RIS data, keyed to billing, can be analyzed to improve competitive capability and pare inefficiency to the nub Physicians and executives are similar to scientists; they like data to assist them in making decisions,” Rob Cercek says. Cercek, vice president of ambulatory services at Rochester General Hospital (RGH), Rochester, NY, adds, “If you can put credible data in front of people, discussions between hospitals and physicians become more meaningful.” He is talking about arming his department representatives with data when they go out to solicit radiology business from referring physicians, but he could just as easily be describing radiologists themselves. Like physicians in any practice, radiologists are indebted to numbers for answers to their operational questions: Are practice RVUs stacking up in a positive way? Are imaging rooms and equipment being used maximally? Are there too many (or too few) technologists on staff? Radiology managers and executives like Cercek are working diligently to create useful reports to guide themselves—and their practices’ physicians—as they make many types of decisions. Should a new clinic be opened? Should certain procedures be marketed, or should they be quietly left to decline? Should productivity be demanded, or will quality suffer if too much pressure is brought to bear? Cercek says that some of the most helpful reports have been those showing patterns in physician referrals to the radiology department at RGH. By comparing referrals from doctors on the hospital staff with those from referrers outside the hospital, Cercek says that he can gauge when a referring physician is falling short of sending the number of cases that might be expected. The department representatives can then be deployed to seek, politely, that missing business. Cercek says, “The people who use these reports feel they are invaluable,” especially for making it possible “to walk into the physician’s office knowing the pattern—knowing what the referral rate should be—and getting them to turn in your direction.” At RGH, which is licensed for 585 beds and conducts about 175,000 radiology exams per year, the RIS is at the heart of the data-gathering effort. The RIS, from the same vendor as the hospital’s PACS, collects patients’ demographic data from the hospital information system (HIS) and correlates them with radiology data from the PACS, such as exam types, exam times, and modalities used. When a radiology report is signed, the RIS also triggers the billing cycle. All these data—exams, referring doctors, technologist times, patient demographics, billing codes, and much more—can be mined from the RIS in the form of either preprogrammed reports or, more pointedly, as customized reports prepared by the RGH staff. Reports like those being done at RGH can be used to shape (or reshape) a radiology practice. Moreover, the cost of the software needed to issue reports is modest, by hospital standards. The return on investment for data-analysis software is so rapid (a matter of months) that many don’t bother to track it. At RGH, the work of turning out most reports falls to Pam Moseley, radiology informatics director. A Data Gold Mine When RGH installed its RIS, Moseley says, the radiology department took over its own billing because the RIS could drive that process. Now, when an exam is completed, the patient data and radiology data, including CPT® code and pricing, are automatically entered into the billing system via RIS. Billers in radiology make sure that all the data match, then send the bill to the hospital’s master billing system for the technical component and to the physician’s billing system for the professional component, Moseley says. “Prior to this, everything was on paper and medical records was billing,” she adds. “They’re not familiar with radiology exams and functions; thus, there were a lot of charges that weren’t being billed.” Since initiating semiautomated RIS-driven billing, Moseley says, the radiology department’s receivables have shown a 30% to 40% improvement. This has added up to hundreds of thousands of dollars in recovered income. Moreover, billing time has dropped from more than two days to a day or less, she adds. Moseley says that the RIS software turns out a number of automated reports, including patterns for the hospital’s top 20 referrers, although this particular task required special programming by the RIS manufacturer. Most of the reports that Moseley submits to Cercek and others are custom reports that she has completed herself, however. Custom reports of this kind provide the real payoff when a hospital or practice wants to make immediate adjustments to workflow, finances, or staffing efficiency. At RGH, Moseley has initiated reports on staff productivity, room utilization, and even exam appropriateness (based on initial indication). “We get a lot of MRI orders from the emergency department,” she says, “but does the indication merit MRI? We have to educate our physicians.” Moseley says that the custom reports that she completes aren’t necessarily difficult, although some can get complicated. She says that her reports have affected staffing, modality efficiency, and scheduling, as well as the recovery of lost billings. She notes that the RIS data can influence purchasing decisions or reveal when imaging equipment is approaching the end of its expected lifetime, which is a