Building Trust: Foundations for Measuring Clinical Data Quality

Do you run a data-driven organization? If so, how do you ensure the quality of that data and the decisions, suggestions, and payments generated from them? According to Dr. Brad Ryan, Chief Product Officer at the NCQA, "Lack of trust and transparency in the data and computations that score and drive payments" is one of the largest barriers to the adoption of value-based contracts. FHIR adoption alone does not mean the data are good; data quality most go beyond schema conformance. The Data Quality Index is a novel method for measuring quality and building trust in double-blind interoperable health data in the FHIR format. The Index leans towards a machine-learning based, rather than rules-based, approach to quantify dimensions such as conformance, completeness, plausibility, and ultimately, fitness for use. This session will introduce the concept of the Index, lessons from its development, and seek feedback for future use cases. While statistical frameworks may be discussed, we lean on visualizations and explainable AI principles to make the framework as intuitive as possible, and no technical expertise is required. After all, the goal is for humans to trust the data, with some help from machines.