Knowledge graphs for enhancing transparency in health data ecosystems1

BackgroundHealth data is one of the most valuable assets in health service delivery yet one of the most underutilized in especially low-income countries. Health data is postulated to improve health service delivery through availing avenues for optimal patient management, facility management, and public health surveillance and management. Advancements in information technology (IT) will further increase the value of data, but will also call for capacity readiness especially in rural health facilities.We aimed to understand the current knowledge, attitudes and practices of health workers towards health data management and utilization.MethodsWe conducted key informant interviews (KII) for health workers and data staff, and focus group discussions (FGD) for the village health teams (VHTs). We used both purposive and convenience sampling to recruit key informants, and convenience sampling to recruit village health teams. Interviews and discussions were audiotaped and transcribed verbatim. We manually generated the codes and we used thematic analysis to identify the themes. We also developed a reflexivity journal.ResultsWe conducted a total of 6 key informant interviews and 3 focus group discussions of 29 participants. Our analysis identified 7 themes: One theme underscored the health workers’ enthusiasm towards an optimal health data management setting. The rest of the six themes resonated around working remedies to the systemic challenges that grapple health data management and utilization at facilities in rural areas. These include: Building human resource capacity; Equipping the facilities; Improved coordination with partners; Improved data quality assurance; Promotion of a pull supply system and Reducing information relay time.ConclusionOur findings reveal a plethora of systematic challenges that have persistently undercut optimal routine health data management and utilization in rural areas and suggest possible working remedies. Health care workers express enthusiasm towards an optimal health management system but this isn’t matched by their technical capacity, facility readiness, systems and policy willingness. There is an urgent need to build rural lower facilities’ capacity in health data management and utilization which will also lay a foundation for exploitation of information technology in health.

This paper investigates the impact of knowledge extraction methods on the effectiveness of RAG (Retrieval-Augmented Generation) systems that utilize knowledge graphs. It highlights that the quality of the knowledge graph, which is formed using various knowledge extraction methods, is crucial for overcoming limitations of large language models (LLMs), such as “hallucinations”. The paper analyzes the architectures of LightRAG and GraphRAG, emphasizing that the selection of an optimal knowledge extraction strategy depends on specific tasks and the subject area. LLMs have advanced significantly, but they have limitations, including generating factually incorrect information (“hallucinations”) and possessing “outdated knowledge”. RAG systems were proposed to address these issues by combining LLMs with external knowledge bases. This approach reduces hallucinations, ensures factual accuracy, solves the problem of outdated knowledge, and increases transparency. Knowledge graphs are powerful tools for structuring information, consisting of entities (nodes) and relations (edges). They enhance RAG systems by enabling more precise and contextually grounded retrieval compared to keyword-based searches. The quality of a knowledge graph depends on the knowledge extraction methods used, which include named entity recognition (NER), relation extraction (RE), entity linking (EL), and event extraction. Different methods, such as rule-based, classical machine learning, and deep learning approaches, have varying trade-offs in terms of accuracy, completeness, and scalability. Entity linking and knowledge graph completion are also crucial for accuracy and richness. LightRAG and GraphRAG are two main graph-based RAG systems. LightRAG uses the knowledge graph as a quick reference, requiring high precision in knowledge extraction to avoid noise degradation. GraphRAG uses the knowledge graph as a domain model, where completeness of extracted knowledge is more critical, though systematic errors are still harmful. Both systems rely on LLMs for knowledge extraction, which makes them dependent on the LLM’s quality and the size of document fragments processed. The theoretical analysis confirms that the effectiveness of RAG systems is critically dependent on knowledge extraction methods. The quality, completeness, and accuracy of the knowledge graph directly influence the RAG system’s ability to provide relevant, accurate, and truthful answers. Different RAG system architectures like LightRAG and GraphRAG have distinct requirements for knowledge graph characteristics, prioritizing either accuracy or completeness in knowledge extraction.

