OIG Home 
Fraud
Corporate Integrity Agreements Overview
Frequently Asked Questions
1. Reporting of Overpayments
2. Selecting an Independent Review Organization (IRO)
3. Reportable Events
4. CIA Claims Reviews
- Background
- Purpose of Claims Review FAQs
- What information should be reported to the OIG in the Claims Review findings, as part of an entity's Annual Report?
- Is there a standard format for submitting the Claims Review findings to the OIG?
- Does the OIG require certain statistical software to be used when performing a Claims Reviews?
- If a reviewer chooses to use RAT-STATS, what features of RAT-STATS should be used to perform a CIA Claims Review?
- How is RAT-STATS used to select a random sample?
- How is the Variable Appraisals component of RAT-STATS used for the Discovery Sample analysis?
- How is RAT-STATS used to determine the Full Sample size?
- What sampling units should be included in the population?
- What sampling unit should be used?
- What guidelines should the reviewer use to evaluate the sampling units?
- What is the purpose of a Discovery Sample for a CIA Claims Review?
- If a sampling unit in the Discovery Sample has been underpaid and a Full Sample is required, how should the underpaid sampling unit(s) be counted in terms of the unit's dollar difference when determining the Full Sample size?
- When conducting a Discovery Sample, how is the error rate determined?
- In what circumstances should the reviewer review a Full Sample?
- How many units must be reviewed as part of a Full Sample under a CIA Claims Review?
- Can the units reviewed as part of the Discovery Sample be used as part of the Full Sample?
- If a sampling unit in the Full Sample has been underpaid, how should this be assessed when determining the amount of the overpayment?
- If the actual precision for the Full Sample does not meet the 90% confidence and 25% precision threshold, do additional sampling units need to be reviewed in order to achieve the specified confidence and precision levels?
- What if the number of sampling units estimated to yield the required confidence and precision levels is unreasonably large?
- Can the Full Sample size be estimated based on paid amounts rather than overpayments (the dollar difference between what was paid and what should have been paid)?
- When performing a Claims Review, if supporting documentation for the unit that was randomly selected is not available, how should this unit be evaluated?
- Is it appropriate to use "alternate" or "spare" records in the Claims Review?
- If a provider receives payment under a Medicare or Medicaid Managed Care plan, should the managed care claims be included in the Claims Review?
- What is a Systems Review?
- When should a Systems Review be performed?
- What should be reported to the OIG as a result of the Systems Review?
5. OIG Site Visits
- Information
- What is the purpose of a site visit?
- Who is subject to an OIG site visit?
- How are entities selected for an OIG site visit?
- Who from the OIG conducts the site visit?
- How long does a site visit typically last?
- What typically occurs during a site visit?
- Who is present during the employee interviews?
- Does the OIG conduct a claims review during the site visit?
6. Provider Obligations When a Contractor is Suspended by the General Services Administration (GSA)
Answers
1. Reporting of Overpayments
What is the OIG's policy regarding the reporting of Federal health care program overpayments by providers under Corporate Integrity Agreements (CIAs)?
Providers under CIAs must promptly notify the appropriate payor of all identified overpayments and must promptly repay the overpayment amount in a manner consistent with the payor's policies. Although all identified overpayments should be refunded to the appropriate payor, a provider under a CIA does not need to report to the OIG all identified overpayments at the time it reports such amounts to the payor. Rather, the provider must report to the OIG the yearly aggregate overpayment amount in its annual report. Further, the provider must report to the OIG within 30 days all "reportable events" as defined by the CIA. "Reportable Event" generally means anything that involves: (i) a substantial overpayment; (ii) a matter that a reasonable person would consider a potential violation of criminal, civil, or administrative laws applicable to any Federal health care program for which penalties or exclusion may be authorized; or (iii) the filing of a bankruptcy petition.
A reportable event may be the result of an isolated event or a series of occurrences.
2. Selecting an Independent Review Organization (IRO)
Does the OIG endorse or assist a provider in choosing an Independent Review Organization (IRO)?
