Physician Satisfaction with EHRs: Finding Balance in a Digital World

Dr. Howard Landa, MD, CMIO Vice Chairman of AMDIS

As Vice Chairman of the Association of Medical Directors of Information Systems (AMDIS), Howard Landa frequently discusses with CMIOs the challenges and opportunities they face in leading informatics and analytics initiatives at their health systems.  We asked Dr. Landa to address physician satisfaction with EHRs and the emerging technologies that may help improve the value of these systems. 

While we have succeeded in providing healthcare systems with a plethora of health information technology pathways designed to improve the efficiency and quality of care, we have also managed to make patient care much more complex. At a recent AMDIS symposium on the evolving role of the CMIO in physician leadership, a discussion regarding the impact of the Electronic Health Record (EHR) on medical practice yielded interesting results.  When polled, many front-line clinicians will tell you that they have a love-hate relationship with the EHR.  They don’t want to give it up completely and return to the ‘old school’ days of paper, but the way it has impacted their workflows has simultaneously improved efficiency and caused a complete shift in the care paradigm.

What are the issues causing the physicians and clinicians to feel this way? One is the transition from Clinician created prose to standardized and structured documentation. EHR templates structure the documentation process by capturing data in drop-down boxes, checked or unchecked boxes and prefilled templates.  The templates provide consistent documentation methods, speed up documentation time and allow rapid analysis of large groups of patients via automatic analytics tools.  Yet they leave little room for interpretation regarding the nuances of the patient’s specific case.  In the analog world, a physician would dictate findings which would later be transcribed and become part of the record.  The dictation process allowed the clinician to share why certain decisions were made, the timing of the care process and “if/then” statements which could guide steps in clinical decision making.  As other care team members provided care to the patient, these aspects of the story were often useful in providing clinical context and understanding.  While the structured data provided by the usage of the EHR is very valuable from a data analytics side, we have lost the physician’s intellectual input into those notes making it difficult to “individualize” the patient.  One patient’s notes-over-time may be almost indistinguishable from each other, raising both cognitive and coding concerns.

So what can we do to swing the pendulum back toward the middle and find a balance between free prose “storytelling” and the high value of structured data?  Natural Language Processing (NLP) is one option. NLP is a practical application which can process free text that is entered or dictated into the record.  NLP enables physicians to extract the data in a more evolutionary and biologic way – i.e., the way a human brain would work as opposed to a computer.  When NLP is combined with nomenclature codification schemas such as SNOMED, structured data can be extracted incorporated into the patient’s record. Leveraging the “OpenNotes” patient engagement/empowerment movement and applying analytics to this data, clinicians will be able to improve the accuracy of the patient story and connect what the patient says with what the physician hears. The patient engagement with the process makes the requirement of clear and individualized information all the more important.

So to find balance and truly deliver the promise of EHRs, NLP and other tools will allow the care team the ability to read the notes and extract data while keeping track of the nuances of patient care due to the specificity of prose.  We will also be able to ask the systems to categorize, store, prioritize, codify, and place structured extracted data where it can be redelivered when needed while still retaining the physician’s intellectual thought processes.  Health systems can then use the findings to improve weaknesses in clinical documentation or care processes, further stratify population health risks and outcomes, provide input into clinical research and more.

Dr. Howard Landa, MD is an accomplished CMIO with a proven record of leveraging innovation, technical expertise and operational knowhow to deliver HIT solutions.  Industry recognition including 20 years as Vice-Chair of the Association of Medical Directors of Information Systems; two years as the chairman of the HIMSS Physician Community, and recipient of  Modern Healthcare’s Top 25 Medical Informaticists Award in 2010, 2011 and 2012.

Beware Best Practices

Almost twenty years ago, in 1996 after publishing “America’s Health in Transition: Protecting and Improving Quality”  the Institute of Medicine launched a long term, ongoing concerted effort on assessing and improving the quality of healthcare.  “To Err is Human” further galvanized the national movement to improve the quality and safety of our healthcare practices by putting the spotlight on how tens of thousands of Americans die each year from medical errors.   The “Quality Chasm” report underscored the importance of a dramatically improved information technology infrastructure to support a 21st century health system.  Building blocks for such a system include an electronic health record system and national standards.  Progress has been made, the federal government has paid out over 30 billion dollars in Meaningful Use incentives as of March 2015 and impressive examples of quality improvements are frequently quoted in the literature.  Yet, most would agree that the results to-date have been underwhelming.

It is important to recognize that most implemented EHRs with a “check-the-box” mentality order to comply with Meaningful Use.   When Meaningful Use was initially launched, our team suggested that we were “enabling the dinosaur”.  And while not prehistoric, the design of today’s healthcare system does have ancient roots.    The Romans constructed buildings called valetudinaria for the care of sick slaves, gladiators, and soldiers around 100 B.C. (Heinz E Müller-Dietz, Historia Hospitalium, 1975).  In the U.S., the number of hospitals reached 4400 in 1910, when they provided 420,000 beds (U.S. Bureau of the Census, Historical Statistics of the United States 1976).  So clinical information technology was about automating existing clinical processes in hospitals (Stead 2005) rather than transforming clinical decision-making and work processes across the care continuum” (Brown, Patrick, Pasupathy 2013). 

 Separately, quality and performance improvement departments focused on deploying best practice – a method or technique that has consistently shown results superior to those achieved with other means, and that is used as a benchmark. (Wikipedia).  While best practices have their place, it is important to recognize the risks associated with emulating others when the practice depends on an antiquated business model such as hospital care. JPGshutterstock_159756653

As health systems transition from 1.0 – Bricks and Mortar Healthcare to 3.0 – Digital, Value Driven Connected Health and Healthcare, we encourage a focus on emerging practice.  A concept born in “systems thinking”, emerging practice assumes:

  • We cannot copy other organizations, use it in our organization and expect it to work given the number of variables at play
  • Intentional design of care management and business models will result in disruption of today’s best practices
  • Collaboration and integration of clinical teams, business leaders, information technology experts and data analyst will create new value
  • Big bang, long term projects are giving way to agile, experimentation where we learn to work in new and different ways
  • Rather than using our intuition or past experience to drive improvement, data driven innovation can often have more remarkable results and new practice will emerge

So, the next time someone mentions “best practice” challenge their thinking.