Minority Report or Moneyball

I have often heard comparisons on the automation of System Status Management to the 2002 Spielberg movie starring Tom Cruise called “Minority Report” loosely based on the 1956 short story by Philip K. Dick. This science fiction action thriller is set in the year 2054 when police utilize a psychic technology to arrest and convict murderers before they commit their crime. The obvious comparison there is to the forecast of future call demand and the eerie accuracy of the reports that allow the right resources to get there in time to make a difference in the outcome. Sometimes in the movie, as in real life, there is a considerable cost to achieve that goal as well. It is easy to get wrapped up in the technology, particularly the virtual reality user interface that Detective Anderton (Cruise) uses to make sense of the premonitions and quickly locate the scene. I like to end the analogy there before we learn the darker side of the way the technology works and can even be manipulated to put a stop to the whole project. Perhaps some EMS providers think they see a similar inherent darkness and hope for an eventual collapse of the whole dynamic deployment paradigm as well. This may be where the art of a story and our reality diverge, especially considering the current economic dynamics even given the admittedly sporadic successes. This may also be why we need a different analogy.

Advertisement

Improving EMS Deployment Performance

I work regularly with agencies that are looking to improve aspects of their operations. Some casual readers may be surprised to know that the focus of those discussions is not always about cutting response times. While response is a simple and common measure, it clearly does not evaluate EMS well and certainly fails to encapsulate many of its complex needs and values. Still, I feel the necessity to address the time objective briefly before going on to other important aspects.  

Where Do We Go Next?

To know where our increasingly limited emergency resources will be needed next, we need to understand where future requests for service will originate. If we knew exactly where the next call would come from, we could proactively dispatch a resource there even before it is requested (watch the movie “Minority Report” for an idea of how that might work.) Unfortunately, the nature of emergency response is not nearly that easy, but that is not to say it is impossible to recognize useful patterns across both time and space. While the 2002 Spielberg movie was set 50 years into the future, it correctly predicted the use of several new technologies that have become reality in less than twenty years. And although we don’t use “precogs” in forecasting demand, the ability of data to show future patterns that effectively influence deployment is also now well established within some agencies.

No one can tell you who will be that very next person to dial 9-1-1; however, it is imperative for the effectiveness of deployment that we concede that people and events often follow certain predictable patterns. Let me explain how this works in just a few steps. First, consideration of the repeatable nature of the temporal distribution of calls has been used for years in making shift schedules. The following chart represents the daily call volume from a specific study, but without a scale along the vertical axis, it could easily be representative of almost any agency regarding their relative hourly volumes.

The daily behavioral routine of individuals perpetuates the collective pattern for the larger community. These daily patterns not only replicate over the years, but across various types of political jurisdictions according to a 2019 Scandinavian study on the �Use of pre-hospital emergency medical services in urban and rural municipalities over a 10?year period: an observational study based on routinely collected dispatch data.� The following graphs from that study represent the relative call volumes of rural, small and large towns, as well as medium and large cities over a decade showing the reproducibility of call volume forecasts by hour of the day.

If we segregate the total call data by weekday, we can capture variations by the hour-of-the-day within each day-of-the-week. The chart of call volumes by day over a twenty-week timeframe, shown below, displays the commonly repeated variation throughout each week. It is the reproducibility of these volumes that allows us to schedule adequate crews to cover these anticipated call volumes.

The next step is to adequately distribute those available resources spatially to address the variation over the geographic area by time which requires an even deeper understanding of the call patterns. The fact that we, as social creatures, often live or work in communities that share similar and predictable risk factors allows us to generalize assumptions of individual activities over larger community groups. Corporations have used targeted demographic profiles to understand local populations for many years. Community profiling has even been recognized by the World Health Organization as an essential skill for all health professionals to help understand the specific and detailed needs of focused populations. (See �Community Profiling. A Valuable Tool for Health Professionals� published in Australia during 2014.) Beyond predictable human variables that focus primarily on medical emergencies are the physical characteristics of our built environment that determine the repeatability of traumatic accidents. A 2009 publication by the Association for the Advancement of Automotive Medicine looked specifically at �Identifying Critical Road Geometry Parameters Affecting Crash Rate and Crash Type� to aide road safety engineers with the challenge of addressing safety issues related to the shape of motorways. The existence of identifiable causes explains the ability to properly forecast the vicinity of calls in addition to their timing.

