Health Interoperability Exposed

Good news or bad? Insurance companies "own" clinical data...

The big news of the week – Aetna’s acquisition of Medicity – understandably has created a buzz throughout the industry.

In theory, the notion of genuine health information exchange (HIE) between providers and payers makes sense. Unfortunately, the “all for one and one for all” rallying cry holds little credence among providers – particularly if payers hold the keys to the castle.

Providers view sharing patient information with skepticism, even when the exchange occurs with a “friendly” entity, like a local hospital. They worry about the security of the information and that the data they offer may be put to nefarious use – to plot referral patterns, for example, or to extrapolate other business intelligence. 

Consider the level of misgiving, then, that accompanies the idea of sharing information with a traditional adversary.

Unfortunate though it may be, providers become wary when payers declare an altruistic dedication to improved clinical care. Cynicism abounds when providers feel payers go to great lengths to minimize reimbursement. 

In other words, the Aetna-Medicity deal may do little to resolve the impasse that exists between payers and providers, despite the homage paid to the importance of clinically integrated HIE – unless, of course, the company commits to genuine two-way information exchange, and uses the information to improve its own processes and workflow.

There’s no question about the potential that claims data holds to enhance clinical care. Providers along the entire continuum of care would have comprehensive visibility into the diagnoses, therapies, treatments and medications their patients receive. Clinical decision-making could be improved and waste reduced.

Conversely, payers could improve their own efficiency by taking advantage of the EMR data that would suddenly be at their fingertips. Imagine how much time could be saved in the pre-authorization process if payers relied upon information already at their disposable, rather than continuing to require physicians to fill out paperwork, submit faxes and emails, and schedule conversations with reviewers?

Here’s another example: My own practice recently was penalized when a payer declared we were not meeting standards for prescribing generic medications – when, in fact, we exceeded the payer’s parameters. We use electronic prescribing through our EMR and our patients often pay cash when picking up their medications. Claims data alone, therefore, did not provide a full picture of our prescribing habits – but information contained in our EMR would have supported our practice.

Bottom line: the idea of clinically integrated health information exchange has potential. But few on the provider side of the equation believe the promise will be borne out in practice if left in the hands of payers. If their deal is to succeed, Aetna and Medicity needs to commit to genuine bidirectional data sharing and begin to regard the provider community as allies in the effort to improve quality and reduce the cost of healthcare. 

In short, dialog must trump dollars. 

 

Semantic Interoperability and Ontology

Physicians, like all professionals, have developed their own language over time. Patients frequently complain about their overuse of “medicalese,” and suspiciously complain that we use “big Latin phrases” to justify high fees. I certainly can’t argue that the diagnosis of Hallux Rigidus sounds much less serious when reduced to its English translation of “stiff toe.”  None-the-less, medical terminology is essential for exact and effective communication between healthcare professionals. Unfortunately, the translation of this knowledge into computer-understandable vocabulary has been dismal at best. Don’t get me wrong, when it comes getting paid, we have done a masterful job—billing codes and disease classification for claims transmission seem to work just fine. Semantics, or the meaning of terms, is where we have difficulty in the clinical world. As I will explain, it is common for similar terms to have vastly different meanings in other contexts. Clearly, there is enough confusion in medicine today. 

Last week, I saw a patient of mine with a persistent mild and non-productive cough that is slowly improving.  He contracted a viral illness while onboard a cruise 5 weeks ago.   I obtained a chest x-ray which was read by the radiologist as “clear except for possible chronic fibrosis or discoid atelectasis at the right lung base.” This got me thinking about clinical semantics and ontologies. My first dilemma was how to explain to my patient that his x-ray was basically normal, and that the radiologist’s comments were more malpractice-driven rather than clinically helpful. There is, however, a more concerning problem here:  how would informatics interpret these results?

