COVID-19 Response Empowered Through Nursing Knowledge Generated Through Existing IT Infrastructure

Citation: Austin, R., Chi, C., Delaney, C., Kirk, L., Michalowski, M., Pruinelli, L., Rajamani, S., & Monsen K. (2021). COVID-19 response empowered through nursing knowledge generated through existing IT infrastructure. Online Journal of Nursing Informatics (OJNI), 25(1)

Abstract

The current COVID-19 pandemic has created a high demand for an informatics infrastructure to support data interoperability and transparency. The use of the existing infrastructure, Unified Medical Language System (UMLS), supports and promotes informatics practice, research and policy. Nursing informatics should leverage the UMLS resource repository to support quality, effective, person-centric health care and respond to this pandemic. In this paper, we describe the use of UMLS by nursing informaticians through three exemplars to showcase the importance of this resource to advance health and health care, knowledge discovery, and formulate rapid responses to COVID-19. The three exemplars include the Nursing Knowledge Big Data Science Initiative, Omaha System Partnership for Knowledge Discovery and Health Care Quality, and Knowledge Representation of Integrative Health. We include recommendations for future work and policy recommendations to advance interoperability and transparency.

Background

The first version of the Unified Medical Language System (UMLS) was published in 1990 from the National Library of Medicine (NLM) (Bodenreider, 2004; National Library of Medicine, 2019). The UMLS provides resources related to terminologies, vocabularies and coding to facilitate system integration of various terminologies and coding systems (Bodenreider, 2004; National Library of Medicine, 2019). The UMLS consists of three components and includes the metathesaurus, semantic work, and lexicon. The metathesarus includes terms and codes from many vocabularies including the Systemized Nomenclature of Medicine Clinical Terms (SNOMED CT) and Logical Observation Identifiers Names and Codes (LOINC)(National Library of Medicine, 2019). Semantic work consists of categorization of concepts and their relationships within the UMLS, and the lexicon includes tools such as Natural Language Processing (NLP) (National Library of Medicine, 2019). The UMLS brings together health and biomedical vocabularies, including standardized nursing terminologies, to enable interoperability among computer systems. The American Nursing Association (ANA) recognizes 12 standardized nursing terminologies (Table 1) (American Nurses Association, 2018)

Table 1 ANA Recognized Nursing Terminologies

UMLS has played an integral part in the alignment of nursing terminologies (e.g., Omaha System), has had a profound impact on the field of nursing informatics, and is foundational to big data science in nursing (Coenen et al., 2001; National Institute of Health, National Library of Medicine, n.d.). The latest version of the UMLS (2019AB) includes the following current versions of the ANA-recognized nursing terminologies: NANDA-I (2018 AA), ICNP (2019AB), CCC (2018AA), NIC (2018AB), NOC (2018AB), PNDS (2018AB), and the Omaha System (2007AC). UMLS resources and information are foundational for information structures within the electronic health record (EHR) and for supporting current research and practice to integrate nursing data into clinical data repositories for common data models and for reuse in big data science (Monsen et al., 2016, 2017; National Library of Medicine, 2019; Topaz & Pruinelli, 2017; Westra et al., 2015).

Objective

The purpose of this paper is to describe the use of the UMLS by nursing informaticians through three exemplars that advance health care, knowledge discovery, and responsiveness to COVID-19. The three exemplars from the University of Minnesota Center for Nursing Informatics are aligned with the UMLS and terminology resources and include: 1) Nursing Knowledge Big Data Science Initiative; 2) Omaha System Partnership for Knowledge Discovery and Health Care Quality, and 3) Knowledge Representation of Integrative Health. Further support of this work was evidenced by a preliminary search that resulted in 82 scientific papers published in PubMed related to UMLS and nursing. While not an exhaustive list, this corpus demonstrated the extensive UMLS-related contributions to the science and implementation of standardized nursing terminologies over the last 30 years.

