The Idea: A Sandbox Hospital To Test Our Systems


Click play for the audio version of this entry:




The Cost of Poor Quality Justifies The Concept of The Sandbox Hospital

Did you realize that the costs of poor quality are incredibly high in most systems?  It turns out that the median amount of cost recovered from a lean or six sigma project is approximately $250,000 in healthcare.  Because the costs of poor quality are so high it is critically important to focus on quality.  Like we have discussed previously on the blog, one of the challenges in healthcare is that routine and classic healthcare accounting systems don’t make these high costs associated with poor quality very clear.  Some of the other issues exist with the fact that reimbursement is not linked directly to quality in many cases.  Here, let’s talk about a useful idea that could really help improve quality:  the sandbox hospital.


Most Quality Issues Are In Fact System Issues

Before we describe the sandbox, consider the idea that most quality issues are in fact due to system issues.  Yet again, I’ll say this concept is very different than what I was taught in the residency and fellowships I performed.  However, on reflection and review of data, this is true.  It’s actually a very common theme in the quality control literature.  Although we in healthcare tend to focus on human error, in fact, the manner in which a system is designed can make human error more or less likely.


Look:  if you saw a train coming at you and you didn’t get out of the way whose fault is that?  Well, similarly, we can see human error coming.  If we continue to see human errors in the system, and have the ability to make them less likely by building the system a certain way (FYI we can do this by the way)…well, you see where this is going.  If we can design a system that reduces human errors it’s sort of silly NOT to design a system, based on DATA, that reduces those errors.  I mean, come on!


…Not Just Human Factors.

After all, we can imagine that having a surgeon who worked 200 hours in 10 days in a hospital, with minimal support staff, in a environment that had more than thirty trauma activations an hour we would be apt to have some delayed diagnosis or quality issues, right?  We are using the extreme case to indicate how silly it is to pretend human errors sit in isolation from circumstances.


Hey, to all the “blamers” out there:  the world is complex and things are multifactorial.  This doesn’t excuse human error, yet the pragmatic way of looking at it helps us reduce it rather than beating ourselves up.  When friction abounds in a system and a human error is produced should we really attribute the issue to a human error?  Labeling things as isolated “human error” is the typical fall back position of physician or healthcare provider (doc you don’t work hard enough or you weren’t smart enough).  In fact, most surgeons and healthcare providers are incredibly dedicated, bright, and sometimes hampered by ineffective systems.  They didn’t suddenly become foolish after being some of the top students in high school and college.  So the truth is:  real process improvement takes work, and labeling things as isolated human error is kind of, well, lazy.  Or, worse yet, labeling things as isolated human error may lead us to turn over our healthcare provider staff (a bandaid solution) rather than allocating resources to improving the system.  You’ll find, if you replace the offending physician, you will have often continue to have similar human errors.  (I promise.) So, FYI, human error is only part of a bigger story.  Again…I promise.


Poka-yoke Is The Design Philosophy “Design The System To Make It Easier For People To Give The Output You Want”

Guess what?  There’s even a design philosophy we use in Lean and Six Sigma called “poka-yoke”.  It means “make it easier to do the right thing”.  If you want people to be in the trauma bay within 15 minutes, for example, make it easier to get there.  Put it close to the call room, get some prehospital notification where possible, or do something to reduce the friction in getting there.  Remember, the solutions you use don’t have to work every time.  They just need to make it more likely that the desired outcome will occur.


Ok, ok.  I’ll get off my soapbox.


Finally…The Concept Of A Sandbox Hospital

Now enter the concept of a sandbox hospital.  If you accept the idea that quality and functional systems are key to deliver excellent care especially with respect to the complex systems in place nowadays, consider how important it is to have effective systems in the hospital.  Now consider how funny it is to roll out things like new electronic medical records and similar large scale programs to a hospital without piloting them first.  Experience with roll outs and innovation literature focus us on multiple ideas to reduce the silliness of rolling out a new system to the entire organization all at once.


There are terms for innovation labs and other entities that work within organizations to build and pilot new ideas.  A skunkworks is an entity within an organization that focuses on being an innopreneur in distinction to an entrepreneur.  A skunkworks facilitates thinking differently than the company at large and can avoid being victimized by the company’s culture.


To set up an effective skunkworks, it is important that the system at large not feel threatened by the skunkworks, the staff within it, or the results from its workings.  Wouldn’t it be interesting, then, to have a “sandbox hospital”?  A sandbox hospital is a microcosm of an entire hospital used expressly to pilot roll outs of new large scale initiatives on a smaller scale and to innovate new systems.  If you’re into Six Sigma, you could even think of it as giving an easier way to perform Design of Experiments (DOE) for healthcare.


This would minimize the cost of poor quality and dropped revenue seen when large scale initiatives (like rolling out electronic medical records) occur.  Hospitals:  spend some time thinking about how to rollout big initiatives to avoid dropped revenue, quality issues, and patient care problems.  Put it in the sandbox and learn your lessons there (!)


