Dollars & Sense: Business Intelligence & Financial Decision Support

As healthcare reform efforts and other factors shift more financial responsibility to healthcare providers, organizations and departments are turning to business intelligence to bolster their financial decisions, which can become patient care issues as well. The regulatory heat is also on, in the form of increased attention to quality improvement efforts, pay-for-performance and cost reporting initiatives, making business intelligence essential for an organization’s fiscal health.

Broadly defined, business intelligence tools are those that glean information from a variety of sources to enable better business decisions. These tools take a variety of forms, from specialized hospital information system modules to add-on software that live on top of internal systems such as EMRs. Clinical decision support tools can be used as business intelligence resources.

“Some clinical systems have some business intelligence features; in other [cases], it’s an add-on component where organizations take a tool and lay it on top of a SQL database that exists within their practice, as an example, and allow themselves to integrate with that tool so they can interface with the information that’s found in that particular database,” says Todd Evenson, assistant director of survey operations at Medical Group Management Association (MGMA). The Englewood, Colo.-based association provides resources and support for professional administrators and leaders of medical group practices.

According to a HIMSS Analytics report released in June, a focus group of 10 CMIOs and chief medical officers at organizations with 300 beds or more indicated that their facilities are collecting and leveraging clinical and/or claims data to reduce patient care costs, and enhance safety and efficiency.

Furthermore, data are looked at on a variety of levels, according to the report, titled “Clinical Analytics: Can Organizations Maximize Clinical Data?” Respondents are not only looking at information on a specific patient, but also are exploring data across population-based metrics, such as data specific to a particular physician; or to a certain condition, such as diabetes or hypertension, the report states.

The goal of the report was to understand how organizations—both providers and payors—can maximize clinical data, says Jennifer Horowitz, senior director of research at HIMSS Analytics, who authored the white paper. The size of the CMIO/CMO group was not intended to be a statistically valid sample, “but it gives us a starting point for discussions around clinical analytics,” says John Daniels, HIMSS Analytics vice president of healthcare organization services.

CMIO’s own findings corroborate those of the HIMSS Analytics white paper. In the recent CMIO Top Health IT Trends Survey, nearly 52 percent of 153 respondents said their facility has implemented business intelligence and metrics tools. And nearly 25 percent indicated they would implement these tools in the coming year.

Although vendors and organizations expect to incorporate business intelligence into more processes in the future, for now, many CMIOs are using analytics mainly for “things that have to be done for compliance reasons, or for the organization’s strategic requirements,” Daniels says.

“The healthcare industry is so information-intensive … that individual organizations are challenged with getting access to all the data, [including] clinical, administrative and financial data,” he says. “The challenge is being able to pull all those pieces together to answer the specific questions they may have that would consider that whole cross section of information that’s available.”

In addition, “there are challenges with the way data are captured that make them hard to overlay a tool on top of that,” Horowitz says. “A lot of data are captured in a free-form method, [such as] a physician’s note in a chart. For an analytical tool to be effective, you need [standardized data capture] for data to be converted to a discrete analysis.”

Measurement is management

According to Evenson, that isn’t stopping organizations from deploying business intelligence. “Practices are starting to implement these tools … this is the direction people are heading. If you can’t measure it, you can’t manage it,” he says.

MGMA studies have shown that providers and others are shopping with certain BI capabilities in mind as well. Whether integrated within a clinical system or available as add-on to a financial management product, there are some features included in business intelligence instruments that are generating a lot of interest, says Evenson.

“Frequently we find a lot of [interest in] the dashboard functions, real-time reporting and key performance indicators,” he says. “All of those items are very hot topics in terms of the reporting features that are found within many of the systems that are out there today.

“Just as we find in evidence-based medicine, it’s true in the same regard in evidence-based management. We’re really seeing that healthcare, [with regard to many of those types of process improvement applications], maybe was latent but is quickly moving as an industry in that direction, to find cost savings, improve performance, and increase the accuracy of their decision-making process.”   Administrators who are inside these group practices, as an example, utilize these tools to communicate more effectively with physicians about what’s going on inside the practice, says Evenson. The nature of the dashboard reporting capabilities in some business intelligence tools enables administrators and physicians alike to condense a lot of different items or metrics that are important to the group, into a singular form that can work almost in real time, he says. “That gives [administrators] a better capability to provide informed decisions to their physicians.”

