|
| Sample Newsletters | MarketPlace AIS Products & Services |
|
AIS's Health Business Daily
Featured Story September 29, 2008 Stakes Are Raised for Compliance Auditing as CMS Deploys Advanced Data-Analysis Tools Reprinted from REPORT ON MEDICARE COMPLIANCE, the nation's leading source of news and strategic information on false claims, overpayments, compliance programs, billing errors and other Medicare compliance issues. By Nina Youngstrom, Managing Editor, (nyoungstrom@aispub.com) Several changes in the regulatory and compliance world — including the loss of free government audit data, the mandatory assignment of present-on-admission (POA) indicators and CMS's deployment of sophisticated data analytics to detect payment errors — are raising the stakes for hospitals to develop their own data capabilities. In January 2009 hospitals face the end of the Program to Evaluate Payment Patterns Electronic Report (PEPPER). Through the Hospital Payment Monitoring Program (HPMP), CMS's vehicle to reduce inpatient payment errors, CMS distributed PEPPER data to all hospitals for 13 risk areas. PEPPER informed hospitals when they were below the 10th percentile or above the 75th percentile in billing for each risk area compared to all other hospitals in that state. This allowed hospitals to expend audit resources on obvious hot spots. Though HPMP ended in July, PEPPER data will be available until Jan. 31, 2009. So hospitals are out one big source of free data at the same time as other pressures are mounting. Medicare and Medicaid program-integrity contractors are intensifying audits, with CMS rolling out the national version of the recovery audit contractors (contract awards are slated to begin at the end of September) and the first set of Medicaid integrity contractors. CMS also is developing the "One Program Integrity System Integrator," known as "One PI." CMS's One PI contractors provide data-analysis tools and use data-analysis methods for Medicare and Medicaid fraud-and-abuse detection. And that's just for starters. CMS in July announced that fiscal intermediaries and Medicare administrative contractors were taking over medical reviews from the quality improvement organizations. There are also seven zone program integrity contractors, which will investigate overpayments and fraud in all parts of the Medicare program for their region. What do all these players have in common? They perform data mining, which enables contractors to sort through huge amounts of data to identify improper claims. "This is a new era of using data in the health care marketplace," says Larry Vernaglia, an attorney with Foley & Lardner LLP. "CMS has always had access to tons of data, but now they have new ways to slice and exploit this data both internally and through Medicare contractors." Some compliance experts think hospitals should be matching computer wits with the government. "There is a simple way of doing auditing, but it's the equivalent of walking around the edge of the lake. You have to data mine if you want to dive into the lake," contends Harriet Kinney, organizational integrity manager for Trinity Health, a large Michigan-based nonprofit integrated delivery system. "You need to do this kind of analysis to have a robust compliance program. You have to be able to look at your data and analyze it and let it help you focus your audit work." Trinity has a data warehouse containing Medicare revenue data, clinical data and claims data, among other things. Everyone in the organization can tap into it. Kinney says you don't have to be an IT expert to use these kinds of tools. Just figure out what information you want, and let the tech people manipulate the data to extract it. For example, "knowing PEPPER was going away, we wanted to continue the success we had with that data and share it across the system. So we wanted to recreate it internally or create a better and more comprehensive set of tools for ourselves," says Kinney. She says the tools also are important for conveying salient compliance information to senior executives. But she emphasizes that you cannot make a value judgment based on the data alone. It requires medical-record review to determine an error. It's dangerous to infer errors solely from the data, Kinney says. Using Data to Audit MS-DRGs Here's an example of how hospitals can use data in their compliance auditing. Pneumonia Medicare-severity DRGs (MS-DRGs) are always ripe for scrutiny. Kinney says hospitals may want to pull data for a six-month period for what used to be DRGs 79 and 89. DRG 89 is now MS-DRGs 193 (Simple Pneumonia & Pleurisy with major complications and comorbidities), 194 (Simple Pneumonia & Pleurisy with CCs) and 195 (Simple Pneumonia & Pleurisy without CC/MCC). DRG 79, also known as "complex pneumonia," is now MS-DRG 177 (Respiratory Infections & Inflammations with MCC), 178 (Respiratory Infections & Inflammations with CC), and 179 (Respiratory Infections & Inflammations without CC/MCC). Hospitals could start by looking at how many discharges they had per month for each set of DRGs. Were there spikes in the admissions from year to year? That would be something to look into, trending from year to year. After this initial analysis, data would need to be parsed into more detail, focusing on the principal diagnosis code, the length of stay and the discharge status-code assignment for the complex pneumonia discharges. For example, what do your data say about pneumonia due to Gram-negative bacteria (482.83)? It'sa fairly generic diagnosis unless the strain of bacterium is specifically identified. Right off the bat, hospitals should question one-day stays for medical necessity, especially if the patient is receiving only oral antibiotics. But compliance gets trickier. Physicians generally try to nail down the type of Gram-negative bacteria by ordering a two-part lab test. The first part, a sputum test, takes three days. The second part, a Gram stain culture, takes five to seven days. If the data reveal that the hospital has billed for a type of Gram-negative pneumonia, such as Klebsiella or Staphylococcus aureus pneumonia, after a one-, two- or three-day stay, then there is the possibility of inappropriate principal diagnosis coding, such as coding from lab results without appropriate physician documentation. This could result in a payment error. So a fact pattern like this would trigger an account review on the grounds that there could be potential coding errors. Analyzing the data from this viewpoint may also reveal there is a sudden higher incidence in the reporting of one ICD-9 code over another. Lastly, data may reveal that there is a higher incidence of patients being discharged home after one to two days. Only with chart review would an auditor be able to determine if there was a payment error caused by inappropriate coding or issues of medical necessity, she says. The bottom line: Hospitals are at risk for billing either too-short stays, which may signal a lack of a medically necessary admission, or stays that are longer but not long enough to justify the assignment of a specific type of bacterial pneumonia as principal diagnosis. Kinney also runs data reports to evaluate POA indicator compliance. CMS requires hospitals to attach one of five POA indicators to all principal and secondary diagnoses, and Kinney recommends checking on compliance frequently. Quality and finances are implicated with POA compliance. Start by running a coder-specific POA report for one day. Depending on the findings, these mini-audits may be a useful daily audit tool. The findings can be presented in a simple box. The left and middle columns list, respectively, the POA indicators and what they stand for. These indicators are "Y" for yes (the condition was POA), "N" for no (the condition was not POA), "U" for unknown (the documentation is insufficient to determine if the condition was POA), "W" for clinically undetermined (the provider is unable to clinically determine whether the condition was POA), and "1" (exempt from reporting). Also, leave a blank to account for charts lacking a POA indicator, which could be a simple coder error of forgetting to enter a POA indicator or a "hiccup in the computer system," Kinney says. CMS says it will return claims to providers unpaid if they lack POA indicators, and the claims then have to be resubmitted. The right column of the box should report the principal diagnosis count for that particular day's audit run. There should be a high number in the "Y" column and a low number in the "N" column. If there are more than a handful of "U"s, Kinney suggests that hospitals track down the particulars. Is there one attending physician responsible for most of them? If so, focus education on him or her that stresses the need for clear, consistent documentation. "This is something you can tackle and improve. If you set it up correctly, you can do this audit every day," she says. She also advises running the report for each coder separately. That helps pin down when errors stem from just one coder's misconceptions. The timing is perfect
since CMS stops paying for 11 hospital-acquired conditions that weren't
POA on Oct. 1. Daily audits will help hospitals fix errors before dropping
claims. |
| ||||||||