The OIG does not endorse any proprietary products. Moreover, the OIG will not indicate which IROs it believes are most qualified. It is up to the provider to determine the most appropriate CPA firm, law firm, or consultant to engage as its IRO. However, most CIAs include language under which the OIG has the opportunity to notify a provider that its choice of IRO is unacceptable within 30 days after the OIG receives written notice of the identity of the IRO. Also, if during the term of the CIA the OIG has concerns about the quality of the review or the qualifications or independence of the IRO, the OIG will make those concerns known to the provider and may request that the provider terminate the existing IRO and retain a new IRO.
3. Reportable Events
The OIG's CIAs require that the provider report all “reportable events.” Providers often ask the OIG whether a certain percentage or dollar threshold is appropriate for determining if a violation is reportable. Does the OIG believe such percentage or numeric thresholds are appropriate for determining if there was a “reportable event?”
No. The OIG defines the terms "reportable event" in its CIAs by noting that it could be (i) a substantial overpayment, (ii) a matter that a reasonable person would consider a probable violation of criminal, civil, or administrative laws applicable to any Federal health care program for which penalties or exclusion may be authorized; or (iii) the filing of a bankruptcy petition. Such a situation may be the result of an isolated event or a series of occurrences. In short, a provider under a CIA must determine if an identified matter requires disclosure under the CIA based on the totality of the facts and the context of the surrounding circumstances. The OIG believes that a percentage or numeric threshold simply constitutes the initial step in determining materiality and therefore reliance solely on such a "rule of thumb" threshold is inappropriate and should not replace a detailed analysis of all the relevant facts and circumstances.
4. CIA Claims Reviews
Background:
Most CIA s require that a Claims Review be conducted, either by an IRO or in some cases by the provider, with a verification review performed by the IRO.
The Claims Review procedures require a Discovery Sample of 50 sampling units to be randomly selected for review. If the net financial error rate of those 50 sampling units equals or exceeds 5%, then a Full Sample must be reviewed and a Systems Review must be conducted. The Full Sample must include a sufficient number of sampling units to yield results that estimate the overpayment in the population within a 90% confidence and 25% precision level.
Purpose of Claims Review FAQs:
The following is a list of questions the OIG is frequently asked by providers or their IROs regarding the Claims Review. These questions are intended to provide guidance to all providers under CIA s, regardless of the specific Claims Review provisions used in the provider's CIA.
Q 4.1: What information should be reported to the OIG in the Claims Review findings, as part of an entity's Annual Report?
The requirements for the Claims Review report are described in detail in the Claims Review Appendix to the CIA. As indicated in the Claims Review Appendix, the Claims Review report must include a description of the claims review methodology, statistical sampling documentation, and the claims review findings (including both narrative and quantitative results).
Q 4.2: Is there a standard format for submitting the Claims Review findings to the OIG?
The OIG does not have a standard format for reporting Claims Review findings as part of an Annual Report. The content and length of reports vary based on, among other factors, provider type, provider size, and the complexity of the required review(s). However, as indicated above, the Claims Review Appendix to the CIA specifies the information that should be included in the report regarding the claims review findings, including both the narrative results and the quantitative results
Q 4.3: Does the OIG require certain statistical software to be used when performing a Claims Reviews?
In the past, the OIG required the reviewer to use RAT-STATS, a statistical software package developed by the OIG's Office of Audit Services. RAT-STATS is publicly available in downloadable form through the OIG Web site. The users manual can be downloaded from the same site. Both the software and the users manual are free.
The OIG no longer requires RAT-STATS to be used when performing Claims Reviews. If the reviewer chooses to use a statistical software program other than RAT-STATS, in its report to the OIG, the reviewer should provide a description of the software package used to conduct the Claims Review and should provide the supporting documentation generated from the software package (e.g., random number printouts and sample size estimate printouts) to support its sampling methodology and results.
Q 4.4: If a reviewer chooses to use RAT-STATS, what features of RAT-STATS should be used to perform a CIA Claims Review?
The following components of RAT-STATS should be used to perform a CIA Claims Review:
- "Random Numbers" to draw random samples for both the Discovery Sample and the Full Sample (See response to Question 4.5 for specific guidance on how to use this feature);
- "Variable Appraisals" to calculate the mean and standard deviation of the overpayment amount in the sample (See response to Question 4.6 for specific guidance on how to use this feature); and
- "Sample Size Estimator" to estimate the size of the Full Sample (See response to Question 4.7 for specific guidance on how to use this feature).