EXEMPLARS

Exemplar 1. Nursing Knowledge Big Data Science Initiative. Now in its eighth year, this initiative aims to develop and advance an action plan for shaping health policy and informatics knowledge discovery based on sharable, comparable nursing data (Delaney & Westra, 2017; UMN Center for Nursing Informatics, 2020; Westra et al., 2015). Hosted by the University of Minnesota Center for Nursing Informatics (UMN), the initiative provides a structure for leaders from various nursing and health sectors including academia, health systems, government institutions, national associations, and software vendors to advance a national plan to capture nursing information for big data research, make progress toward achieving the Institute for Healthcare Improvement’s Quadruple Aim, and improve health outcomes. As a national working conference and year-round working infrastructure, the initiative continues to develop and guide consistent documentation and data collection practices to support nursing big data research for transforming health care. The initiative encompasses workgroups actively advancing the UMLS vision specifically for nursing. The current workgroups span coordination, context of care, data science and clinical analytics, encoding and modeling, e-repository, knowledge modeling, policy and advocacy, mobile health for nurses, nursing informatics education, nursing value, social behavioral determinants of health, and transforming nursing documentation (UMN Center for Nursing Informatics, 2020).

In response to COVID-19, the 2020 working conference shifted to four virtual sessions held from June to September 2020. The focus of the virtual sessions ranged from expanding data-driven COVID-19 best practice discoveries to leveraging nursing big data science. The sessions brought together national leaders from various domains, such as clinical practice, academia and industry perspectives and included breakout rooms to enable discussion between participants. The breakout room discussions provided the opportunity to create a shared understanding of informatics application technologies used during the pandemic. Hosting the working sessions in this format further facilitated conversations related to the current state of big data, the use of real-time data, and the use of big data to support pandemic efforts where appropriate. Furthermore, several working groups shifted the focus toward pandemic support. For example, the Nursing Value Working Group developed methods for measuring nursing contributions to individual patient outcomes and the cost of delivering care from a single nurse perspective (Moon et al., 2019; J. Welton & Harper, 2016, 2017). This value-based financial model integrated Nursing Management Minimum Data Set (NMMDS), SNOMED CT, LOINC, administrative codes, and nursing terminologies including Nursing Intervention Classification (NIC) and Nursing Outcomes Classification (NOC). This workgroup is also working on incorporating artificial intelligence (AI) for better clinical nursing reasoning, a key measure to demonstrate the value of nursing care.

Similarly, the Knowledge Modeling Workgroup developed information models (IMs) derived from EHR flowsheet data and used these IMs to extract clinically useful information from the EHR to advance data science knowledge generation. As a result, an IM for pain was built. The IM pain model has been nationally validated and is linked to LOINC and SNOMED CT (Westra et al., 2018; Westra, Christie, et al., 2017). This approach provided new directions to understand the context within which pain occurs in the management of opioid use, thus supporting the work of the Data Science and Clinical Analytics Workgroup to educate and expand the use of analytics techniques for nursing knowledge generation (Pruinelli et al., 2018, 2020). The process of developing IMs has enabled the workgroup to use this knowledge and apply it to documentation and data sharing during the pandemic.

The initiative also supports a publicly available resource that provides access to all materials developed and shared throughout the last eight years (UMN Center for Nursing Informatics, 2020). A unique communication strategy leveraging the web and LinkedIn functionalities are available for workgroups. This web-LinkedIn tool is a searchable, topic-driven repository for big data articles, workgroup reports, presentations, and other documents. It is paired with a LinkedIn group to allow discussion of posted information on big data topics with colleagues and peers. Members have shared more than 115 publications that showcase the impact of this initiative, with more publications in development and review. Members of the initiative have testified at the National Academies, the World Health Organization, and other policy-setting conferences and organizations. Initiative members engage with other professionals across nursing specialties and with other interprofessional peers and use the UMLS to advance standardized terminologies and promote interoperability.

