Journal Pre-proof Time is Money: Can Punctuality Decrease Operating Room Cost? William C. Chapman, Jr., MD, MPHS, Xun Luo, MD, MPH, Majella Doyle, MD, FACS, Adeel Khan, MD, FACS, William C. Chapman, MD, FACS, Ivan Kangrga, MD, Jackie Martin, Jr., MD, MBA, Jason Wellen, MD, MBA, FACS PII:
To appear in:
Journal of the American College of Surgeons
Received Date: 30 June 2019 Revised Date:
21 October 2019
Accepted Date: 22 October 2019
Please cite this article as: Chapman Jr WC, Luo X, Doyle M, Khan A, Chapman WC, Kangrga I, Martin Jr J, Wellen J, Time is Money: Can Punctuality Decrease Operating Room Cost?, Journal of the American College of Surgeons (2020), doi: https://doi.org/10.1016/j.jamcollsurg.2019.10.017. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Inc. on behalf of the American College of Surgeons.
Time is Money: Can Punctuality Decrease Operating Room Cost? William C Chapman Jr, MD, MPHSa; Xun Luo, MD, MPHa; Majella Doyle, MD, FACSa; Adeel Khan, MD, FACSa; William C Chapman, MD, FACSa; Ivan Kangrga, MDb; Jackie Martin Jr, MD, MBAc; Jason Wellen, MD, MBA, FACSa a Washington University School of Medicine, Department of Surgery, St Louis, MO b Washington University School of Medicine, Department of Anesthesiology, St Louis, MO c Barnes Jewish Corporation, Division of Perioperative Services, St Louis, MO
Disclosure Information: Nothing to disclose. Presented at the American College of Surgeons 105th Annual Clinical Congress, Scientific Forum, San Francisco, CA, October 2019. Protocol approved by Washington University Institutional Review Board (IRB# 201802053). Corresponding Author: William C Chapman Jr, MD MPHS Washington University School of Medicine Department of Surgery Campus Box 8109, 660 S Euclid Ave St Louis, MO 63110 Email: [email protected]
Phone: (314) 362-4533
Short Title: Punctuality and Operating Room Cost
Background: Inefficient operating room (OR) utilization wastes resources. Studies have suggested “first case on-time starts” (FCOTS) reduce OR “idle time,” yet no direct association between FCOTS and markers of OR efficiency like “last case on-time end” (LCOTE) or overtime costs have been reported. We performed this study to evaluate factors associated with FCOTS, LCOTE, and OR overtime costs. Study Design: In April 2017, our medical center launched an FCOTS improvement initiative. Prospectively-collected data concerning cases performed in the 6-month pre- (10/2016–3/2017) and post-intervention (10/2017–3/2018) periods were retrospectively analyzed. Elective, nontraumatic cases performed by orthopedics, gynecology, urology, minimally invasive surgery, or colorectal surgery were eligible. Univariate and multivariable analyses were utilized to evaluate three outcomes of interest: the association between FCOTS and LCOTE (primary), the change in FCOTS rates after intervention implementation (secondary), and estimated overtime cost savings associated with FCOTS (secondary). Results: 12,073 cases (6,095 pre vs. 5,978 post-intervention) performed over 2,631 OR days (1,401 pre vs. 1,230 post) were analyzed. FCOTS rate increased after intervention (76.1% vs 86.6%, p<0.001), with post-intervention cases twice as likely to start on-time (adjusted odds ratio (aOR) 2.07; 95% CI 1.73–2.46, p<0.001). Additionally, starting on-time was associated with a higher likelihood of LCOTE (aOR 1.76; 95% CI 1.38–2.24, p<0.001) and 21.8 fewer overtime minutes (95%CI 13.7–29.8, p<0.001) per OR day. Post-intervention estimated savings of $87,954 in direct OR costs over 6 months were associated with the FCOTS initiative. Conclusions: The FCOTS initiative was associated with higher frequency of FCOTS, which was independently associated with LCOTE. This achieved an estimated 6-month cost savings over $80,000 in direct OR expenditures.