The following animation demonstrates several spatial demand forecasts in quick succession that are normally separated in the real world by hours. Your existing historical CAD records contain the necessary information to build such dynamic views in real-time.

The demonstrated reliability of demand forecasts, both spatially and temporally, is well known to MARVLIS users and proven to provide the critical information necessary to make decisions in prepositioning resources to reduce the time of emergency responses and limit the distances travelled in emergency mode to enhance the protection of crews and citizens. Furthermore, the Demand Monitor has the capability of grading demand hotspot calculations specific to your service by comparing actual call locations � as they are being recorded � with the forecast probability surface to highlight both the accuracy and precision of our demand forecasts over time that is specific to your agency data and query parameters. The following screenshot shows comparisons of various forecast models.

The percentage of calls that correspond with each shaded area over the selected timeframe quantifies the query accuracy while the hotspot size denotes the relative precision. Accuracy could be increased easily by enlarging the hotspots, but this would be at the cost of precision. A well-balanced query should result in a relatively small-sized hotspot that properly captures a significant portion of actual calls.

Still, knowing when and where to anticipate calls is not enough in itself to determine resource deployment. Some number of outlier calls will likely occur outside of the forecast hotspots, so it is critical to also develop a strategy for managing the risk of covering demand versus geography as weighted factors in any deployment decision. Where we need to be next is well beyond the simple strategies we typically employ now and must fully leverage the depth of our data for deeper understanding and action.

Toward a Better Understanding of Dynamic Deployment

I recently had two articles published by EMS1 as a couple of “mythbusting primers” on the topic of dynamic deployment. The articles were Dynamic deployment: 5 persistent myths busted and Dynamic deployment: 5 more persistent myths busted. My intention was not to convince anyone of a position that opposes their current EMS world view pertaining to deployment models, but I had hoped to extend the work Dave Konig began in The EMS Leader defining the terms of EMS resource deployment in 2013 and to have an open discussion about it. My hopes of engaging in dialog fell somewhat short of my expectations. But after watching the presidential debate last night, I understand that the idea of a robust “give and take” may be more difficult to achieve in public interaction than simply setting a stage with opposing actors.

One comment I received the first week after publication of my articles was a posting that basically just left a link for an article by Dr Bryan Bledsoe from 2003 entitled “EMS Myth #7: System Status Management Lowers Response Times and Enhances Patient Care.” The assumption being that the topic was settled long ago. While I have great respect for the man who calls himself “The EMS Contrarian” and his robust body of writings (including by first EMS textbook), I respectfully disagree with the finality of some of his assertions. A great deal has changed in the past 13 years. Some readers may actually recall that MySpace debuted the same year that his opinion was written. For those who do not recall that social media phenomenon, MySpace was a precursor to Facebook that was once the largest social networking site in the world – even surpassing Google as the most visited website in the US. This was also a time when almost every patient was administered high-flow O2 because it was considered safe, even if not always effective. Fortunately, the evidence-based movement in EMS has caused many practices to be re-evaluated both for inclusion as well as exclusion. And computer technology has also made great developmental strides from the 2003 introduction of the first wristwatch cellphone named the Wristomo. At that time, engineers were still thinking of wearable technology as a cross between the 2-way wrist radio device that became iconic for Dick Tracy in the 1940’s comic strip and the modern flip phone of the day. Naturally, the device was designed to be easily unclipped in order to hold it to the ear like a traditional cell phone. It even offered an optional cable allowing it to exchange data with a computer. The development of Bluetooth freed designers to reconsider how a smartwatch could interact in an entirely different way with a user’s smartphone. The evolution of dynamic deployment has followed a similar trajectory.

