This section will discuss Health Informatics and targeted Quality Measures for ACOs.
a. Performance Standards
b. Quality Metrics
c. Analytics and Big Data
Health Informatics includes resources and methods to manage health information and the tools and infrastructure to analyze data. Analytical tools may come in the form of EHRs, data exchanges, data warehouses, database repositories, paper documents, networks and protected internet access.
Due to the nature of data collected by ACOs and the need to share data among providers, there can be significant challenges. Some of these challenges come in the form of "Big Data."
Big data essentially implies a database schema that is suitable for answering queries that may or may not have been anticipated at the time of schema design.
One of the challenges in a big data implementation is recognizing that numerous design considerations are far different than what normally apply to a transaction system.
- For example, transaction systems, usually have a relational database schema in what is called 3NF or what is called Third Normal Form. Whereas data warehousing systems generally have a star schema.
- Although these two types of design system are as different as night and day, a big data implementation is more akin to a data warehouse on steroids.
- Very few people understand how to do star schemas correctly. Star schema design is usually performed by specialized consultants working within particular niches. If an organization gets the database design wrong, then it is likely that the entire project fails. DOA.
- Further, in a data warehouse has more moving parts than a transactional system. Usually you are required to integrate data from various systems and therefore data identify issues arise and must be resolved. For example different transaction systems often call the same data element by different names.
- Further you have to build a robust scalable extract transform and load process that is capable of dealing with exceptions in an effective manner otherwise you have your traditional garbage in garbage out and users quickly begin to doubt the veracity of the data.
Performance Issues
- Transaction systems are optimized for updates whereas data warehousing and data analytics are optimized for query performance. Given the size of data warehousing and data analytic systems hardware and parallelism and other advanced techniques may need to be applied.
- In short data warehousing an analytic systems are complex and expensive to develop an the information highway is littered with failed projects. Primary due to teams not having the necessary talent available to build these types of complex systems.
- It takes an entirely different skill set to build these types of systems and most IT organizations do not have the necessary in-house talent.
Big data implies much more than data warehousing.
- BD systems are by definition generally orders of magnitude larger in terms of data than DW or analytic systems.
- BD systems usually have to deal effectively with both structured and unstructured data which means that you are likely going to need more than a relational database to meet your objectives. BD may require NOSQL databases to deal with unstructured data, yet another level of complexity.
- To implement successfully, BD require world-class engineers and even that is not enough.
Why?
- Because engineers generally do not have the industry expertise to build the systems without world—class analysts that can bridge the gaps between IT and Clinical staff.
- If past is prologue, then for every successful BD project you can see dozens of failed projects.
- Although BD will transform healthcare it is no silver bullet and there is no predicting who the winners and losers might be.
Comments