There are many elements of business intelligence in play within the healthcare IT continuum, but without a specific focus, the initiative is stuck in neutral. In fact, that is where many healthcare providers find themselves with their BI programs, system specialists say - wondering how to proceed and in which direction to move.
But despite whatever paralysis may be going on, it can be eased by keeping an eye on the "end game," says Mike Garzone, solutions director for technology at Buffalo, N.Y.-based CTG Health Solutions.
"We think the end game is the ability of hospitals to improve - whether it is through outcomes, a decrease in readmissions, or identifying hospital-borne conditions before they occur, like sepsis or infections," Garzone said. "It comes down to how BI functions with care delivery and how it relates to reimbursement. The end game is about improving how to get there."
In order to steer the BI component in the right direction, hospitals need to make it a priority, which Garzone admits is a tall order in an environment where personnel are preoccupied with meaningful use and other directives. Still, the pieces of a strong BI system are largely in place within the hospital sector and are mostly ready to be deployed, he says.
"We've enabled it - we haven't codified everything and it isn't interoperable, so there is still work to do, but it is there," Garzone says. "Now is the time to act and leverage it. We are getting to the point where accountability and performance will really start to matter and that is when BI will bubble to the top."
For the past two years Houston-based HealthPost has deployed a cloud-based search and booking platform that enables BI through the gathering of data and intelligence around the healthcare industry. CEO Omar Alvi describes it as "covering the spectrum and connecting all the points in the healthcare ecosystem."
Engineering a system that resembles both Google and Expedia.com in form and function, HealthPost's core business is improving access to care by increasing hospitals' availability to partner with other providers in the community. The search-and-book digital method replaces what has traditionally been a "slow, cumbersome process" with phone and fax, Alvi said.
"Now all these transactions come through a single point, creating complete data intelligence on who is booking with providers," Alvi said. "With the hospital as the sponsor of these opportunities, the entire outpatient provider network is in alignment. For the first time ever, they have complete visibility into booking activity."
BI capability also allows for data mining on a number of fronts and in fact is superseding the decision support function, says Taylor Moorehead, partner for Carmel, Ind.-based Zotec Partners' west region.
"Imagine a world where there are 425 data points to bill insurance, ranging from ICP to ICD-9 to other points across the board - you can dissect all of those data points in a granular way and correlate them to the billable amounts," he said. "BI can take all the different aspects of services and relevant points of the claim and run them in various forms and fashions to mine for all the right data."
Applying analytics
BI technology has advanced to the point where it can handle larger volumes of data, larger analytical models and put them in the hands of layman users, says Dan Foltz, vice president of Blue Bell, Pa.-based Anexinet.
"Data visualization is getting better," he said, "but a lot of organizations just aren't there yet."
Anexinet works with other industries on BI and Foltz says healthcare lags behind in utilizing its full potential. By contrast, the financial services sector is well ahead of the pack, he says.
"Things are starting to happen in healthcare, but there are some barriers," Foltz says. "Currently, analytics reporting is retrospective, charting care that was delivered yesterday. Healthcare needs to move toward prospective and predictive analytics."
Prospective analytics, he says, is determining why a patient didn't fill a drug prescription and acting proactively while predictive analytics is looking at a population to determine co-morbidities and complications among chronic disease patients.
For healthcare to truly transform, the industry needs to use BI as a measuring stick for results, not activities, Foltz says.
"Currently hospitals are being paid for activity and not outcomes - the incentives need to change," he said. "We need to work toward a system where everyone is incented by outcomes, including payers, drug companies and patients."
Ann Arbor, Mich.-based Arbormetrix has been conducting analytics around the population and is generating some interesting statistics based on its evaluations, says CEO Brett Furst. For instance, because the majority of the chronic disease population is likely to experience a co-morbidity or episode, there are inconsistencies within the acute care setting, he said.
"There is a 30 percent cost variation from hospital to hospital and they need to look for trends on the reason for this difference," he said. "They need to consider demographics and co-morbidities and their influence on outcomes. And while the disease state management piece is an important piece, the majority of spending remains inside the acute care setting. Population health and DSM are here to stay, but for the near term, the concern is more on acute care."
Device tracking
BI isn't just for tracking activities and outcomes - gauging the performance of medical devices is also an important part of the equation, says Peter Witonsky, president and chief sales officer for Panama City, Fla.-based iSirona.
By combining its apps with greater intelligence and algorithms, the iSirona system enables hospitals to anticipate and prevent potential sentinel events related to improper device usage, such as ventilator-induced pneumonia.
"The staff is alerted if a vent patient's head angle is not 30 degrees," he said. "It also issues an alert if is not rotated properly to prevent skin ulcers. These are things that are not typically included in electronic medical record tracking - it is all about predictive modeling."