The Gartner Hype Cycle is a graphical and conceptual presentation that describes the maturity of emerging technologies through five common phases. Each year, the organization follows several technologies through this consistent cyclical journey. While EMS deployment was not one of these tracked technologies, I would submit that the initial technology trigger in the case of dynamic deployment would have certainly been the work of Jack Stout on System Status Management in the 1980s. His publications in the Journal of Emergency Medical Services (JEMS) throughout the decade inflated the expectations for performance returns. Implementation issues however, contributed to it sliding down into the trough where many disillusioned system providers left it for dead around Y2K. But the story doesn’t end there. The combination of his economic theory with Geographic Information Systems (GIS) provided a new operational view of both demand as well as current positions of available vehicles reported in near real-time with growing bandwidth. The advancement of computer processing has allowed some of these same Stoutian concepts to now be performed in real-time. With practice in modifying the parameters, the concept of Dynamic Deployment has become, as one comment to the article stated, effectively SSM 2.0. The benefits are no longer theoretical or even limited to Public Utility Model services, but are being realized by both public and private EMS providers climbing the slope of enlightenment or who are content with the productivity gains they have already reached.

One of Stout’s assumptions that has changed since the Bledsoe article is the “20 week” rolling window for analysis. This is too broad of a query that effectively combines different seasonal impacts throwing off focused projections not improving them. Experience shows that just a few weeks backward or forward from the current date for only a few previous years gives the best demand  forecast. Tests conducted at BCS show that MARVLIS correctly forecasts 80-85% of calls in the next hour by identifying hotspots that are limited to approximately 10% of the overall geography. Going back too many years, as Bledsoe was led by a consulting statistician, can actually unfairly weight more established neighborhoods while undervaluing newer communities. The clinical significance of shorter response times is not always in the “37 seconds” that are saved or even in meeting an arbitrary response goal, but in reducing response to a meaningful 4-minute mark. Achieving this milestone has had a proven impact on ROSC in New Jersey for instance. And beyond clinical significance is contractual obligation. Like it or not, EMS is often judged (and even purchased) similar to fire protection – by compliance to a time standard. Software makes a difference in meeting those goals. Running a system so that it performs well in most cases means it is more likely to perform well in the cases where it really does matter to the long term health of the patient.

The increase in maintenance costs of 46% as claimed by Bledsoe has also been disproven with services showing a reduction in the number of unloaded (non-reimbursed) miles driven and even a reduction in the number of post-to-post moves in favor of post-to-call dispatches. By reducing fines for late calls, some services have found significant cost savings compared to previous operations.

In trading station lounges for the cramped cab of an ambulance, there has been a genuine cost to the paramedics and EMTs. However, the argument they make is not about fixing the plan, but rather it becomes an attempt discredit the foundation of that plan completely. Consider the fact that most field providers in a closest vehicle dispatch operation describe a “vortex” that traps them in an endless cycle of calls if they do not escape it in time. They find ways to try to beat the system rather than suggest that recommendations account for the unit hour utilization by vehicle and allow busier units to leave the high call volume area and move to less call prone posts to complete paperwork and recuperate. It is not that the strategy is inherently evil or wrong, but is designed to support a business philosophy that is not properly balanced, so the outcome becomes skewed. It is time to stop challenging the core notion and focus on specific concerns of the implementation that will make the system work better for all participants. As long as we demonize the idea, we will not be able to impact how it works.

Much like the polarization of the presidential debates, I have learned from experience that when we perceive only bits and pieces of the world around us, our minds fill in the blanks to create the illusion of a complete, seamless experience, or knowledge of a system in this case. Sometimes that interpolated information is no longer correct and it can keep us from participating in the crafting of a solution that truly works for everyone.

We Need Some New Stories

We always hear that EMS is still a relatively new discipline. And in the scheme of medicine, or even public safety, that is certainly true. But we shouldn’t let the fact of its youth keep us from acknowledging that it has already been around long enough to accumulate some of its very own antiquated dogma. If you have any doubt, consider the reaction to changes in protocol – even those with good evidence to support some new practice. Working cardiac arrests on scene, for instance, was not met, at least in my experience, with enthusiasm at the prospect of improving patient outcomes. What I heard were excuses for why something different wouldn’t work. I thought about that exchange this week as I was listening to a recent Medicast podcast on an entirely different topic. Near the end of that recording, Rob Lawrence remarked that we really need to do away with the old stories that start out with “back in my day…”

The stories of some grizzled professionals include not just memories of MAST pants or nitrous oxide, but the idea that tourniquets take limbs, not save lives. More recently stories have been spun about the movement away from the long-held reliance on the long spine board as an immobilization splint during transport or even the value of therapeutic hypothermia for cardiac arrests.

