Time Is Money

A tale in Indian folklore describes a group of blind men, each touching an elephant for the first time. Each man touches one part of the elephant and firmly believes that he is right about the form of the whole; but a person with sight who sees the entire picture can understand how each man is right yet misses the entirety of the animal. The quest to develop models to improve operating room (OR) efficiency proceeds in a manner akin to the blind men struggling to describe the elephant. ORs are the economic engines of many hospitals, contributing up to 40% of the total hospital revenue.1 This resource-rich, high fixed-cost environment accrues considerable expenses. Efficient use of OR resources is a focus of recent operational research endeavors. Every minute of non–value-added care in the OR must be eliminated in the face of unsustainable rising health care costs. Anesthesiologists, as leaders of the Perioperative Surgical Home,2 are striving to develop methods to increase OR efficiency while maintaining quality care and patient satisfaction. Efficient OR planning and scheduling requires accurate estimation of operative time (time when a patient is physically present in the OR), including both anesthesia-controlled time and surgeon-controlled time. In this issue of Anesthesia& Analgesia, Maheshwari et al3 present an attempt to develop a tool to predict the time required for the anesthesia team to prepare a patient for surgery. Their retrospective cohort study examines patient characteristics and the planned surgical procedure to create a model to predict presurgical anesthesia preparation time (termed “anesthesia ready time” by the authors). They hypothesized that such a model would enhance appropriate resource allocation and improve OR efficiency at their institution. However, the final model predicted only about 25% of the variance in anesthesia time. Thus, the authors concluded that patient and procedural factors alone are not sufficient to reliably predict anesthesia ready time. Maheshwari et al3 should be applauded for their work, because it is a significant finding that the focus on patient characteristics and the surgical procedure alone does not allow for a model to improve efficiency. Failure of their model is due in part to limitations in 3 areas of the study, which the authors highlight in their discussion. First is the problem of patient selection. The study population included all noncardiac patients over a 5-year period in a tertiary center. The wide spectrum of surgical procedures, patient types, and clinical scenarios may dilute the effect of any patient subgroup where intervention may be focused. Notably, the results showed that anesthesia ready time for most patients is less than 20 minutes. This significant finding indicates an opportunity to obtain more granular data on those subsets of patients requiring more time. Moreover, 60% of the intended study population was excluded due to incomplete data—the reader has no way to determine if lack of documentation occurred because of production pressure in high-volume rooms or due to preoccupation with a complicated patient. The absence of documentation and anesthesia preparation time may have been interdependent. Second, the authors omitted the influence of the wider OR environment in their investigation of anesthesia time. Anesthesia coverage models, resident level of experience, distractions in the OR, availability of assistants at induction, airway management, time of day of the procedure, and emergency versus elective procedure status should be considered in predicting anesthesia time. While some of these factors have been studied, limited quality data exist regarding their importance to anesthesia-controlled time.4–6 Further, changes in OR management or clinical direction, nursing staff, team composition, and support staff are not addressed, falsely simplifying the complexity of the OR system. Finally, the authors used the Health Quest score, a predictive algorithm from a computer-based questionnaire developed by anesthesiologists at their institution. The ability of other institutions to adapt this knowledge to local preadmission testing clinics may limit the application of this model. Further, the patient-specific factors selected for analysis are those shown in the literature to affect hospitalizations, morbidity, and mortality; but these same factors may not accurately predict surgical time. Clinical outcome and surgical risk predictor models, such as the National Surgical Quality Improvement Program calculator, may be inadequate to predict operational metrics. The findings of Maheshwari et al3 align with recent literature examining the role of anesthesia in OR utilization. A number of studies have examined anesthesia-controlled time, including the ability of the attending anesthesiologist to predict anesthesia time, but none have achieved a reproducible prediction model.4,6 Notably, efforts made to assign a defined duration for anesthesia-controlled time have consistently failed.1,4,6 In fact, focusing solely on anesthesia time has not been shown to improve OR efficiency.7 Variables must include both anesthesia- and surgeon-controlled time,8,9 in addition to team performance10 and overall communication.5,11 Addition of this study to the literature should herald a move away from attempts to use patient factors in large, heterogeneous populations to improve resource utilization by predicting anesthesia time. Rather, our goal must be to improve overall efficiency to increase throughput with minimal wastage of resources and excellent patient/provider satisfaction. Understanding that anesthesia-controlled time may be 20% to 33% of surgeon-controlled time reinforces our need to examine our role as anesthesiologists in the broader