Moving Towards FAIR Data Practices in Pharmacy Education

Recommendations

We must conduct FAIR needs assessments. Assessing data practices within pharmacy education will elucidate current practices, needs, infrastructure, and barriers, and enable educators to identify gaps between current and FAIR data practices. This would provide a comprehensive understanding of the current state of data practices, including the requirements and infrastructures provided by journals, funding agencies, data repositories, and other institutions; the met and unmet data needs of key stakeholders of FAIR data practices; and the real and perceived barriers to FAIR data practices for various stakeholder groups. As a starting point, this type of needs assessment could be conducted for an individual research team/laboratory, an entire pharmacy school, or across a specific pharmacy education sub-discipline (eg, OSCE data). Ultimately, this deep understanding of the current state of data practices and related gaps will provide the foundation for an evidence-informed roadmap to FAIR data practices for the scholarly community and help craft specific recommendations for educational and research endeavors tailored to individuals and institutions.

We must train educators and researchers in FAIR data principles. Educators and researchers must be equipped with the knowledge and skills needed to teach and apply FAIR data principles. As noted by Mons and colleagues, “FAIR compliant data stewardship will require many different skills that are not traditionally covered by the research curricula of contemporary students and researchers. Therefore, extensive training capacity and training materials are needed, and in need of development.”⁵ Furthermore, FAIR practices rely heavily on humans to format and communicate data, tools, and findings in a way that enables understanding and mitigates ambiguity associated with natural language barriers (eg, jargon). Although some researchers believe access to shared data will advance research, training will need to address strategies for overcoming identified FAIR barriers, such as: unfamiliarity with data sharing; strategies for handling the effort, time, and cost of preparing data to share; uncertainty regarding benefits to oneself or others; and ethical and legal issues associated with FAIR practices.^9,13

We must build FAIR communities of practice in pharmacy. Since the FAIR principles were first described in 2016,⁴ an estimated 1,000 communities of practice have implemented them across various disciplines.⁵ These communities have explored ontology mapping, machine learning, automation, and annotation and curation, among other FAIR issues.¹¹ Within pharmacy education, several strategies could be used to identify and connect with those interested in building communities aimed at advancing these efforts. Professional organizations, such as the American Association of Colleges of Pharmacy (AACP) and Research Data Alliance (RDA), often include special interest groups (SIGs) with shared interests related to assessment, data, and use of FAIR principles (eg, AACP Assessment SIG, RDA Raising FAIRness in Health Data and Health Research Performing Organisations Group, RDA Health Data Interest Group). As an increasing number of funding agencies adopt and advocate for improved data practices, such as the National Science Foundation’s Harnessing the Data Revolution,¹⁴ individuals may also find potential collaborators among lists of funded projects.

Pharmacy educators might also find individuals and/or committees within their departments, schools, or universities focused on relevant data issues, such as data governance, data science, and information science. Because student pharmacists are taught by experts from various disciplines, such as biomedical science, education, library science, and healthcare practice, it is natural and necessary to create interprofessional FAIR communities that embody relevant expertise. Together, pharmacy educators and researchers must leverage this diversity of expertise to identify and engage collaborators who can identify and implement FAIR strategies.

We must integrate FAIR practices into educational practice and research. Studies of first-generation implementation of the FAIR data principles revealed disciplinary differences, including legal and workflow issues.¹⁵ The use of FAIR data principles in education could include navigating legal restrictions associated with accessing and sharing data (eg, Family Educational Rights and Privacy Act), adhering to ethical principles (eg, deidentifying data), and consistently operationalizing commonly used constructs. For example, a single construct, such as wellness, may be measured differently by different institutions, challenging the interoperable nature of the data.

The strategies and related effort required to implement FAIR principles can vary widely depending on the goals of the individual or institution, and the complexity of the data. Some strategies may be achieved with minimal additional effort, such as making a full survey instrument and related de-identified data available. Others might require more effort, such as publishing qualitative transcripts, given the unique processing required to ensure confidentiality of participants.¹⁶ Even more complex would be data subject to FERPA restrictions, in which case data instruments may be easily shareable yet the data itself could not be shared. As noted by BMC Medical Education, “it is not always possible to share research data publicly, for instance when individual privacy could be compromised.”¹²

As a discipline, we must elucidate how FAIR principles intersect with education and how to best position pharmacy educators to integrate them into their research and practice. A survey of Society for Directors of Research in Medicine Education members, for example, indicated that data sharing was strongly supported, yet few reported actually using data repositories.⁹ Clearly, the challenges facing implementation of the FAIR principles in education are not trivial and will likely require sustained effort and strategy across a wide range of expertise.

We must leverage machine learning in pharmacy. Data-intensive research and decision-making rely on the ability of humans to discover, access, integrate, and analyze task-appropriate scientific data.⁴ As the amount and diversity of data related to pharmacy education (such as aforementioned data on student selection, performance, and job placement) grows and becomes available in FAIR-compliant datasets, their manual, “eyeballing” analysis of such data becomes increasingly unattainable. Thus, computational tools such as machine learning are growing in use for education data analysis and trend elucidation.¹⁷

Machine learning relates to a large group of computational and statistical algorithms and respective software tools that can explore the relationships and correlations between various characteristics of digital objects. For instance, machine learning models can help predict student performance in pharmacy curricula based on student admission data.¹⁸ Data that are compliant with FAIR principles are a “lame substrate” without workflows that empower the formatting, communication, and consumption of that data.⁵ Thus, we should continue to leverage machine learning tools to make the most efficient use of education data to improve student selection, advance their education, and help them find the most satisfying employment. Achieving these goals is, arguably, a common objective of all educational institutions, and the use of bigger datasets typically improves the accuracy and reliability of machine learning models. Thus, promotion of FAIR principles of data use in pharmacy education will enable data standardization and integration across multiple schools and, consequently, growth of respective educational databases. We surmise that these advances will catalyze data-driven decision support and improve the efficiency of pharmacy education.