:15 am 6. 24-hour update to the surgical history and physical by 7:15 am (if applicable)
All the process measures showed significant improvement from the preintervention to postintervention period except the “Signed anesthesia consent in the patient’s chart” measure (Figure 1; p values displayed are corrected for multiple hypothesis testing). There was no consistent pattern of anesthesia staffing. In fact, a different anesthesia team (anesthesiologist and nurse anesthetist or anesthesiologist and anesthesia resident) cared for patients on each day of the observation. Therefore, data for anesthesia teams are reported in aggregate as the large quantity of variables prevents meaningful interpretation of data. However, granular anesthesia data were reported to anesthesia leadership so they could be referenced in discussions with individual providers. To evaluate whether the improvement in on-time starts was maintained after the postintervention period of daily monitoring had concluded, we conducted a retrospective audit of case starts using EMR data from the five months following the postintervention monitoring period. This revealed that the monthly percentage of on-time starts during this period follows a similar trend to the postintervention monitoring period (Figure 2). When graphed on a statistical process control chart using preintervention data as a baseline, the postintervention and retrospective audit data together meet criteria for special cause variation using Western Electric rules. This suggests that the improvement in on-time starts cannot be accounted for simply by the variation established by the baseline data and indicates that the underlying process has changed sufficiently to establish a new baseline.
4
Dan B. Ellis, MD, et al.
Improving and Maintaining On-Time Start Times
Table 1. Number of Cases by Type by Observation Period Case Type
Preintervention n (%)
Postintervention n (%)
Active Surveillance n (%)
Wound debridement PEG Lap cholecystectomy I&D Lap appendectomy Lymph node biopsy Exploratory laparoscopy Amputation Exam under anesthesia Groin cyst excision Lap colectomy Muscle biopsy Temporal artery biopsy Fixation of rib fracture Low anterior resection Skin graft Pancreatic debridement Exploratory laparotomy Hernia repair Cholecystectomy (open) Tracheostomy Excision of mass Creation of colostomy Diagnostic laparoscopy
7 (18.9) 7 (18.9) 5 (13.5) 4 (10.8) 3 (8.1) 2 (5.4) 2 (5.4) 2 (5.4) 1 (2.7) 1 (2.7) 1 (2.7) 1 (2.7) 1 (2.7) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
9 (22.5) 1 (2.5) 12 (30.0) 3 (7.5) 6 (15.0) 0 (0) 3 (7.5) 0 (0) 0 (0) 0 (0) 0 (0) 2 (5.0) 0 (0) 1 (2.5) 1 (2.5) 1 (2.5) 1 (2.5) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
17 (18.3) 7 (7.5) 17 (18.3) 10 (10.8) 9 (9.7) 1 (1.1) 5 (5.4) 0 (0) 3 (3.2) 0 (0) 1 (1.1) 1 (1.1) 1 (1.1) 1 (1.1) 1 (1.1) 1 (1.1) 2 (2.2) 7 (7.5) 3 (3.2) 1 (1.1) 2 (2.2) 1 (1.1) 1 (1.1) 1 (1.1)
PEG, percutaneous endoscopic gastronomy; I&D, incision and drainage.
DISCUSSION
To improve on-time start times for nonelective, nonemergent cases, we assembled a multidisciplinary team to address the different elements associated with first-case delays. Our process (identifying metrics, setting expectations, and feeding data back to individual providers) was associated with a 20 percentage point improvement in on-time starts. Although many groups have identified and addressed perioperative delays,21–23 our initiative is unique in terms of the patient population we studied. To date, there has not been a published example of improving on-time starts in nonelective, nonemergent surgical patients. This may be in part due to unpredictability and the heterogeneous nature of these types of cases as compared to scheduled, elective cases. However, the success of our initiative could potentially be applied to other services. We believe the success of our initiative is due to a number of factors. First, we established a multidisciplinary team to develop mutually agreed-upon metrics and group norms. After the metrics were established with buy-in from relevant stakeholders, positive and negative feedback could confidently be relayed to the involved clinicians and staff. Second, we agreed as a team to challenge the previously held belief that the Trauma and Acute Care Surgery service was inherently different from any other service line and unchangeable. Believing in our team and our process was key to this project’s success.
