Coding for Codes

In primary care, we spend the majority of time on prevention (primary, secondary and tertiary) intermixed with acute illness.  The pace is furious and staying up on this aspect of medicine, backed by evidence, can be all absorbing.  Therefore, although practices vary, by necessity overall practice scopes have narrowed.  A shrinking number of primary care doctors outside of remote areas are performing a broad spectrum practice with emergency care, obstetrics, and inpatient/ICU.

The code.  Adult cardiac arrest, pediatric respiratory arrest, major trauma, status epilepticus.  We will all be confronted at sometime by emergency situations, where there is no pocket reference available, no time for phone calls beyond 911, noone to turn to with more experience.  Ten minutes to paramedic arrival at the clinic, or an hour of stabilization at a rural ER before air transport to a tertiary center.  We have the opportunity, here more than the 4000 other patients we will each see this year, to make a difference.   The opportunity to deliver  optimal care based on early recognition of diagnoses and rapid initiation of life saving measures.

Here is the tenuous chasm we face in primary care: the discrepancy between a shrinking breadth of practice and the always unexpected true emergency.   To prepare we take ACLS, PALS, NRP, ALSO, ATLS – courses that present best practices in emergency management.  But between certifications, in 2 months or 2 years, how prepared are we for that emergency that is far from routine; the mnemonics and tricks that seemed so obvious in the resuscitation course mega-codes seem distant now.

And here lies the opportunity: Coding the simulation program.  Pilots train in flight simulators.   Soldiers use simulators for war games.  In medicine, we have the SimMan group mock codes – you just need 5-8 others for a quorum.  We have Heart Code software from Leardahl for ACLS and PALS.  These are intended for one time use for testing (and for sustaining AHA coffers).  These programs are scripted more like a play then a simulation.   They are built in isolation of other resuscitation scenarios; children are seperate from adults are seperate from trauma are seperate from pregnancy..

This has been my challenge for 15 years – the perfect resuscitation simulation.  1988 – we used  physiology simulation program in College Biology.  1993 – I contact the author, Tom Coleman – he kindly sends me the code and books.  I convert it to Pascal.  It works, but the math is far beyond my understanding and the model is suited more for intricate dry lab experiments, then for continuously cycling simulations.  1999 – I find myself the course leader and instructor for ACLS and PALS.  To practice, I write my first version of Virtually Resuscitated – a winForm application in Visual Basic.  It works – object oriented, a graphical interface, a hover over exam and a full crash cart.   2001 – the program is already outdated with the new ACLS, PALS guidelines.  2009 – I start re-training in the emergency related courses,  Resuscitate the Virtually Resuscitated program and start the process of  re-write for C#, WPF/Silverlight with an underlying more physiologic model – more control over parameters.  2010 – Frustration.  cardiovascular modeling proves impossible.  I start a review of computational biology. Peskin and MatLab are interesting, too complicated for me, and  a dead end for the integrated simulation I intend.   2011 – re-resuscitate the program by writing unit tests and integration tests.  At 200 of 400 methods tested, another brick wall:  modeling the pulmonary and circulatory systems.  Look again at Tom Coleman’s work – great stuff and updated, but still the physiology models are not adaptable to my purpose.  In the process, I did learn about Differential Equations again (thanks to the great online video school – Khan Academy).

And then, my first breakthrough in physiology modeling:  The constant physiologic linch pin.  Set this first and everything else follows.  In the respiratory system, it is oxygen consumption, which leads to CO2 production, which results in reflex changes in the  respiratory rate.  In the cardiovascular system, it is the End diastolic volume (which could be manipulated as preload) and whose normal value can be estimated for any age based on Body Surface Area (LV EDV index).  From this, with known ejection fraction and inotropy, we can estimate stroke volume based on the Starling Curve.  With heart rate, we know cardiac output.  If we know cardiac output for a certain BSA, and we know the normal Mean arterial pressure for a given age, we can fudge the Normal Systemic Vascular Resistance for a given age. 

The project is moving again.  Onward in Coding for Codes, Virtually Resuscitated 2012 here I come.

This entry was posted in Coding, Uncategorized. Bookmark the permalink.