While there is no denying, or even stopping, a rapid state of change in EMS, we must be sure that it is not just change simply for the sake of change or even resistance for the same reason. Change must be meaningful change that is guided by reasoned thought and scientific evidence, not personal anecdote. And new practices should be carefully modified to address current issues or new understandings of the problem.

Another sacred, yet unjustified, belief among too many providers is that the dynamic deployment of resources (commonly referred to as “SSM”, or System Status Management) is an unmitigated failure of cost-consciousness that actually leads to increased expenses and provider dissatisfaction. The evidence, however, from many of the services who now employ some facet of dynamic deployment has proven that while it can be tricky to implement well; the savings in time, money, and lives are definitely real. And those savings need not come at the cost of provider safety or comfort either. Whether you have had bad experiences in the past, or just heard about it from others, it is time to set aside the old stories and take a new look at the current technology and practice in every aspect of EMS that leads to improved performance.

To advance our profession, we must completely ban the expression,  “but that’s how we’ve always done it” and look toward “how we can do it now!”

 

Static v. Dynamic: A Continuum of Cost

In our recently published book, “Dynamic Deployment: A Primer for EMS“, John Brophy and I established a dichotomy between the standards of static deployment and dynamic deployment in the very first chapter.  Fortunately, that strong polar perspective has spurred some interesting discussions for me. While the check-out lane analogy was effective in distinguishing some of the differences of static and dynamic deployments, its simplicity only recognized the extreme ends of the spectrum and failed to acknowledge what I would describe as a “Continuum of Cost” between them.

Few systems (at least those with more than just a few ambulances) probably function exclusively at either extreme. The static model will necessitate some flexibility to provide “move-ups” to fill holes, just as dynamic systems will have reasons to keep specific posts filled as long as enough ambulances are available in the system. The reasons for moving, or even fixing locations, may have something to do with demand necessity or even the political expedience of meeting community perceptions.

While there are many differences between static and dynamic deployments that we could discuss, there are also some elementary misconceptions. For instance, dynamic deployment does not mean vehicles are constantly in motion. The term dynamic refers to the nature of their post assignments which can vary between, and even within, shifts. As alluded to in the book, proper post assignments also reduce, not increase, operational expenses. In at least one example we stated, the dynamic deployment strategy was shown to significantly reduce the number of unloaded miles actually driven, which in turn increases the percentage of overall miles that can be billed. This situation not only increases revenue while simultaneously reducing expenses, it also reduces fuel costs and wear on the vehicles (and crews) too which potentially extends their useful life. All this is still in addition to reducing response time and improving crew safety by positioning ambulances closer to their next call so that fewer miles need to be driven under lights and sirens.  The inherent efficiency of this management strategy allows a system to achieve response compliance at the 90th percentile with the smallest possible fleet.  To achieve the same compliance level with a static deployment of crews and posts, the fleet must grow significantly larger. Another recent sample calculation showed that both staff and fleet size would need to grow by well over double in order to reach the same goal. The resulting cost continuum, therefore, clearly shows that a static fleet has operational and capital expenses multiple times the costs of the dynamic deployment model without burning crews out with excessive and unhealthy UHU figures.

For the sake of validating my argument, it is unfortunate that these examples are from private ambulances companies who do not wish to openly share details of their calculations at this time for competitive reasons. It would be safe, however, to assume from these competitive reservations that these results are not automatic, but dependent on proper management and the use of good tools. There are certainly numerous examples of poorly managed systems or ineffective operational tools. To achieve similar positive results in your own system requires certain knowledge, an underlying reason for having written the book in the first place, and an assurance that the deployment tools are proven to be effective.  Just as managers should have references checked during the hiring process, vendors of operational deployment tools should be able to provide ample references for successful implementations of their technology in comparable systems to your own. It is also important that any solution be able to address a continuum that includes your specific objectives to find a balance between geographic coverage with anticipated demand coverage at an acceptable workload and schedule for your staff.

