Accuracy of Medicare Claims Processing
The objective of the audit was to examine the accuracy of Medicare claims processing, including the adequacy and operation of relevant manual and system processes. The audit assessed the:
- adequacy and operation of relevant manual and system controls used to support the reliable processing of Medicare claims, and
- accuracy of the assessing and processing of Medicare claims, using Computer Aided Audit Techniques (CAATs).
1. The Medicare programme (‘Medicare') was introduced in 1984 to provide affordable and accessible health care to eligible Australian residents. It provides access to medical and hospital services for all Australian residents and certain categories of visitors to Australia.
2. Medicare Australia (MA), which is part of the Human Services Portfolio1, is the agency responsible for administering the delivery of Medicare2. In 2006–07, there were 21.1 million people enrolled in Medicare and a total of $11.8 billion in Medicare benefits were paid with the processing of claims relating to 257.9 million Medicare services3. This number of services and customers makes the Medicare claims processing system one of the highest volume transaction systems in Australia.
Processing of claims
3. To facilitate ready access to Medicare, MA has expanded over time the range of claiming and payment methods (or channels) that are available to both patients and service providers. To achieve this, MA has been a relatively early adopter of emerging technologies and, as a result, the agency has a dynamic Information Technology (IT) environment. The introduction of new systems to support additional claiming and payment methods has often required the retrofitting of, and/or integration with, the existing older systems.
4. Notwithstanding that the number of systems has expanded to accommodate new claim submission and payment methods, the actual assessment of a claim, irrespective of how the claim was submitted or how the payment is to be made, is undertaken using a common mainframe based processing system. The design of this processing system, and of the underlying Medicare Claims History File (MCHF) records that it produces, has remained relatively unchanged since its inception.
Audit objective and scope
5. The objective of the audit was to examine the accuracy of Medicare claims processing, including the adequacy and operation of relevant manual and system processes. The audit assessed the:
- adequacy and operation of relevant manual and system controls used to support the reliable processing of Medicare claims, and
- accuracy of the assessing and processing of Medicare claims, using Computer Aided Audit Techniques (CAATs).
6. The design of the IT systems comprising the Medicare processing system was analysed to determine if it promoted accurate assessment and processing of Medicare claims. CAATs were used to assess the accuracy of a sample of the Medicare claims processed in a two week period, involving some 6.5 million claimed Items4.
7. The focus was on the mainframe based common assessing processing system, and the supporting processes, that are used to assess all Medicare claims irrespective of what method was used to submit or pay the claim. The validity of the Medicare Consumer Database, which is used to determine whether a patient is a ‘valid Medicare' patient, was not tested by this audit.
8. The overall accuracy of Medicare claims processing was demonstrated by the results of the ANAO's testing of a sample of claims. The ANAO's sample was drawn from Medicare claims submitted and processed in a two week period in late 20065 and consisted of some 3.7 million claims, which covered 6.5 million Items. The testing found all claims in the sample to be for valid patients and for valid Medicare Items. The schedule fee identified by MA's system as payable for each Item claimed, with one exception6, was also found to be correct7, both before and after the annual change to fees that occurred on 1 November 2006.
9. The majority8 of Medicare claims are assessed automatically using system based business rules, with less than two per cent of claims processed with manual intervention by SOs9. The ANAO found that the relevant system controls are generally adequate to support reliable processing of Medicare claims. Manual intervention is required for claims where the patient's particular circumstances mean that it is either difficult or not possible to code system based rules that can by themselves assess the claim. SOs are supported by system controls and guidance material10 to mitigate the risk that they will make an incorrect assessment or other error. However, there is a need to improve the controls relating to the configuration of business rules so as to ensure that the system does not allow for the manual override of any business rules which are legislatively based.
10. MA has established a number of activities that aim, over time, to ensure, monitor, maintain, and improve the quality of processing of Medicare claims. These quality activities are of particular importance where manual processing of a claim is involved. While these quality processes are generally effective, there are opportunities for MA to improve the robustness and, possibly, the efficiency of these quality processes.
11. Given the scale of transactions processed by MA it is critical that the underlying systems process transactions accurately and in a timely manner. The results of audit testing of claims11 processed over a two week period indicate that the MA information technology systems and complementary support and quality activities achieve this end.
Key findings by chapter
System Assessed Claims – Chapter 2
12. All but a very small proportion of Medicare claims are assessed automatically by the Medicare processing system and require no manual assessing intervention by a SO. This is the case whether the claims are submitted electronically, scanned into the system using Optical Character Recognition (OCR) technology, or are manually keyed into the system by a SO. In the audit test sample, 98.92 per cent of claims that were submitted either electronically or scanned into the system, were processed automatically and without any manual SO assessing intervention12.
13. The processing of Medicare claims by the Medicare processing system relies on a set of system based business rules that are defined in a Medicare database known as the Item Fee File (IFF). The business rules contained in the IFF extend beyond the Medicare rules that are explicitly defined by legislation.13 The IFF also contains a number of guidance rules added by MA to cover policy and operational requirements. The explicitly defined legislative rules can generally be easily applied by the system, whereas the guidance rules often require manual assessment of the patient's circumstances. For each Item in the IFF a number of settings defines whether a business rule associated with the Item can be solely system applied or must also be considered manually.