Exemplar 2. Omaha System Partnership for Knowledge Discovery and Health Care Quality. This partnership is a practice-based research network within the University of Minnesota School of Nursing Center for Nursing Informatics (UMN Center for Nursing Informatics, 2020; Westra et al., 2015). The UMLS metathesaurus lists the Omaha System, a multi-disciplinary standardized interface terminology, as a source vocabulary in the UMLS. The UMLS metathesaurus provides metadata and detailed statistics regarding the counts and distribution of terms, supporting the Omaha System as a known terminology for health and health care (Martin, 2005). This UMLS recognition, foundational for the uptake and use of the Omaha System software in the 1990s continuing to the present, has had a positive impact toward realizing the anticipated and desired goals of practice-generated data being reused to evaluate the quality of care and discover new knowledge. Further, the UMLS SNOMED CT browser enables public searches for Omaha System terms, which are embedded within SNOMED CT and LOINC.

The Omaha System Partnership consists of scientific teams, community partners, and a collaborative of de-identified clinical Omaha System data (Monsen et al., 2012). Researchers and clinicians use these data to generate new clinical knowledge related to health care quality and the health of diverse populations, and to develop new methodologies for reusing standardized clinical data. Findings from this research have implications for clinical practice, research and health policy. Since its inception in 2010, more than 100 partnership studies involving community partners in practice organizations, student investigators, and researchers at numerous institutions in the United States and internationally were completed or are in progress (Monsen et al., 2012; UMN Center for Nursing Informatics, 2020).  In response to COVID-19, the partnership launched an international effort to deploy evidence-based guidelines for using the Omaha System to support COVID-19 for disease prevention and control activities and documentation at the individual, family and community levels (Monsen, 2020). This work has resulted in an evidence-based guideline that includes 108 interventions for 45 COVID-19 response roles ranging from community members to acute care providers (Monsen, 2020). The existing infrastructure of the UMLS has been pivotal to the adoption of the Omaha System and will be a key factor in sustaining the growth and impact of partnership research.   

Exemplar 3. Knowledge Representation of Integrative Health. As the value of non-pharmacological therapies is increasingly recognized, these therapy modalities need to be represented in standardized terminologies (Clarke et al., 2015; Han et al., 2018; Skelly et al., 2018). During COVID-19, non-pharmacological approaches have been examined to augment current evidence-based guidelines (Alschuler et al., 2020). For example, traditional Chinese medicine (i.e., acupuncture) shows promise as a potential adjunctive therapeutic option to COVID-19 (Chan et al., 2020; Li et al., 2020). As new evidence continues to emerge, it is imperative to document and represent non-pharmacological therapies within the EHR.

Before the pandemic, this interprofessional research team evaluated the representation of non-pharmacologic therapies within the comprehensive international standardized clinical health care terminology, SNOMED CT (SNOMED International, 2019). UMLS was leveraged to examine the non-pharmacological term coverage within SNOMED-CT. This work was performed in response to the opioid epidemic and recommendations from several national organizations on the use of non-pharmacological therapies that found EHRs lacked representation of these therapies (Centers for Disease Control and Prevention, 2017; Cockayne et al., 2005; Joint Commission, 2017, 2018; Porter, 2017; Stan et al., 2018).

A total of 1,209 non-pharmacological therapy terms from multiple selected foundational references and resources in integrative health were extracted (Bauer, 2017; Kreitzer & Koithan, 2019, 2014; Leibowitz & Smith, 2006; Lindquist et al., 2018; National Center for Biotechnology Information National Institute of Health, 2018; National Center for Complementary and Integrative Health (NCCIH) 2019; Wieland et al., 2011). Of the 578 terms validated through consensus, over half (51.9%) of the terms were not found within SNOMED CT. This research was the first step to identify non-pharmacological term coverage within SNOMED CT and, consequently, in the UMLS. The next steps include broad dissemination of these findings with interprofessional organizations and associations (e.g., American Medical Informatics Association (AMIA) and Academic Consortium for Integrative Medicine and Health) to develop consensus for new non-pharmacological terminology content to be included in SNOMED CT. Future research includes analysis of clinical EHR non-pharmacological therapy data to compare with the non-pharmacological therapy terms mapped/non-mapped to SNOMED CT to develop core data elements for use across health care organizations and settings. The team proposes to add a non-pharmacological therapy SNOMED CT subset to existing UMLS resources to ensure consistent use of the existing and new terms. This work has high value given the need for reliable quality evidence-based practices during and after the pandemic.

