A common question in biomedical informatics circles these days is: What role does the traditional, academic biomedical informatics enterprise play in the increasing computerization of healthcare? Current signals are mixed. Some point to the essential quality (and track record) of informatics in building successful health information technology (HIT) applications. Others make it look as if informatics were irrelevant. In this post, I will take a look at the state of current affairs. In a future installment, I will discuss two other important questions: What role should informatics play in the continued development of health information technology (HIT)? And, how should it get to play that role?
Those of us who have been in informatics for a while have been part of an amazing story arc. It is rare to get to witness something close to the birth of a discipline (medical informatics in the 60s and 70s), and then see that discipline grow to exert a larger and larger influence in real life. Yet, this is exactly what happened with the modern discipline of biomedical informatics. From its humble beginnings with a few local computer systems in the medical space to its pervasive influence in healthcare today, biomedical informatics has had a truly astounding trajectory. Morris Collen’s History of Medical Informatics in the United States documents that trajectory very nicely.
At the same time, who works in informatics and how we work has changed significantly. Early on, it was the discipline of a few pioneers, many of them with a medical, some with a computer science or engineering background, and several with a combination. Those pioneers often developed visionary solutions that applied early computer technology to important, but highly specific problems, in a local context. Most or all of them had an association with and/or were funded by the National Library of Medicine (NLM), an institution who played a key role in the development of medical informatics. Many of the early medical innovations live on in healthcare computing applications to this day, whether in concept or actual implementation.
But, something changed along the way. Did traditional informatics become irrelevant? Depending on who you are, you may scoff at this notion. The American Medical Informatics Association, our major professional organization, now boasts a membership of over 4,000 informatics professionals. Informaticians continue to produce important and seminal research findings, many of which get published in high-quality, (mainly) informatics journals. We sometimes make it into the press (but maybe not as often as we like). We help set important policies and standards.
Yet, as an informatician it is hard not to feel somewhat lost in today’s large and complicated health information technology landscape. Most academic software development, once focused on computer-based physician order entry, decision support and similar applications, has either atrophied or ceased. The number of NLM-funded training programs in biomedical informatics, as well as the number of trainees in them, has shrunk dramatically. Most informatics departments/centers continue to compete heavily for a largely shrinking federal pool of funds for grants. The NLM’s budget (~$350m) is a fraction of what many in the informatics community think it should be.
Instead, the HIT industry is now dominated by multi-billion dollar giants, such as Epic, Cerner and Siemens. A torrent of medical computing innovations, such as apps, devices, applications and services is gushing forth from an ocean of entrepreneurs, startups, and small and midsize companies, many of whom have never heard of “biomedical informatics.” Technology leaders, such as Apple and Google, are helping reinvent healthcare.
Is it time for biomedical informatics to let one of its children, health information technology, grow up and move on to other things? It may seem to be the reasonable thing to do. The real-world innovations that basic and foundational informatics research have produced are quite a success, any way you look at it. Many would say that industry has a firm understanding of the innovations and advances that healthcare needs, is able to drive the necessary developments, and has the capital to do so. There might be little to no place for academic informatics research, basic or applied.
However, in my opinion doing so may be the equivalent of walking away from a child when they need you most. I base that assessment on four major points:
- Few in the HIT industry would contend that the development of clinical applications is finished. Many drivers require continual evolution and advancement of clinical applications. Among these are regulation, the continuing rapid expansion of biomedical knowledge, the continually changing healthcare workflow and disruptive innovations.
- Most of HIT’s users are far from happy with the tools at hand. Satisfaction ratings with clinical applications are low across the board. Cumbersome, barely usable and, often, less than useful, these applications have contributed to a significant reduction of the time that clinicians are able to spend with their patients, as well as increased job stress.
- Truly disruptive innovation is unlikely to come from big vendors or institutions. Many large suppliers of HIT suffer from the “installed base” problem: As their customer bases grow, more and more resources must be directed to maintaining the existing base of code and functionality, and the path from idea to implementation gets longer and longer. In addition, the larger your customer base, the higher the degree of resistance to fundamental and disruptive changes in the software. As many historical examples show, disruptive innovations tend to emerge in small places in a bottom-up fashion.
- With the shuttering of most academic software development shops, an important wellspring of innovative ideas in the space has degenerated. The great thing about being an informatician is that we get “paid to dream.” No stockholder is waving a 10-k report in our face to demand an increase in the stock price or next quarterly dividend. We are able to look at problems from more fundamental and theoretical viewpoints. What emerges can become entirely new research and application fields, such as artificial intelligence, which arose from the early work on medical decision support at Stanford and other places.
So, whether it realizes it or not, the world needs biomedical informatics. There are signs that some stakeholders in HIT have caught on to that. But, what do you think? I know that strong arguments can be made on either side of the debate. Let’s hear them!
PS: What could the informatics community do (or do better) to enhance its contributions and significance to HIT? I will examine this question in a subsequent blog posts in a few weeks.
PPS: You may wonder why I included some pictures from the Hudson River Valley in this post. Today, my wife Alida and I are off to a two-week vacation, bicycling in New York’s Hudson River Valley. On the program is a self-supported tour from Bear Mountain to Rhinebeck and beyond.
Titus Schleyer, DMD, PhD
Clem McDonald Professor of Biomedical Informatics
Director, Center for Biomedical Informatics
Regenstrief Institute, Inc., 410 West 10th Street, Suite 2000, Indianapolis, IN 46202-3012
Skype: titus.schleyer, Ph: (317) 274-9204 (direct), cell: (412) 638-3581, E-mail: schleyer (at) regenstrief (dot) org, Web: http://www.regenstrief.org/cbmi/, Blog: https://titusschleyer.wordpress.com, General: http://about.me/titusschleyer