In February 2007, William Winkenwerder Jr. announced he was stepping down from his post as assistant secretary of defense for health affairs following a press conference in which he downplayed the Walter Reed scandal as a mere "quality-of-life experience." In the months that followed, it seemed clear that Winkenwerder’s negligence may have been partly to blame for the deplorable conditions at the military hospital.
Now, more than a year and half after his departure, Winkenwerder’s legacy lives on in a multibillion-dollar Defense Department electronic medical-records (EMR) system that many military doctors believe is fatally flawed. One military physician, speaking anonymously, calls it "another Walter Reed-type scandal."
And now, as I noted in a piece that Mother Jones magazine posted this morning, it turns out that the Defense Department’s foray into the world of healthcare IT, a system dubbed AHLTA, is going to cost taxpayers somewhere in the realm of $20 billion—four times what the government had originally budgeted.
~~~~~~~~~~~~
Over the past few months I’ve twice posted about AHLTA, the poorly-designed, unreliable, and costly EMR system that the Department of Defense introduced in November 2005. In my first post on the issue in June, I noted that one fundamental problem with AHLTA is that it shouldn’t exist: in contracting with an IT firm called Integic to develop the AHLTA software, the Defense Department has actively ignored the Veterans Administration’s successful VistA system as a promising option for building up the military’s EMR capacity.
Last month, my second post focused on the unhappy military clinicians who are forced to use AHLTA to manage their patient records. Over the summer, the DoD held an online town hall to collect the comments and thoughts of military doctors on AHLTA, and the response was overwhelmingly negative. Participants said that they were “completely disappointed” with AHLTA, and that the system is “a debacle,” too slow and unreliable to be anything besides an “impediment to…seeing patients in an expeditious manner.” The message of the town hall was crystal clear: the Defense Department had spent over $5 billion in taxpayer money to develop an EMR system that its own doctors don’t want to use.
In the piece Mother Jones magazine just posted,
I’ve taken my analysis of the AHLTA fiasco a bit further. Health Beat
readers should be familiar with most of the details I describe in the
article, but I want to call special attention to one important fact
about the AHLTA controversy: it has all the hallmarks of a classic Bush
Administration scandal.
Indeed, all of the key features of a Bush-style muck-up are present and accounted for in the sad story of AHLTA. These include:
(1) A private sector acolyte in a position of governmental power. Winkenwerder,
who oversaw the introduction of AHLTA into the military health system,
was a vice president of insurance giant Blue Cross Blue Shield before
coming to the DoD–and proud of his corporate pedigree. During his
nomination process, he told the Senate that "coming from the private
sector, I am…confident in the ability of private health care
contractors to [provide]…high quality services.”
Winkenwerder’s corporate background stands in stark contrast to the
record of his successor, Samuel Casscells, who took over after
Winkenwerder stepped down: Casscells was a cardiologist at the
University of Texas and a member of the Army Reserve who served a
three-month tour of duty in Iraq. That’s a record of service, not
profit.
(2) Stubborn refusal to admit mistakes. After AHLTA was
rolled out in 2005, military doctors quickly voiced their
disappointment and outrage with the system. But Winkenwerder plowed
forward with implementation, all the while doing what he could to
sugarcoat AHLTA. As I note in Mother Jones,
Winkenwerder “even took the unusual step of hiring the PR firm Edelman
to drum up media attention for the system and brushed off complaints
from clinicians."
(3) Private sector outsourcing gone wild. Recall that despite
having access to VistA code, the DoD chose to develop AHLTA through
private contractors. As usually happens in government projects,
contracts begot contracts: AHLTA’s $5 billion price tag includes $61
billion to Integic, “a $67.7 million follow-up contract for
‘monitoring’ and ‘management’ with Northrop Grumman…and another $12.3
million to Northrop and Booz Allen Hamilton…” amongst many others.
As I pointed out in a previous post, the result of all this
contracting is to grant corporations a lot of clout when it comes to
AHLTA. Earlier this year, NextGov magazine reported that, with
billions at stake, “AHLTA contractors, including Northrop Grumman
Corp., would use their clout to resist any change in the status quo.”
(3) Suppression of dissent. As doctors grew more frustrated
with AHLTA, they began turning to other alternatives to manage their
patient records. One of these was a patient tracking system developed
by Lt. Colonel Mike Fravell in 2003, which was much more efficient to
use—and much cheaper to maintain—than AHLTA. The DoD, fearing a new
competitor to AHLTA,blocked soldiers’ access to the new system and
transferred Fravell, first to South Korea and then to a post just
outside of Washington, DC—"or, as one congressional source put it to Government Executive, ‘bureaucratic Siberia.’”