Discussion

Three UMLS-related exemplars advance the science of nursing informatics to improve health and health care for patients, families, and communities. The University of Minnesota Center for Nursing Informatics leveraged the power of UMLS in the integration of diverse nursing data across settings and stakeholders to support knowledge discovery. As nursing informatics science continues to evolve, the influence and continued support of the UMLS in promoting standardized nursing and interprofessional health care terminologies is essential for advancing national health policy, adoption into EHRs, and use in clinical care. These examples demonstrate the ability to shift the focus to advance health outcomes resulting from the standardization and integration of nursing data, information and knowledge during the pandemic (Bakken, 2020; Monsen, 2020).

We have several recommendations. First, nursing informatics leaders must advocate for the use of standardized terminologies within EHRs through national health policy work by serving as members of task forces on interoperability for quality, consumer data sharing, and national governance structures. For example, in the Nursing Knowledge Big Data Science Initiative and the Omaha System Partnership, nurse informaticians work collaboratively across health organizations, academia, vendors, and national associations on policy recommendations to optimize information systems through the use of data standards that are aligned with the nursing process and support the practical application at the point of care (Delaney & Westra, 2017; Monsen et al., 2012; Westra et al., 2015). Numerous studies conducted by the initiative and partnership demonstrate that it is critical that informatics science and UMLS promote accurate representation of nursing data and work to ensure that the voice of nursing and those that nurses care for – individuals, families and communities – are represented within health data (Gephart et al., 2018; Topaz & Pruinelli, 2017; Westra et al., 2015). Given the importance of representation in health data, a national policy is needed for the adoption and use of standards to occur.

Second, we recommend that health systems and EHR vendors use standardized terminologies in value-based care and population health management solutions (Minnesota Department of Health Office of Information Technology, 2014; Office of the National Coordinator, 2017). The shift toward value-based care underlies the need for structured, interoperable data that can be shared across stakeholders, systems and settings (Centers for Disease Control and Prevention, 2020; Office of the National Coordinator for Health Information Technology, 2020). The UMLS makes sense of this complex environment and promotes interoperability across all systems and platforms, as we described in the non-pharmacological core data elements case study. During this pandemic, nursing informatics experts and researchers are well-positioned to drive this widespread adoption as they understand the value of structure and codification of data in conjunction with nursing information in larger datasets (Bakken, 2020; Chow et al., 2015; Jiang et al., 2017; Topaz & Pruinelli, 2017; Westra, Sylvia, et al., 2017).   

Finally, in alignment with the UMLS's stated purpose to "facilitate the development of computer systems that behave as if they ‘understand’ the meaning of the language of biomedicine and health" (National Library of Medicine, 2019, para. 1), we recommend the reuse of standardized data for evaluation and knowledge discovery. Three examples - Nursing Knowledge Big Data Science Initiative, Omaha System Partnership and SNOMED mapping support consistent documentation of the interventions for nursing and all health care roles for future discovery of what worked during the pandemic. This work begins with defining the data elements as described in the non-pharmacological case study exemplar and is advanced through collaboration described in the initiative case study. The partnership case study highlights knowledge and wisdom to improve practice and inform the science through research. Nursing informaticians are committed to leading this work across the health professions to advance the UMLS agenda.

Conclusion

This paper highlights the influence of UMLS in supporting nursing informatics science, specifically during COVID-19. Over the last 30 years, the UMLS has been a shared resource for health and nursing informatics science. Given the current global pandemic, these three exemplars from the University of Minnesota Center for Nursing Informatics illustrate the use of the existing UMLS and showcase the importance of this resource in continued knowledge generation. To further advance the science, we recommend and advocate for nursing terminology standards through national health policy, use of standards in EHRs, and reuse of standardized data for research.

Funding: This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
Competing interest statement: The authors have no competing interests to declare.
Contributorship statement: All authors on this manuscript meet the International Committee of Medical Journal Editors’ criteria for authorship.  
Acknowledgments: The authors have no acknowledgments to disclose.