Leading the charge

Can business intelligence derived from cost capture capabilities improve the quality of patient care? Yes, says Alistair Erskine, MD, associate dean of medical informatics at Virginia Commonwealth University (VCU) School of Medicine, and chief of clinical informatics of VCU Health System (VCUHS), based in Richmond. VCUHS includes the 779-bed Medical College of Virginia Hospitals, a 600-physician faculty group practice and several outpatient clinics.

VCUHS decided to implement PatientKeeper in October 2006, and piloted deployment in internal medicine, hospital service and ICU medicine service, to find out if use of electronic charge capture—the primary module of interest at the time—would improve overall charge capture for these units, says Erskine.

“We also were interested in mobilizing the physicians’ access to clinical data as well as using PatientKeeper as a shadow EMR … something that we could use on top of our current [Cerner] EMR where it may have some holes or need further enhancements,” he says.

“We were able to do an analysis of charges that were captured the year prior to and after we had gone live with PatientKeeper. Adjusting for changes in payor contracts and adjusting for volume, we were able to determine those changes attributable to PatientKeeper that were captured in addition to what we normally would have captured.

“What we found in these [pilot] departments was that we were missing about 30 percent of charges, which was much more than we anticipated: We had thought the paper-based system was a reasonably good system, but it was a very difficult system to audit,” Erskine says. For example, immediately after a charge was dropped, the system required collection of cards over a period of a month or two, then adjudication of charges with the census and eventually, for those charges that were missed, the physician might get a call months later. “That process was so distant from the actual activity of the charge that we had poor success in capturing them,” he adds.

“In the electronic workflow, we set up a process by which the next day, if a patient had been in a hospital bed and there was no charge, [the physician] would get a page. So the billers that in the past were focused on interpreting physician handwriting on paper cards were now making sure charges were being posted.”

In addition, the VCUHS billing department developed “almost a new relationship with those physicians in helping them with the actual coding process. The physicians now were in a position where they were able to ask questions and get clarification on billing rules they need to know to post their charges,” says Erskine.

That relationship ended up providing more accurate, more comprehensive charges in less time, “so we had both a boost in revenue in terms of how quickly the charges come in—now they come in as soon as they’re posted—as well as the number of charges that are arriving. [We’re seeing] an influx of cash three months earlier than we were seeing before as well as a higher volume,” he says.

“With that success we went to the practice plan and said ‘let’s do this for everybody else.’… We now have all the clinical data from Cerner, so there are clinical notes, labs, medications, allergies, radiology reports, echo results, all that kind of clinical data are in the same portal that we use for the charge capture portal, [and are] available on the web and on Blackberry devices.

“We integrated a lot of the components of the medical record into the PatientKeeper charge-capture system so the patient list they were using and seeing was the exact same patient list as they were seeing in the Cerner system. The problems that they would enter into the Cerner system would translate from SNOMED dictionary into ICD-9 dictionary as they would show up in PatientKeeper. Conversion is made automatically, many-to-one, from the interface itself,” he says.

In addition to faster, more accurate charge capture, the PatientKeeper tool’s reporting capabilities have resulted in VCUHS physicians inputting their notes into patients’ in a more expeditious way, according to Erskine.

“About a year after we put in PatientKeeper, we went to full electronic documentation. One thing that came out of that is we created a report that basically identified every patient on the census, [whether] he or she had a charge and if there was electronic documentation [on a] particular day. With 30,000 discharges a year, … a majority of patients end up with a charge and an electronic note, [although] not all of them had electronic note and electronic charge for that particular day,” he says.

“That ended up being an extraordinarily useful report, not only to identify charges, but also in terms of quality of care, to make sure that all notes that were supposed to be written were being written in a timely fashion. So it ended up going beyond just charge capture toward other aspects of auditing quality of care.