system.1 We can conceptualize OR utilization in terms of 2 variables: human factors and system factors. As such, anesthesia time is influenced by both human factors, including patient, physician, or procedural needs and team composition, and system factors, including scheduling or staffing models and resource availability. A 2010 review of OR planning by Cardoen et al12 recognizes the human interactions at play, but argues for shifting operations research away from patient factors toward system factors. The human factors must be examined in the context of the larger hospital system. Modern electronic medical record systems provide a wealth of data on individual patients, procedures, and providers. Responsible use of this information involves using individual patient, surgeon, and anesthesiologist data to identify opportunities for improvement. Eijkemans et al13 demonstrated that patient factors are of limited value in determining OR times, but can be incorporated into a larger model including procedure and team details to improve OR planning. Optimization strategies may also include evaluations of team performance regarding specific procedures or patient populations. For example, fixed anesthesiologist-surgeon combinations for a given procedure may be correlated with historical times per anesthesiologist.10 Unlike studies examining all patients over a long period of time, more granular data about specific types of patients having distinct operations may highlight areas for improvement that can be individualized to the team responsible for those patients. Van Veen-Berkx et al8 created a successful time prediction model using “anesthesia time packages.” These personalized packages, considering patient, surgeon, and procedure, reduced the case cancellation rate by nearly 20%.8 Reducing cancellations improves patient satisfaction as well as patient throughput, improving utilization. The findings of Maheshwari et al3 underscore those of van Veen-Berkx et al.8 We must use proven quantitative techniques and not intuition or experience to develop accurate OR schedules and staffing plans. Moreover, the OR environment has both predictable and unpredictable sources of variability complicating decisions on staff and equipment resource allocation. Standardizing processes in the OR will decrease predictable variability. Accurately predicting operative time is essential to allow for a buffer to account for unpredictable variability. Maheshwari et al’s3 findings thus serve as a scaffold for anesthesiologists to apply operational research to OR efficiency.14 An ideal example of anesthesiologist-led system level change is the Perioperative Surgical Home.2 Following this model, what can a practicing anesthesiologist do to increase OR efficiency? Use locally available statistics to determine provider-specific data regarding key performance metrics in clinical, operational, organizational, and financial arenas. Sophisticated electronic medical record keeping systems and mathematical prediction models allow analysis of individuals, procedures, and teams in the OR. Because many large hospital systems have discipline-specific teams for surgeons, nursing, and anesthesiologists (ie, orthopedic team, urology team, etc), individual-level data may be expanded to examine the performance of such dedicated teams. Review of team performance metrics, displayed on an OR dashboard, allows opportunities for self-improvement and enhances communication. Data in the electronic medical record may allow in-depth examination of specific patient populations shown to consistently overutilize OR block time. Maheshwari et al3 identified that their prediction model was off by 2 to 3 categories in 3% of their studied population. While a small overall percentage, this 3% corresponded to patients who were of higher American Society of Anesthesiologists status with more comorbidities. Scheduling and planning can focus on this 3% of population who may need additional anesthesia time. Targeted interventions can be designed for resource-heavy procedures in this subset population. ORs must function as a part of the larger hospital environment—patient flow through the OR must be integrated with other facility operations, necessitating process improvement both before and after the procedure aided by external resource scheduling.12 Personnel, facilities, and the availability of postoperative care (intensive care, postanesthesia recovery beds, floor beds) influence scheduling and utilization.14 The correlations between the efficiency of different units improve global hospital performance.12 Applying variability methodology prioritizes the system over the individual to produce efficient quality care. Predicting anesthesia time alone has not been clearly demonstrated to improve OR utilization. This does not mean that anesthesia time should be ignored. Rather, it should be examined in the context of global OR scheduling. Maheshwari et al3 attempted to create a tool to predict “anesthesia ready time” but were not able to account for more than a quarter of the variance. OR scheduling is an active, dynamic process requiring both individualized and system-level data. Achieving efficient OR utilization is not simple, and the vision needed to achieve this is not apparent. Like the blind men touching separate parts of the elephant, each study looking at individual components of OR utilization is important. Now it is time to address the elephant in the OR, with eyes wide open. DISCLOSURES Name: Kelly A. Machovec, MD, MPH. Contribution: This author helped with editorial content and manuscript preparation. Name: Deepu S. Ushakumari, MD. Contribution: This author helped with editorial content and manuscript preparation. This manuscript was handled by: Nancy Borkowski, DBA, CPA, FACHE, FHFMA.