14. Changes to the Medicare programme often result in a need to change the system applied business rules that are stored in the IFF. The use of controlled processes to only allow authorised and appropriate changes to be made is therefore essential to ensuring the ongoing reliability of the system applied business rules. The ANAO reviewed the processes used to make changes, and a sample of changes that had been made, and found that the processes were adequately designed to ensure only authorised changes were made, and this was supported by the sample of changes tested.
15. Where a business rule relating to an Item is configured as ‘guidance only' the system allows for a rule to be overridden manually. ANAO's review of the IFF database identified instances where the configuration of some business rules indicated that they were ‘guidance only' when, in fact, the rule was based on an explicit legislative requirement applying to the relevant Item. In the sample of claims tested by this audit, there were 42 instances where claims were made for Items where an associated business rule was incorrectly defined as ‘guidance only'. A review of each of these claims showed that the claim was valid and no instances were found where the incorrectly categorised rule had been manually overridden. However, the incorrect configuration of business rules that are legislative requirements as ‘guidance only' increases the risk that claims may be incorrectly processed by SOs. The correct configuration of the relevant business rules is a control that could prevent this.
16. In addition to business rules the IFF also contains assessing rulings. Assessing rulings are sections of text associated with some business rules that can be referred to by SOs to provide guidance about how and when to override the specific business rule and pay the claim. Assessing rules are configured as warnings within the IFF, and when triggered result in an onscreen message directing the SO to refer to prior assessing rulings for further guidance when assessing the Item.
17. ANAO's review of the assessing rulings in the IFF identified that the assessing rulings stored in the IFF generally provide clear guidance, which promotes the correct and consistent interpretation and application of the assessing rules by SOs when assessing a claim. However, their currency and validity is not consistently maintained when new rules are added to an Item. For example, ANAO found instances where current assessing rules made reference to other Items that have since been abolished.
Service Officer Assessed Claims – Chapter 3
18. To maintain and improve the quality of processing of Medicare claims MA has established a number of activities that aim, over time, to ensure, monitor, maintain, and improve the quality of processing of Medicare claims. These quality activities are of particular importance where manual processing of a claim is involved. Manual processing can include the manual entry of a claim by a SO and or the manual assessing intervention by a SO that can be required even for claims entered using systems14.
19. The quality activities undertaken by MA include:
- a well defined and understood process to support SOs with manual claims assessing and processing;
- daily testing of a system selected sample of claims from prior day processing (Quality Control System process);
- a systematic review of certain classes of manually overridden business rules (Medicare Data Validation); and
- a defined process to assess and improve the quality of claims, known as the Continuous Data Quality and Improvement framework (CDQI framework).
20. A number of mechanisms exist to support a SO during the assessing and processing of a claim, which include: documented guidance; support from their supervisors; and state level helpdesk support. If a state level helpdesk is unable to clarify an issue it can be referred to a helpdesk run by the Medicare Policy Team in the national office. However, if an issue is resolved by a state level helpdesk it will not be communicated to the national office, or to other state offices, and may exist nationwide but go unidentified by other state offices.
Quality Control System
21. The Quality Control System (QCS) is used by MA to automatically select a sample of Medicare transactions, processed by a sample of SOs on the prior work day, for review by their supervisor. Summary reports from QCS review results are then used by MA management to provide assurance over the accuracy of Medicare claims where manual entry and or operator intervention (assessing) occurred.
22. For the results from the QCS process to be a valid indicator of overall quality of manual claims processing the number of SOs selected daily, and the number of their transactions selected, must be statistically relevant. The statistical relevance of the sample size was not known by MA. As a consequence, the sample size may be too small to reliably indicate overall quality, or alternatively, if the sample is larger than it needs to be, there may be inefficiencies due to wasted effort.
23. Since January 2006, a quality initiative has been underway to monitor the consistency of the QCS manual reviews of claims undertaken by supervisors. The ‘Aim-for-Accuracy' initiative has involved the conduct, progressively, in each Medicare office across Australia of a non routine review to assess the consistency of the reviews performed in the QCS process. Progress reports on the ‘Aim-for-Accuracy' reviews indicated that, during the period the initiative was underway, error rates reduced. The 'Aim-for-Accuracy' initiative highlighted that the quality of the QCS process would benefit from the ongoing monitoring of the Medicare team leaders' QCS reviews.
Medicare Data Validation
24. The Medicare Data Validation (MDV) process aims to provide assurance that information entered on a patient's history is accurate and, if necessary, enables a claim record to be corrected (via a latter day adjustment (LDA)) before being archived. The MDV process contributes to the quality of MA claims processing by adding a timely mechanism to detect age, sex, fee or date of birth anomalies that are confirmed for validity by following up each anomaly with the originating Medicare office. The NSW state office staff members who perform the daily MDV checking for all of MA have a good understanding of the process, which is embedded into their daily operational activities and is performed consistently. However, the MDV process is not integrated with other quality processes in MA.