key factor in a certificate-of-need state such as New York, where major equipment purchases must be justified to regulators. “We’ve also adjusted staffing patterns—hours, lunches, and break times,” Moseley adds. “The hospital thought that we were overstaffed in radiology, but in actuality, we’re not. Report data are based on user input; in our case, the technologists were putting in the incorrect time for how long it took an exam to get done. They thought it was scan time, but it should have been time with the patient. Once the staff was re-educated, a true cycle time was reflected. Overall, our throughput-cycle time has shown an 80% improvement since we went to the RIS.” Moseley is now tracking patient-transport times via RIS in an attempt to justify using radiology staff to transport patients in beds or wheelchairs, instead of a relying on the hospital’s centralized transport service. She says that too much downtime occurs when central transport is called; sometimes, its personnel fail to bring patients on schedule, resulting in a schedule backlog, even though this is the favored approach of hospital administrators. According to Cercek, RIS data mining has been especially useful for quick marketing response to changes in referral patterns. Efficiency inside the department has also made the hospital more attractive to referrers, he adds. He notes that a third of patients admitted to RGH come through the emergency department. Those emergency patients who need imaging are now entered into the RIS on a fast track. “We were looking at a three-hour visit in our emergency department for fast-track patients, and our competitors were at 90 minutes. Of that three-hour visit, 52 minutes were spent in the imaging department,” Cercek says. “Now, we have a target of less than 30 minutes to perform imaging on fast-track patients to hit a visit time of 90 minutes or less.” RGH now averages 28 minutes for imaging studies in the emergency department. The order issued for imaging triggers a time clock. Cercek says, “Before, we really didn’t know how to hit that time. Now, we can put a time stamp on how long those patients are in radiology.” Both Cercek and Moseley note that RIS capabilities also play into the hospital’s larger recordkeeping effort, along with the HIS data and the lab reports, all of which appear in the patient’s electronic medical record. Moseley notes that the city itself is forming a regional health information organization (RHIO) as a central base for shared knowledge. When the RHIO is finished, the radiology reports will be available there “so that the whole city will be connected,” Moseley says. Cercek agrees that some skepticism needs to be applied to the reports that flow out of data repositories. “Whenever you receive a report to review, you have to gauge what it’s really telling you,” he says. “The RIS will never describe the benefits or nuances of a 64-slice CT study versus an interventional procedure. You have to finish the story.” King’s Daughters Hospital A midsize hospital such as RGH is an obvious fit for RIS data mining because it generates so much information, but even a small hospital can make profitable use of the technology. King’s Daughters Hospital (KDH) in Yazoo City, Miss, has six doctors on staff and 25 acute care beds. Stefanie Dendy, a former mammography supervisor at the University of Mississippi Medical Center in Jackson, is the KDH director of radiology. Coming back to KDH five years ago was a homecoming for Dendy in the most basic sense; she was born at the hospital. The radiology department at KDH includes diagnostic radiography, nuclear medicine, fluoroscopy, ultrasound, and CT. A mobile MRI unit appears on a regular schedule. The department installed a PACS in 2006 and a companion RIS from the same vendor a year later. There are no radiologists resident at KDH; the single radiologist on staff reads from a site in Louisiana. He was hired when the hospital installed its RIS. Dendy says that KDH completes about 15,000 exams per year, and that single radiologist reads them all. “No radiologist will come to our small hospital anymore,” Dendy says. Nonetheless, she adds, competition is stiff among outside providers. “Now, with PACS, everybody wants your business,” she says. At night, if the contracted radiologist can’t be reached, the hospital will send exams (CT only) to an after-hours teleradiology service for preliminary readings, Dendy adds. She credits the PACS/RIS not only with making this possible, but for making it quite workable. “He’s been a great find for us,” she says of the contracted radiologist. “We contact him all day long, and he even calls the emergency department with critical results. He hasn’t been sick, he works holidays, and he has his laptop with him. He can give us preliminary readings from that, and when he gets back to the PACS monitor, he will dictate the final interpretation. That satisfies the doctors on our medical staff.” Despite the comparative simplicity of the KDH radiology department, Dendy says that she routinely pulls data from the RIS and creates her own custom reports to guide operational and strategic decisions. The RIS will track workflow averages and generates stalled-procedure reports. “The workflow averages help me; they tell me where the breakdowns are,” Dendy says. Is the problem in registration or elsewhere in the exam cycle? Dendy says