Dr. James Clavin

Ph.D. in Information Systems, 2024

James Clavin was advised by Dr. Karuna Joshi from 2020-2024.  His general research interest is in distributed systems and consensus protocols and how they can improve healthcare data security, privacy, compliance, and interoperability through blockchain, Byzantine Fault Tolerance, and Health Informatics.

He worked on the Medical Data Polygraph project

He is currently the Chief Technology and Compliance Officer of Hilltop Institute at UMBC.

Publications are available at Google scholar:
https://scholar.google.com/citations?user=oVprICwAAAAJ&hl=en&oi=ao

Publications

  1. J. Clavin and K. P. Joshi, “Policy Integrated Blockchain to Automate HIPAA Part 2 Compliance,” 2023 IEEE International Conference on Digital Health (ICDH), Chicago, IL, USA, 2023, pp. 307-314, doi: 10.1109/ICDH60066.2023.00052.
  2. Clavin, James and Duan, Sisi and Zhang, Haibin and Janeja, Vandana P. and Joshi, Karuna P. and Yesha, Yelena and Erickson, Lucy C. and Li, Justin D., “Blockchains for Government: Use Cases and Challenges”, ACM Digital Government Journal, vol 1, number 3, 2020
  3. Clavin, James, Mandlem, Satyasai, Joshi, Karuna, “Addressing Improper Payments in Government Healthcare through Blockchain and Generative AI”, under review at Journal: Digital Government: Research and Practice
  4. Mandlem, Satyasai, Clavin, James, Joshi, Karuna, “Measuring the Compliance Costs of Exchanging Part 2 Healthcare Claims Data Through Blockchain”, under review at ACM DLT Journal

James successfully defended his Ph.D. Thesis on 21 November 2024.

Proposal: A Medical Data Polygraph for Examining Healthcare Claims

Committee: Dr. Karuna P Joshi (Chair), Dr. Ian Stockwell, Dr. Patricia Ordoñez, Dr. Jianwu Wang, Dr. Sisi Duan (Tsinghua University), Dr. Jon Hirshon (UMB)
Abstract:
Healthcare organizations exchange claims data across peer-to-peer networks in either standard Electronic Data Interchange formats or ad hoc data formats. We create the ability to use AI and Byzantine Fault Tolerant (BFT) systems to securely explore claims data in combination with other ontologies to identify outliers in the data and measure the cost of out of compliance data exchange.
This work presents the Medical Data Polygraph, which introduces a novel approach to quantifying the expense of exchanging sensitive healthcare claims data and measuring whether the claims follow data use agreements. It enables both real-time interaction with claims data through secure chat functionality and a unique framework for evaluating the financial costs associated with compliance as measured in dollars.
To ensure robust security and fault tolerance, we introduce BFT GraphRAG, a novel integration of Byzantine Fault Tolerance and Graph-based Generative AI Retrieval Augmented Generation (RAG). This framework upholds distributed system security through leveraging two-layer consensus mechanisms and enforcing BFT consensus on RAG-augmented responses from large language models. Operating in both batch and real-time modes, the system enables batch processing to verify adherence to agreed-upon data use policies between entities, focusing on detecting unauthorized data sharing within claims data; calculates and monitors compliance costs for claims data exchange, converting Ethereum-based measures into dollar values; and, provides a secure, real-time chat interface for claims adjudicators to query claims data and receive insights on sensitive health information.

James successfully defended his Ph.D. Proposal on 1 December 2022.

Proposal: Integrating Knowledge Graphs With Byzantine Fault Tolerance for Insuring  Compliance: Medical Data Polygraph
Committee: Dr. Karuna P Joshi (Chair), Dr. Sisi Duan (Tsinghua University), Dr. Patricia Ordoñez, Dr. Ian Stockwell, Dr. Jianwu Wang

Abstract: Healthcare organizations often have to exchange sensitive health records across peer-to-peer networks, and it is challenging to proactively find and fix compliance issues and incorrect data transfers. The Healthcare industry anticipates a growing need to identify faulty medical data when linking individuals’ electronic health records (EHR) across systems and to audit data having been shared outside the scope of data use agreements. We have developed a novel methodology that integrates semantic reasoners with Blockchain technology to track and correct health data exchange automatically. We propose a network layer that identifies errors in health information exchange between organizations that has at its core a two-layer Byzantine Fault Tolerant (BFT) protocol that uses smart contracts and machine learning to identify medical errors. Generally, we consider these Byzantine medical faults. Using instances of existing data use agreements and from the medical literature, we measure the effects of introducing errors into the system. We consider our identification and recommended fixes to such data to be insurance against error by enforcing compliance with regulation and enabling apparent cause analyses before adverse events occur.

Paper presented at the ACM DEBS conference