Extracting Value from Health Care Big Data with Predictive Analytics Gregory Veltri, CIO, Denver Health Mical DeBrow, PhD RN, Siemens Clinical Strategic Consulting DISCLAIMER: The views and opinions expressed in this presentation are those of the author and do not necessarily represent official policy or position of HIMSS.
Conflict of Interest Disclosure Gregory Veltri, BS, FHIMSS Has no real or apparent conflicts of interest to report. 2013 HIMSS
Conflict of Interest Disclosure Mical DeBrow PhD RN Siemens Healthcare Employee 2013 HIMSS
Learning Objectives Explain the value of predictive analytics and the potential benefits that attendees can achieve by using it to gain insights into healthcare big data. Describe recommended best practices that attendees can employ to evolve from a traditional to a predictive analytics environment. Plan your organization s path toward the adoption of predictive analytics by identifying potential starting points for this initiative.
Definitions Big Data is a collection of data sets so large and complex that it becomes difficult to process using standard database management tools. Consider all the pieces of information about each patient throughout their lifetime. Predictive Analytics encompasses a variety of statistical techniques from modeling, machine learning, and data mining that analyze current and historical facts to make predictions about future events.
Analytics Wins ATLAS study proving the value of 10 years of tamoxifen therapy vs current recommendation of 5 years. Results showed at 56% reduction in recurrence of breast cancer. Antibiotic Assistant at LDS Hospital ranking effectiveness and cost of antibiotic therapy. No change in clinical outcomes but, costs dropped from $123 per dose to $52 per dose.
The tough analytics choices ahead 32-year old sedentary patient who smokes and has a BMI of 30 is at risk for multiple chronic illnesses. Patient, over 65, living alone, post CABG is at high risk for readmission. What is our willingness to intervene? No algorithm, that exists today, is going to predict the impending stroke of a 34-year old patient who is a competitive triathlete with no family history of cardiovascular disease. CMS estimates claims data for Medicare and Medicaid to exceed 800 terabytes by 2015.
The data avalanche is here: Will you be buried in it?
Data Management Technology alone cannot resolve management of data. A cultural shift in how data is perceived and managed is required. Data management will change the way healthcare organizations work.
Meaningful Use If you are in the measurement-data governance fray to meet Meaningful Use objectives and get incentive funding, you are missing the big picture what those objectives are driving toward.
Is the Triple Aim for Health Care High Enough? Better quality care for patients. Better care for populations. Lower per capita cost for health care. What comes next?
Medication Administration Check Clinical Documentation Dashboard Single Sign-on Enterprise Master Patient Index Results Repository Analytics / BI Dashboard Patient and Provider Workflow Data Warehouse PACS/Imaging Systems Enterprise Document Management Immunization Tracking CPOE and Clinical Rules
Steps in Data Management Define measures in a data dictionary. Collect data. Measure and analyze the data. Make the data actionable.
Building a Predictive Analytics Model Web Reports External Reports Internal Reports Executive Reporting Portal Design and Implementation Decentralized Reporting / Training Disease Management Reporting Clinical Data Validation and Report Development Clinical Interface Development and Testing Financial Data Validation and Report Development Financial Interface Development and Testing Maintenance, Upgrades and Support Foundation Cubes / Data Structures Development Security and Auditing Tool Implementation Basic Application Structure / Reporting Tool Implementation Network and Hardware Infrastructure
Goal of Measurement The goal of measurement is to turn data into knowledge and knowledge into ACTION not for the sake of collecting or reporting it.
How Do We Know Something How Do We Know What Data We Have We measure. Assess against known standards and benchmarks for retrospective analysis. Looking prospectively requires predictive analytics.
What Managing does it mean? Expectations Administrative data to meet strategic business objectives Customer / Patient / Family data for patient satisfaction Clinical data for clinical quality and safety All data to drive down cost, while improving outcomes
Managing Performance Manage based on real-world, empirically-based, scientific evidence. It is necessary to find a balance between the theoretical and the practical in looking at data. The goal of health care data must be to ultimately predict the behavior of the system and its consumers.
Delivering Success Meaningful and actionable insights come from predictive analytics. Predictive analytics drive competitive advantage.
Predictive Analytics Predictive Analytics will drive efforts to improve clinical quality and financial performance in health care and tie costs to outcomes.
Reliable and Accurate Data Data quality is vitally important to the success of any healthcare organization. Most people never check that data is complete and accurate before sharing it. Most people never seek to understand what the data means. Costs and risks associated with poor data quality.
One Source of Truth There can only be One Source of Truth. Master data repository. Must contain at least those data attributes and values that are required to make the right decisions. No longer do clinical, financial, and operational data operate independently of each other.
Data Knowledge Action Reporting is organizing data into information. Analytics turn data into knowledge and insights. Analytics drive decisions that lead to considered conclusions and recommendations for ACTION.
Mining and Storing the Data You will be required to mine and store the data for a lifetime (not the artificial political limits currently set) AND probably beyond. Consider this, 315 Million US population (US Population Clock, US Census Bureau) x 70 average years of life x every healthcare encounter and result. All storage considerations must anticipate future data storage and retrieval including changes in technology.
What The Future Holds? Data becomes paramount with DNA analysis and genomic surveillance being used to predict best treatment plans, specific drug value, and interventions prior to development of chronic illness. One must consider genetically-engineered drugs for the patient s specific genetic complement.
Challenges The right technology. The right data governance. The right data transparency. The right provider engagement. Are you ready for data as a Disruptive Technology?
Insanity? The problems that exist in the world today cannot be solved by the same level of thinking that created them. Albert Einstein
Questions
Thank You! Contact information Gregory Veltri gregg.veltri@dhha.org Mical DeBrow mical.debrow@siemens.com