The majority of hospitals are currently using data analytic software, but most are using it for simple functions and want to make more meaningful use of health data analysis, according to a new survey conducted by CHIME and the eHealth Initiative.
More on data analytics use in hospitals
Expert discusses what hospitals
CIO talks about her experience with analytics
Test your knowledge of business intelligence and mobile health
Results from the College of Healthcare Information Management Executives (CHIME) and the eHealth Initiative (eHI) survey showed about 77% of respondents said they currently use data analytic software. However, the 90 hospital chief information officers who responded to the survey said they most frequently use analytics for financial management, operational efficiency initiatives and reporting data for national quality measures. Few respondents said they use analytics in direct patient care.
Furthermore, health data analysis is typically retrospective, rather than prospective. A total of 58% of respondents said the majority of their analytic resources go to retrospective data tracking. Only 17% use analytics for real-time decision support and just 3% have predictive analytics programs. While 93% said analytics is important to the future of care, 28% said they have the resources they need.
Jennifer Covich-Bordernick, CEO of eHI, who spoke in an Aug. 30 Web conference outlining the results, said the responses indicate a disconnect among the types of analytic tools available. While many experts talk about the importance of predictive analytics and real-time decision support tools, these systems are rarely used in actual practice.
"There's a lot of discussion of these areas, but there are small numbers associated with them," Covich-Bordernick said. "This is really the tip of the iceberg in exploring this area, because while we see that 93% say [predictive analytics] is very important to the future of their organization, less than a third actually feel like they have what they need."
The way we do analytics retrospectively is how we will position ourselves to do it prospectively.
vice president of business intelligence
Catholic Health Initiatives
Evon Holladay, vice president of business intelligence at Catholic Health Initiatives, who also spoke at the Web conference, said the most important thing providers can do now to support predictive health data analysis is develop more precisely defined quality standards. She said currently there is little agreement in health care about what constitutes a quality outcome. Without this agreement, it will be difficult for providers to look at data to evaluate their effectiveness and use this information to guide treatment in real time.
"The way we do analytics retrospectively is how we will position ourselves to do it prospectively," Holladay said. "We skip being aligned on what we're going to measure, which makes it virtually impossible from an infrastructure standpoint to be successful to drive measureable performance improvements."
Improved quality performance needs to be the ultimate goal of any analytics initiative, Holladay said. Currently many organizations do ad hoc queries and data mining, but there is no clear set of criteria that define the type of information providers should be looking for or how to assess the quality of the data.
Developing more effective ways of using analytics has never been more important, according to Fred Bazzoli, senior director of communications at CHIME, who spoke at the Web conference: Health reform and the growing adoption of accountable care models demand it.
"CIOs know their organizations have to get better at assessing data to improve their clinical performance," Bazzoli said. "Reform initiatives such as accountable care will place a higher emphasis on using data for other purposes, such as predicting what will happen with patients and real-time decision support."
While he said providers' use of analytics is mostly "utilitarian" now, focusing mostly on financial management and data reporting, Bazzoli said providers are starting to see health data analysis as crucial to their long-term survival. This could lead practices to start analyzing data at much deeper levels.