Q 4.5: How is RAT-STATS used to select a random sample?
Before RAT-STATS can be used to generate the sample, the sampling unit and the population from which the sample will be selected must be defined. Once the sampling unit and the population have been identified, each sampling unit in the population should be assigned a number (these assigned numbers will correspond to the numbers generated by RAT-STATS). Begin at the "Random Numbers" tab of RAT-STATS and select "Single Stage Random Numbers." The program then asks for:
- the seed number (this field should be left blank unless reproducing a previous sample);
- the name of the audit;
- the number of random numbers to be generated;
- the number of spares to be generated (generally, there is no reason to use spares, because if documentation to support payment of a sampling unit cannot be produced, it should be considered an overpayment);
- the universe range, the number of the first sampling unit (e.g., 1) and the number of the last sampling unit number;
- choose "disk" to save results to a disk or "printer" to print the results.
The generated results will be a list of numbers. The numbers generated by RAT-STATS are the numbers of those sampling units that have been selected for review.
Once the sampling units have been reviewed, a text file detailing the amount of the overpayment for each sampling unit must be created. If there is no overpayment or if there is an underpayment, a zero should be associated with the corresponding unit. Each line of the text file should include the sample unit number, a space, and the amount of the overpayment (dollar sign is not necessary).
Q 4.6: How is the Variable Appraisals component of RAT-STATS used for the Discovery Sample analysis?
After creating the text file of overpayments based on the results of the Discovery Sample, select the Variable Appraisals component of RAT-STATS to calculate the mean and standard deviation of the overpayment amount in the sample. To calculate the mean and standard deviation of overpayments in the sample, the sampling units from the Discovery Sample should be reviewed and a dollar difference determination should be made on each sampling unit (i.e., the difference between what was reimbursed and what should have been reimbursed based on the reviewer's determination). The mean and standard deviation of the overpayment amount in the Discovery Sample is used to determine the Full Sample size.
To use the Variable Appraisals feature:
- choose the "Unrestricted" option;
- input the name and location of the text file (e.g., c:/probe.txt);
- enter the universe size;
- choose "Difference Values" for the data file format (the text file created is composed of the dollar difference of what was paid and what should have been paid); and
- choose the format of the data output (text file, printer, or screen).
Q 4.7: How is RAT-STATS used to determine the Full Sample size?
Once the mean and standard deviation of the overpayment amount in the Discovery Sample have been calculated, the Full Sample size can be determined. Select the "Sample Size Determination" tab on the main menu. Select "Variable Sample Size Determination," "Unrestricted," and "Using a Probe Sample."
RAT-STATS will ask for the following information:
- probe sample format (e.g., text file, excel file, etc.);
- universe size (i.e., the total number of sampling units from which the Discovery Sample was selected);
- required confidence and precision levels; and
- select the format of the data output (text file, printer, or screen).
The results indicate the estimated number of units that should be sampled to achieve the specified confidence and precision levels. The number of sampling units to be selected for the Full Sample should correspond to the 90% confidence and 25% precision levels, as required by current CIAs. Note: Sample sizes vary according to the variability of the overpayments in the Discovery Sample and the population size. Therefore, the results of one Discovery Sample may not be applied to another population.
Q 4.8: What sampling units should be included in the population?
The reviewer determines the sampling unit, keeping in mind that the sampling unit chosen should be consistent with the objective of the Claims Review. For example, if the objective is to examine only a certain type of service, sampling units for other types of services should not be included in the population. Further, only services for which the provider has submitted a bill/claim AND has received reimbursement (full or partial) should be included in the population.
Q 4.9: What sampling unit should be used?
A sampling unit is any discrete unit that can be sampled (e.g., code, line item, beneficiary, patient encounter). For any sampling methodology used, each sampling unit in the population should have a known chance of selection.
The sampling unit is selected and defined by the reviewer and should be consistent with the objective of the Claims Review. The manner in which a provider maintains its records may influence the sampling unit to be used. For example, if a provider keeps all information in the patient's chart and cannot easily retrieve the information by claim or line item, the better sampling unit is the beneficiary (e.g., patient chart or record), because an accurate population based on another characteristic could not be easily determined.
