For those of us who've been toiling in the trenches of health information exchange for a number of years, we're finally living the dream. According to a 2011 KLAS report and a more recent Chilmark report, the HIE market is poised for spectacular growth over the next couple of years. Most of this growth will be driven more by "private" HIE efforts (enterprise efforts usually driven by a hospital system and/or physician organization) than by "public" ones (cross-organization regional or state collaborations usually seeded with government funds), but, regardless of what is driving it, the reality is that HIE is sprouting all around us.
I'm delighted that we're moving rapidly in this direction, but one concern keeps nagging away at the back of my mind, and that is the propensity to pursue over-architected HIE solutions.
This history goes back to the ill-fated community health information networks (CHINs) of the 1990s, continued through the highly-publicized failure of the Santa Barbara Care Data Exchange, the difficulties experienced by the Massachusetts eHealth Collaborative pilot projects and many of the regional health information organizations established by New York's HEAL-NY program. And it continues into the present-day with the demise within the last year of CareSpark (Tennessee) and the Minnesota Health Information Exchange, and the recent challenges experienced by Cal eConnect. Many of the HITECH-funded HIE programs carry this same risk.
What is an over-architected HIE? Put simply, it's one that tries to do too much for too many with not enough money and time. It tries to establish an all-encompassing infrastructure and service to meet multiple, heterogeneous current and future requirements of multiple, heterogeneous current and future customers. It tries to do all of this with a shoestring budget and staff. And worst of all, it focuses more on long-term potential "big-bang" value at the expense of short-term, realizable, incremental value. Or as one HIE organization's promotional material put it, the value proposition is to be a "one-stop shop for Clinical and Administrative Information."
The counter to the over-architected HIE is the incremental or phased HIE, which focuses specifically and radically on concrete, discrete, value-generating and self-standing steps and does not tie its fortunes to a specific future end-state whose horizon is further than the range of our ability to navigate. I was recently describing my concern to a health care system executive, and he said, "Yes, well, but we just want to jump to the end." By that he meant, build the final solution infrastructure and services right away to solve the big problem of creating a "one-stop shop," and assume that by-products of that long-term effort will keep everyone motivated along the way. My concern reached new heights after that conversation.
It's totally understandable how this happens and, interestingly, both "public" and "private" initiatives are led down this same path, albeit for different reasons. For many public HIE efforts, "waste" in health care spending feels like low-hanging fruit. Don Berwick says that 30% of health care spending is "waste", and the CEO of Geisinger recently stated that 40% of health care spending is "crap".
If "overuse, underuse, and misuse" of resources driven by fragmentation of care is the problem, the thinking goes, then surely the answer is simple. Create a single infrastructure that all participating organizations can share. Bring the data together (either logically or physically) into a single data schema with a common vocabulary. Then repurpose the information for lots of beneficial things, like giving access to disparate providers to improve continuity of care, reporting to public agencies, searching for patient records across hospitals and doctors' offices, decision support for case management, population health, disease management, claims adjudication, fraud detection, medication adherence, personal health records, patient-recorded outcome measures, pharmacosurveillance, clinical trials recruitment, predictive modeling, a learning health care system ... the potential list goes on and on.
On the private side, where the real HIE action is increasingly happening, accountable care is the main driver, but it leads down the same perilous path to over-architected solutions. Rather than chasing waste per se, private HIEs are focused on aggregating information for risk management, which requires comprehensive data aggregation for the application of analytic tools to assess and mitigate risk. With clearer focus, more resources, seasoned IT expertise, and more nimble decision-making, one would expect these private HIEs to be much better able to rapidly build complex HIE capabilities than their public counterparts.
Yet, I'm concerned that too many efforts, both public and private, resemble the Pequod's search for the elusive White Whale, and we all know how that turned out. (Spoiler alert: Captain Ahab dies.)
Why is that? There are three main concerns.
First, there is a more or less rate-limited logical progression in HIE development that no reasonable amount of money or individual talent can leapfrog. Moving from pushing basic clinical documents back and forth, to searching and retrieving clinical information, to aggregating and normalizing disparate clinical, administrative and financial data, each of these is a discrete threshold that would be easy to traverse if it could be solved by whiz-bang software alone. In reality, though, the pace of progress is determined by a complex interplay of legal, business, governance, organizational, cultural, clinical, and yes, technical factors that no software vendor is going to be able to resolve.
Second, though progression of HIE development is rate-limited, technology and business move very rapidly and know no such barriers. By trying to architect and execute a solution now for an end-state that is three to five years from now, there is a very high risk of making a solution that is brittle to the unpredictable but inevitable changes that will take place. Consider that the term "accountable care organization" was first mentioned in policy circles in December 2006. Consider further that the standard now known as NwHIN Direct was not even conceived until 2010. Consider, finally, that the first iPad was released just over 2 years ago. You see what I mean.
A third and final concern is that, in case you haven't noticed, health care CIOs are under just a little bit of stress these days. Even the best and the brightest are feeling the pressure. A recent survey of North American CIOs found that the top three concerns that keep them up at night are (in order):
- Speed of execution to meet business goals;
- Inadequate budgets; and
- Lack of requisite staff capacity and skill sets.
Over-architected HIE approaches hit a CIO where it hurts most. Jumping to the end doesn't speed execution, it delays gratification. Better to proceed with achievable steps that deliver incremental value along the way.
The pushback I always get to this perspective is that the first incremental steps -- pushing clinical documents whether by NwHIN Direct or other more widely used industry standards -- are of low value and therefore not worth the effort. Well, the value may be lower, but so, too, is the cost, and both are easier to estimate for directed exchange than for more speculative longer-term steps.
In addition, one should never underestimate the ingenuity of human beings -- systems only get better through use. The statewide HIE efforts in New Hampshire and Massachusetts have come up with well over 50 use cases that can be enabled simply through push technologies, and I am confident that many many more will arise once clinicians begin using the system.
Finally, intentionally focusing on incrementalism allows one to design the first step to be a high confidence one, rather than allowing confidence to be a variable. Under-promising and over-delivering is as true in HIE as it is in all other walks of life.