Wednesday, October 28, 2020

What is the Work of Informatics? Integration of Recent Workforce Analyses

In 2006, I published a paper entitled, Who are the Informaticians? - What We Know and Should Know [1]. As my interest in developing educational programs was growing at the time, I was also thinking about the nascent growth of biomedical and health informatics as a profession. Since I had entered the field in the late 1980s, informatics was mostly a research discipline. But with the growth of information technology (IT) adoption in healthcare and other aspects of biomedicine, informatics was becoming a profession that included operational, research, and academic roles and activities.

The last decade (2010-2020) has seen progress in answering my questions from 2006 paper. There has been progress in defining the field on several fronts, although not all efforts have been fruitful. Beginning with the rollout of the HITECH Act, the health IT workforce was viewed as an essential component for success [2]. Professional certification in informatics emerged, first for physicians [3] and soon for others working in the field [4]. The work in preparing for the physician subspecialty led to the development of a core content outline for the field, which would be used among other things for the content of the physician board certification exam [5]. Other efforts led to definitions of competencies for graduate study in biomedical informatics broadly [6] and more focused in master’s-level applied health informatics [7]. One unsuccessful effort was the attempt to become defined in US federal labor statistics through the designation of a Standard Occupational Classification (SOC) code [8]. Health informatics was included in the initial 2018 update of the SOC, but was ultimately left on the proverbial BLS cutting-room floor [9].

A more recent effort has been led by AMIA in conducting two parallel analyses focused on defining the knowledge, skills, and tasks of people who work in informatics [10, 11]. The results have been published in two papers that focus on one narrow and one broader group. The narrow group consisted of physicians in the clinical informatics subspecialty (CIS) [10]. One goal of this effort was to update the core content that was now nearly a decade old and still being the “study guide” for the board certification exam. The larger group consisted of all who work in health informatics (HI), which was defined broadly to include those who work in informatics focused on individual health, healthcare, public health, and research [11].

Interestingly, the workforce analyses conducted by AMIA for the CIS and HI groups led to very similar results. Each found five domains of practice that define the required knowledge, skills, and tasks of informatics practice and research. One interesting yet unsurprising result of the analysis was that four of the domains were relatively similar to those of the original CIS core content [5], while the fifth domain showed the growing importance of issues related to data, including its capture, governance in organizations, and analytics. New uses of the data existed during the time of the original CIS core content but played a small role in the field, such as machine learning and predictive analytics.

I recently took a deep dive into both analyses, with a major aim of identifying the similarities. I describe my findings here (and take any blame for any misrepresentation of this impressive body of work). Both analyses describe the first domain of fundamental knowledge and skills, which include a common vocabulary, basic knowledge across all informatics domains, and understanding of the environment(s) in which the workforce functions. Depending on where an individual works, this may include consumer health, health care, public health, or research settings.

The second domain differs somewhat between the HI and CIS analyses but can be integrated into an overall focus on enhancing health decision-making and improving health care delivery and outcomes. Informatics practice should support and enhance decision-making by clinicians, patients, policy makers, researchers, and public health professionals. It must also analyze existing health processes and identify ways that health data and health information systems (HIS) can enable improved outcomes. Informatics work should also evaluate the impact of HIS on professional practice as well as pursue discovery and innovation. More clinically, informatics practice should be able to develop, implement, evaluate, monitor, and maintain clinical decision support while also supporting innovation in the health system through informatics tools and processes.

The third domain of each analysis can be combined into an overall category of health and enterprise information systems. Informatics practice should include planning, developing or acquiring, implementing, maintaining, and evaluating HIS that are integrated with existing information technology systems across the continuum of care. This should include the clinical, consumer, and public health domains and address security, privacy, and safety considerations. This domain should also include the development, curation, and maintenance of institutional knowledge repositories, also while addressing security, privacy, and safety considerations.

A critical domain is the new addition to the previous four domains of the CIS core content, which can be integrated as data governance, management, and analytics. Practice should include establishing and maintaining data governance structures, policies, and processes. The workforce should be able to acquire and manage health-related data to ensure their quality and meaning across settings and to utilize them for analysis that supports individual and population health while driving innovation. It is also critical to incorporate information from emerging data sources, ensure data quality and meaning across settings, and derive insights to optimize clinical and business decision-making. Although not explicitly mentioned in the overall descriptions of this domain (but covered in the details of practice) are the ability to identify and minimize biases in data and mitigate their impact as well as to implement and evaluate machine learning and artificial intelligence applications in all health-related settings.

The final domain reflects the organizational and management aspects of informatics, with required abilities in leadership, professionalism, and transformation. Informatics practice should be able to build support and create alignment for informatics best practices as well as lead informatics initiatives and innovation through collaboration and stakeholder engagement across organizations and systems.

Although it is valuable to have the requirements for the workforce well-elucidated, the results of the new analyses are hardly surprising. We have known for many years that biomedical and informatics is a sociotechnical discipline, i.e., influenced by the interaction between social aspects and use of technology. We also know from prior explorations of competencies for master’s-level education that foundational knowledge and skills are required from health sciences, social sciences, and information sciences [7]. This new work demonstrates more clearly the work of informatics, and future work will hopefully quantify the different types of professionals and their require knowledge and skills.