Q 4.10: What guidelines should the reviewer use to evaluate the sampling units?
As part of its Claims Review, a reviewer should ensure that the sampling units were correctly coded, submitted, and reimbursed. Documentation to make these determinations include, but are not limited to: national policies, local policies, CMS program memoranda, Medicare carrier or intermediary manuals or bulletins, medical records, claim forms, and any other supporting documentation. Only those sampling units for which payment (either full or partial) has been received should be included in the population from which the sample is selected.
Q 4.11: What is the purpose of a Discovery Sample for a CIA Claims Review?
The purpose of conducting a Discovery Sample as part of the Claims Review is to determine the net financial error rate of the sample that is selected. If the net financial error rate equals or exceeds 5%, the results of the Discovery Sample are used to determine the Full Sample size. The Full Sample size is based on an estimate of the variability of the overpayment amount in the population from which the sample was drawn. The results of the Discovery Sample allow the reviewer to estimate how many sample units need to be reviewed in order to estimate the overpayment in the population within certain confidence and precision levels (e.g., generally, a 90% confidence and 25% precision level).
Q 4.12: If a sampling unit in the Discovery Sample has been underpaid and a Full Sample is required, how should the underpaid sampling unit(s) be counted in terms of the unit's dollar difference when determining the Full Sample size?
For purposes of calculating the Full Sample size, sampling units that have been underpaid in the Discovery Sample should be considered as a zero (note: if conducting a Discovery Sample, a Full Sample size is only required if the net financial error rate of the Discovery Sample equals or exceeds 5%). This is because the objective of the Claims Review is to determine the amount of overpayments in the population. Given this objective, neither a dollar difference of zero nor an underpayment are considered to be overpayments, thus a zero should be entered into the calculation for that particular sampling unit.
Q 4.13: When conducting a Discovery Sample, how is the error rate determined?
In a Discovery Sample, each sampling unit is evaluated to determine the dollar difference between the amount that was reimbursed and the amount that should have been reimbursed. Once all sampling units have been reviewed, the results of each sampling unit are added together (underpayments may be netted or offset from overpayments). The resulting calculation is the net overpayment. The reviewer divides the net overpayment by the total dollar amount of the sample. The resulting calculation is the error rate.
Q 4.14: In what circumstances should the reviewer review a Full Sample?
When conducting a Discovery Sample, if the net financial error rate of this sample equals or exceeds 5%, the reviewer must conduct a Full Sample. The Full Sample should consist of a sufficient number of sampling units to yield results that estimate the overpayment in the population to be within a 90% confidence and 25% precision level. Please note, if the net financial error rate of the Discovery Sample is below 5%, the review is complete.
Q 4.15: How many units must be reviewed as part of a Full Sample under a CIA Claims Review?
There is no set number of units that the reviewer is required to examine in the Full Sample. The Full Sample size is based on the mean and standard deviation of the overpayment amount as calculated in the Discovery Sample. As a result, the Full Sample must include a sufficient number of sampling units to yield results that estimate the overpayment in the population to be within certain confidence and precision levels (e.g., 90% confidence and 25% precision). The sample size will vary according to the variability of the Discovery Sample and the size of the population.
Q 4.16: Can the units reviewed as part of the Discovery Sample be used as part of the Full Sample?
The OIG will allow, if statistically appropriate, the Discovery Sample (as a whole) to be used as part of the Full sample.
For example, if the reviewer must examine 200 sampling units in the Full Sample, the reviewer may use the results from each of the 50 units in the Discovery Sample. Therefore, the reviewer only has to randomly select and review an additional 150 sampling units. The results of all units reviewed as part of the complete Full Sample should be reported, i.e., 200 sampling units.
Q 4.17: If a sampling unit in the Full Sample has been underpaid, how should this be assessed when determining the amount of the overpayment?
For purposes of reporting the overpayment to the OIG, underpayments may be netted (or offset) from overpayments. However, in terms of repaying the overpayment to the appropriate payor, the provider should consult with that payor as to whether it will allow underpayments to be netted from overpayments for collection purposes.
Q 4.18: If the actual precision for the Full Sample does not meet the 90% confidence and 25% precision threshold, do additional sampling units need to be reviewed in order to achieve the specified confidence and precision levels?