There is no “magic bullet” to achieving operational nirvana, but the combination of effective management with operationally proven tools has shown that cutting costs while improving performance is an achievable goal in most any size system. It is also fair to say that performance can be enhanced with less skill through the application of significant sums of money; but honestly, who can afford that sort of strategy in the competitive arena of modern mobile integrated healthcare.

It is our desire to produce yet another, even more extensive, volume on the topic of dynamic deployment to make the achievement of efficient and effective high performance EMS a reality for more systems. Stay tuned for future details!

Could Busier be Better?

There is plenty of talk about “evidence-based procedures” in EMS lately. Well, today I read an interesting article that shows a link between being busier and better patient outcomes.

Okay…, now after reading that statement, what just happened to your heart rate? Was your automatic response to click the link in order find fault so you can dismiss the finding, or did it pique a genuine interest to read the article and find what might be of value to you personally in hopes of possibly achieving a better understanding of even one aspect in a very complex patient/care giver dynamic? It is interesting to see how we respond to “evidence” we don’t necessarily like, or evidence that contradicts with our own longstanding personal stereotypes.  I know that whenever I talk about Dynamic Deployment, or System Status Management, I immediately hear complaints from those who work in the field that it is all about the numbers and is often driven by greedy consultants forcing “snake oil” math on all too willing administrators who have forgotten their “street experience.” I usually try to combat the stereotype perception with facts about more progressive experiences with creating high performance systems, but I will admit right here that everyone is at least partially right – it really is about the numbers. However, it may not just be the same numbers you are thinking (but I will stick to my assertion that the logic is probably much less nefarious than suspected.)

Time is an easy thing to measure, but in itself, it is seldom very important. In fact, it can be much like a single vital statistic from a patient taken out of context. Still, time is a pretty fair proxy measure of performance on the aggregate.  And, like good base line vitals, it becomes especially useful when combined with other numbers.  Now, before writing your comment, please note that I never said anything about a 7:59 response standard, I was only talking about measuring time in the abstract.  I believe the argument over response time standards is very similar to arguing that everyone should have a BP of 120/80. Sometimes it is the right goal, but for others, or depending on the situation, the target may be higher or lower.

Each of us measures our work shift in terms of hours.  System Status Management extends that basic idea by measuring everyone’s time in a shift along with the work they accomplish and balance it against the public’s perception, reasonable risk, and the actual needs of individual patients and their providers.  There are plenty of bad examples out there and I refuse to justify them, but at the same time there are good examples of systems that are improving and taking the right measures into account.

The key is not UHU, TOT, response times, compliance, ROSC, patient outcomes, employee satisfaction or budgets – it is all of those things and much more. Those numbers are no more definitive in themselves than BP, pulse, O2 sats, capnography, skin condition, ECG, GCS or anything else we measure is a truly accurate indication of a person’s overall health. Similarly, it is no less fair to view SSM as a static group of measures than to believe the components of our patient assessment are unchanging. If some medic had overly emphasized, or even ignored, some measures in an assessment, that specific experience should not condemn a process that has been proven valuable in many other cases.

It may seem that I have ventured pretty far from the question with which I started this post about how busy we should be in order to be most effective. You may have even thought I was promoting an idea to maximize every minute.  As for the clinical interpretation of the answer, I will leave that to the authors of the particular study I referenced.  Instead, I will suggest that we all must be a little busier in understanding how our collective time and actions impact the performance of the systems in which we work. It doesn’t matter if your service is private, non-profit, fire-based or whatever; money and resources are always finite while demand and expectations are often increasing.  I would ask that you don’t simply rely on the assessment from “vitals�? of SSM taken years ago, but reassess with an open mind and set aside the prejudices of previous assessments. After all, very little in our business is truly static. Like a “routine�? interfacility transport, we can assume nothing has changed regarding the patient’s condition, or we can get busy and engage in our profession looking to have a positive impact on potential outcomes. Don’t leave leadership to the administrators, but take initiative to at least understand, if not improve, your corporate mission. You may be caring for patients, but the care of your career is part of your job too. Get even busier and improve that outcome for yourself.

Is 'SSM' Still a 'Bad Idea'?