Abstract: Challenges with health data management in the healthcare industry include security, privacy, and interoperability. Most traditional centralized systems tend to be insecure, inefficient, and deprive the patient of considerable control over their personal health record. HealthLink introduces a decentralized approach to the management of health data employing a Merkle tree-based recovery of these issues. With the application of blockchain technology, HealthLink provides the immutability and security of data while allowing for its integrity. The Merkle tree-based architecture is used for health record verification in a way that is both efficient and secure without exhausting the dataset's content. This system makes it possible to have seamless interoperability among different healthcare providers, thus making health information exchange between them both secure and efficient. By allowing the patients control over their health data, access permissions are provided on demand, thus incorporating enhanced privacy and trust. This would thus place the HealthLink method at the forefront of health data management, providing a robust, secure, and user-centric solution that might be tailored to meet the needs of the healthcare sector as a means of improving patient care and also reducing the efficiency in managing data

Increasingly, across geographies, citizens are requiring access to and control of their health data. This paper examines the "Logging Component" proposed by the European Health Data Space (EHDS) regulation and its crucial role in facilitating secure and transparent access to electronic health records (EHR) and health data. We analysed the proposal for the five elements of the Logging Component (LC): identification of data accessors, identification of data subjects, categorisation of accessed data, temporal logging, and data origin tracking. Explored how these elements contribute towards enhanced accountability and compliance in health data management. We experimented with distributed ledger technology (DLT) to support the "data origin tracking element", reaching the demonstration level which can be presented. We used hybrid DLT to develop a system for immutable storage of access logs, and for using smart contracts to maintain a self-governing decentralised access control list (ACL) directly integrated with EHR and PHR systems. We found that the LC is more than a regulatory requirement. It can serve as a framework for the integration of advanced technologies, e.g. DLT and others, increasingly mature and potentially foundational to building new networks of trust among stakeholders, while ensuring data privacy in cross-border and intra-border healthcare scenarios. The study also identified shortcomings of the LC, such as the absence of "purpose logging", which was conceptualised and proposed. This study contributes to the understanding of how logging mechanisms can enhance transparency and accountability in electronic health record systems within the European healthcare landscape, but with potential usefulness for the "Global EHR". In conclusion, our findings suggest that the successful implementation of the five elements of the Logging Component are mandatory and can benefit from mature advance technologies, but the sixth element proposed by us would be critical for achieving the EHDS's broader objectives of harmonised health data sharing in Europe and beyond while maintaining robust security standards.

The development of information technology offers great opportunities to improve public health services, particularly in managing health data for pregnant women and toddlers. However, at the community level, challenges remain such as manual recordkeeping, limited digital literacy, and the absence of integrated data systems. This community-based digital health data mentoring program aims to enhance the capacity of local cadres and residents to manage health data digitally, accurately, and sustainably. The methods included socialization sessions, training on health data recording applications, and field mentoring over a three-month period in community health posts. The results showed a significant improvement in cadres’ ability to use digital applications (from 45% to 90%) and a 60% reduction in data reporting time. Moreover, the quality of health data collected became more valid and easily accessible to medical staff and local health centers. This program has proven effective in fostering community independence in technology-based health data management and has the potential to be replicated in other regions with cross-sectoral support.

IT providers offering services based on genetic data face serious challenges in managing health data in compliance with the General Data Protection Regulation (GDPR). Based on a literature research and our experiences, an overview of GDPR compliant processing of sensitive data is given. The GDPR requirements for processing sensitive data were specified for a use case concerning a service provider of a pharmacogenomic decision support system. Start-ups who want to enter into the health market also have to comply with the Medical Device Regulation (MDR). The associated efforts for legal compliance constitute an impediment for many start-ups. We created a comprehensive overview, which aligned the requirements of the GDPR with the life-cycle of a medical device. This overview shall help start-ups to grasp and overcome the regulatory hurdles faster.

Extracting and organizing knowledge from the agricultural crop disease research literature are challenging tasks because of the heterogeneous terminologies, complicated symptom descriptions, and unstructured nature of scientific documents. In this study, we developed a multi-stage natural language processing (NLP) pipeline to automate knowledge extraction, organization, and integration from the agricultural research literature into a domain-consistent crop disease knowledge graph. The model combines transformer-based sentence embeddings with variational deep clustering to extract topics, which are further refined via facet-aware relevance scoring for sentence selection to be included in the summary. Lexicon-guided named entity recognition helps in the precise identification and normalization of terms for crops, diseases, symptoms, etc. Relation extraction based on a combination of lexical, semantic, and contextual features leads to the meaningful generation of triplets for the knowledge graph. The experimental results show that the method yielded consistently good results at each stage of the knowledge extraction process. Among the combinations of embedding and deep clustering methods, SciBERT + VaDE achieved the best clustering results. The extraction of representative sentences for disease symptoms, control/treatment, and prevention obtained high F1-scores of around 0.8. The resulting knowledge graph has high node coverage and high relation completeness, as well as high precision and recall in triplet generation. The multi-stage NLP pipeline effectively converts unstructured agricultural research texts into a coherent and semantically rich knowledge graph, providing a basis for further research in crop disease analysis, knowledge retrieval, and data-driven decision support in agricultural informatics.