If the sample size of the Claims Review was determined by RAT-STATS or another statistical software package to reach a 90% confidence and 25% precision level, and if the actual precision level exceeds 25% at the 90% confidence level, the reviewer does not need to continue the review. The number of sampling units identified by statistical software for review is an estimate that is based on the results of the Discovery Sample. Because this is an estimate, some samples will achieve a precision better than 25% and some worse than 25%.
Q 4.19: What if the number of sampling units estimated to yield the required confidence and precision levels is unreasonably large?
When there are few errors in the Discovery Sample, it is not uncommon for the Full Sample size to be quite large. The reason for this is because, generally, the fewer errors in the sample, the greater the variability of the sample. When the variability is large, the reviewer needs to review more units to reach a reasonable level of confidence and precision. If the provider believes the Full Sample size is unreasonably large, it should contact its OIG monitor for further guidance.
Q 4.20: Can the Full Sample size be estimated based on paid amounts rather than overpayments (the dollar difference between what was paid and what should have been paid)?
No. Estimating the sample size based on paid amounts will yield an estimate of the sample size needed to determine the actual amount paid, a figure already known. Instead, the objective is to estimate the amount of the overpayment, thus the figures entered into the Full Sample size calculation must be overpayment amounts.
Q 4.21: When performing a Claims Review, if supporting documentation for the unit that was randomly selected is not available, how should this unit be evaluated?
If the provider cannot produce supporting documentation for a unit that was selected, this unit should be considered an error and the total amount paid should be considered the overpayment. It is NOT permissible for the provider to remove this unit from the sample, to replace this unit with a spare, or to consider that the service was properly coded, billed, and reimbursed.
Q 4.22: Is it appropriate to use "alternate" or "spare" records in the Claims Review?
Generally, the use of alternates or spares is not permissible for CIA Claims Reviews, as stipulated in current CIAs. If a record or supporting documentation cannot be located, the reviewer should consider this item (e.g., claim, line item, encounter) to be an error/overpayment.
Q 4.23: If a provider receives payment under a Medicare or Medicaid Managed Care plan, should the managed care claims be included in the Claims Review?
No. Because the payment methodology of managed care plans typically differs from the traditional Medicare and Medicaid payment methodologies, the reviewer should not include managed care claims in its population of claims to be reviewed in the context of a CIA Claims Review.
Q 4.24: What is a Systems Review?
The purpose of the Systems Review is to identify problems and weaknesses that resulted in overpayments. A Systems Review is a 'walk through' of the systems(s) and process(es) that generated the sampling unit in error.
For example, if a sampling unit was overpaid, the reviewer should begin by determining whether it was initially coded correctly. If it was incorrectly coded, the provider should discuss the incorrect code assignment with the person responsible for assigning the code and take appropriate actions (e.g., training). If the code was correctly assigned on the source document, the reviewer should verify that it was correctly entered into the system. If correctly entered into the system, the reviewer may need to review the Chargemaster or other computer software to ensure the code entered was correctly 'cross-walked' to the code that should have been billed. The goal of the Systems Review is to identify at what point the error that resulted in the overpayment occurred and to determine why. The provider should then take any necessary steps to prevent such problems in the future.
Q 4.25: When should a Systems Review be performed?
If a Full Sample is required based on the results of the Discovery Sample, a Systems Review also is required. See Question 4.14 to determine when a Full Sample is required.
Q 4.26: What should be reported to the OIG as a result of the Systems Review?
The Systems Review Report should include the reviewer's observations, findings and recommendations on possible improvements to the system(s) and process(es) that generated the overpayments.
5. OIG Site Visits
Information
The OIG monitors hundreds of providers, practitioners, suppliers, payors and other entities that operate under CIAs. Since 1999, the OIG has conducted site visits as part of its CIA monitoring efforts. The following are frequently asked questions regarding these site visits.
Q 5.1: What is the purpose of a site visit?
The primary purpose of a site visit is to verify the entity's compliance with the terms of its CIA and to provide the OIG with an opportunity to observe an entity's compliance program in practice. The first hand observations obtained while on site provide the OIG with a more accurate and comprehensive assessment of an entity's compliance program. The site visit also offers the entity the unique, one on one opportunity to educate the OIG regarding the entity's operations. Additionally, the OIG has found that site visits help foster increased and more effective communication between the entity and the OIG.
