Data Analytics: Buy, Grow, or Rent Says eHI-CHIME Report

While analytics is becoming increasingly important for healthcare organizations, a shortage of expertise persists, leaving hospital CIOs three options: buy, grow, or rent.

This was one of the conclusions of a newly released report on “Data & Analytics to Improve Care, “a survey of provider organizations carried out jointly by the e Health Initiative (eHI) and CHIME (College of Healthcare Information Management Executives). Results of the survey were detailed in a public webinar on December 19.

The survey was conducted of some 90 hospital executives, representing a mix of academic medical centers, multi-hospital systems, and community hospitals.

According to Jason Goldwater, vice president of research and programs at eHI, only 67 percent of respondents are currently using any type of software to analyze the data collected within their information system.

The survey found the top uses of data were:

  • Ad hoc queries (87 percent)
  • Data mining (61 percent)
  • Data warehousing (57 percent)
  • Exploratory data analysis (35 percent)
  • Online analytical processing (31 percent)
  • Predictive modeling (27 percent)

According to Paul Thompson, director of commercial health and life sciences strategy and business development for Hewlett-Packard, the lack of lack of skilled personnel is a major barrier to implementing an effective data analytics program that can improve care.

Thompson pointed out that when it comes to managing an analytics program, “There are two separate skill sets. One is the ability to synthesize and analyze the data; the second is to drive change and results within the organization.”

Buying a software program or hiring a consultant to implement an analytics program can be expensive. However, one possible solution for hospitals to partner with other organizations such as other hospitals, medical groups, or health plans.

Another key challenge in managing analytics is the changing landscape of measurement. Thompson noted that URAC, JCAHO, and other organizations are changing their reporting requirements. The trend is more reporting rather than less, he noted.

Another challenge in effectively managing analytics is that most data is currently limited by geography, usually by state. Thompson cited the example of a patient who lives in New Hampshire, but is treated at a hospital in Boston. If the Boston hospital collects and reports the patient’s data, it might be assumed it the patient comes from Massachusetts, failing to provide an accurate picture.

Most data analytics programs are used for revenue management rather than in improving patient outcomes through clinical care improvements. That will change as new forms of payment, such as accountable care organizations, emerge.

The survey presenters noted that behavioral health presents a particular challenge since there is a lack of reporting of behavioral health status. For example, substance abuse is rarely self reported. In addition, few metrics or models exist for behavioral health outcomes. There is little agreement on the definition of a quality outcome, Thompson noted.

There is also no single set of metrics to consider when measuring the ROI of data analytics programs, Thompson noted. One common measurement is the hospital readmission rate. “Did you reduce the rate? Did you avoid penalties?” Thompson said.

These benefits must be weighed against the total cost of ownership, which should include both the licensing and maintenance of the software. The latter can be very expensive, Thompson said.

A recorded playback of the webinar including slides is available from the e Health Initiative. It is free to eHI members and costs $25 for non-members.

This article was contributed by James Harris, president of, a healthcare technology marketing agency.

↑ Back to top

Founding Sponsors

Platinum Sponsors