CIO survey: data variety, obstacles are on the rise in health analytics

 - data mining

Healthcare CIOs are using increasingly diverse sources of data on which to perform analytics, according to results of a survey co-sponsored by the nonprofit eHealth Initiative and the College of Healthcare Information Management Executives (CHIME). Many, however, have not committed significant resources or time to the activity.

Not surprisingly, the primary data sources being mined include clinical data from EHRs, (95 percent), and both pre-adjudicated (91%) and post-adjudicated (69%) administrative, billing and financial data, according to “The Landscape of Data & Analytics in Healthcare,” a report based on the survey and released earlier this year.

Beyond those primary sources, about 45 percent of the 97 respondents said they were pulling data from patient portals and health-risk assessments, but 32 percent said they do not currently analyze patient-generated data.  Respondents also reported using the following novel sources of data: remote monitoring devices (29 percent), health information exchanges (22 percent), and mobile applications (11 percent).

Most analytic activity (94 percent) is still in the category of descriptive analytics, described as mining data for historical or retrospective analysis; 94 percent of respondents reported that descriptive analytics are most commonly used, with 38 percent reporting daily use and 20 percent reporting monthly use.

While almost 70 percent reported that they use predictive analytics (forecast ing outcomes, trends or performance), the frequency was monthly for 25 percent, and quarterly for 20 percent.  Two-thirds of the respondents said they do not use prescriptive analytics (use of predictive modeling to recommend specific actions).

Important, but…

While the success of value-based initiatives are thought to be dependent on analytics for success by nearly all respondents (94 percent), just 42 percent have made the investment of time and resources to implement a flexible and scalable plan for their analytics platform to adapt accommodate the growing volume, liquidity and availability of healthcare data.

Staffing remains a critical impediment (79 percent) to collecting, processing and analyzing data, as do cost and interoperability.  Fully 33 percent said they had jobs but can’t find sufficiently trained candidates; 64 percent reported that inconsistent data sources and interoperability were their biggest challenges; and lack of funding for critical activities such as cleaning, integrating and data quality assurance, was a problem for 68 percent of organizations.

New barriers, however, are emerging. Complaints about access to external data, lack of funding or ROI, patient privacy and regulatory and compliance requirements doubled and quadrupled.

Finally, survey results indicate that healthcare organizations have belatedly begun to commit analytics resources to patient engagement activities. Two-thirds of respondents reported applying analytics to consumer engagement (for the purposes of patient satisfaction scores), but fewer are applying analytics to personalized patient communications (29 percent), customer acquisition and ret4enction (28 percent) or targeted behavior change programs (24 percent).

In the discussion segment of the report, Chuck Christian, CIO, St Francis Hospital, Columbus, Ga., compared analytics to essential infrastructure: “I don’t know too many businesses that do an ROI on their parking lot, but they’d better have one if they want customers. If we don’t have these things, we can’t compete.”