(4) A ballooning budget. As I reported on Health Beat last
month, Casscells has finally admitted that AHLTA is a bust, and that
it’s time for the DoD to pursue alternatives. The alternative of choice
will be a “converged evolution” of AHLTA and VistA—i.e. the creation of
a universal platform that allows both to work together. The price tag
of this convergence? $15 billion over the next several years, putting
the grand total cost of AHLTA at $20 billion.
All in all, the AHLTA story is a microcosm of the broader themes of
misgovernance that we’ve seen played out over the past eight years–and
that’s troubling: if there’s any sector that you’d think would be
untouched by politics, it’s military health care IT. Everyone agrees
that soldiers should have the most efficient, modernized health care
possible. Even the Bush Administration, which has bungled AHLTA so
badly, thinks so: the need for a greater adoption of EMRs was a key
point of Bush’s 2004 state of the union speech.
Some may suggest that AHLTA failure shows that the task of
developing Healthcare IT should not be left in the hands of government.
But as I’ve pointed out
in the past, VistA, a product of the Veterans Administration, has
greatly improved productivity while reducing medication mix-ups.
What the AHLTA controversy shows is that, even when there’s consensus on what needs to be done, it really does matter how
the government goes about pursuing its goals. The Bush methodology of
governance—commitment to the private sector, intolerance and neglect of
criticism and debate, and a mismanagement of resources—has taken
another victim.
Now the Defense Department finds itself exactly where it was ten
years ago when it first decided to update its EMR system and set itself
on the long, winding course that eventually spawned AHLTA in 2005:
facing a multi-billion dollar medical record modernization project.
The full Mother Jones piece is available here.
Sigh. While it is a laudable goal to expedite the sunsetting of AHLTA, to paste the internet with rhetoric simply makes it more likely that the next cycle of wishful thinking about EHRs will repeat the mistakes of the current cycle.
A critical aspect of VistA’s apparent success has been reasonable expectations and robust operational policies and procedures that help balance its shortfalls. A critical aspect of AHLTA’s failures has been, as you do accurately capture, great effort by senior-most leadership to suppress problems and blocking opportunities to learn from mistakes.
CHCS II’s failures may be a “classic Bush II failure” but not in the way you describe. The failures arose from a DoD apparatus that pre-dated Bush II but was certainly exacerbated by a decision-making structure that could not hear voices from the trenches. CHCS II pre-dated Bush II and was a direct result of an effort to make sure that there would be standardized health data on service personnel deployed to war, before and after, as a “lesson learned” from the Persian Gulf War part I. Furthermore, if AHLTA was the only multi-billion dollar EMR project failure, it would merit even more abuse, but there have been at least two or three others (depending on who you talk with) in the private sector. It is not just the DoD that is trying to figure out how to skin this cat, it is the entire world. Since we in the US have not (quite) yet chosen to nationalize health care, we have so far elected to also not take the Western European expeditious option of a unitary “take it or leave it” system where all equally suffer systems and their failures. Until then, let’s be a bit more circumspect about what is possible and what are reasonable expectations today.
Rhetorical flourishes are useful to help bring attention to a failing program and accelerate its replacement, but don’t put down the AHLTA electric Kool-Aid and simply pick up the VistA pitcher. If we don’t learn from CHCS II/AHLTA, we will be doomed to repeating the same mistakes. Furthermore, if we’re in too much of a hurry to move on to whatever is next, we will also suffer from today’s reality that technical and functional standards for EMRs remain spotty and are rapidly evolving. As one who has served for going on 5 years with the HL7 EHR Technical Committee’s Records Management-Evidentiary Support Profile Workgroup and now in my third year of voluntary service to the Certification Commission for Health Information Technology, I can tell you with absolute assurance that nobody could build a “best EMR” today based on standards because sufficient standards do not yet exist.
Sunset AHLTA indeed, but there are ingenious people who have figured out how to make its defects less damaging, if only the brass would quit suppressing repairs instead of silencing critics and analysts who have provided solid recommendations on mitigating key defects while we await the further evolution of better standards, tools, and infrastructure.
Furthermore, all around our great country are people who have meaningfully addressed and sometimes solved individual pieces of the “best EMR” puzzle. Can we return to an earlier era of commitment to get it right step by step rather than the current get ‘er done fast, regardless of how bad “quick and dirty” screws things up? The current implicit assumptions seems to be that the perfect EMR is achievable today. This in turn enriches electric kool aid salespeople regardless of what vendor they work for.