For instance, there exists an ICD-9 code for the fibrosis reported above: 515 – Postinflammatory pulmonary fibrosis, but this also includes things like cirrhosis of the lung chronic or unspecified. The same for the atelectasis: 518 – which essentially translates atelectasis to pulmonary collapse. Clearly, ICD doesn’t really represent my patient at all.  In fact, it is hard to conceive of a controlled medical vocabulary that represents “schmutz on the x-ray that probably has no clinical significance but the radiologist didn’t want to say ‘normal’ because he doesn’t want to get sued.”  

If you think about it, the basic problem here is how the receiver of information imperfectly uses language to reconstruct the patient at presentation. After all, reality is in the eye of the beholder.

Speaking of “eye,” the perception of this term varies markedly depending on whether you are a potato farmer, a hurricane expert, or an ophthalmologist. This is known as semantic heterogeneity and human beings are capable of understanding context and nuance, where computers are not. For instance, a logic based computer model might recognize that the fundus in part of the stomach, the stomach is part of the GI system, therefore the fundus is part of the GI system. However, if a clinician was referring to the fundus of the female uterus, the inferential logic as being part of the GI system is silly.

Here is the role of an ontology—a construct that can delineate the semantics of a particular terminology system in an unambiguous manner and lead to effective communication. This communication is not just human-to-human, but also human-to-computer-to computer-to human.

 

Semantics and Interoperability: The Foundation of the Personal Health Record

Much has been written already about the sustainability and practicality of PHRs. 

While a current generation of healthcare users frets about security and privacy, the next generation is much more concerned with availability.  

 Take Mint for example. Since its launch just a few years ago, this on-line banking service is incredibly popular.  In a time of well-publicized identity theft, it is interesting that a younger generation is all too eager to give out banking and credit passwords in exchange for the convenience of a single source of financial information. According to Mint, 65 million people in the U.S. use online banking today, and the average American has more than four bank relationships. Despite individual banks’ long-established on-line offerings, Mint’s success derives from its ability to provide an aggregated view from differing sources.  The analogy to health care is obvious, simple portals to payers or health systems are not sufficient, an interoperable view is an obvious necessity to today’s consumer.  But here’s the problem- unlike the US banking industry, our medical systems do not deal in a single domain (money), with a well-established standard (the dollar). 

What then is the role of semantics in this brave new world of the Personal Health Record?  Let me illustrate some common issues:

1.  Disparate medical records result in similar and/or duplicative data using different vocabulary standards. This ‘raw’ data, presented to the layperson in a PHR will result in misunderstanding.
2.  Lab data from differing facilities may have differing reference ranges. This will likely cause patients to question “normal” results, and limit their ability to accurately follow trends.
3. “Medical-ese” will need to be balanced by layman’s terminology.  A chest X-ray report with the phrase “pulmonary vasculature and cardiac silhouette appear unremarkable, however, one cannot rule-out neoplasm on this limited study” would probably better be listed in the PHR as “normal”.
4. Common laboratory panels with archaic or non-contributory values will need to be reassessed.  These are tests that should only be ordered in rare clinical instances, but for historic reasons are part of automated panels.  Clinicians have long ago learned to ignore such items in the complete blood count panel (eg. RDW, MCHC) or chemistry panel (eg. BUN/Cr ratio).  Patient’s frequently call me when they see their own reports—alarmed by “abnormal” results that have no significance.

I have previously opined that disease-specific clinical views of harmonized and appropriate information will be necessary for clinicians to interpret and utilize voluminous data in a meaningful work-flow. It is even more imperative that the layperson have analogous view in a PHR which will allow them to knowledgeably participate in their health care process.

Finally, we must remember that PHR data will undoubtedly flow both directions. EMRs must be capable of consuming home device or patient-entered data, in a presentation that does not confuse source to the caregiver.

Learn more about how the role semantics and interoperability play for PHR success by registering for this free webinar: Engaging Patients and Families with a Google Health PHR

Won’t the UMLS Solve Everything?