Abbreviation Master List: FCOTS: First Case On-Time Starts LCOTE: Last Case On-Time End OR: Operating Room ASA Class: American Society of Anesthesiologists Physical Status Classification System BMI: Body Mass Index MIS: minimally Invasive Surgery
Introduction Operating rooms (ORs) generate significant revenue for medical centers yet are also high-cost facilities to build, staff, and operate due to their complex personnel and equipment requirements.1,2 Maximizing utilization and minimizing costs of these capital-intensive resources, therefore, is paramount for healthcare systems. To that end, strategies for optimizing “First Case On-Time Start” (FCOTS) rates have become increasingly popular across institutions of all sizes and geographic settings.3-6 Mechanisms to increase FCOTS rates have been well studied.7 Factors associated with successful FCOTS programs include: specifying a discrete definition of FCOTS,8 early identification of factors associated with delayed starts,9 establishing a system of data collection and monitoring,3 and short-interval feedback to stakeholders.10 Additionally, multiple groups have found that successful FCOTS initiatives can improve surrogate measures of OR efficiency such as FCOTS rates,6,10,11 duration of “first start” delays,4 and perioperative staff satisfaction.12 However, no prior study has investigated the association of FCOTS with end-of-day endpoints such as “Last Case On-Time End” (LCOTE) rates or overtime OR costs. Two years ago, our medical center implemented an FCOTS improvement initiative across all perioperative areas. Recognizing the dearth of data regarding FCOTS effects on end-of-day outcomes, we undertook the following study to determine the effectiveness of our FCOTS program and evaluate the association between punctual first-case starts and last-case on-time finishes. Methods Description of Intervention
In April 2017, the Division of Perioperative Services at our tertiary-care medical center launched a multifaceted FCOTS improvement initiative. Based on best practices3,4,10-13 and investigation of institutional pre-intervention workflow, the policy comprised three main components: defining FCOTS and stakeholder education regarding its importance; assignment of a discrete, standardized process for pre-operative workflow known as “swim-lanes” (see eFigure 1); and daily reporting of FCOTS rates and sources of failure to hospital leadership and surgical, anesthesiology, and nursing stakeholders. All physicians, nurses, and ancillary staff working within perioperative areas received role-directed education on this initiative in the months preceding roll-out. Implementation of this three-part initiative was performed by Perioperative Services with little overhead cost. An educational slide-deck, authored by Perioperative Services leadership, guided a 15-minute orientation delivered by temporarily re-assigned nurse-educators at surgical conferences, anesthesia departmental meetings, and nursing huddles prior to implementation. The perioperative data tracking system was programmed to create a daily report identifying all first-start cases failing to achieve FCOTS and then email an explanation request to the surgeon, anesthesiologist, and circulating nurse from each late case. Explanations were then compiled by Perioperative Services office staff and reviewed during regular data tracking meetings within the Division to identify systemic opportunities for improvement. The primary marker of implementation success was monthly FCOTS rate. Implementation required no additional staff, equipment, or software beyond the resources already allotted to Perioperative Services. Study Period To study the impact of the initiative, we performed a retrospective observational analysis comparing cases performed in eligible operating rooms within the 6-month pre (October, 2016 –
March, 2017) and post (October, 2017 – March, 2018) intervention periods. A 6-month washout period immediately following initiative implementation allowed for full adoption of the initiative and controlled for any seasonal fluctuation in case mix or volume. Outcomes The primary purpose of this study was to evaluate the relationship between OR day start times and late finishes at the end of the day. Secondary outcomes included FCOTS rates preand post- intervention and estimated cost savings from overtime reductions associated with FCOTS. First Case On-Time Start FCOTS was defined as the first case “wheels in” time no later than 5 minutes after the scheduled start time for that case. Elective non-traumatic cases performed by orthopedic, gynecologic, urologic, minimally invasive, or colorectal surgery from OR days with scheduled start times prior to 9:00am were included. For most days, the scheduled start time was 7:30am. To assess the association of FCOTS rates with our initiative, we first compared FCOTS rates between study periods (pre- vs. post-intervention) using standard univariate statistical analyses. Then, a logistic regression model was constructed to evaluate the association between on-time starts and study period while adjusting for patient-level covariates. Variables were included based on clinical relevance and availability within the perioperative database; study period (a binary variable representing relationship with the FCOTS intervention), age (categorized as those younger or older than 65), body mass index (categorized as under or over 35kg/m2), American Society of Anesthesiologists Physical Status Classification (ASA Class 14), surgical service, and admission status (inpatient, outpatient, and same day surgery) were covariables in the adjusted model.