Neural symbolic reasoning with knowledge graphs: Knowledge extraction, relational reasoning, and inconsistency checking

This paper aims to improve health outcomes research and data management practices. Typically health care records are very large and cumbersome to manage, and the quality of the data is often overlooked because the volume is thought to be large enough to overcome issues arising from missing data. However, simply removing observations with missing data is problematic because the distribution of missing information is non-random, thus the sample used for analysis becomes biased. We propose a method for evaluating and addressing bias in the data cleaning process. Specifically, we identify where bias exists within data and address the bias using sub-sampling or discarding data. We present a case study analyzing data from a level 1 trauma center to establish how bias in health registries exists and how this bias can have downstream implications for evaluating hospital performance. Our method utilizes a two-tailed z-test to compare subgroups in the data set, which demonstrates how missing data in these subgroups can lead to bias. We demonstrate how to localize the bias in particular subgroups and provide corrective actions to handle the bias. We also exhibit how failure to account for bias can distort performance, illustrating the importance of the proposed method.

Blockchain technology is a forefront innovation that drives multi-user data accessibility with patient ownership, traceability, transparency, security, and privacy protection. However, optimizing blockchain latency and its relationship with security remains underexplored in health data management. This study aims to optimize security and latency in healthcare data management through a novel Blockchain-based framework. The objectives include: (1) developing an innovative EMR management architecture that integrates key stakeholders using Blockchain and smart contracts, (2)creating a multi-objective optimization model for Hyperledger Fabric Blockchain in medical health management, and (3)implementing and evaluating advanced optimization techniques using block Size and consensusLevels as key decision variables. This study bridges the gap by proposing a novel methodology that integrates the Hyperledger Fabric framework with smart contracts to optimize security and storage in health data systems. A key contribution of this research is the development of a multi-objective optimization model employing two advanced non-dominated sorting genetic algorithms, NSGA-II and NSGA-III, to configure system parameters such as blockSize and consensusLevels. Evaluation results show that NSGA-II outperforms NSGA-III, achieving 50% higher Hyper-volume and five distinct non-dominated solutions compared to two, demonstrating its superior ability to balance conflicting objectives and provide optimal configurations. The proposed framework tackles critical challenges in health data interoperability, privacy, and security while ensuring robust, patient-centric data management. By enabling multi-user access, enhanced security, and optimized storage, this work advances Blockchain’s role in healthcare and positions the industry for the Digital Health era. Future work will expand objectives and enable real-time deployment.

ObjectiveTransitioning from digital health applications to digital health ecosystems, leveraging the advances in technologies and informatics, could be the next revolution in digital health. This includes offering centralized access to various health services and improving citizens’ well-being, delivery, clinical processes, and data management. However, a limited understanding of citizens may impede adaptation. Therefore, this study investigates citizens’ attitudes within digital health ecosystems, differentiated by their characteristics, to support health service-providers and governmental policymakers in establishing user-centered solutions.MethodsThis study follows a three-step sequential mixed method methodology: (1) a literature review. (2) Qualitative thematic analyses based on semi-structured qualitative interviews. (3) Quantitative analyses based on a web-survey (descriptive statistics, one-way analysis of variances, Tukey-honestly, and Cohen's d tests).ResultsN = 15 citizens were interviewed and n = 1289 responded to the web-survey, to our knowledge the largest survey on this topic. Citizens desire a more convenient management of health services and data (M = 5.2, SD = 1.59). Services with peer-to-peer interactions (M = 3.7, SD = 1.81) and lower involvement of health professionals (M = 3.8, SD = 1.75) are less demanded. Data protection is critical (M = 6.2, SD = 1.23). Public payers are mandated as orchestrators (M = 4.3, SD = 1.99), while private companies receive lower acceptance (M = 3.0, SD = 1.42).ConclusionsHealth service-providers could follow a three-staged approach to establish digital health ecosystems: (1) Increasing the convenience for citizens by enabling online management of health services and data. (2) Extending the citizen–healthcare provider partnership through online interactions. (3) Fostering preventative behaviors and quicker recovery by personalizing health services and interactions. Governmental policymakers should integrate an electronic health record.