I recently submitted a speaking proposal on a practical approach to achieving semantic interoperability, and was shocked to see the response from one of the reviewers.  He intimated that the Unified Medical Language System (UMLS) is all that is needed to solve this complex problem. Furthermore, he indicated a disdain for any proprietary solutions. (Somewhat disingenuous, given that much of UMLS is proprietary—SNOMED and CPT amongst others). I thought I would take some time to discuss the common misconception of the UMLS as a panacea, for this is a banner that is commonly raised by well-meaning, but naïve individuals who dream of simple solutions to complex problems. I certainly understand the appeal of the National Library of Medicine’s open source efforts.

Unfortunately, while the UMLS may be an important tool, it cannot be a solution, and this is borne out by its failure to attain this goal despite its creation nearly 25 years ago. Now to be fair, it was never intended for this purpose.

The UMLS is not software. It is a system that was designed for information retrieval. The idea was to get information from different databases and present it to users—a noble goal to be sure.

Real clinicians need to do much more. They need to integrate information and knowledge (not merely data points) into their workflow in order to enhance the care of their patients in an efficient manner.   

As I have stated in previous blogs, standards are important. So many of them exist—HL-7, CCR, CCD, ICD-9 and 10, SNOMED, LOINC, CPT, etc. Standards are necessary, but not sufficient, to obtain true interoperability.

I do not profess to be an expert on this subject. I remain a Family Physician dedicated to establishing a practical solution. I would like to share a criticism of UMLS, however, by someone who is very much recognized as an expert. 

Dr. Barry Smith is the Director, National Center for Ontological Research, and Julian Park Distinguished Professor of Philosophy, at the University at Buffalo. I highly recommend taking a look at his blog. When writing about the Top 10 Errors To Avoid When Building Ontologies, he specifically calls out the UMLS for the sin of mixing ontology with epistemology. The example he cites is "cardiac output is a Laboratory or Test Result or Diagnostic Procedure”

Having said this, let me be sure to clarify—my goal is not to put down the UMLS— it does not deserve this, nor am I qualified to try. Instead, I wish to promote more enlightened conversations in an attempt to solve problems. Like other complex issues affecting health and wellness, there are rarely panaceas.  

I have ranted enough… and feel much better!

Canadian Healthcare and HIE

 

I have been fortunate to travel and interact regularly with my physician and HIT colleagues in Canada, and am always surprised by the difference between political rhetoric and real life.

For instance, frequently cited waiting lists just don’t apply to emergency and urgent care.  It is true that delays exist for elective procedures, but statistics showing increases in cataract surgery, coronary bypass, and joint replacement seem to negate claims of healthcare rationing.   

The cultural difference between American and Canadians seems to have a profound effect on the perception of problems.  When I order a non-urgent MRI for my US patients, they are confronted by a choice of several imaging centers, and delay of a few days scheduling is due frequently to a desire for more convenient location, hours, or open magnet.   

When I ask my friends in Canada about week-long (sometime month-long) waits, I almost always receive a response like “well, yes it’s annoying, but I suppose it’s a reasonable price to pay knowing that children in our society all have equal access to healthcare.”  

There appears to be a community ethos in Canada that profoundly influences reactions to their universal care system. The concept of sharing seems to have a lot to do with this, and I wonder if Information Exchange has a different societal context.

How will this emphasis on “community” effect Canadians’ adoption of HIE?  With their impressive recent ramp-up of HIT, it seems that meaningful sharing of key medical information via technology is at the core of several important initiatives.  These include community care plans and registries for chronic disease management, insuring safety and continuity during transitions of care, quality measure compliance, and basic access to a complete patient data set at all venues of care.

A recent article in the Ontario Globe and Mail discusses concerns about an aging population and its potential negative effects on healthcare delivery efficiencies.  Given the Canadian emphasis on “we” rather than “me” the benefits of HIE will markedly influence this debate.

I am privileged to be working with the Manitoba eHealth Initiative.  Part of their mission statement sums all this up nicely: "We are all on the same team, we win and lose together… win/lose is not an option."