Last Case On-Time End LCOTE was defined as the conclusion of the last scheduled case of the day by the scheduled end time. An operating room day (OR day) was defined as the period between scheduled first case start and actual last case end in one OR during a single calendar day. Analyses of LCOTE were limited to OR days that were scheduled to start first cases prior to 9:00am and end last cases before 5pm, performed at least two cases during the day, and experienced no gaps between cases longer than 90 minutes. OR days that included operations performed by different services were also excluded. Modeling of factors influencing LCOTE was performed with both logistic and linear regression. Multivariable regression of LCOTE required construction of models at the OR level while accounting for patient-level variables from the multiple cases included in each OR day. Attempts to build a true multilevel model failed because individual ORs experience significant day-to-day variability in surgical services, staff, and cases. Instead, we elected to collapse patient-level variables from multiple cases into a single representative value for each OR day. We hypothesized that using averages to consolidate patient-level variables would wash out the true effect on LCOTE; therefore, we instead consolidated each patient-level variable as the maximum value of that variable from that OR day. For example, a hypothetical OR day with 3 cases with ASA Classes of 4, 1, and 1 would take the “ASA Class maximum” of 4. Compared to the average (2), the maximum value (4) much more appropriately accounts for outlier values within patient-level variables. OR day covariables analyzed in this way included maximum age, maximum BMI, maximum ASA Class (categorized as 1-2 vs 3-4), number of cases (categorized as ≤2 vs >2), surgical service, and admission status (categorized by the inclusion of at least one inpatient case per OR day).
Cost Evaluation We were unable to directly measure true overtime costs. Instead, overtime costs were estimated based on changes in overtime patterns observed from pre- to post-intervention periods. We first calculated individual OR day overtime by comparing actual case end times to the scheduled end times; any OR working beyond the scheduled end time was considered to have missed LCOTE. Observed overtime minutes were summed across both time periods. Cases ending prior to the scheduled end time were considered to have ended early and counted as negative minutes; doing so normalized the distribution and also more accurately “credited” cases for finishing early. The number of cases, OR days, and patient mix varied between study periods; therefore, we could not directly compare the overtime observed pre- and post-intervention. Instead, we used a previously described prediction model method14,15 to calculate the attributable time saved by the FCOTS initiative. First, a linear regression model derived from pre-intervention data calculated the “expected” amount of overtime among post-intervention OR days based on relevant covariates – age maximum, BMI maximum, ASA grade maximum, inpatient status, number of cases per day, and surgical service (see eDocument 1). The “expected” overtime per OR day was then summed across all OR days. This post-intervention “expected” overtime estimate was then subtracted from the “observed” overtime. The difference between the “estimated” and “observed” overtime in the post-intervention era was calculated. Assuming that the only change across the study periods was implementation of the FCOTS initiative, we then attributed the cumulative overtime difference to the FCOTS initiative. To estimate cost savings, we then multiplied the attributable overtime minutes saved by an estimated OR cost of $13.70 per minute. This institutional estimate of OR cost does not 8
include many additional costs such as anesthesia charges or the added costs of overtime wages. Common sources of service-specific variability in base OR cost such as special equipment or specialized anesthesia-related costs are excluded from this cost estimate. Statistical Analysis The Wilcoxon rank sum test was used to compare continuous variables. Pearson χ2 test was used to compare categorical variables. Linear regression modeling was used as previously described to produce an adjusted estimate of overtime minutes saved by FCOTS. Logistic regression was used to assess the relationship between FCOTS and LCOTE with calculated adjusted odds ratios (aOR), while the association of FCOTS and overtime minutes was estimated using linear regression. Since OR days from the same OR rooms might be correlated due to similar staff and administrative systems, the variance from these regression models can be biased due to clustering of OR rooms. To account for this bias and adjust the variance estimate for the models, we applied a “sandwich estimator” to account for such correlation to all regression models described above.16 All analyses were performed using Stata 14.0/MP (StataCorp, College Station, TX). Results 12,073 operations (6,095 pre- vs. 5,978 post-intervention) performed over 2,631 OR days were eligible for analysis (Table 1). Of these, 5,065 were first-start cases (2,580 pre- vs. 2,485 post-intervention). Pre-intervention cases tended to have a lower ASA Class (Overall ASA Class distribution p <0.001) and median BMI (28.9 vs. 29.4, p <0.001) compared to postintervention operations, although these differences are small and may not be clinically significant. Orthopedic surgery performed more cases both before (47.0%) and after (44.6%) implementation of the FCOTS initiative than any other service, followed in descending order by