that the RIS reports answer that question and more. She can look at RIS reports to see where exams are stalling and then respond to physicians who ask why a report is not showing up in the patient’s chart. Is the delay being caused by the radiologist, the transcriptionist, or the clerk? “The nurses blame radiology, and radiology blames the nurses,” Dendy says. “That’s another advantage of the RIS: If the report is not in the chart, the clerk can call and print one. On the units, they have access to the reports.” The RIS also allows Dendy to track and respond to modality-use patterns, on which she reports monthly. After studying these trends, she can make adjustments. She uses the example of a bone-density scanner. “We got that machine at the request of medical staff, but all of a sudden, nobody was ordering the test anymore,” she explains. In such cases, if use of a modality is unusually low for three months, she can remind the medical staff that the hospital has this capability. “The doctors do go through cycles,” Dendy says. “Sometimes, all it takes is a reminder.” Dendy has also used RIS data to track the referral sources of patients who don’t show up for appointments. No-shows compromise patient flow by tying up schedules, she notes. Most of the KDH no-shows turned out to be coming from clinics with high volumes of indigent patients. KDH then worked with those clinics. “We said, ‘Talk to your patients,’ and we actually have cut no-shows,” Dendy says. She also relies on RIS data in a way that may be unusual: She creates the contract radiologist’s monthly invoice. The invoice lists patient names and the code for each procedure that the radiologist interpreted. “I break it down by modality, and by how many studies (and what types) he read most,” Dendy says. What she is doing is creating the radiologist’s bill. She sends it to him, and then he approves it. “It is backwards, but it’s black and white,” she says. Large-practice Models Data mining, as a management tool, can be even more useful when applied to a large radiology practice. J. Keith Radecic is CEO of Infinity Management, LLC, the management arm of Radiology Alliance, PC, in Nashville, Tenn. Radiology Alliance, with 48 radiologists, is the largest private radiology practice in Tennessee. Its clients include three hospitals and more than half a dozen outpatient imaging centers. The RIS used by Alliance is interfaced with the HIS and PACS at Alliance’s various clients, and an IT staff of four at Infinity makes sure that those interfaces are performing, Radecic says. For analytical purposes, he relies primarily on a RIS reporting tool that assembles the financial data in conjunction with the imaging and demographic data that flow to the RIS. While customized reports get a lot of attention because they can focus on specific patterns or details, Radecic emphasizes the importance of the template-based reports that the RIS generates automatically. He has more than 150 of them that he can choose from monthly. A particularly important template report, he says, is the aged trial balance. This report tracks accounts receivable, showing delinquent payments, where the receivables are with each payor, and what the payment cycles have been. Radecic adds that studying receivables has allowed the practice to lower its turnaround time for receipts from payors from the 50-day range to the mid 40s. “Our goal is the high 30s for 2009,” he says. He also notes that the receivables data, because they can be tracked to the line-item level, have made it easier to prove to insurance carriers that some payments should be given immediate approval, rather than having to go through an appeals process. Other template reports from the RIS reveal specific modality patterns or CPT codes in the aggregate, Radecic says. Among other things, these can be used to assess work RVUs and other efficiencies. “There are such powerful tools in today’s market from the various vendors,” he adds. Like others, Radecic tracks referral patterns. “They go a long way in telling us where our procedures come from,” he says, but the gross figures for a referrer don’t tell the whole story. He drills down to understand why the numbers show what they do. “Do all of a physician’s referrals come from Medicaid? Is he or she giving us all the low reimbursables? We’ll take any patient, but we need to manage our practice. If all we get is Medicaid or self-pay, we’ll go to that physician and say, ‘We’re happy, but all you’re giving us is self-pay. We want all your business.’ We try to convince that physician that if we’re good for some of his or her business, we’re good for all of it,” Radecic explains. He adds that Infinity tracks referrals by ZIP code to see whether new centers should be opened or a joint venture with a hospital should be initiated. “These are decision-driving models,” he says. Infinity Management also works with its RIS vendor to turn regularly needed reports from the custom variety into those that the RIS does automatically or nearly automatically. Radecic uses the example of a report that combines charge counts, charge dollars, work RVUs, and actual cash receipts. “I used to have to run four reports, coordinate them with one giant spreadsheet based on the physicians’ work at multiple locations, and then calculate their average productivity per day. Now, I have a single reporting tool that gives me