Specifically, a sandbox hospital would be a free standing facility that could be utilized by different healthcare systems with their providers, culture, and systems.  In other words, it would be a healthcare facility in which they could experiment with how things will look as they roll out a new system.  It would be a sandbox for precisely the reason that it would be a safe area to play.


Ethical Ramifications of the Sandbox Hospital


There are some interesting questions which are ramifications of this idea.  For example, could the Sandbox Hospital treat patients?  My answer to this is yes definitely.  The Sandbox Hospital would be staffed by providers and other participants from the larger healthcare system that seeks to pilot the new initiative or dream up a new system.  The staff would be equipped with laboratory tests and all the things required to practice medicine.  In fact the only thing that would vary is the new roll out or new initiative being tried for a mode of care delivery.  We could tweek other things to simulate patient flow through the ED and beyond, although this would be more difficult owing to the ethics of potentially turning away patients to control flow rate.  But it’s ok, we could figure it out.


The medicine, technical prowess, and cognitive dimensions of practice would not change substantially.  Therefore, patients could be treated at the Sandbox Hospital and it would be disclosed to them that they would be in hospital during the roll out of a new initiative.  However, the Sandbox Hospital would be continually rolling out new initiatives.  We could tell the patients that too.  The interesting thing, again, is that it would be staffed with a portion of the various specialties or systems involved from a ‘mothership’-type healthcare system.  No matter how you feel about rolling out a new system to patients in the Sandbox Hospital, think of how much better it will be than just trying your new thing on ALL the patients who come to the larger healthcare system.


I feel the Sandbox is much better than the alternative which currently occurs.  That is, new initiatives are rolled out to health care systems as a whole.  It is challenging to roll out these initiatives while maintaining revenue and, more importantly, quality of care.  And this is fairly high stakes and high risk!  What if the new initiative, for example, grinds the ORs to a hault?  That new initiative suddenly becomes very expensive.  A sandbox hospital, and any costs associated with one, are sort of a hedge against that level of risk.  Better to spend money on a small experiment than a big failure!


Consider the ethical ramifications of the Sandbox Hospital.  In fact, it seems that the concept is actually more ethically satisfying than what we typically do in healthcare, in that it focuses on safety and could have failsafes, etc., built in with a stop-the-line patient protection measure.  Currently, when a new initiative is rolled out in the hospital, it is often up to the hard work of the residents, nurses, advanced practitioners, or other physician staff to get around the fact that the health record or physician order entry system just plain doesn’t work.  A Sandbox Hospital would help staff anticipate those opportunities and focus on expected issues to eliminate the problems with a larger roll out.


At the end of the day, a Sandbox Hospital would likely be cost effective in that it would decrease the risk of catastrophic failure of roll out for a new system to a hospital.  Sandbox Hospital could help with roll out to operating rooms, etc.


Some possible downsides of the Sandbox?  Yes, it is possible that staff from the mothership hospital running the experiment will act differently in the new venue–it’s still better than the alternative of “hey let’s try this in real life first”.  Yes, there would be some variation compared to the larger system, but a small scale model like this may help us avoid the cost of poor quality seen with larger roll-outs.


In other words, the type of errors being watched and tracked in the larger hospital system may not exist once the electronic record comes, but very real other errors do exist and a Sandbox would let us see them.  And, of course, the systems can be sold as cost saving (“hey where’d all the unit clerks go?”) and quality improving (“we don’t have to read all that handwriting anymore but hmmm now the notes say the same thing everyday…).  We feel that near-miss events and medical errors of a different type may actually be at play and again we see a justification to try a new system in the Sandbox first.


At the end of the day, the Sandbox Hospital would be a highly ethical, cost effective, small-scale “experimental” hospital. When we say “experimental”, we don’t mean the medical care delivered to patients would be anything less than excellent. When we say “experimental”, we do not mean that we would be experimenting with medical care.  When we say “experimental”, however, we mean that the Sandbox Hospital would be constantly rolling out new initiatives at the behest of the medical centre supplying it with staff and culture for a certain time.  This would allow the collection of data regarding the failure modes into which the new system deployment may enter.  The rationale for the sandbox hospital is that it is much cheaper to fail quickly and effectively on a small scale than to fail catastrophically or partially with a roll out to a major healthcare system.  The Sandbox Hospital may minimize pain of deploying new systems and would improve patient safety on the whole by obviating patient care issues that occur during major rollout.


Call For Your Comments and Thoughts

If you are reading this, you may have participated in one of these roll-outs, such as the often-cited new electronic medical record rollout, we would love to hear from you with respect to how this influenced down time, physician order entries, and near-miss type events.  For example, although we are often instructed that physician order entry decreases medical error, some of us feel that an electronic medical record actually changes the nature of the medical errors.


As physicians and healthcare providers, have you seen major problems during the roll out of new administrative plan or system? We would love to hear from you below. Experiences like yours give us the foundation on which to build the sandbox hospital in the modern day.