5 Strategies for Implementing Data Analytics in Hospitals

Since the advent of the Internet, people have been able to access a significant amount of information with the click of a computer mouse. As technology becomes more ingrained in healthcare delivery, providers are facing the quandary of how to glean useful information from the enormous amount of data available through electronic health records, computerized provider order entry and other databases. Steve Nitenson, RN, BSN, MS, MBA, PhD, adjunct professor at Golden Gate University, San Francisco, and senior healthcare solutions architect for Perficient Healthcare, shares five strategies hospitals can use to implement healthcare data analytics.

Goals and challenges
"Healthcare analytics is the Holy Grail with respect to healthcare," Dr. Nitenson says. The conversion of loosely net patient data — data that is loosely associated with other data — into meaningful, actionable trends holds the promise of increased coordination of patient care, patient safety, quality of care and cost-efficiency for not only individual patients, but also for patient populations at large. Gathering and using meaningful information on patient populations is one of the major goals of accountable care organizations. "Part of the accountable care organization model requires the ability to analyze healthcare data," Dr. Nitenson says. However, turning patient data into usable information is challenging; furthermore, once healthcare providers can access the information, determining how to act upon it poses more problems.

Dr. Nitenson differentiates between data — the facts entered into a database — and information — the interpretation of the facts in a meaningful context. "Just because you have a lot of healthcare data doesn't mean you can do anything with it. It becomes mission critical, especially today with all changes occurring in healthcare, to get as much data converted into information that is actionable," he says. According to Dr. Nitenson, the healthcare industry is currently able to access 80 percent of available healthcare data and can convert only 20 percent of that data into meaningful information.

Strategies

1. Establish a governance structure. Hospitals should set up a governance structure to manage implementation of healthcare data analytics capabilities. The CEO, CMO, CNO and CIO should all be involved, Dr. Nitenson says. The CMO and CNO need to communicate the kind of information they want to the CIO, who has the IT knowledge to conduct the actual implementation. "You need to have a governance structure [in which] the CIO takes the lead but has the [CMO and CNO] to always ensure that whatever he or she is doing is going to meet their needs," he says.

Hospitals should also consider partnering with a professional organization experienced in data analytics, Dr. Nitenson says. In this situation, the IT organization would develop analytics and the CIO would implement it with the support of the CEO and guidance of CMO and CNO. "[Providers'] business is delivering healthcare, not developing healthcare analytical data models for themselves," Dr. Nitenson says. He says the learning curve would be much steeper and the time to implementation longer if the providers try to create an analytical toolset on their own. "It's a major undertaking that takes time and a good deal of effort. [If providers do it on their own] it usually ends up on the back burner and never gets done."

2. Define the desired information. Due to the large amount of healthcare data available to hospitals, from everything from lab data and physician notes to insurance claims to medical records, hospital leaders need to define what information they want. "You have to be able to filter what's really important to you based on the hospital and specialty [you're interested in]," Dr. Nitenson says. Leaders should also determine when they want the data, how they want it presented and who they will share it with.    

3. Format the data appropriately. One of the keys to analyzing healthcare data is presenting it in an appropriate format. For example, Dr. Nitenson says if the hospital wants to understand the lab data for someone whose blood is drawn twice a week for five weeks, simply looking at the 10 data points would not yield any useful information. "It's meaningless if there's no reference point," Dr. Nitenson says. Instead, the hospital would need to trend the data and benchmark it against the demographics of the region the hospital is located as well as national averages.

4. Secure the information. Once hospitals have converted the healthcare data into meaningful and actionable information, they need to decide who to grant access to and establish security protocols to ensure access is available only to those individuals with "a need to know," Dr. Nitenson says. He adds that at this point, sharing healthcare data becomes an ethical question. "The more information you get, the higher order of ethical behavior one has to have." For example, he says exposing the information to insurance companies could have consequences for hospital reimbursement. Access to information does not have to be all or nothing, however. While patients should be able to obtain their own information, hospitals should consider what information might be inappropriate. "Some information may be deleterious to patients without understanding the context," he says.

5. Share information effectively. Even if healthcare data has been converted to information and the information has been secured, data analytics cannot produce benefits of improved quality and reduced costs if the information is not shared effectively. Dr. Nitenson suggests using a "push" rather than a "pull" technique for sharing information with physicians and nurses. "Push the information to the professionals," he says. "Present it to them in their daily work. Don't [make] them try to find it." The difference between "push" and "pull" is similar to opt-out and opt-in systems. In a "push" environment, physicians would be automatically presented with information that they would have to consciously ignore or dismiss — or opt out of. In contrast, a "pull" environment would require physicians to find the information themselves, or opt in. The former method increases the likelihood that the healthcare professional will be aware of the information they can use to improve patient care.

Learn more about Perficient.


Related Articles on Health IT:

Top 5 Reasons to Move Patient Data — System-Wide — Into An EMR
Are Scribes the Missing Link Between Physicians and EHRs?

7 Best Practices for Hospitals Implementing Health Information Exchanges



Copyright © 2024 Becker's Healthcare. All Rights Reserved. Privacy Policy. Cookie Policy. Linking and Reprinting Policy.

 

Featured Whitepapers

Featured Webinars