Q 5.2: Who is subject to an OIG site visit?
Any provider, practitioner, or entity currently under a CIA or other integrity agreement with the OIG is potentially subject to a site visit. Additionally, the OIG occasionally conducts site visits to assess an entity's compliance program during the course of settlement agreement and CIA negotiations. Entities previously visited include hospitals, physician offices, nursing facilities, laboratories, third party billing companies, Medicare contractors, ambulance companies, DME suppliers, pharmaceutical manufacturers, and home health agencies.
Q 5.3: How are entities selected for an OIG site visit?
Entities selected for a site visit are chosen both at random and based on specific criteria developed by OIG. Factors considered by the OIG in determining whether to conduct a site visit include: issues raised in CIA annual reports, whether there is a deferred prosecution agreement, the reporting of deficiencies, comprehensiveness of the compliance program, size of operation, provider type, and degree of cooperation when reporting or responding to OIG requests for information.
Q 5.4: Who from the OIG conducts the site visit?
The visits typically are conducted by attorneys and/or program analysts from the Office of Counsel to the Inspector General.
Q 5.5: How long does a site visit typically last?
The average site visit lasts between one and a half to two days. Occasionally, for large entities, site visits may last for two or three days.
Q 5.6: What typically occurs during a site visit?
Each site visit is tailored to reflect the particular structure of the entity and the issues identified by the OIG in monitoring the entity's compliance with the CIA . Therefore, no two site visits are exactly alike. In general, discussions and issues during an OIG site visit often include:
- A presentation by the entity that includes: an overview of the entity's corporate structure and operational organization; an overview of the entity's coding and billing systems; an overview of the entity's compliance efforts; and a discussion of compliance efforts specific to coding, billing and medical record documentation;
- A tour of the pertinent portions of the facility;
- A review of the disclosure log, training sign in sheets, Ineligible Persons screening results and other pertinent documentation;
- Employee interviews regarding the entity's compliance program and its adherence to the terms of its CIA;
- A discussion of the entity's annual report, CIA deficiencies, and corrective actions; and
- A discussion of the OIG's site visit observations.
Because one of the primary purposes of a site visit is to provide the OIG with an opportunity to observe the daily operations of the entity and its compliance program, the OIG does not expect or encourage entities to spend excessive time or resources developing slide presentations or preparing employees for interviews. Instead, the OIG's main recommendation regarding site visit preparation is that the entity arrange to have employees reasonably available for discussions or to answer questions. Typically, the entity's Compliance Officer and often one other member of senior management accompany the OIG representatives throughout the site visit and are available to answer questions. Through this less formal approach, the OIG has found that its site visits cause minimal disruption to the entity's normal business operations.
Q 5.7: Who is present during the employee interviews?
During its site visits, the OIG encourages a candid exchange of information. Thus, the OIG often conducts employee interviews in a one on one format. Employees may, if they so choose, have a representative from the entity present during the interview. The OIG often requests that the entity representative be someone other than the Compliance Officer because questions asked during the interview may specifically relate to the performance of the Compliance Officer and the Compliance Department.
Q 5.8: Does the OIG conduct a claims review during the site visit?
The OIG typically does not conduct claims reviews during site visits. However, the OIG or a duly authorized representative does have the authority in accordance with the OIG Inspection, Audit, and Review Rights section of the CIA to conduct a claims review at any time during the course of the CIA.
6. Provider Obligations When a Contractor is Suspended by General Services Administration (GSA)
If a provider under a CIA is engaged in Federal health care program business with a contractor that is subject to a government-wide suspension imposed by the General Services Administration (GSA), what are the provider's obligations?
As a general rule, the suspension of a contractor by GSA invokes the "Ineligible Persons" provision of a provider's CIA . The CIA would require a provider to remove an Ineligible Person from "responsibility for, or involvement with" the provider's "business operations related to the Federal health care programs." A provider is also required to remove an Ineligible Person from any position for which its compensation would be "paid in whole or part, directly or indirectly, by Federal health care programs or otherwise with Federal funds."