Ideas often take time to saturate a market. Even if the idea is generally recognized as a good one, complete with compelling evidence, change can still take time.  As a current example, how many agencies still have a protocol for complete spinal immobilization on a long spine board for “any fall” or “significant impact”?  On that very point, Dr. Ryan Jacobsen puts forth a lengthy argument in this recording of a  presentation at a NAEMSP conference.  The process of acceptance can be even worse yet if the idea has been controversial – as in the case of “System Status Management” introduced by Jack Stout in 1983. This distinction means it takes longer still in order for it to receive a “fair hearing” even if the evidence now shows a positive impact. In an ideal world, the best ideas would always be automatically and universally adopted, but that simply isn’t how the world works.  And for any professional industry it is a good thing that ideas are properly “vetted”over time to determine what is truly “best” before wholesale adoption or, in the case of “bad ideas”, that they are discarded only when a fair reading of the evidence discredits them.

Gartner, Inc. of Stamford, Connecticut, has built both a reputation as an information technology research and advisory firm and a booming business of annually publishing their signature “hype cycle�? graphs by industry segment.  For those unfamiliar with these charts, the basic structure starts with a technology trigger near the origin of time and is visibility followed by a quick rise to the “peak of inflated expectations” that is often driven by a combination of unrealistic claims by proponents and the hopes of users desperate to believe those claims.  The exaggerated peak of hype is inevitably followed by a crash of popularity into the so-called “trough of disillusionment.”  Many ideas just die here and drop off the curve, but for others, a more realistic set of expectations develop as ‘believers’ (the “early adopters” according to Everett Rogers’ “Diffusion of innovations”) begin to experience measurable benefits and serves to push the idea (sometimes with changes) up the “slope of enlightenment.” This gradual advance passes an important point of inflection on the performance “S” curve known as the “attitude confirmation” identified by Joon Shin.  The next landmark is crossing a social “chasm” identified by Geoffrey Moore at another critical inflection point called the “attitude plateau.”  Once an idea successfully crosses the chasm, it plateaus as a generally recognized productivity concept for that industry. Some ideas fly quickly along these curves passing other older ideas that seem to just plod along at a much slower pace.

So, is “SSM” still on the curve? And if so, where is it?  We must first realize that ideas evolve and sometimes morph into other names (just as “Emergency Medical Services” is known by some as “Mobile Integrated Healthcare” now.)  One apparent synonym for “SSM” is a broader idea of “dynamic deployment.”  If we look at the literature and practices of emergency ambulatory services, we find that the underlying concept is still quite popular despite attempts of detractors to further discredit or simply ignore it.  One such potentially damning article was written by Bryan Bledsoe back in 2003 after a crash of industry expectations for the idea.  This could easily be explained as the time that SSM passed its own pivot point where its value was questioned in the trough of disillusionment. (Some may also claim that hypothermia treatments for cardiac patients was also recently in this trough.)

Computing performance has increased dramatically since the 1980’s (or even the early 2000’s) and algorithms are discovering patterns in many human activities.  Demographic data show socioeconomic clustering that leads to similar health issues and traffic patterns with road designs that see more accidents than they should. These patterns are proving to be key in forecasting demand for EMS services. Automated Vehicle Location systems allow far better tracking than ever before and traffic patterns are being used to calculate more realistic routes. These are some of the advances that help explain the numerous agencies that are significantly improving response performance and making use of resources. Where field providers take an active part is developing strategies, there are also reductions in post moves, unloaded miles driven, and better disbursement of work loads.  The efficiency gained by its use in mainstream agencies beyond the initial public utility model organizations seem to vindicate Stout’s early vision and research as the concept moves up the slope of enlightenment toward the plateau of general acceptance.

Ideas are not static entities, so our understanding must continue to evolve and incorporate new thoughts.  As the iconic American social commentator, Will Rogers once said, “even if you’re on the right track, you’ll get run over if you just sit there.”  So, to honestly argue an idea, proponents of either side must continue to evolve their understanding and witness the current thought and evidence of an idea.  There is little point in continuing to attack past grievances which have been addressed while ignoring the mounting evidence out of sheer disbelief.  If “SSM” is not a “good idea’ yet, it is certainly moving in that direction all the while being shaped by those who are concerned over the future of EMS (or MIH.)