In this study, we apply one of the explicit purposes of developing analytic tools to monitor and elucidate the operational efficiency of perioperative services. We determined that the performance frontier defining the efficiency of UVMMC is more efficient than the performance frontiers defining the efficiency of Stanford Hospital and UAB Hospital (Fig. 1). Taken altogether, the results obtained here and external readings confirm the intuitive thought that as institutions increase in size (and complexity), the accuracy of tactical decisions decreases. As a result, clinical directors must more correctly predict the tactical block allocations– the assumption made here that the set of decisions creating the most efficient operating room constitutes such a set of correct decisions –to achieve operational efficiency. The question persists – what choices lead to operating room efficiency? As it stands, this determination relies not only on theoretical frameworks, but also the practice of operating room management. For now, normative decisions must rely on the judgment of clinical directors who intuitively understand complex systems and the myriad of interacting agents and processes component. Although performance frontier analyses, in its primitive infancy, must be applied post hoc to determine the efficacy of those changes, the results of those analyses then inform future changes until a robust set of institutional set of guidelines can be established.
There are several explanations for study’s results. First, UVMMC is a smaller institution with lower capacity than the other institutions. As the capacity for an institution increases, the absolute values for over-utilized time and under-utilized time will increase. As a result, the performance frontiers for the larger institutions will naturally trend towards lower optimization. Second, the patient volume at UVMMC is also lower. Here, operational objectives may be achieved with less interference because fewer physical and labor resources need to be expended to complete tasks, contributing to a down- and left-ward shift in the performance frontier. Third, differing patient populations and surgical case mix may contribute to different performance frontiers. As health care systems continue consolidate and adopt capacity-based service models, then it stands to reason that over-utilized time, at least, should be more likely [13].
With performance frontiers, tactical and operational optimization of a perioperative service becomes a trade-off between under- and over-utilized time, which represents fixed and variable costs, respectively. In retrospect, Strum perfectly described OR efficiency when he demonstrated that operational performance is a cost minimization analysis [14]. Organizational differences in perioperative services between institutions may be compared to then make normative statements regarding optimal resource utilization. Ideally, a dialectical analysis of the trade-offs at play would result in the ability to adjust demand patterns, staffing concerns, and inventory issues, among other considerations of normal operating room function, and model or replicate the specific aspects of operating room management that lead to a more optimal service. The results from this current study demonstrate that these approaches are potential limits to the ability to create operational efficiencies as perioperative services increase in size.
Previous research has demonstrated that performance frontiers can differentiate various specialties and capacity-based services. Haimes et al. argued that the block allocations for orthopedic trauma services should be 24 hours, not 8, 10, or 12 hours [15]. Similarly, Tsai et al. showed that mixed inpatient services, outpatient ambulatory, and non-operating room anesthesia service lines have different performance frontiers [16]. These insights inform how operations should be managed. For example, hospital administrators should stray from modeling mixed-inpatient settings when possible as they are the most inefficient [17]. When hospitals resemble mixed inpatient/ambulatory settings, administrators should decide to make tactical decisions that model the frameworks provided by ambulatory surgical centers and direct some of the mixed inpatient workload to ambulatory centers. When hospitals resemble NORA services, they should aim to fill the excess time not being utilized by provided staffing. Overall, the synthesis of these lessons gives credence to a heuristic for changes necessary in management tactics – operating room managers should create different lanes to manage each service under their purview, fine-tune the available ambulatory processes that help streamline the services, and adding resources to increase throughput.
Current quantitative metrics (e.g. first case on-time starts, tardiness, turnover times, OR utilization) are predicated on benchmarks and comparison rates. The fallacy of this framework is simply that an average utilization rate has little bearing on fixed and variable costs – meeting an institutional metric or national benchmark does not necessarily translate into a profit. Different institutions, specialties, and teams must employ variable management strategies that address the material conditions of their operational contexts, but the question of efficiency can expand to include a variety of metrics such as start-time tardiness, turnover time, and staffing costs [18], further complicated by the reality that institutions and literature often employ different definitions of these indicators [19]. Reconciling the differences between the performance frontiers of comparable teams may provide insight into what organizational structures lead to more efficient operating room schedules.
Again, performance frontiers can be used in comparative analyses of multiple institutions in operating room management. Although a wide variety of metrics contributing to operating room performance may exist, a targeted approach in which relevant metrics are adjusted for study followed by evaluation of impact on the breadth of metrics for operating room performance; performance frontiers are perfectly suited for this manner of multi-objective data visualization [20]. Further, different perioperative services can understand the limitations to their own operations, and perhaps, consider redesigning them rather than aiming to meet the minimal standard of an OR management metric. In short, performance frontiers support a resource-based view of the perioperative services, reflecting the nature of objective measurements to increase and decrease in response to other metrics in the same system.
To further elucidate the differences between specific institutions, qualitative studies involving management and labor forces involved in the perioperative services must be conducted. From an organizational development perspective, it might behoove perioperative service leaders to create smaller, more agile teams [21, 22]. As human organizations grow larger, Dunbar et al. demonstrated that groups larger than 150 tend to be less cohesive and collaborative [23]. Although there might be less than 150 individuals working in the perioperative services each day at large health care systems (e.g., UAB and Stanford), typically more than 150 individuals are involved in the overall delivery of surgical care, which may contribute to inefficient organizational outcomes. Similarly, building larger, more expansive ORs to minimize the costs of shared infrastructure (e.g., central sterilization, supply chain, pharmacy, radiology), might actually make perioperative services less efficient because of the increasing distances health care providers need to cover to move patients and materials through the system.
Our current study presents several limitations. First, we recognize a present inability to normalize the analysis between large and small institutions. Future analyses may consider normalizing over-utilized times and under-utilized times to the number of operating rooms in function in order to tease out details beyond institution size. Second, our analysis is limited to two operating room metrics; expanding the analysis to include metrics such as turnover time and/or pure productivity metrics such as Relative Value Units may reveal different results and provide further insight into the tactical decision-making necessary to optimize operating room management. Third, the study does not account for any major or minor tactical changes that occur during the time for which data was collected; these changes, although unlikely, may change the performance frontiers and conclusions drawn from the study should their impact be significant. Similarly, the data procured for the study can only represent the operational environment of each institution as they were, despite currently unidentified inefficiencies contributing to unaccounted differences in what under-utilized and over-utilized data represent between each dataset; again, expanding the analysis may help to address this concern as more details and variables are examined. More specific analysis accounting for case diversity, acuity, emergent cases, schedule changes, staffing ratios, and others may be conducted. Moving forward, we expect to expand the analysis to cover more institutions. As more macro-data points are analyzed, clearer delineations for normative judgments should arise.
This study serves as an initial foray into comparative analyses using performance frontiers, which may be expanded to a multi-objective framework that includes the breadth of indicators in the future. The inherent difference in operational efficiencies implies that there might be a limit to scale for organizations with large perioperative services. Future studies should elucidate these limits. In other words, hospital administrators and clinical directors will need to design different organizational systems as health care systems continue to consolidate.