all that I need to run that report. I can run, in 25 to 30 minutes, a report that took days before,” Radecic says. There are, of course, some reports that Radecic can get only by creating the data-analysis parameters himself. At present, he says, Infinity is trying to cut down on the after-hours readings that its doctors must perform. While Alliance Radiology uses a night-coverage service from 11 PM to 7 AM, the evening hours before 11 are unpopular with radiologists. “We know that on a typical day, we need 31 doctors to manage the business, but of those, should we schedule two or four in the evening and 29 or 27 in the day?” Radecic asks. He’s using RIS data to find the most efficient evening staffing pattern that also requires the fewest radiologists. This report, he says, will answer a quality-of-life question that will fit into larger staffing puzzles when doctors want varying workloads and schedules. Another area that Radecic continually investigates is the profitability of certain procedures, including radiofrequency ablation and uterine-artery embolization. “We know that uterine-artery embolization is a good service that generates revenue because we’ve run the numbers,” he says. “If you can’t measure it, you can’t manage it.” External Management Not every data-mining effort is based directly on billing or RIS-driven data. At Advanced Radiology Services (ARS) in Grand Rapids, Mich, data from seven different health systems are pulled from the RIS, HIS, and PACS at client facilities. From there, the data flow to servers at the practice’s management service, Strategic Administrative and Reimbursement Services (STARS), LLC, Grand Rapids, Mich, where they are analyzed using business-intelligence reporting software. ARS is the outgrowth of several radiology practice mergers in the Grand Rapids/Kalamazoo area. Representing 115 radiologists, it is one of the largest radiology practices in the country. STARS handles billing and management for ARS. Bill Ziemke, JD, LLM, MBA, CPA, is CFO at STARS. According to Ziemke, incoming medical reports are entered into coding software, where the reports are coded either manually or using electronic intelligence and entered into a billing system. All interventional radiology procedures are manually coded, whereas some of the diagnostic reports are coded using electronic intelligence. Since the billing system itself has limited template-reporting capacity, STARS routes the data from the billing system into a data warehouse. “There is a nightly extract of data fields from the billing system, which are transferred into a data warehouse. The data warehouse can then be used for reporting purposes using business-intelligence reporting software,” Ziemke says. Financial analysts further segment the data, organizing subsets—for example, daily patient, location, physician, and billing figures—into smaller, more manageable data cubes. Cubes, he says, allow data to be analyzed more quickly than if they had to be pulled off the master file. “When the data warehouse is updated, the cubes are also updated,” he explains. Ziemke says that almost all STARS reports are customized. They are designed, created, and updated by financial analysts on staff. Currently, analysts are evaluating the costs of having radiologists on-site after hours. Among data-analysis successes, Ziemke lists stepped-up productivity in the billing department. “We used cost-accounting principles to measure productivity for each area,” he says. “By tracking productivity, we were able to reduce staffing by 10%.” The study took more than number crunching. “We had a cost accountant sit down with the staff doing the work, watch them to measure what they were doing, and discuss outliers to the standard measures with the staff performing the work,” Ziemke says. “Then, we set the productivity standard, which is constantly being reviewed.” The analysts are also searching for innovative ways to measure physician productivity. Simply applying RVUs to radiologists’ output is too imprecise, so ARS is looking at numerous items. Other factors that need to be considered are the amount of administrative responsibility a physician has; where the interpretation is being performed (home, hospital, or off-site location); the modality being interpreted; and the degree of acute need for the studies being interpreted. Another factor that ARS is measuring is the quality of the work. “Our doctors are committed to quality,” Ziemke says, noting that an electronic peer-review system has been implemented to give feedback on quality. “All of these factors need to be considered when looking at and evaluating physician productivity.” Drowning in Data There are so many benefits of data mining that it’s hard to notice the pitfalls, but they exist. In drilling down, there is always the fear of suffocating on data. “You research yourself into a hole,” Radecic says. For example, a practice can study the numbers on a projected service expansion, he says, but until the move is actually made, nobody can answer the most important question: Are the physicians in the new neighborhood going to give the practice referrals? “Excess data make you want to analyze down to the minute level, but you can stymie a business decision because you get so minute you never act. You overanalyze yourself out of projects,” Radecic says. Additional Reading - Is Custom Reporting Difficult?