Data Quality: Practice, Technologies and Implications Kathy Grise, Senior Program Director, IEEE Future Directions DaSH 2017
Data Information is only as good as the quality of data that supports it. Data quality is only as good as the observations, documentation and standardized coding of the details that report the key parameters of the patient condition. A focus on data quality is essential to assuring that we have the information needed to improve healthcare value. Joe Nichols, MD, Principal, Health Data Consulting, Inc. DaSH 2017 19 October 2017
Data Sources Within 15 Years or less, Health Data Goes Up to 1 Yottabyte! Just where is all this data coming from? Lower cost sensors More sensors Lower cost equipment Lower cost data storage Easier access to technology High Performance Computing IoT Cloud Everything in between DaSH 2017 19 October 2017
Data Sources Personalized from the every day consumer Social networks DaSH 2017 19 October 2017
Data Implications How to ensure data quality? Must first address questions of, Where is the data source? Do you trust the data? How old is the data? Are there multiple instances of same data? Which instance is more reliable? Is the data secure? Will the data be here one day and gone tomorrow? Traceability Reproducibility All valid questions to consider, which may directly impact one’s ability for informed healthcare decision making! DaSH 2017 19 October 2017
Examples of Data Implications Consumer generated data and social media What safeguards are in place to guarantee data integrity and accuracy? DaSH 2017 19 October 2017
Examples of Data Implications Healthcare generated data by clinician EHRs stored on cloud Data is monetized What safeguards are in place to guarantee data integrity and protection from hackers? How will you even know data has been hacked? Healthcare records for sale on Dark Web: Clinic in Baltimore had its records stolen which went on sale Dark Web for less than one cent per record. DaSH 2017 19 October 2017
What is IEEE doing? Establishing consistency in data formats via standards Mobile Health Data Data Privacy Process Algorithmic Bias Considerations Big Data Metadata Governance and Management DaSH 2017 19 October 2017
Who should participate: Wearable device makers Medical device makers Goal: Define specifications for a mobile health data application programming interface (API) and standardized representations for mobile health data and metadata. Mobile health data encompasses personal health data collected from sensors and mobile applications. No mobile health data or metadata standards currently exist. Who should participate: Wearable device makers Medical device makers Health data aggregators Health information technology systems managers Health information infrastructure providers Mobile health app developers Biomedical researchers Clinicians Data scientists Government Why get involved: Mobile apps and sensors are an increasingly valuable source of health-related data. Standardization of mobile health data and metadata will make data aggregation across multiple mobile health sources easier and more accurate, and will reduce the costs of using mobile health data to make biomedical discoveries and to improve health and manage disease. DaSH 2017 19 October 2017
Goal of these standards Define medical-specific 3D printed and virtual reality model applications for clinical diagnosis and practice based on real human information, and approach new frameworks for 3D medical application established on visualization, data management, simulation and 3D printing. Who should participate: Manufacturers of 3D devices including 3D display and HMD systems Medical 3D model developers Software programmers for 3D volume imaging Medical device developers and manufacturers Healthcare data managers Medical researchers Technical experts Clinical practitioners Additive manufacturing companies Other parties that have a material interest in developing these standards IEEE P3333.2 IEEE P3333.2.2™, Standard for Three-Dimensional (3D) Medical Visualization IEEE P3333.2.3™, Standard for Three-Dimensional (3D) Medical Data Management IEEE P3333.2.4™, Standard for Three-Dimensional (3D) Medical Simulation IEEE P3333.2.5™, Standard for Bio-CAD File Format for Medical Three-Dimensional (3D) Printing DaSH 2017 19 October 2017
White paper(s) that frame the problems and identify issues Governance and metadata management poses unique challenges with regard to the Big Data paradigm shift. The governance lifecycle needs to be sustainable from creation, maintenance, depreciation, archiving, and deletion due to volume, velocity, and variety of big data changes and can be accumulated whether the data is at rest, in motion, or in transactions. Furthermore, metadata management must also consider the issues of security and privacy at the individual, organizational, and national levels. From the new global Internet, Big Data economy opportunity in Internet of Things, Smart Cities, and other emerging technical and market trends, it is critical to have a standard reference architecture for Big Data Governance and Metadata Management that is scalable and can enable the Findability, Accessibility, Interoperability, and Reusability between heterogeneous datasets from various domains without worrying about data source and structure. Desired Outcomes: Workshops to collect, analyze, and identify relevant use cases, requirements, and potential solutions, including documentation on the findings White paper(s) that frame the problems and identify issues Reference architecture(s) concepts and solutions from relevant best practices in big data governance and metadata management to formulate data interoperable infrastructure to enable data integration/mashup between diversified domain repositories, including those maintained by participating entities and IEEE Dataport A proof-of-concept reference implementation would be welcomed Identification and initiation of potential new standards DaSH 2017 19 October 2017
What is IEEE doing? Societal implications with data and technology Ethics Health and safety implications of technology Professional responsibility DaSH 2017 19 October 2017
bigdata.ieee.org web portal Contact me: k.l.grise@ieee.org bigdata.ieee.org web portal @ieeebigdata DaSH 2017 19 October 2017