‘The ATM of the health system’

BI tools that pull data organization wide have given a big boost to financial decision support at Duke University Health System, based in Durham, N.C., according to Scott Williams, administrative director of billing, collections, and customer service. Duke University Health System includes the 924-bed Duke University Hospital, the Duke Raleigh Hospital, the Durham Regional Hospital, and Private Diagnostic Clinic which encompasses 1,200 physicians.

GE Healthcare’s Centricity Business Enterprise Task Manager (ETM) application, which was implemented in 2007, “took us a quantum leap forward in terms of operational capabilities and system support, to allow us to build a better workflow,” says Williams. “[This application] has been the one that has allowed us to make some of the farthest advances in terms of efficiency of our operations.”

Handling the billing and collections of insurance for the system’s physician organization, Williams’ group bills slightly over $1 billion a year in services and charges for about 1,200 physicians, he says.

“We are, for lack of a better term, the ATM of the health system: It’s our job to take the great care that our physicians provide to patients and turn that into cash, primarily from insurance carriers but also the patient responsibility portion. In many respects, we’re overhead or administration, and we would like to minimize the amount of resources [that have] to go to administration as opposed to clinical care. So any opportunities we have to streamline that process and make it more cost effective is certainly appreciated,” Williams says.

The task manager’s advantages include integrated access to multiple systems, he says. Very few operations have the luxury of having everything in a single-source system, and we’re the same way, so as we’re trying to deal with multiple other systems, internally or externally, we need it in an integrated way. Sometimes we need to access information from medical record in the [physicians’] EMR. ETM allows direct access to the EMR, for example, to determine medical necessity or get a copy of the EMR. ETM also can access the stand-alone imaging system, making users much more productive, Williams says.

Although industry-standard coding conventions are used—Current Procedural Terminology (CPT) to identify services as well as ICD-9 and (soon) ICD-10 diagnosis codes—each insurance plan and employer may have a different benefit structure in terms of what services they cover, says Williams. Enterprise Task Manager allows denied claims to be reported back to physicians: “We categorize them in three different groups; registration or eligibility, when insurance information is captured incorrectly; coding-type denials that relate to procedure codes, diagnosis codes or modifiers that are assigned by their physicians or representatives; and the third category is everything else that falls in the claims and billing world.

“The piece that physicians have most input on and ability to change—and therefore the most interest in getting feedback on—are those [denials] related to the types of codes that they or their representative select. So as we give them feedback, they can look at opportunities to improve or better understand the process. Sometimes it’s a function of additional training that might be available. In some cases, though, it’s a function of dealing with the myriad of different insurance companies, which also change on an ongoing basis. At times, it seems like they have a different set of rules.”

Typically, 7 percent to 8 percent of claims are denied, and it takes a fair amount of effort to remediate these denials, extending turnaround time, Williams says. Within the ETM system, “reporting tools evaluate all those denials; sort and segment them, and break them down by categories such as payor, by physician group, by diagnosis codes, etc. That segmentation facilitates a much greater amount of analysis and then going upstream to try to figure out how to prevent those in the first place,” he says.

“These tools provide a lot of reporting in a lot of different ways. We can and do gather where patients are coming from, what referring physicians are sending them, and use that information for a variety of marketing as well as understanding disease prevention and disease causes. It’s integrated with clinical systems in that respect, and we use that significantly.”

The BI future is promising, according to MGMA’s Evenson. Healthcare reform and other initiatives have put the focus on quality improvement and outcomes, and how to accomplish them. “Tomorrow’s key performance indicators won’t just necessarily look at financial metrics, total medical revenue, operating costs, etc. … They’ll revolve more around actual outcomes, patient satisfaction, some of the PQRI [Physician Quality Reporting Indicator] measures that are being measured currently in health arena,” he says.

“As a result, these models will become more and more important in utilization in terms of understanding how their organization’s performing because ultimately that understanding and those outcomes will lead to the revenue that they receive.”

HIMSS Analytics’ Daniels concurs. “I think there’s a desire to do that, and more so in the future. I’m not so sure it’s as widespread as we would like it to be. I think that will increase with the advent of ARRA; hopefully we’ll see some increased emphasis on the ability to do that.”