Viewing Medical Information on a Mobile Device: Why Semantics Matter

The recent hoopla surrounding the release of the Apple iPad has been fascinating. Many of the most reticent PC users are beginning to see how revolutionary this is. Even in the very early days of EMR use everyone clamored for portability. The problem is not with the technology of rendering the EMR information to a Blackberry or even an iPhone, it’s the issue of screen real-estate. To put it simply, the amount of scrolling, clicking or font-adjustment, just makes this impractical. The iPad is on the right track, but since it doesn’t fit in my pocket, it will never be the device I use while on-call. The truth is, my phone is always with me and this is how I will continue to receive calls and pages when I am out of the office. And why shouldn’t I? My wife reads books on her iPhone. My kids spend time on Facebook on their phones. I don’t think it’s too much to ask for a busy physician to have the ability to see vital patient information when he or she is making important care decisions.

Now let’s rethink the conventional wisdom a bit. Suppose my Blackberry (or iPhone) could display relevant information that is specifically rendered for small screen. Take things a bit further.  If I was paged with a request for Coumadin orders, wouldn’t it be cool if I could easily see just the most recent lab results (PT/INR), diagnoses (a.fib, DVT, prosthetic valve), and start/stop dates?  Now that would be game-changer. How about chronic disease specific summaries and dashboards?  You get the idea. 

In my opinion, we don’t wish to replicate the medical chart on a hand-held device. We want to serve up the most useful information in a way that takes advantage of the limited screen real estate. 

This is where knowledgeable vocabulary management, coupled with a sophisticated medical ontology, facilitates the understanding of semantic neighborhoods.

My good colleagues at UPMC have embarked on this journey by uniquely displaying harmonized data from multiple systems on the Blackberry.  (See the upcoming webinar: Delivering a SOA-Based Community Record to Caregivers Everywhere at UPMC with BlackBerry® Smartphones on April 28, 2010).

Many challenges exist—security and patient identity amongst them.  More difficult, I suspect, will be changing old operational models of on-call communication.

HIMSS 2010

I am looking forward to the annual Healthcare Information and Management Systems Society (HIMSS) meeting this year. From my vantage point, interoperability is at the forefront. There are literally thousands of vendors present. The incredible display of varied and new technologies only serves to underscore the ever-deepening silos of medical information.

The HIMSS Interoperability Showcase is one venue to see Information Exchange in progress, but in my opinion, it represents only the fledgling state of what is needed to serve the day-to-day workflow of a busy practitioner.

This year will be an important milestone for healthcare IT, and interoperability in particular. There is so much buzz about “the stimulus” and “meaningful use,” and one feels as if we have finally overcome the coefficient of friction needed to push EMR adoption (finally) into widespread use.

When talking to hard-nosed, daily grind, community physicians, they remain extremely skeptical of receiving a single penny for IT adoption.  More importantly, they worry that physicians and hospitals will adopt IT to “get the checkboxes” for meaningful use, but no great improvements in actual patient care will result.

Much of this has to do with a compelling need to practice connected healthcare—a need to improve collaboration and act as a true care team without walls.  Additionally, a population oriented approach to chronic disease management is becoming an ever increasing priority.  However, to accomplish these noble goals, overworked and often underpaid physicians demand tools to markedly improve efficiencies in work-flow, communication, and “closing the loop.”I know that it seems as if I continually beat the dead horse of semantic integration on this blog, but I will again stress that a complete and clinically relevant presentation of data must be the foundation for managing the increasing complex transitions of care that we are seeing in healthcare today.

So if you are attending HIMSS this year, try not to observe these wonderful, yet disparate technology jewels, shining as individual displays; instead try to see the bigger picture and imagine them dazzling vibrantly together. Interoperability as a technology will play a role, but cooperation, and acceptance for change will ultimately lead to a sum greater than its individual parts.

The Medical Home and Interoperability (Follow-Up)

This post is directed to the comment from my previous dialogue on Medical Home and Interoperability as a deeper explanation to encourage a continuing discussion. 

In response to a real world example, diabetes mellitus serves as an excellent model for the need for semantic integration and population-based decision support.

