HXR 2016: Data Insights: Mining, Modeling, and Visualizations- Jennifer Gamble

Editor's Notes

• #3: The Department of Health and Human Services says 50 percent of all Medicare payments will be tied to quality or value through alternative payment models by 2018. Around 90 percent of all remaining Medicare fee-for-service payments are scheduled to be tied to value or quality by then as well. http://www.beckersasc.com/asc-turnarounds-ideas-to-improve-performance/5-things-for-surgeons-to-know-about-episodes-of-care.html
• #5: What’s wrong with this picture: 1) Hypothesis driven: Susceptible to physician bias and limited by unrepresentative guidelines. 2) Not Comprehensive: Limited awareness of prevailing practices – and often blind to good variation, what a doctor knows, not what ALL doctors know 3) Manual: Difficult to build and maintain. Exceptionally resource and time consuming 4) Limited Adoption: Because of the subjective nature of the process adoption and adherence are hampered, they are often out of date by the time they are published
• #6: So what’s different about the Ayasdi approach to understanding and managing clinical variation. Our software uses an advanced analytical technique called machine intelligence to automate the discovery of clinical pathways – it rapidly groups similar patient procedures to help surface clinical pathways It uses your hospitals data and so it reflects your physicians expertise and experiences It uses quantitative methods to help you understand the sources and impact of ALL variations – both good and bad And finally the software provides you with an unbiased mechanism to track adoption and adherence - because these pathways are developed using your own data, it reduces physician bias, which makes it much more likely to be adopted and adhered to Creation piece vs operational/improvement piece. Modeling & refinement Doing it faster not really a key selling point. Data behind it is better. So how does this process work? The analysis starts with all of your EMR and financial data You could be analyzing thousands of patient procedures and millions of individual events/ We use the entire timeline for each patient procedure: for instance, the pre, intra and post operative information - all the labs, meds, and orders related to that patient procedure The software then generate pathway candidates It does this automatically surfacing groups of similar patient procedures It then lets you examine these similar procedures further by overlaying a common proxy for quality of care, for instance – like the length of stay, costs , or re-admission rates The software then automatically surfaces subsets of procedures that are highly correlated with that desired outcome It then mathematically aligns common events across this subset of best case procedures to generate a candidate pathway You can then examine and refine these pathways and model the impact of adding or deleting events to the outcomes The recommended pathway can then be sent through for review, approval and deployment The software provides you with intuitive dashboards that let you continuously monitor adoption and adherence And the pathways can be easily updated as new data arrives The net is that this approach lets you rapidly create objective, data-driven pathways (in the total knee replacement, each row represents a patient - we’re looking at 1300 patient procedures, and over twenty thousand variables – after transforms)
• #10: To summarize:   To summarize, Ayasdi’s machine intelligence software helps hospitals tackle challenges with managing clinical variation. It uses your patient data to design a data-driven care path to optimize patient outcomes. By standardizing on best practices, you can minimize unwarranted variations and lower hospital costs. And further more, the adherence dashboards allow a clinician to monitor the performance and track adoption. This feedback loop helps with continuous improvement and the discovery of other best practices. It helps hospitals achieve the triple aim of improving patient outcomes, increasing patient satisfaction, and lowering costs.
• #13: To summarize:   To summarize, Ayasdi’s machine intelligence software helps hospitals tackle challenges with managing clinical variation. It uses your patient data to design a data-driven care path to optimize patient outcomes. By standardizing on best practices, you can minimize unwarranted variations and lower hospital costs. And further more, the adherence dashboards allow a clinician to monitor the performance and track adoption. This feedback loop helps with continuous improvement and the discovery of other best practices. It helps hospitals achieve the triple aim of improving patient outcomes, increasing patient satisfaction, and lowering costs.