Bottom-Up Modeling of Permissions to Reuse Residual Clinical Biospecimens and Health Data

Abstract

Consent forms serve as evidence of permissions granted by patients for clinical procedures. As the recognized value of biospecimens and health data increases, many clinical consent forms also seek permission from patients or their legally authorized representative to reuse residual clinical biospecimens and health data for secondary purposes, such as research. Such permissions are also granted by the government, which regulates how residual clinical biospecimens may be reused with or without consent. There is a need for increasingly capable information systems to facilitate discovery, access, and responsible reuse of residual clinical biospecimens and health data in accordance with these permissions. Semantic web technologies, especially ontologies, hold great promise as infrastructure for scalable, semantically interoperable approaches in healthcare and research. While there are many published ontologies for the biomedical domain, there is not yet ontological representation of the permissions relevant for reuse of residual clinical biospecimens and health data. The Informed Consent Ontology (ICO), originally designed for representing consent in research procedures, may already contain core classes necessary for representing clinical consent processes. However, formal evaluation is needed to make this determination and to extend the ontology to cover the new domain. This dissertation focuses on identifying the necessary information required for facilitating responsible reuse of residual clinical biospecimens and health data, and evaluating its representation within ICO. The questions guiding these studies include: 1. What is the necessary information regarding permissions for facilitating responsible reuse of residual clinical biospecimens and health data? 2. How well does the Informed Consent Ontology represent the identified information regarding permissions and obligations for reuse of residual clinical biospecimens and health data? We performed three sequential studies to answer these questions. First, we conducted a scoping review to identify regulations and norms that bear authority or give guidance over reuse of residual clinical biospecimens and health data in the US, the permissions by which reuse of residual clinical biospecimens and health data may occur, and key issues that must be considered when interpreting these regulations and norms. Second, we developed and tested an annotation scheme to identify permissions within clinical consent forms. Lastly, we used these findings as source data for bottom-up modelling and evaluation of ICO for representation of this new domain. We found considerable overlap in classes already in ICO and those necessary for representing permissions to reuse residual clinical biospecimens and health data. However, we also identified more than fifty classes that should be added to or imported into ICO. These efforts provide a foundation for comprehensively representing permissions to reuse residual clinical biospecimens and health data. Such representation fills a critical gap for developing applications which safeguard biospecimen resources and enable querying based on their permissions for use. By modeling information about permissions in an ontology, the heterogeneity of these permissions at a range of levels (e.g., federal regulations, consent forms) can be richly represented using entity-relationship links and embedded rules of inference and inheritance. Furthermore, by developing this content in ICO, missing content will be added to the Open Biological and Biomedical Ontology (OBO) Foundry, enabling use alongside other widely adopted ontologies and providing a valuable resource for biospecimen and information management. These methods may also serve as a model for domain experts to interact with ontology development communities to improve ontologies and address gaps which hinder successful uptake.