Simple problems, such as comparing blood work from differing lab systems can be troubling. Comparing “blood sugar” to “serum glucose” to “glucometer” for instance; or “HBA1c” to “glycohemoglobin” to “glycosylated Hb” to “hemoglobin A1c” is in itself a daunting task. Let alone the myriad of other tests that need to be directly and indirectly compared.

What about the hurdles of simply identifying new patients in the diabetic population. Take for instance a patient who has a history of gestational diabetes diagnosed during a remote pregnancy. This information might only be present in the obstetrician’s office records or in the labor and delivery information system. If this patient were to have an eye exam years later, which demonstrated retinopathy, clearly we would like to be able to “connect the dots.” The problem, of course, is that the decision making is obfuscated by information silos and translational difficulties that result from differing medical vocabularies.

More difficulty arises from identifying negations of care. It is easier to use information technology to see if a patient has satisfied a particular quality measure, but providing a list of diabetics who “haven’t had a urine microalbumin level checked in 6 months” is technically challenging on many levels.

I believe that the times cry out desperately for improving diabetic care through the use of this kind of advanced technology.

The now famous UKPDS study showed that for every one point drop in a diabetic’s HbA1c level, there was upwards of a 14% decline in morbidity (i.e., kidney disease, blindness, loss of limb, stroke, etc.). It seems obvious to me that we must direct our most intensive efforts on treating those outliers with poorly-controlled disease, if we are ever to have any hope for a healthier, and thus less economically burdensome, population.

As a primary care physician, I am passionate about this subject, and I know my dedicated colleagues compassionately yearn to help this population. What are needed are the proper tools to accomplish this. I submit that a semantically organized interoperability platform, coupled with advanced population surveillance technology is a critical, yet difficult and poorly understood component.

The Medical Home and Interoperability

 

The patient-centered medical home concept has generated tremendous discussion during this dynamic period of proposed healthcare reform. Whether or not this model will finally take hold remains to be seen. It is clear, however, that a new paradigm in chronic disease management must occur. The combination of an aging population, increase in co-morbidity, and an emphasis on ambulatory management, places a heavy burden on technology to come to the rescue.

For those not familiar with the medical home model, it goes like this: a personal physician oversees a team that takes care of chronically ill patients at the practice level. Care is coordinated across the continuum (acute, subacute, ambulatory, long-term care, home care). Adherence to evidence based guidelines are the foundation for a payment model that may support $40-$51 per member per month, depending on illness severity and the physicians level of medical home designation.

The National Committee for Quality Assurance (NCQA) has emphasized the following:

• Access and communication
• Patient tracking and registry functions
• Care management
• Patient self-management support
• Electronic prescribing
• Test tracking
• Referral tracking
• Performance reporting and improvement
• Advanced electronic communications

Success of this crucial shift in healthcare delivery is incumbent on significant financial incentives.Clearly, the infrastructure and resources needed to accomplish the above tasks can seem formidable to primary care clinicians. The ability to automate much of this work can mean substantial dollars remaining in the hands of physicians. 

Electronic data is only a small piece of the solution. Aggregation of disparate information is another small piece. Semantic normalization, however, is crucial for difficult burden of advanced analytics and disease-specific dashboards.

Most important, in my opinion, is the ability to monitor specific populations in real-time, using advanced and meaningful decision support coupled with appropriate messaging and alerts. Semantic interoperability is thus the cornerstone for the long-term success of the patient-centered medical home model.

Avoiding the Seven Deadly Sins

 

I am anticipating the upcoming Webinar entitled Avoiding the Seven Deadly Sins That Can Derail Your Interoperability Strategy. I like this theme as a metaphor. After all, it seems that "Lust, Gluttony, Greed, Sloth, Wrath, Envy, and Pride” all eventually find their way into HIT projects. It is the recognition and management of these traits of human frailty that ensure successful implementation.

“Sloth” seems to me as the most glaring sin today. How long must we continue to wait for meaningful information exchange while those responsible for health and wellness continue to decide whether or not they can adapt to new technology? As a “deadly sin,” the Institute of Medicine’s now famous report on deaths due to medical error certainly places the slow pace of technology acceptance in the limelight. But perhaps I am a bit harsh. Maybe, like my own display of impatience, the slow pace of change is merely a “venial” sin—meaning that they can be forgiven.