The third important factor was that key team members had recent outside experience. Having worked at a variety of other institutions prior to joining our hospital, team members could identify alternative approaches to the established patterns of care in our OR. Often, members of an institution can become entrenched in existing processes and have difficulty finding innovative solutions to long-standing problems. Recent experience with other systems and different processes allowed for a more critical appraisal of our current system. Given the paucity of literature surrounding this challenging patient population, this experience proved invaluable. A fourth factor contributing to our success was that we used the data for individual feedback in the setting of full-group performance. Although neither rewards nor punishments were tied to individual performance in this initiative, providers repeatedly commented on how well they performed relative to their peers. Finally, we were able to deploy a consistent set of providers to collect data. Over the 12-month project period, seven surgeons (one attending surgeon left the institution, and another surgeon joined the practice during the initiative), two new surgical residents per month, and a rotating cast of anesthesiologists, anesthesia residents, and certified nurse anesthetists were involved in these cases. Given the size and complexity of our teaching institution, it is nearly impossible to have a stable group of providers. However, two circulating nurses were consistently assigned to
Volume 000, No. , November 2019
5
Figure 2: This p-chart displays the percentage of cases starting on time by month throughout 2017. † Points that meet one of the Western Electric rules for identifying special cause variation.
the Trauma and Acute Care Surgery service and remained on service throughout the entire initiative. Limitations
There are several limitations to our study. First, our QI initiative was hampered by a limited sample size. It is possible that we would observe a more statistically significant signal if the observation period was increased and a greater number of cases was included. However, our preintervention data are consistent with our institutional performance— 51% to 58% of cases at our hospital started on time during the time frame in question. Second, it is possible that a general culture shift toward QI contributed to our success, as a variety of other perioperative process improvement projects were taking place concurrently with our intervention. Third, the second phase of this QI initiative took place during the summer, when new trainees were joining our hospital. Perhaps having fresh trainees that were willing to try a new QI initiative on the surgical and anesthesia services during the intervention phase of this study affected our results. Finally, it is possible that our results were influenced by the Hawthorne effect, as the postintervention
period was actively monitored using checklists to track processes and outcomes. However, this is less likely in light of our findings that the increase in on-time starts was maintained in the five months after the circulating nursing staff ceased tracking process measures.
CONCLUSION
By bringing together a multidisciplinary group of champions, including surgery, anesthesia, and nursing, we were able to change long-standing institutional practice and culture. The success of the initiative depended on a commitment from the circulating nursing staff to collect data, and support from anesthesiologists, nurse anesthetists, and attending surgeons. By working together for a common goal, the entire team benefited from better communication, teamwork, and sustained improvement. Improvement efforts centered on provider feedback may help hospitals improve the rate of on-time starts for nonelective, nonemergent surgeries, reduce patient wait time, and improve OR capacity.
6
Dan B. Ellis, MD, et al.
