AMDIS 2014: A View From The Edge
With more acronyms than you can shake an EHR at, the Association of Medical Directors of Information Systems (AMDIS) Physician-Computer Connection Symposium was a conference with an edge – at times edgy with dissonance, at times pushing to the edge of our knowledge and capabilities, at times the cutting edge of growth and change.
Held June 18-20 in Ojai, Calif., the 23rd annual program was two and a half days where national experts in clinical computing and healthcare information technology defined the latest issues and advances in the direct physician use of healthcare information systems and technology.
Meaningful Use and eCQM
Peg Meadow, Director of Government and Industry Affairs at Siemens, provided the legislative update and noted what many other conference participants were experiencing first hand, that the move from manually extracted clinical quality measures to electronic clinical quality measures (eCQM) can be staggering. Peg identified further quality challenges, specifically that nearly 100 percent of Stage 2 Meaningful Use quality measures contained errors, that the measures were immature and more time was needed for testing.
A panel of Harris Stutman, MD, Charles Sawyer, MD and Michael Zaroukian, MD further addressed Meaningful Use. Reflecting the feeling of conference attendees, the panel identified the rushed nature of the Meaningful Use timeline and also highlighted the eCQM reporting problems. The panel also addressed the HIT Policy Committee recommendations to the ONC, emphasizing the recommendation to focus more on interoperability and clinical quality measures. It remains to be seen how the ONC will respond to these concerns of IT professionals and clinicians. This divergence between those creating CMS Meaningful Use standards and those tasked with achieving standards has potential for substantial dissonance.
This tension could also be seen in coding discussions centering on ICD-10. As we all know, the CMS mandated change to ICD-10 has been delayed yet another year. A three-physician panel, Dr. Harris Stutman, Dr. John Handler and Dr. Alastair Erskine, spoke respectively on the merits of ICD-10, ICD-11 and SNOMED-CT.
Dr. Stutman reluctantly advocated ICD-10, but shared that the panel could identify no speaker interested in defending ICD-10. When the audience of some 200 participants was polled, there was not one vote for continuing to implement ICD-10. A significant majority preferred to implement SNOMED-CT for coding.
Beyond this dissonance, we also found ourselves at the edge of our capabilities. At AMDIS 2014, we heard the success stories of health information exchanges (HIEs) that had been providing patient-specific health information for years to a region of care providers. Unfortunately, the information provided did not include patient imaging. This should come as no surprise, as only a very few operating HIEs currently provide images.
Clinicians and patients are best served by the images, not diagnostic descriptions. Practicing clinicians are profoundly aware of the clinically significant variations present in the chest x-ray diagnosis of, for example, “right pleural effusion” and EKG diagnosis of “nonspecific ST/T wave changes.” A recurring message and recommendation of AMDIS 2014 is that HIT supports and facilitates, not interrupts and delays, the clinician’s workflow. Not supplying images is a clinical workflow “hard stop.”
An additional challenge of making HIEs work is to have them add net value to care delivery. Unfortunately, at this time, almost 75 percent of HIEs cannot support themselves.
We seem to be at the edge of what our technology and business models can do with HIEs. To achieve what HIT offers in cost reduction care improvement, we will need working, effective and efficient HIEs along with the next level of big data and business intelligence. The contents of our HIEs must mature to the next level, as should security and ease of access. As we approach these next levels, with the quality and cost reductions in these levels, the business case makes its own argument.
Wearables and data analytics
Patient engagement is another topic that is at the edge of our knowledge and capabilities. Probably surprisingly to some, patient engagement has been part of medicine since the time of Galen (“He favored listening carefully to the patient’s own words…”) and predates Stage II Meaningful Use.
One standard for Stage II Meaningful Use requires that 5 percent of users view, download and transmit their own health data. The concept is straightforward: When patients are involved in their care, both provider and the patient benefit. Wearables are a direct link to the Stage 2 standard of patient engagement. Besides patient engagement, wearables and their flood of data have been seen as tools “to prevent unnecessary readmissions, reduce costs and promote long-term health” and further bring the promise of “new methods of communication and more robust data for patients and providers alike.”
The “more robust data” is also bringing us to the edge of EHR knowledge and capabilities. There are hundreds of EHRs and there are, or will be, hundreds of wearables. Sounds like a lot of interfaces. Beyond this difficult although solvable technology exercise where do we put this data? How do we value this data? What can we do with it?
Dr. Larry Ozeran addressed this concern in his “Patient Managed Health Data” talk at AMDIS. Dr Ozeran reported a study where commercially available sleep monitoring wearables were worn during a conventional sleep lab study. Using the sleep lab study as the gold standard, the commercial wearables left much to be desired. Wearables should be standardized and regulated if the data will be used in clinical EHRs.
Concerns for the provenance of the EHR data is not limited to wearables and patient-provided data. EHRs have only sharpened our awareness of a long-standing problem. From the time of paper, entry of time documentation in clinical records has been a scary evolution. Going forward, the sensitivity to the need of data provenance will bring good results. But to achieve these results, much work, direction and standardization must happen first.
The high point of this conference was when Sameer Badlani, M.D., of the University of Chicago, discussed complex event processing (CEP) during the analytics presentation. Dr Badlani invited us to go a step further from retrospective studies on big data and analytics, to make data actionable in real time and discover unforeseen relationships.
With complex event processing, we continuously monitor diverse streams of data to identify relationships and forecast events. A combination of low-frequency, low-intensity events may combine to forecast or even cause a highly significant event. These previously unnoticed` predictive or causal relationships provide actionable information to the decision-maker.
CEP is not new. It has been with us in our cars (smart skid/slip braking) and credit cards (fraud detection). I believe CEP is the cutting edge of IT and healthcare. Coming soon, CEP at an ICU near you! Harnessing this analytic is the gift of fire.
Want more health IT insights? Subscribe to our blog below.