gynecology, urology, minimally invasive general surgery, and colorectal surgery. Age, sex, and patient admission status did not significantly vary between groups. First Case On-Time Starts The frequency of FCOTS increased significantly after implementation of the initiative (76.1% vs. 86.6%, p <0.001) (Figure 1, eFigure 2). Among first cases that failed to start punctually, the degree of delay did not significantly vary between study periods (median 12 minutes pre-intervention vs. 14 minutes post-intervention, p=0.06). In the logistic regression model of FCOTS, cases from the post-intervention period were more than twice as likely to achieve an on-time start (aOR 2.07, 95% CI 1.73 – 2.46; p <0.001) compared to pre-intervention cases once adjusted for multiple clinical variables (Table 2). Patients ≥ 65 were also more likely to achieve FCOTS (aOR 1.16, 95%CI 1.02 – 1.32; p=0.03). Factors associated with failed ontime starts included ASA Class > 2 (aOR 0.78, 95%CI 0.63-0.97; p=0.03) and inpatient (aOR 0.25, 95% CI 0.19 – 0.35; p <0.001) or same-day surgery (aOR 0.66, 95% CI 0.52 – 0.82; p <0.001) admission status versus the reference admission status group (outpatient). BMI and surgical service were not significantly associated with FCOTS. Last Case On-Time End For analysis of covariates associated with end-of-day outcomes, 1,401 pre-intervention OR days and 1,230 post-intervention days (eTable 1) were analyzed. OR characteristics were not significantly different between groups; maximum age, BMI, and ASA class did not differ. The unadjusted percentage of LCOTE improved from 33.9% to 39.7% post-intervention (p=0.002). After adjusting for a number of factors including maximum age, BMI, ASA class, number of cases per day, surgical service and admission status, an on-time start for the first case was associated with a 76% higher odds of LCOTE (aOR 1.76, 95% CI 1.38 – 2.24; p<0.001)
(Table 3). Simply being in the post-intervention timeframe further increased the odds of LCOTE by 22% more (aOR 1.22, 95% CI 1.04-1.44; p=0.02). This effect was independent of FCOTS on the day of surgery. Cost Evaluation The unadjusted number of cumulative overtime hours decreased from 1088.6 to 906.4 (p = 0.001) post-intervention. In the multivariable linear regression model, FCOTS was associated with 21.8 (95% CI: 13.7, 29.8; p<0.001) fewer overtime minutes per OR day compared to OR days where the first case did not start on time (Table 4). Using the constructed overtime prediction model described previously, we estimated 663.7 overtime hours would have occurred post-intervention in the absence of the FCOTS initiative (see eTable 2, and eDocument 1). Only 556.7 overtime hours were actually observed. This difference of 107 hours represented an estimated savings of $87,954 in direct OR cost savings over the 6-month followup period. Discussion At our tertiary medical center, we found that the frequency of First Case On-Time Starts increased to over 85% after implementation of our low-cost, evidence-based program. Most importantly, we identified a positive association between FCOTS and OR day on-time finishes, leading to reduced OR operating costs. The positive association between FCOTS and on-time conclusion of the OR day (LCOTE) is the most consequential finding of this study. Proponents of punctual OR starts have often invoked this “common-sense” relationship,2,11 but these claims have previously lacked supporting evidence. Our data, however, suggest that starting on time – regardless of other sources of delay that occur throughout an OR day – does significantly reduce the risk of late
finishes and their associated costs. While this relationship appears to be worth an estimated $87,000 over 6 months, we suspect this estimate to vastly under-value the overall savings of punctual starts since it does not include multiple other perioperative costs such as overtime staff pay or anesthesia costs. The success of our FCOTS initiative is rooted in several previously-described components. First, establishing a common definition of an on-time start and then widely educating operating room staff, physicians, and nursing personnel is critical. Whether in critical access hospitals or tertiary care institutions, prior studies have almost uniformly noted the fundamental need of a clear, common goal in any punctuality improvement program.3,4,8,9 Second, programs must establish methods of tracking progress and reasons for failures – preferably in real time.3,4 In our center, this function is enabled by our perioperative data management system (SIS; Surgical Information Systems, LLC, Alpharetta GA), which automatically generates a daily report of late starts across all perioperative areas. Email followup with the surgeon, anesthetist, and circulating nurse in each late OR are then used to determine a cause of delayed start. Third, we established multi-tiered feedback channels that included daily, monthly, and annual reporting to various stakeholders including physicians, nurses, perioperative staff, department chairs and hospital leadership. As in other facilities, this multidisciplinary, immediate feedback enabled rapid correction of protocol, equipment, or personnel causes of delayed starts. In addition to these validated components, our initiative also included designated “swimlanes” to reduce delays and bottlenecks in the critical hour prior to scheduled case start times. These discrete time assignments for each part of the perioperative team streamlined the preoperative process of patient registration, anesthesia evaluation and preparation, consent and
