PLAY HARDBALL WITH SOFTWARE
The amount of fraud in the health care system is staggering, but computer analysis can help reduce the level. Here is one company’s strategy.
Fraud in health care can assume diverse forms. According to the National Health Care Anti-Fraud Association, the most common is billing for health care services that were never rendered, either by adding charges to legitimate claims, or by using actual patient names and health insurance information to fabricate claims. Upcoding is the second most common; third is the deliberate provision of medically unnecessary services, which include tests, surgeries, and other procedures.
Other forms of provider fraud include waiving patient deductibles or copayments, soliciting or receiving kickbacks for referrals, unbundling charges, falsifying medical records, and fraudulent diagnosis or misrepresentation of the diagnosis to justify payment. Medical-equipment manufacturers have also gotten into the act. Some have been known to offer “free” products to individuals and then charge insurers for products not needed — and, in some cases, never delivered.
In addition, “rolling labs” have become a concern: Unnecessary and sometimes fake tests are given at retirement homes, shopping malls, health clubs, and other venues, and then billed to plans.
Despite the myriad schemes that exist, the reality is that most health care providers are honest and hard-working. It is important then that any antifraud program distinguishes between those who are making honest mistakes in filing claims and those engaged in unscrupulous behavior.
Despite the advantages of using current technology to combat fraud, few insurers or governmental entities have actually embraced it. Some health insurers rely on their patients to report suspicions of fraud, although reliance on this method alone is a highly flawed tactic. Not all insurance companies, for example, share billing statements with insured members. When those people do receive a statement, some are not willing to take the time to review the details, and others simply lack the ability to do so.
Many insurance companies or their third-party administrators still deal with paper claims. When reviewing and processing paper claims, the discovery of fraud is dependent solely on the watchful eye of a claims processor or even on lucky accidents. While many insurers have begun moving to computerized claims, few analyze the data to flag potential fraud.
Recently, Penny Thompson, program integrity director for the Health Care Financing Administration, testified on Medicare and Medicaid fraud before a house committee.
“Better data systems are key to improving efforts to fight Medicaid fraud, waste, and abuse, but many states have inadequate technological infrastructures and a basic inability to interrogate data bases efficiently to ferret out improper claims,” she said. “A number of states indicate that they need better, more targeted data, to pinpoint areas most likely to foster problems, as well as guidance and technical assistance on acquiring new data systems and other fraud-and-abuse detection tools.”
At Doral USA, technology has proved to be critical in targeting fraud, which is approached as a statistical deviation from the norm. Using “fuzzy logic,” we analyze claims data for trends and aberrations that indicate fraud, waste, and abuse.
The effort begins in the offices of medical providers. They receive proprietary software for filing claims.It not only reduces errors but also provides Doral with the information it needs to monitor claims for key indicators of fraud. Providers without access to a computer may submit paper claims, which are entered, edited, and reviewed by claims staff. This process provides a comprehensive data base and allows for effective monitoring and for analysis of information.
Doral measures quality and quantity components for individual providers and offices, and compares them to other providers in the same state or HMO. This approach employs the concept of community standards, which is readily accepted in the medical and legal communities as a comparative measurement tool to profile providers.
In addition, Doral’s process incorporates the “three-strike rule,” which gives providers three opportunities to change their behavior. This approach is used primarily to identify quality issues, such as treatment that is not consistent with stated symptoms or the improper use of diagnostic tests to determine the existence of a specific medical condition. It has, however, also proved useful in identifying instances of fraud and abuse.
In evaluating practice patterns, we use three statistical analyses that look at provider behavior in a weighted format, individually as well as a collectively. Providers who score high in one report or have a high cumulative score are examined more closely. Borderline variances that may prevent a provider from being identified on any individual report would be flagged on the cumulative report. This approach ensures that all inappropriate tendencies are identified.
The first statistical evaluation measures 31 relationships between medical procedures, comparing the outcomes of various treatments that an office performs with the average outcomes seen by all offices in the network. Offices are assigned a variance level that indicates to what degree the office varies from network norms.
For example, when physicians are evaluated, one measurement counts the number of times a blood test for diabetes is given to a patient at the same office in a specific time period. Another measures how many glucose tests are routinely administered. Some glucose tests, for example, are often available at local drugstores, and the difference between the cost of a self-administered test and one conducted in the physician’s office can be significant.
