“Hospitals are more like battlefields than Toyota production lines”. I wrote that years ago in response to a Lean consultant who had challenged the need for discrete event simulation in health care. Simply put, I said, health care systems are more complex, variable, and interdependent than anything encountered in modern manufacturing systems. Thus, why should we merely copy the same manufacturing improvement methodologies to this very different environment? More to the point, why not deploy different analytical tools and methods more appropriate to the far more complex world of health care? While we have certainly made some strides in improving health care processes using traditional manufacturing approaches, we clearly have a very long way to go. For instance, a recent meta-analysis of manufacturing-based improvement efforts of some 47 Emergency Departments shows only trivial performance improvements after two years of effort and expense, with some showing no improvement at all. Thus, as I have stated for years, traditional manufacturing approaches need additional tools and different approaches in health care if it is to successfully evolve.
And even now, care systems continue to grow even more complex. Hospitals are challenged with the proactive, holistic management of population health rather than the treatment of disease and ailments after they erupt. Furthermore, the continued concentration of health resources into single regional provider networks challenges the ability to effectively consolidate and streamline service provision to improve cost and quality rather than merely consolidating pricing power. Thus new questions need to be asked. For instance, how does the arrival patterns of ED patients impact the patterns of post-hospital discharges to nursing homes in the community? How does the surgical schedule impact the need for inpatient clinical resources on weekend shifts? How can we allocate beds within a geographic region to match the demographics of our patient population? And how can we better use non-clinical communal resources to drive down costs while improving clinical outcomes?
All this means that the current operational models need to change, and change dramatically. Yet, again, the common current process improvement methodologies are woefully ill equipped to quantitatively, accurately, and holistically create the necessary new operational models and predict the outcomes of the dramatic changes we require. Rather than tweaking the existing operational models, we must be able to understand the range of impacts of new performance-altering parameters as we develop our future-state business and care systems. And in order to do this, we must be able to fully discern and quantify the increased variability and complex interdependencies of the larger, more dynamic care systems we seek to optimize.
Thus our tools and approaches must become ever more sophisticated and reliable. A process map, no matter how many walls it covers, simply won’t do if we are designing a new community-wide care system. Current health care resources may simply be unable to envision the bold changes that must take place in their work environments, thus requiring a means of visualization of the future state from which new task allocations can be derived. And the overall capacity of the new business and care systems must be assessed and optimized with great accuracy prior to implementation so as to ensure success and aid with change management.
This is why I continue to purport the use of discrete event simulation in health care. We routinely use simulation in our Capacity Optimization work to aid in, for example, discerning and quantifying the interdependencies that drive systemic up- and downstream operational constraints; developing new care and business models that expand capacity; creating resource allocation algorithms that enable better care at a lower cost; and predicting the key demand and capacity patterns that promote efficiency, operational stability, and cost management.
All the sticky-notes in the world cannot demonstrate the outcomes of large-scale changes in these complex systems, nor can they predict how new operational models will function under varieties of circumstances. Simulation can. Thus, simulation has become increasingly important as hospitals continue to try to break down the departmental “silos” and create the system-wide operational scenarios that promote change and staff buy-in and enable the future of care delivery.
Whether it is the design of our “60-Minute ED” in which all patients are dispositioned within sixty minutes or an optimized community care model that encompasses the breadth and depth of the community’s resources, simulation is a vital tool for the development of new operational models and the prediction of the parameters of systems performance that will enable ongoing optimization under a variety of future circumstances. Without simulation and its predictive analytical power, we are left with walls of sticky notes and hopes that our guesses and assumptions are correct. Healthcare needs so much more than that!