marking by the surgical team, and final evaluation by the operating room nursing staff. While we did not specifically study the independent effects of each aspect of the FCOTS initiative, the swim-lanes anecdotally organized the pre-operative workflow and allowed for earlier identification of potential delay-causing problems (missing surgical consent, need for anesthetic block placement, etc.). 12 months after implementation, the implemented changes remained in use and produced FCOTS rates of almost 90% across our multiple operative areas. The remarkable durability of these improvements is likely due to multiple factors, but the most impactful appears anecdotally to be the short-interval feedback provided via daily emails. This system allows for rapid identification of systemic barriers to on-time starts (such as patient registration delays or improper packaging of certain surgical trays), which in turn empowers resolution in almost realtime. Multiple examples of these systemic, addressable barriers to FCOTS have been identified over the past two years by this system. In short, the data collection and monitoring of this process appears to have created an agile system capable of making evidence-based decisions to optimize on-time starts. However, assessing the specific effects of each component of the FCOTS initiative were beyond the scope of this study; further examination is needed to determine component-specific impact. Our conclusions must be considered in the context of several limitations. Of the included service lines, case volumes among the orthopedic and gynecologic services accounted for almost 75 percent of total procedures. Therefore, trends noted in the larger dataset may in fact be specific to one of those particular services. The uneven case contribution may therefore limit the applicability of this study to medical centers with different service volumes. Additionally, variables statistically associated with FCOTS (for example, the estimated 21 fewer overtime
minutes per OR day) may not be clinically significant in some inpatient settings; therefore, extrapolating the findings presented herein must be done with the attention to the particular goals and capabilities of individual institutions. Second, several important covariates were not analyzable within our data, particularly type of surgical approach (robotic versus open, for example) and utilization of anesthetic blocks. Both factors represent choke points where limited bandwidth or complex equipment may delay operative starts or turnovers. Anecdotally, performance of local blocks has been a substantial source of delay at our medical center. Third, the cost savings presented in this study is an estimate based on the difference between expected and observed overtime after implementation of the FCOTS initiative. While we consider these estimates very conservative, the calculations may not reflect actual savings recouped by our medical center. Finally, the retrospective nature of the study prevented causal assessment of FCOTS and LCOTE; we were only able to identify an association between the two. Conclusion By implementing a program to improve punctuality with first start cases, our medical center increased the frequency of on-time starts. On-time starts were shown to be significantly associated with finishing the last case by the scheduled end time, and we estimated that the FCOTS initiative saved over $87,000 in direct OR costs over the 6-month post-intervention period. We therefore highly recommend a tailored FCOTS initiative for medical centers seeking to reduce costs.
We would like to acknowledge the work of Jim Thomas, RN and Matthew Geiser for their critical assistance in procuring and analyzing the data presented herein.
Denton B, Viapiano J, Vogl A. Optimization of surgery sequencing and scheduling decisions under uncertainty. Health Care Manag Sci. 2007;10(1):13-24.
Achieving operating room efficiency through process integration. Healthcare financial management: Journal of the Healthcare Financial Management Association. 2003;57(3):suppl 1-7 following 112.
Lin G, Theisen B. Analysis of First Case Start Time Delays in the Operating Room of C.S. Mott Children's Hospital: Pre-Med, History and Physical, and Consent Issues. Ann Arbor, MI: Univeristy of Michigan; 12/12/2006. URL: http://www.umich.edu/~ioe481/ioe481_past_reports/f0612.pdf. Accessed 05/15/2019
Zitsch R, Wakefield D, Waterman M, Baker D, Brown M. OR First Starts Case Study. Columbia, MO: Univeristy of Missouri; 2015. URL: https://www.himss.org/sites/himssorg/files/FileDownloads/MU%20Health%20Care%20 Davies%202015%20-%20OR%20First%20Starts.pdf. Accessed 6/05/2019.
Ang WW, Sabharwal S, Johannsson H, et al. The cost of trauma operating theatre inefficiency. Ann Med Surg (Lond). 2016;7:24-29.
Wright JG, Roche A, Khoury AE. Improving on-time surgical starts in an operating room. Can J Surg. 2010;53(3):167-170.
A. Halim U, A. Khan M, Ali A. Strategies to Improve Start Time in the Operating Theatre: a Systematic Review. J Med Syst. 2018; 42:160. 16
Valentine K, Fleeger R. Staffing: Start Every First Case of the Day On Time. Outpatient Surgery. Vol XVII. Malvern, PA: AORN; 2016.
Volpin A, Khan O, Haddad FS. Theater Cost Is pound16/Minute So What Are You Doing Just Standing There? J Arthroplasty. 2016;31(1):22-26.
Mathews L, Kla KM, Marolen KN, et al. Measuring and Improving First Case On-Time Starts and Analysis of Factors Predicting Delay in Neurosurgical Operating Rooms. J Neurosurg Anesthesiol. 2015;27(3):203-208.
Prejeant D, Hattle R, Parker M, Stock G. Starting the First Surgical Case on Time to Cut Delays. URL: https://www.isixsigma.com/new-to-six-sigma/dmaic/starting-first-surgicalcase-time-cut-delays/. Accessed 5/31/2019.