Therefore, if a physician consistently administers a test that is available at local drugstores at substantially lower cost, and consistently charges significantly more for that test, a red flag would be raised.
Doral also measures 75 medical-procedure codes for frequencies per 100 members. Those codes account for 95 percent of all procedures submitted. The measurements are compared to the average number of procedures in the network. Providers are then assigned a level of variance.
|BENCHMARKING: Frequency of medical procedure codes per 100 patients|
PROVIDER NO. 778900
|Procedure||Count||Code per 100 patients||Code per 100 patients:
All providers in HMO
|Variance code (%)|
|For each medical-procedure code, providers are assigned a level of variance. A high ranking would indicate that an office is doing significantly more tests per 100 patients than other offices in the network. Providers that score high in this report or who have a high cumulative score are examined more closely.|
Doral also measures the relationships between the various treatment codes that are submitted for payment. This measurement looks for upcoding by analyzing the number of procedures for which claims are submitted, and their relationships to other procedures performed in the same office. They are compared to the overall averages of network providers.
For example, one measurement looks at the number of Medicare billings received for orthotic body jackets in relationship to wheelchair pads. A high ranking indicates that a medical equipment supplier may be billing for a significantly higher number of body jackets in relationship to wheelchair pads, and thus may be upcoding wheelchair pads to orthotic body jackets, when compared to the average for all suppliers in the network.
The three-strike approach
We rank providers by their variance level. After each analysis is performed, the five offices with the highest aggregate numbers are selected for further review. We contact the providers and request the charts for cases that have failed the tests performed by the system. Then we review these charts to verify the accuracy of treatment and evaluate whether the treatment is the result of a valid variation in treatment or is consistent with a potential quality concern.
If a medical quality concern is raised, Doral will advise the provider’s office of its findings and recommend a strategy that would change the clinical behavior of the office.
Office utilization is reviewed over a three-month period to determine if the recommendations have resulted in changed behavior. If behavior has improved, no further action is taken. If no change in behavior occurs, the matter is referred to a peer-review committee that will examine the facts and recommend corrective action.
Office utilization is again reviewed over a three-month period to determine if the provider’s behavior changes. If no change occurs, the peer review committee meets with Doral’s chief medical officer to discuss appropriate action. A recommendation may then be made to remove the provider from the network.
Poor management or fraud?
At every stage of these analyses, the provider is given the benefit of the doubt. In some cases, variances are simply a matter of poor practice management or misunderstandings concerning the filing of claims. All providers must be given the opportunity to explain any statistical variances. However, several situations may warrant further investigation for fraud.
Each statistical analysis enables us to flag strong indicators of fraud, such as repeated upcoding or the provision of procedures that appear unwarranted. If a review of sample patient charts fails to yield a justifiable explanation, further investigation is needed.
In some cases, data analysis may reveal continuing practice patterns that indicate fraud. In other cases, the variances found through statistical analysis may be low enough to avoid routine notice. However, continuing variances at a low level may be an attempt to stay off the radar screen and could command a closer look.
Other provider behavior can also give rise to a suspicion of fraud. For example, refusal to explain practice patterns, or failure to enact corrective measures, in conjunction with other indicators, may provide the impetus for further investigation.
All fraud investigations are strictly confidential. Until there is conclusive proof of fraud or abuse, providers and their patients receive no indication that an investigation is under way. An important advantage of using computer technology to control fraud is that the confidentiality of the data can be effectively safeguarded.
Despite the industry’s slow acceptance of technology as a means of combating fraud, waste, and abuse, technology is becoming essential to stopping wrongful behavior and recouping losses.
The recovery of dollars lost to fraud will help maintain affordable premiums for those with private insurance, and will provide more services for those in government-sponsored programs.
Recovered losses alone provide a compelling reason to take advantage of the benefits technology provides.
Our most popular topics on Managedcaremag.com
Paul Lendner ist ein praktizierender Experte im Bereich Gesundheit, Medizin und Fitness. Er schreibt bereits seit über 5 Jahren für das Managed Care Mag. Mit seinen Artikeln, die einen einzigartigen Expertenstatus nachweißen, liefert er unseren Lesern nicht nur Mehrwert, sondern auch Hilfestellung bei ihren Problemen.