But let’s not talk about the sins of omission.  Institutions and clinicians that are embarking on the initial steps of interoperability must be wary of two of the deadly sins—“Greed” and “Gluttony”

When I talk to physicians about information sharing, the truth is that most want “everyone to share with me,” yet many are unwilling to contribute data back. Much of this is due to paranoia that still exists regarding the potential misuse demographics and referral patterns for unfair, competitive business purposes. 

In addition to “wanting without sharing,” most physicians are unwilling to change their own practice in order maintain an HIE. Workflow seemingly demands that each maintain records in his or her own vernacular, perhaps using a proprietary EMR, often with free text. In other words, complex vocabularies and disparate data formats must adapt to their needs.

Think about the consequences of “Gluttony” when it comes to medical data. Why wouldn’t we “want it all?” Yet like many things in life, overconsumption can hurt you. Differentiating critical data can be like finding the proverbial needle in a haystack, if this overwhelming data collection is not semantically organized.  Because it is now “there,” many clinicians continue to worry about the new medico-legal consequences of being responsible for “all the data.” Finally, many conscientious caregivers feel compelled to review the entire record, and worry that interoperability will drastically cut into their efficiency.

The Webinar will be held on October 20 from 1-2 pm (EDT),  is co-sponsored by dbMotion and Initiate Systems and will feature Dr. William A. Fera VP, Medical Technologies & Medical Director, Interoperability The University of Pittsburgh Medical Center (UPMC) and Lorraine Fernandes, RHIA Vice President, Healthcare Industry Ambassador, Initiate Systems as moderator.

I look forward to Dr. William Fera’s take on this.

For more information and to register for the webinar, visit http://www.dbmotion.com/register%20for%20webinar.aspx

 

“Meaning” Full Use: How Many MS’s do you know?

           

Take Ms Magazine, Master of Science, Microsoft, just to name a few.  Now step into an EMR.

I have seen the abbreviation MS used in the medical record so many ways.  Here are just a few:

    • Morphine Sulfate   
    • Magnesium Sulfate 
    • Ms.  (as a term of address like Mr. and Mrs.)
    • Mississippi 
    • Millisecond
    • Mental Status
    • Mandibular Series (X-ray order)
    • Mitral Stenosis
    • Musculoskeletal
    • Medical Student
    • Multiple Sclerosis
    • Motor Strength

To a computer, the same 2 letters can mean vastly different concepts, including medications, demographics, diagnoses, procedures, and measurements.  Furthermore, the 4000 mg of morphine sulfate used to treat preterm labor, would be a fatal overdose of morphine sulfate in a routine post-op patient (where 10 mg is often a sufficient dose to treat severe pain).

We’ve certainly come a long way in diminishing the use of unsafe abbreviations (thanks in large part to JCHO regulations), but the fact remains that clinicians continue to mix standard and vernacular terms in their daily lives.  Even if we could successfully map disparate terms, what happens when a new instance of “MS” is used—Menkes Syndrome, motile sperm, mass spectrometry?

The point I am raising is not about abbreviations, it is about how we use terminology and symbols to represent reality in vastly different ways.

It is fascinating to realize how the human brain can actively discern truth through the use of context.  By understanding the association of terms or abbreviations to specific metadata, a semantic neighborhood can be constructed that allows machines to begin delineation of context as well.

For instance “MS” in the context of an echocardiogram report clearly represents “mitral stenosis,” and “MS” in a nerve conduction study might represent “milliseconds.”

It is vital that we begin to address the huge challenge of semantics in healthcare. 

There are over 540,000 words in the English language (5 times more than during Shakespeare’s time), and technical information is doubling every two years.  

Meanwhile, the desire to share data is rapidly becoming a mandate in healthcare.  The term “meaningful use” is becoming a buzzword.  Let’s just make sure that in this avalanche of data, that we find “meaning in full use.”