Heslin MJ, Doster BE, Daily SL, et al. Durable improvements in efficiency, safety, and satisfaction in the operating room. J Am Coll Surg. 2008;206(5):1083-1089; discussion 1089-1090.
St Jacques PJ, Patel N, Higgins MS. Improving anesthesiologist performance through profiling and incentives. J Clin Anesth. 2004;16(7):523-528.
Grams ME, Kucirka LM, Hanrahan CF, et al. Candidacy for kidney transplantation of older adults. J Am Geriatr Soc. 2012;60(1):1-7.
Massie AB, Gentry SE, Montgomery RA, et al. Center-level utilization of kidney paired donation. Am J Transplant. 2013;13(5):1317-1322.
Williams RL. A note on robust variance estimation for cluster-correlated data. Biometrics. 2000;56(2):645-646.
Pre-intervention (n = 6,095) 56.0 (15.9) 3,565 (58.5) 28.9 (24.8-34.4)
Post-intervention (n = 5,978) 56.0 (15.9) 3,445 (57.6) 29.4 (25.1-35.0)
Age, y, mean (SD) Female sex, n (%) BMI, kg/m2, median (IQR) ASA Class, n (%) 1 469 (7.7) 339 (5.7) 2 3,199 (52.5) 3,132 (52.8) 3 2,276 (37.4) 2,389 (40.2) 4 144 (2.4) 75 (1.3) Surgical service, n (%) Colorectal 384 (6.3) 400 (6.7) Gynecology 1,287 (21.1) 1,175 (19.7) MIS 515 (8.4) 629 (10.5) Orthopedic 2,863 (47.0) 2,666 (44.6) Urology 1,046 (17.2) 1,108 (18.5) Admission status*, n (%) Inpatient 541 (8.9) 520 (8.7) Outpatient 3,289 (54.0) 3,287 (55.0) Same day surgery 2,265 (37.2) 2,171 (36.3) Case distribution, n (%) First 2,580 (42.3) 2,485 (41.6) Second 2,043 (33.5) 1,957 (32.7) Third 1,048 (17.2) 1,053 (17.6) Fourth or later 424 (7.0) 483 (8.1) FCOTS, n (%) All services 1,963 (76.1) 2,153 (86.6) Colorectal 125 (75.8) 145 (90.6) Gynecology 462 (83.5) 489 (94.2) MIS 181 (78.0) 227 (87.6) Orthopedic 922 (73.7) 973 (84.0) Urology 273 (72.0) 319 (82.0) Table 1. Demographic and Clinical Characteristics of Eligible Cases
p Value 0.7 0.3 <0.001 <0.001
<0.001 <0.001 <0.001 0.004 <0.001 0.001
Admission status describes the origin of the patient on day of surgery; outpatient denotes a patient presenting from outside the hospital who will be admitted after surgery, while same day surgery describes a patient who is admitted neither prior to or after surgery IQR, interquartile range; ASA, American Society of Anesthesiologists; MIS, minimally invasive surgery. FCOTS, first case on-time starts
Table 2. Association Between First Case On-Time Starts and Clinical Covariates from a Multivariable Logistic Regression Model Adjusted odds ratio* (95% Variable p Value CI) Study period Pre-intervention Reference <0.001 Post-intervention 2.07 (1.73 – 2.46) Age < 65 y Reference 0.03 ≥ 65 y 1.16 (1.02 – 1.32) BMI < 35 kg/m2 Reference 0.1 ≥ 35 kg/m2 0.86 (0.71 – 1.05) ASA Class ≤2 Reference 0.03 >2 0.78 (0.63 – 0.97) Surgical service Colorectal Reference Gynecology 1.19 (0.67 – 2.12) 0.5 MIS 0.89 (0.41 – 1.94) 0.8 Orthopedics 0.68 (0.40 – 1.16) 0.2 Urology 0.63 (0.37 – 1.08) 0.09 Admission status† Outpatient Reference Inpatient 0.25 (0.19 – 0.35) <0.001 Same-day surgery 0.66 (0.52 – 0.82) <0.001 *
Displayed adjusted odds ratios refer to the odds of achieving FCOTS. Admission status describes the origin of the patient on day of surgery; outpatient denotes a patient presenting from outside the hospital who will be admitted after surgery, while same-day surgery describes a patient who is admitted neither prior to or after surgery. FCOTS, first case on time starts; ASA, American Society of Anesthesiology Physical Status; MIS, minimally invasive surgery