Electronic medical records (EMRs) of patients with lung cancer (LC) capture a variety of health factors. Understanding the distribution of these factors will help identify key factors for risk prediction in preventive screening for LC. We aimed to generate an integrated biomedical graph from EMR data and Unified Medical Language System (UMLS) ontology for LC, and to generate an LC health factor distribution from a hospital EMR of approximately 1 million patients. The data were collected from 2 sets of 1397 patients with and those without LC. A patient-centered health factor graph was plotted with 108,000 standardized data, and a graph database was generated to integrate the graphs of patient health factors and the UMLS ontology. With the patient graph, we calculated the connection delta ratio (CDR) for each of the health factors to measure the relative strength of the factor's relationship to LC. The patient graph had 93,000 relations between the 2794 patient nodes and 650 factor nodes. An LC graph with 187 related biomedical concepts and 188 horizontal biomedical relations was plotted and linked to the patient graph. Searching the integrated biomedical graph with any number or category of health factors resulted in graphical representations of relationships between patients and factors, while searches using any patient presented the patient's health factors from the EMR and the LC knowledge graph (KG) from the UMLS in the same graph. Sorting the health factors by CDR in descending order generated a distribution of health factors for LC. The top 70 CDR-ranked factors of disease, symptom, medical history, observation, and laboratory test categories were verified to be concordant with those found in the literature. By collecting standardized data of thousands of patients with and those without LC from the EMR, it was possible to generate a hospital-wide patient-centered health factor graph for graph search and presentation. The patient graph could be integrated with the UMLS KG for LC and thus enable hospitals to bring continuously updated international standard biomedical KGs from the UMLS for clinical use in hospitals. CDR analysis of the graph of patients with LC generated a CDR-sorted distribution of health factors, in which the top CDR-ranked health factors were concordant with the literature. The resulting distribution of LC health factors can be used to help personalize risk evaluation and preventive screening recommendations.

Health information management deals with the collection, storing, analysis, and management of health data. It consists of various fields such as computer science, information science, information management, medicinal, business and data analytics. Machine Learning (ML) in healthcare evolving as an emerging field for healthcare industry. They help in analyzing the health data more effectively and timely. ML is one of the vital regions in the field of software engineering. It gives a streamlined answer for this present reality issues by utilizing past learning or past experience information. There are distinctive kinds of machine learning calculations present in software engineering. This chapter gives the outline of some chosen machine learning methods, for direct relapse, straight discriminant examination, bolster vector machine, innocent Bayes classifier, neural systems, and choice trees. Every one of these techniques is described in detail, which thus helps the reader to create our own answers for the given issues.

Triplétoile: Extraction of knowledge from microblogging text

We would like to welcome you to the 11th ACM International Conference on Knowledge Capture: K-CAP 2021. Driven by the increasing demands for knowledge-based applications and the unprecedented availability of information from heterogenous sources, the study of knowledge capture is of crucial importance. Knowledge capture involves the extraction of useful knowledge from vast and diverse data sources as well as its acquisition directly from human experts. K-CAP 2021 aimed at attracting researchers from diverse areas of Artificial Intelligence, including knowledge acquisition, knowledge representation and reasoning, intelligent user interfaces, problem-solving, planning, agents, information extraction, machine learning, natural language processing, information enrichment and visualization, as well as researchers interested in cyberinfrastructures to foster the publication, retrieval and reuse of data. Nowadays knowledge is derived from an increasingly diverse set of data resources that differ with regard to their domain, format, quality, coverage, specificity, viewpoint, bias, and most importantly, consumers and producers of the data. The heterogeneity, amount and complexity of the data allow us to answer complex questions that could not be answered in isolation - requiring the interaction of different scientific fields and technologies. A goal of K-CAP is to develop such synergies using systematic and rigorous methodologies. The call for papers attracted 123 submissions from all over the world, covering a diverse range of topics spanning ontology development, information extraction, knowledge graphs, natural language processing, reasoning, entity linking, querying and knowledge-based applications. From a competitive set of high-quality submissions, we accepted 31 long research papers, 8 short papers, and 2 demo papers. We encourage everyone to attend the keynote talks that we have planned for K-CAP 2021. The highly anticipated talks by Leila Zia (Wikimedia Foundation) and Ian Horrocks (University of Oxford) will guide us to a better understanding of the future of knowledge graph and knowledge representation technologies, as they become prevalent in real world applications.