Reed D. Gelzer, MD, MPH, CHCC
Advocates for Documentation Integrity and Compliance
(former AHLTA functional analyst for SAIC under contract to the U.S. Navy’s Bureau of Medicine and Surgery)
I wish I had the time to respond to the article posted in Mother Jones in its entirety, and perhaps as time avails itself I will come back to this (though probably not, as the article is very obviously slanted and my posts would most likely be unwelcome anyway). It seems very convenient for people to criticize the AHLTA system of systems (“system of systems” is how it was described by Winkenwerder during the “unvailing” that was referenced in this article, not a stand alone system but the CHCS II code and all its feeder systems). Even more so there seems to be a continuing desire to compare this system to that of the VA (ViSta). ViSta has been around longer than the DoD’s legacy CHCS and is based on the same MUMPS code . Point of fact, CHCS was the borrowed base code of ViSta. The primary reason for DoD’s desire to upgrade its EMR system was in part due to that antiquated technology. The availability of MUMPS programmers across industry is almost to a point of extinction. This creates a potential patient safety issue as updates to address security vulnerabilities or new routing protocols (IPv6) cannot be created due to lack of a viable resource pool. If we are to do an accurate side by side analysis of clinical EMR systems we should compare ViSta to CHCS, not to AHLTA.
I do not know who the author is, but the article is a real disservice to those reading it, and reflects a total misunderstanding of computerized patient records and where they are heading. First of all, the military and the VHA are not the same organizations and have different missions. Hence it is no surprise that each organization took a different approach to building its computerized patient record. Second, the author fails to realize that the military’s computerized physician order entry system, CHCS, is based on VISTA. AHLTA expands upon VISTA by providing structured MEDCIN clinical encounter data at the point of care for all beneficiaries. AHLTA allows free text entry, which is particularly useful for the S/O portions of the record. Clinical observational data collected at the point of care is also stored in a single central Clinical Data Repository which employs a 3M Health Data Dictionary which cross references more than 20 types of terminologies. A beneficiary’s record can be called up anywhere in the world that beneficiary seeks care. This is important given the mobile military beneficiary population. In addition, 18 months of lab, pharmacy, and radiology data from CHCS were brought from the 140 host sites and normalized against the AHLTA CDR. The VHA system, VISTA, is highly decentralized, and it is still working on its CDR equivalent, called the HDR. VISTA contains large amounts of free text, which are not really computable, unless one applies sophisticated natural language processing to it to turn it into codes. With that said, VISTA has a good front end that may be more useable than the AHLTA front end. Many clinicians prefer using free text, and do not like to use structured terms, so it is no surprise that the clinicians may not like the AHLTA interface. The tradeoff is between usability and computability. If we want the computerized record to move from simply a collector and documentor of information to a software tool that can assist the clinician with automated clinical practice guidelines and decision support, software developers must provide for structured data. VISTA has a long way to go in this regard. Even Kaiser, which has implemented EPIC, is struggling with the balance between free text and structured data input. Many efforts are underway to re-design the AHTLA interface, and implement innovative user interface tools such as speech recognition and natural language processing to deal with this issue. Perhaps a combination of the best features from each system will be good for both organizations. As for each system’s ability to communicate with each other, great strides have been made in the development of the DoD/VHA Bi-Directional Health Information Exchange (BHIE). DoD is able to send HL-7 CDA messages to the VHA, which is an emerging standard to promote semantic interoperability, but the VHA VISTA system has to dumb-down those message for display in VISTA. In time, I am confident that the VHA VISTA system will establish a more robust HDR with standard data model and terminology services that can accommodate the DOD messages. Also, great strides are being made in the integration of the DOD CDR and the VHA HDR, which will support semantic interoperability. DoD and VHA are also collaborating on common analytical frameworks. The Clinical Data Mart is DoD’s analysis platform for certain extracts of AHLTA Clinical Data Repository data. The VHA has a Corporate Data Warehouse. A recent prototype of a new Clinical Data Mart for the military used the Kimball fact and dimension modeling employed in the VHA’s warehouse model.
It seems to be a negative story for one developing an EMR software. It doesn’t tell or show that all EMRs are like that. I think it just tells us that these can happen if not given importance.
-nj
Furthermore, all around our great country are people who have meaningfully addressed and sometimes solved individual pieces of the “best EMR” puzzle. Can we return to an earlier era of commitment to get it right step by step rather than the current get ‘er done fast, regardless of how bad “quick and dirty” screws things up? The current implicit assumptions seems to be that the perfect EMR is achievable today. This in turn enriches electric kool aid salespeople regardless of what vendor they work for.
A great article written with great hard work…i must say….a great work of your which shows…I like your site its quite informative and i would like to come here again as i get some time from my studies.And I will share it with my friends
New York city just recovered $500 million dollars from SAIC, they proved fraud on the $700 million dollar City Time payroll systems contract. We see now where DoD has spent over $20 BILLION Dollars and over 15 years of development with SAIC and the AHLTA system is not complete and in fact about to be ported into EMR or EHR!!!??? I have intimate knowledge of SAIC and AHLTA…..SAIC is totally unethical and equivalent to ENRON in corporate culture. Not to be trusted….trailblazer, Virtual case file, future combat system, kelly afb cleanup, DOE, MDA, 5 million military paying records lost in San Antonio and much, much more…. they are a train wreck…..