Improving and Maintaining On-Time Start Times
9. Asfour L, et al. In surgeons performing cardiothoracic surgery is sleep deprivation significant in its impact on morbidity or mortality? Interact Cardiovasc Thorac Surg. 2014;19:479–487. 10. Taffinder NJ, et al. Effect of sleep deprivation on surDan B. Ellis, MD, is Assistant Division Chief, Division of General Surgery Anesthesia, Department of Anesthesia, Critical Care and Pain geons’ dexterity on laparoscopy simulator. Lancet. 1998 Oct Medicine, Massachusetts General Hospital (MGH), Boston. Jason San10;352:1191. toro, RN, is Operating Room Nurse, Department of Perioperative 11. Grantcharov TP, et al. Laparoscopic performance after one Nursing, MGH. Dale Spracklin, RN, is Registered Nurse, Department night on call in a surgical department: prospective study. of Perioperative Nursing, MGH. Vanessa Kurzweil, PhD, is Project BMJ. 2001 Nov 24;323:1222–1223. Manager, Department of Anesthesia, Critical Care and Pain Medicine, 12. Komen N, et al. After-hours colorectal surgery: a risk MGH. Stephanie Sylvia, CRNA, is Certified Registered Nurse Anesthetist, Department of Anesthesia, Critical Care and Pain Medicine, factor for anastomotic leakage. Int J Colorectal Dis. MGH. Peter Fagenholz, MD, is Surgeon, Department of Surgery, MGH. 2009;24:789–795. Aalok Agarwala, MD, MBA, is Chief Medical Officer, Massachusetts 13. Ricci WM, et al. Is after-hours orthopaedic surgery associEye and Ear, Boston. Please address correspondence to Dan B. Ellis, ated with adverse outcomes? A prospective comparative [email protected]. study. J Bone Joint Surg Am. 2009;91:2067–2072. 14. Chacko AT, et al. Does late night hip surgery affect outcome? J Trauma. 2011;71:447–453. 15. Fechner G, et al. Kidney’s nightshift, kidney’s nightmare? REFERENCES Comparison of daylight and nighttime kidney transplanta1. Clark A, et al. Early endocrine attending surgeon prestion: impact on complications and graft survival. Transplant ence increases operating room efficiency. J Surg Res. Proc. 2008;40:1341–1344. 2016;205:272–278. 16. Wright MC, et al. Time of day effects on the inci2. Centers for Medicare & Medicaid Services. Nadence of anesthetic adverse events. Qual Saf Health Care. tional Health Expenditures 2017 Highlights. Ac2006;15:258–263. cessed Oct 17, 2019. https://www.cms.gov/Research- 17. Wright JG, Roche A, Khoury AE. Improving on-time surgical Statistics- Data- and- Systems/Statistics- Trends- and- Reports/ starts in an operating room. Can J Surg. 2010;53:167–170. NationalHealthExpendData/downloads/highlights.pdf. 18. Vardaman JM, Cornell PT, Clancy TR. Complexity and 3. Cima RR, et al. Use of Lean and Six Sigma methodolchange in nurse workflows. J Nurs Adm. 2012;42:78–82. ogy to improve operating room efficiency in a high-vol- 19. Associates in Process Improvement. Home page. Accessed ume tertiary-care academic medical center. J Am Coll Surg. Oct 17, 2019. https://www.apiweb.org. 2011;213:83–92. 20. Benjamini Y, Hochberg Y. Controlling the false discovery 4. Franklin J, Franklin T. Improving preoperative throughput. rate: a practical and powerful approach to multiple testing. J Perianesth Nurs. 2017;32:38–44. J R Stat Soc Series B Stat Methodol. 1995;57:289–300. 5. Fowler PH, et al. Perioperative workflow: barriers to effi- 21. Overdyk FJ, et al. Successful strategies for improving opciency, risks, and satisfaction. AORN J. 2008;87:187–208. erating room efficiency at academic institutions. Anesth 6. Garner P. Complexities in the operating room. In: Lim G, Analg. 1998;86:896–906. Herrmann JW, editors. 62nd Annual Conference and Expo 22. Sarin P, et al. Specialized ambulatory anesthesia teams conof the Institute of Industrial Engineers 2012. Red Hook, NY: tribute to decreased ambulatory surgery recovery room Curran. p. 807–814 https://pdfs.semanticscholar.org/a256/ length of stay. Ochsner J. 2012;12:94–100. 269cee9d586a7bd324c83e47efb3b899d98e.pdf . 23. Wong J, et al. Delays in the operating room: signs of an 7. Deldar R, et al. Improving first case start times using Lean in imperfect system. Can J Surg. 2010;53:189–195. an academic medical center. Am J Surg. 2017;213:991–995. 8. Yount KW, et al. Late operating room start times impact mortality and cost for nonemergent cardiac surgery. Ann Thorac Surg. 2015;100:1653–1659. Conflicts of Interest. All authors report no conflicts of interest.