Table 3. Association of First Case On-Time Starts with Last Case on Time End from Multivariable Logistic Regression Model Adjusted odds ratio* (95% Variable p Value CI) FCOTS <0.001 No Reference Yes 1.76 (1.38 – 2.24) Study period 0.02 Pre-intervention Reference Post-intervention 1.22 (1.04 – 1.44) Age maximum 0.7 <65 y Reference ≥65 y 1.03 (0.85 – 1.25) BMI maximum 0.2 < 35 kg/m2 Reference ≥ 35 kg/m2 0.91 (0.78 – 1.06) ASA Class maximum 0.7 ≤2 Reference >2 0.97 (0.80 – 1.17) Admission status† 0.7 No inpatient case Reference ≥ 1 inpatient case 0.91 (0.61 – 1.37) Cases per day 0.2 ≤2 Reference >2 1.16 (0.93 – 1.46) Surgical service Colorectal Reference Gynecology 1.46 (1.03 – 2.06) 0.03 MIS 1.16 (0.79 – 1.71) 0.4 Orthopedic 1.38 (0.94 – 2.03) 0.1 Urology 1.37 (0.91 – 2.06) 0.1 * Displayed adjusted odds ratios refer to the odds of achieving last case on time end. † Admission status describes the origin of the patient on day of surgery; outpatient denotes a patient presenting from outside the hospital who will be admitted after surgery, while same day surgery describes a patient who is admitted neither prior to or after surgery FCOTS, first case on time start; ASA, American Society of Anesthesiology Physical Status; MIS, minimally invasive surgery
Table 4. Risk Factor Association with Overtime Minutes per Operating Room
Effect on overtime (minutes)
FCOTS <0.001 No Reference Yes -21.8 -29.8 -13.7 Study period 0.3 Pre-intervention Reference Post-intervention -3.1 -8.9 2.73 Age maximum 0.998 < 65 y Reference ≥ 65 y 0 -6.2 6.2 BMI maximum 0.08 < 35 kg/m2 Reference ≥ 35 kg/m2 5.2 -0.7 11.1 ASA Class maximum 0.02 ≤2 Reference >2 7.6 1.2 14.0 Admission status 0.2 No inpatient case Reference ≥ 1 inpatient case 6.4 -3.1 16.0 Cases per day 0.006 ≤2 Reference >2 -8.3 -14.3 -2.3 Surgical service Colorectal Reference Gynecology -22.6 -35.8 -9.3 0.001 MIS -10.3 -24.8 4.1 0.2 Orthopedics -21.9 -34.3 -9.6 0.001 Urology -19.8 -34.1 -5.5 0.007 This table outlines the association of multiple clinical variables with overtime minutes accumulated per operating room day. Negative values indicate OR overtime (minutes) saved; positive values indicate overtime worked. For example, FCOTS is associated with 21 fewer overtime minutes per day compared to late starts after adjusting for the effects of study period, maximum age, BMI, ASA status, surgical service, and high-volume operating rooms. FCOTS, first case on time start; ASA, American Society of Anesthesiologists Physical Status. MIS, minimally invasive surgery
Figure Legend Figure 1. Overall first case on-time start (FCOTS) rate by month. Line graph depicting the monthly FCOTS rate during the pre and post-intervention study periods. Dashed lines indicate the average FCOTS rate per study period; the vertical red line delineates pre- and postintervention periods. The numbers included at each dot report the total number of FCOTS achieved that month. eFigure 1. Preoperative “swim-lanes”. This pictogram illustrates the “swim-lanes” concept implemented as part of the FCOTS initiative. The preoperative period was divided into discrete time periods assigned to each component of the perioperative team – preoperative nursing, surgical personnel, anesthesia, and in-room nursing staff – as well as the core tasks that must be completed in each time window. For example, the surgical team is allotted a 15 minute “swim lane” from 0645 to 0700 to update the history and physical, confirm consent, and mark the patient; anesthesia is then allotted the next “swim lane” to review the updated history and write preoperative orders (preoperative antibiotics). Note that no assigned time slots overlap, and that all preoperative tasks should be completed by 0725 in order to allow for patient arrival in the OR by 0730. eFigure 2. Monthly first case on-time start (FCOTS) rates by surgical service CRS, colorectal surgery; GYN, gynecologic surgery; MIS, minimally invasive surgery; Ortho, orthopedic surgery. This line graph depicts the monthly FCOTS rates subdivided by surgical service during the pre and post-intervention study periods. The vertical red line delineates preand post-intervention periods.
