In the last decade or so, a number of new names have emerged for the kinds of activities historically studied, disseminated, and taught by those who work in biomedical and health informatics. Each of these areas has emerged as a "hot topic" in biomedicine and healthcare, with resulting academic funding opportunities, new courses or degree programs, and even academic centers or institutes bearing their names.
I suppose I have some skin in this game because I have spent my life's work among those developing the field that I consider to go by the over-arching term of this all, biomedical and health informatics. My concern for the new areas and their names has been their often ignoring the historical and current work of informatics, and with it the decades of research, development, implementation, and evaluation of methods and tools that aim to improve health and healthcare through the use of data and information.
Another ramification of this proliferation of terms is that those not directly working in the field, who may include the leaders of academic and other institutions, may not appreciate the historic role of informatics. Thus I write this post to provide my elaboration of the problem.
One early new term that started this trend about a decade ago was data analytics. As the growth of large amounts of digital data started with the growth of the Internet, the large-scale implementation of electronic health records, and the emergence of wearable and mobile devices, there were new sources of data that could be analyzed for interesting discoveries. From this started the proliferation of academic courses, certificate/degree programs, and centers/institutes devoted to healthcare (and related areas) data analytics.
With the proliferation of machine learning, deep learning, and other methods emerged the new discipline of data science. Again, this was followed by the proliferation of academic courses, certificate/degree programs, and centers/institutes focused on data science. One thoughtful perspective of the relationship between informatics and data science was a paper by Payne et al. in JAMIA Open in 2018, which showed there was not only substantial overlap but also aspects of informatics that go beyond data science. Infomatics includes implementation, evaluation, human-computer interaction, and workflow, to name a few. To use the language of the paper, these were the "efferent" processes of applying knowledge discovered from the "afferent" processes of data science (and informatics) to the real world.
In the meantime, especially with the growth of telehealth during the COVID-19 pandemic, there emerged a new term, digital health. Similar to data science, there was substantial overlap between digital health and informatics, with that overlap focused on many on the efferent processes of Payne et al. However, for many, the work of digital health is really no different than what informatics has historically done and continues to do.
And now there is the new term on the block, which is actually not a new term at all but a re-emergence of an area that was the focus of informatics a half-century ago. This term is of course artificial intelligence (AI). Clearly the advances of modern AI are very real and important. The new AI will likely have profound impacts on biomedicine and health, although we also need to see real-world implementation and evaluation to know what truly works.
My goal here is not to discount any of the important advances that have emerged under the guise of data analytics, data science, digital health, or AI. But rather, to express my concern that plenty of people tout these new disciplines sometimes with ignorance of the historic role that informatics has played in them all. I understand that sometimes funders, institutional leaders, and others want to put their resources into things that are "new." But I also am disappointed when these new areas re-invent the methods and tools of informatics.
The best of informatics imposes a rigor of thinking that these newer disciplines not always apply. Informatics is driven by problems in the domains of health, healthcare, public health, and clinical research. Health and clinical realism is imposed by informatics on the systems as they develop, implemented, and evaluate. I hope that these new areas and what emerges next will not ignore the lessons learned by informatics and incorporate them into their work.