1. Hersh, W., 2006. Who are the informaticians? What we know and should know. J Am Med Inform Assoc 13, 166–170.

2. Hersh, W., 2010. The health information technology workforce: estimations of demands and a framework for requirements. Appl Clin Inform 1, 197–212.

3. Detmer, D.E., Shortliffe, E.H., 2014. Clinical Informatics: Prospects for a New Medical Subspecialty. JAMA 311, 2067–2068.

4. Gadd, C.S., Williamson, J.J., Steen, E.B., Fridsma, D.B., 2016. Creating advanced health informatics certification. J Am Med Inform Assoc 23, 848–850.

5. Gardner, R.M., Overhage, J.M., Steen, E.B., Munger, B.S., Holmes, J.H., Williamson, J.J., Detmer, D.E., AMIA Board of Directors, 2009. Core content for the subspecialty of clinical informatics. J Am Med Inform Assoc 16, 153–157.

6. Kulikowski, C.A., Shortliffe, E.H., Currie, L.M., Elkin, P.L., Hunter, L.E., Johnson, T.R., Kalet, I.J., Lenert, L.A., Musen, M.A., Ozbolt, J.G., Smith, J.W., Tarczy-Hornoch, P.Z., Williamson, J.J., 2012. AMIA Board white paper: definition of biomedical informatics and specification of core competencies for graduate education in the discipline. J Am Med Inform Assoc 19, 931–938.

7. Valenta, A.L., Berner, E.S., Boren, S.A., Deckard, G.J., Eldredge, C., Fridsma, D.B., Gadd, C., Gong, Y., Johnson, T., Jones, J., Manos, E.L., Phillips, K.T., Roderer, N.K., Rosendale, D., Turner, A.M., Tusch, G., Williamson, J.J., Johnson, S.B., 2018. AMIA Board White Paper: AMIA 2017 core competencies for applied health informatics education at the master’s degree level. J Am Med Inform Assoc 25, 1657–1668.

8. Request/Recommendation for New Health Informatics Practitioner Standard Occupational Classification (SOC), 2016.

9. AMIA Washington Download: 12.18.17 Government Equates Informatics with Registrars New Occupation Codes, 2017.

10. Silverman, H.D., Steen, E.B., Carpenito, J.N., Ondrula, C.J., Williamson, J.J., Fridsma, D.B., 2019. Domains, tasks, and knowledge for clinical informatics subspecialty practice: results of a practice analysis. J Am Med Inform Assoc 26, 586–593.

11. Gadd, C.S., Steen, E.B., Caro, C.M., Greenberg, S., Williamson, J.J., Fridsma, D.B., 2020. Domains, tasks, and knowledge for health informatics practice: results of a practice analysis. J Am Med Inform Assoc 27, 845–852.

Thursday, October 1, 2020

Welcoming the 4th Edition of Information Retrieval: A Biomedical & Health Perspective

I am pleased to announce the publication of the 4th edition of my book, Information Retrieval: A Biomedical & Health Perspective. Published by Springer, the book updates the content, methods, results, and research in the use of search systems for knowledge-based biomedical and health information.

I am gratified to be active in a number of areas of research in biomedical and health informatics, but my original and still most active area is information retrieval (IR), also sometimes called search. The appeal of getting information from a computer by entering a query or question held appeal to me from early times, including when I was dabbling with computers in medical school and residency in the mid-1980s. Upon entering formal training in the field in my postdoctoral fellowship in 1987, this appeal persisted, even as the thrust of research in the field was still focused on the first era of artificial intelligence.

My introduction to the field came through a monograph by Prof. Bruce Croft, which then led me to discover the work of Prof. Gerard Salton. I had the opportunity to meet Salton when he visited Harvard University during my postdoctoral informatics fellowship there. Salton literally invented the IR field and it is unfortunate that he passed away in 1995 before he could see the true impact of his work on IR systems in the world. The approach of Salton and his legions of graduate students he trained in “automated” IR was quite different than the main biomedical focus in the 1980s and 1990s, which was the set-based Boolean retrieval approaches used to search MEDLINE. My earliest work attempted to marry the automated approaches to the controlled vocabularies being developed and collated in the National Library of Medicine (NLM) Unified Medical Language System (UMLS) Metathesaurus.

Another early interest of mine in IR concerned evaluation of systems and users. A perspective of evaluation has guided a great deal of my informatics research, based on the premise that what we do, whether building systems or advocating their use by people, should be studied for its value to human health and healthcare. My foray into IR research led me to recognize the importance of the relevance-based metrics of recall, precision, and their aggregated combinations, but I also felt dissatisfied that they did not evaluate the entire IR experience, especially for users. I was fortunate to be able to attend the very first Text Retrieval Conference (TREC), and then become involved in organizing a number of its tracks in subsequent years.

I never would have imagined in my early days that we would be able to carry around the Internet - and access to the world’s knowledge - in our pockets via mobile devices. I could not fathom that essentially all scientific publishing would become electronic, and that it would include not only articles but the underlying data. I also would never have imagined that searching would be so ubiquitous by all Internet users, or that the name of a search engine would become a verb (Googling).

The world of IR has certainly changed. The basic task of “ad hoc” searching is pretty well a solved problem. There are still, however, challenging problems in IR to solve, such as some of those on which I currently work.

Both the eBook and hardcover editions of the new edition are now available through Springer and Amazon. If you or your institution have access to SpringerLink, the eBook version can be accessed there.