Precis Within operating rooms at a tertiary medical center, we found that starting first cases on time was associated with finishing last cases on time after adjusting for patient and system factors. We estimated that a modest increase in frequency of first case on-time starts resulted in a savings of $87,954 in direct operating room cost over 6 months. We therefore recommend optimization of first case punctuality as a low-cost savings measure within American medical centers.
eTable 1. Operating Room Day† Demographic and Clinical Characteristics§ Variable
Preintervention OR days (n = 1,401)
Post-intervention OR days (n = 1,230)
Age maximum, n (%)
p Value 0.5
<40 y 40-49 y 50-59 y 60-69 y >70 y BMI maximum, n (%) <25 kg/m2 26-35 kg/m2 36-40 kg/m2 >40 kg/m2 ASA Class maximum, n (%)
48 (3.4) 99 (7.1) 230 (16.4) 440 (31.4) 584 (41.7)
59 (4.8) 84 (6.8) 202 (16.4) 371 (30.2) 514 (41.8)
61 (4.3) 659 (47.1) 334 (23.8) 347 (24.8)
49 (4.0) 553 (45.0) 285 (23.2) 343 (27.9)
1 2 3 4 Surgical service, n (%) Colorectal Gynecology MIS Orthopedics Urology Admission status, n (%) No inpatient case ≥ 1 inpatient case Total cases, n (%) 2 3 ≥4 LCOTE, n (%) All Services Colorectal Gynecology MIS Orthopedics Urology Total overtime, h All services
12 (0.9) 481 (34.3) 848 (60.5) 60 (4.3)
0.054 9 (0.7) 408 (33.2) 782 (63.6) 31 (2.5) 0.2 87 (6.2) 331 (23.6) 139 (10.0) 695 (49.6) 149 (10.6)
76 (6.2) 279 (22.7) 156 (12.7) 564 (45.9) 155 (12.6) 0.6
1249 (89.2) 152 (10.8)
1105 (89.8) 125 (10.2)
759 (54.2) 474 (33.8) 168 (12.0)
677 (55.0) 401 (32.6) 152 (12.4)
475 (33.9) 25 (28.7) 118 (35.6) 51 (36.4) 235 (33.7) 46 (32.1)
488 (39.7) 24 (31.6) 119 (42.6) 49 (31.4) 230 (40.7) 66 (42.9)
0.002 0.7 0.08 0.4 0.01 0.048
Colorectal 85.5 77.8 0.3 Gynecology 236.9 180.0 0.008 MIS 116.2 139.0 0.9 Orthopedic 516.4 396.5 <0.001 Urology 133.6 113.0 0.1 OR day describes a single operating room on a single day during the study period. Since multiple cases occur in each OR per day, patient-level factors such as age and BMI are categorized by the maximum value recorded of all patients in that discrete OR day. Of ORs initially included in the FCOTS analysis, 300 OR days were excluded from LCOTE analysis due to extended (>90 minute) gaps between cases. These extended gaps were assumed due to either off-service or non-elective cases being performed between scheduled operations. ASA, American Society of Anesthesiologists Physical Status; LCOTE, last case on-time end MIS, minimally invasive surgery
eTable 2. Risk Factors for Overtime Minutes from a Linear Regression Model of PreIntervention Operating Days Effect on overtime, Risk factor 95% CI p Value minutes Age maximum ≥ 65 y 4.4 -4.1 12.9 0.3 BMI maximum ≥ 35 kg/m2
ASA Class maximum > 2
At least 1 inpatient case
More than 2 cases
Reference -17.3 -12.0 -17.8 -14.8
-35.5 -32.2 -34.7 -34.5
0.8 8.2 -0.9 4.9
0.06 0.2 0.04 0.1
Surgical service Colorectal Gynecology MIS Orthopedics Urology
This table reports the effect of each risk factor on overtime minutes as modeled by linear regression among all pre-intervention OR days (n = 1,401). Using these effects as coefficients, we constructed a second linear regression model to calculate the total expected overtime minutes from post-intervention OR days (see eDocument 1). ASA, American Society of Anesthesiology Physical Status; MIS, minimally invasive surgery
eDocument 1. Linear Regression Equation for Calculating Expected Post-Intervention Overtime Minutes per Operating Room Day уovertime minutes = 42.9 + 4.4*(Age maximum ≥ 65) + 1.5*(BMI maximum ≥ 35) + 5.7*(ASA Class maximum > 2) – 0.6*(≥ 1 inpatient case) – 4.6*(> 2 cases per day) – 17.3*(Surgical service Gynecology) – 12.0*(Surgical service MIS) – 17.8*(Surgical service Orthopedics) – 14.8*(Surgical service Urology) + εi This equation, derived from the pre-intervention OR days (see eTable 1), was used to calculate the expected overtime minutes per OR day in the post-intervention period. The calculated expected overtime minutes for each OR day were then summed, yielding the total expected overtime minutes. By this method, we calculated an expected cumulative overtime of 663.7 hours. ASA, American Society of Anesthesiology Physical Status; MIS, minimally invasive surgery; εi, random error term of the regression model.