Blog

Usability Profile: Physician-led Health Information Technology

Undergraduate studies at the Florida State University focused on my then dual interests in music and science.  My first brush with computers came during an undergraduate project that required calculation of bond energies and angles for organic compounds.  The University had an early (1960s) Control Data “Supercomputer” that enabled this task.  After numerous iterations of punched paper cards and trips to the computer center the project was completed.  I was to repeat similar exercises writing and debugging PL1 while at Duke a decade later.  Utility yes (I could never have done those calculations on my slide rule) but incredibly tedious and error prone.

After Medical School at the University of Wisconsin in Madison I returned south for internal medicine training at the Ochsner Clinic.  Alton Ochsner was then retired from the operating room but made highly popular daily teaching rounds.  I connected with the Biophysics Lab, which was later to be nuclear medicine at Ochsner and my clinical home.

Nearly all physicians during the Viet Nam era served in the military.  I spent two years at the Patuxent River Naval Hospital in Maryland.  Those two years provided invaluable experience in adult primary care and hospital medicine, and while there I created and certified the clinical laboratory.  Following this experience I turned to nuclear medicine and a fellowship at Duke University.

From Duke I returned to Ochsner in New Orleans where my clinical responsibilities were split between the endocrinology clinic and nuclear medicine.  My primary interest was the development of an endocrine clinical laboratory, which ultimately evolved into the radioimmunoassay laboratory.  Shortly after I arrived Ochsner deployed computers in the clinical laboratory and in nuclear medicine.  I wrote many image manipulation programs and successfully translated radioimmunoassay routines from FORTRAN to DEC PDP-11 BASIC.  This period led to my earliest notions that computers might be able to help doctors take better care of patients.

By 1980 Apple Computer emerged as the leading producer of desktop computers (the IBM PC would come several years later).  VisiCalc, an Excel forerunner, along with capabilities to support BASIC made the Apple II an increasingly attractive platform.  A dozen or so Ochsner physicians began meeting regularly to discuss possible ways computers might help us take better care of our patients.  The group included general surgeons, an orthopedist, an obstetrician, a pediatric nephrologist, pediatricians, internists, a radiologist, an emergency medicine physician, a pathologist and an anesthesiologist.  We proposed to build among other things a program to handle patient flow in the ED and a program to automate operating room scheduling.  We asked for and received several Apple II machines, BASIC (not yet compiled), and VisiCalc and set about to create our own programs.

Our medical records were on paper, our registration system was a card file, our appointment scheduling was done in ledgers laid out on large lazy susans to facilitate coordination of appointments, and our billing automation produced no claim forms.  A group of Clinic leaders wondered if we could automate these processes.  Our physician group and soon a similar group of interested nurses provided clinical input.  We eventually automated all those applications and deployed them from a mainframe computer.  We successfully used the practice management and electronic medical record that derived from this effort over a span of nearly 30 years.

Ten years ago Hurricane Katrina altered the life of our city and changed the course of the Ochsner Clinic.  In the immediate aftermath, we found ourselves fortunate indeed to have a fully deployed electronic medical record.  As the healthcare system struggled to recover from the storm’s aftermath Ochsner executed a bold plan to resurrect hospital care across the region.  We realized that we needed electronic access for other community providers and for our patients so we built Internet-based portals.  From pre-Katrina large Clinic with a single acute care facility Ochsner now owns 12 hospitals in a 25 hospital system and operates about 45 clinics.  We employ over 1000 physicians and another 1500 admit into our community facilities.  The proposition of operating an enterprise of this magnitude on what was largely a clinic-based in-house EMR was of course not sustainable.  We have deployed Epic across all of our clinics and hospitals over the past five years and today our patients have a single record shared by all of our care providers and used in all of our facilities.

I was fortunate to lead the efforts that created our in-house systems and was Chief Information Officer for 12 of those years.  More recently as our clinical focus turned to doing better and Meaningful Use I moved to a Chief Medical information Officer role.

At the origins of Ochsner’s health information technology journey we were concerned with helping our care providers take better care of patients.  Those early efforts focused on utility – initially assembling an electronic synopsis that enabled any Ochsner provider to know something about every patient encounter no matter where in our System it occurred.   Access to ancillary results and documentation soon followed.  One might argue about the usability of mainframe green screens but immediate access to content rather than depending on a fragmented paper chart was a large step forward.  Utility continued – automation of orders processing, electronic prescribing, documentation tools, messaging, in-box functions and the like.  As the electronic medical record evolved from simply providing access to information to supporting workflow usability increasingly became of concern.  Of course the evolution of technology helped – the adoption of a standardized graphical user interface, service architecture and web applications improved usability.  From the earliest of times our development specification and design were guided by those who had to use the end product.  We didn’t get it right all of the time but we learned a lot about what works and what causes frustration.  Much of that learning and experience has informed our Epic efforts; most particularly how important it is to include physicians in all decisions that will impact them.

So do computers help doctors take better care of patients?  In the context of a large, complex health system, the answer is yes, without question.  All of any patient’s record is available wherever they seek care in our System.  In addition to the value of automated process support our current system offers extensive clinical decision support.  Not just alerts and reminders but documentation support, access to supporting information, knowledge retrieval in context, guidelines, order sets, and more make today’s electronic medical record both safer and predictably more effective.  Reports and analytics provide a much richer information environment not just about one patient but populations of patients.

Still there are significant challenges.  Getting to where we are took nearly 40 years, substantial investment, and a culture that has grown to appreciate our now electronic health record’s value while working to improve its shortcomings.  The quest for efficiency, usability, and ultimately greater clinical value will continue for a time to come.

For more on the quest for value and usability in electronic health records please see Robert Wachter, The Digital Doctor – Hope, Hype, and Harm at the Dawn of Medicine’s Computer Age.  For an informative and entertaining discourse on design and usability see Donald Norman, The Design of Everyday Things.


HIMSS User Experience Community

This community is a space for like-minded colleagues to share best practices, come together around key issues related to user experience, and take action through participation in HIMSS initiatives.

HIMSS Physician Community

Visit the Physician Community to see how you can get involved and find the resources you need.

Keywords: 
clinical informaticsPhysician ProfileusabilityMedical Informatics