JJ'Studio - Fotolia
The healthcare industry is somewhat behind other industries when it comes to data analysis capabilities. One big reason for this quandary is the difficulty of getting data out of electronic health record (EHR) systems, which tend toward proprietary data storage methodologies and are notorious for lacking the ability to share data. But healthcare organizations have been developing end-runs around these problems, and the efforts are paying off.
The EHR system, from vendor Aprima Medical Software, came with some pre-packaged healthcare data analytics functions, but they were rather limited, and Seema Gai, CrescentCare's CIO, wanted to do more. Gai said most EHR vendors don't readily give information about their data schemas to customers, but CrescentCare worked with Aprima to get its schema documentation. Gai found that the system is basically just built on top of a SQL database -- nothing too exotic. Now she and her team are able to layer more sophisticated analytics applications on top of the data store.
"A lot of what we do on a daily basis generates a lot of data," Gai said. "Having access to that data schema has enabled us to access the data we need."
Healthcare data analytics ranks risks
For clinical analytic purposes, CrescentCare turned to SAS Institute's Visual Analytics software. One of the primary purposes for which it uses the tool is to stratify patients according to their health risks.
Seema GaiCIO, CrescentCare
Gai's team has used the tool to set up an automated reporting system that analyzes patients' records by looking at past lab results, vital signs and other information, and creates daily reports for clinicians that tell them which patients on their schedule are at high risk of experiencing a complication. This information helps doctors know whom they should follow up with after visits and whom they should recommend more intensive treatments plans to, and generally keeps them alerted to potential problems.
"We're trying to identify these high-risk patients," Gai said. "It's about having the right tools in addition to the right skills."
Obtaining its EHR system's data schema helped CrescentCare get to a point where it can do meaningful analytics on patient data. But this approach isn't an option for all providers. CrescentCare has the advantage that all its offices are standardized on a single EHR. That's not the case at every health system. Today, many health systems are a conglomeration of numerous hospitals and doctor's offices, many of which use their own EHR systems. It would be overly laborious for most healthcare providers to build custom connections between data stores and analytics applications for each of the EHR systems they maintain. That's why data platforms and integration tools are becoming more popular.
A more unified approach to analytics
Tina Esposito, vice president of the Center for Health Information Services at 12-hospital, Illinois-based Advocate Health Care, described the health system's analytics efforts prior to this year as being mostly one-off projects. There was no unified approach to how clinicians or analysts used data, and each project required pulling data together from scratch.
"We really hadn't tied all those threads together to understand how we're delivering care," Esposito said. "We needed to get more robust with how we were looking at data."
This year, however, the health system implemented a data platform from Cerner Corp. called HealtheIntent. It pulls data from multiple sources into a cloud-based store at which analysts point analytics applications. For example, they can now analyze patient records to determine the most successful transition-of-care plans based on certain patient characteristics. Moving patients from one care setting to another, like an in-patient hospital to home-based care, can often lead to complications, and determining the most effective strategy is considered a top priority for many hospitals.
Esposito said Advocate couldn't have developed its more coherent approach to analytics without the common data store. But she cautioned that simply implementing this kind of tool won't address all of a provider's problems. Her team still had to spend significant time making sure connectors worked properly. EHR systems still record data in proprietary formats and use unique database systems. Until this setup changes, providers will look for whatever fix they can find. "It's not an end, it's a means to an end," Esposito said.
Why healthcare data analytics remains a challenge
How big data is impacting healthcare
Providers like cloud analytics options