National Continuous Data Quality Improvement Framework
25. MA has a defined National Continuous Data Quality Improvement (NCDQI) Framework to provide a consistent approach to the analysis and resolution of quality issues that are identified. This NCDQI Framework is supported by a National Continuous Data Quality and Improvement (NCDQI) Section, which provides assistance to the national and state offices through the establishment, implementation and monitoring of projects and initiatives that flow from the NCDQI Framework process. Another key support for the NCDQI Framework is the Medicare programme Continuous Data Quality and Improvement Working Party (CDQI WP).
26. Much of the NCDQI Framework analysis work is initially undertaken by the cross-functional/cross-divisional CDQI WP, which includes staff from all teams involved with the delivery of the Medicare programme. The broad membership of the CDQI WP often enables the working party to identify the true nature of issues during its meetings, and develop strategies for their resolution. The NCDQI framework, through the support of the NCDQI Section and CDQI WP, has contributed to maintaining and improving the overall quality of the Medicare programme.
27. All of the quality activities (as discussed in Chapter 2) were found to contribute towards maintaining the quality of the Medicare programme. However, this was achieved through mitigating specific quality risks that were relevant to a specific part of the Medicare programme. It was found that the activities were undertaken without formal consideration of whether the risks covered by a specific activity were already covered by other quality processes. Indeed, the various quality activities are undertaken in a somewhat isolated manner and without necessarily having regard to each other. There is no mechanism in place to ensure the overall mix of quality activities is the most efficient and adequate mix.
28. The Medicare programme would benefit from MA monitoring whether the overall coverage provided by the current mix of quality activities is adequate and effective. This would address the risk that some quality risks may currently be mitigated by numerous and possibly overlapping controls and assurance mechanisms, while other quality risk areas may be either unmitigated or are only weakly controlled.
Information Systems Audit Analysis – Chapter 4
29. This audit utilised two Information System (IS) audit techniques to contribute towards the assessment of the accuracy of Medicare claims processing; data analysis and systems design analysis.
30. Data analysis was used to test a sample of prior Medicare claims. The claims selected for testing included Medicare claims from all submission methods and from all MA offices. The sample spanned the annual November update to the Medicare Item fees, as this was considered to introduce specific risks that should be tested. The practicality of the computer processing and storage capacity resource demands required to cross-match and test the selected Medicare claim records was also considered. A sample of some 6.5 million15 claimed Medicare Items were selected for testing.
31. The analysis required the validation of Medicare Claims History File (MCHF) records against Medicare's Consumer Directory (CD) and Item Fee File (IFF) databases and a Department of Health and Aging (DoHA) Medicare Benefit Schedule (MBS) dataset. This required obtaining copies of relevant parts of these databases and datasets for analysis on an ANAO database. The records obtained by the ANAO contained no readily identifiable patient information.
32. The analysis of the test sample of claims showed that all claims were made for valid patients, as defined in the Medicare CD. Similarly, all Items claimed were valid Items that existed in the DoHA MBS. The testing of Items did however identify one Item where the fee amount was not updated during the annual November 2006 MBS fee update. It was found that this error was not due to a fundamental breakdown of a process, but rather resulted from a set of circumstances which are unlikely to be repeated.
33. In addition to the data analysis of a sample of claims, system audit techniques were also used to examine the design and operation of the systems used to process Medicare claims, with a view to identifying possible underlying design limitations and logic errors. The information technology systems were found to be capable of adequately supporting current Medicare programme processing requirements, and generally do. However, a design limitation with the current Medicare processing system can result in some information supporting a claim decision being lost.
34. When a Medicare claim is processed some of the details of the claim, including processing information, is recorded in the MCHF. The MCHF was designed to be a record of a patients claim history and not a record of the processing of the claim. The design review of the MCHF identified that where multiple processing indicators and reason codes are generated during the assessment of a claim, they overwrite prior codes and indicators, and as a result only the last of each is retained. These codes represent the record of why a claim may or may not have been paid, and as such are important administrative records that should be retained.
35. The Chief Executive Officer of Medicare Australia provided the following response to the proposed audit report:
Medicare Australia welcomes the assurance provided by the ANAO's report that our information technology systems and complementary support and quality activities achieve the processing of large scale Medicare claims transactions accurately and in a timely manner. This is a positive outcome in providing confidence to the community about the integrity of the Medicare program.
Medicare Australia agrees with the recommendations and audit findings. We are actively taking steps to implement the recommendations. We are committed to continually seeking to improve our business processes, including the electronic lodgement of Medicare claims and the quality of Medicare claims processing.
Department of Human Services
36. The Secretary of the Department of Human Services provided the following response to the proposed audit report:
The Department of Human Services (DHS) welcomes the report by the ANAO and recognises that the Medicare claims processing system is one of the highest volume transaction systems in Australia. DHS notes that, overall, the proposed audit report recognises that the relevant Medicare Australia system controls adequately support reliable processing of claims and that Medicare Australia works continuously to improve access to its services.
DHS notes that Medicare Australia agrees with the recommendations and has already taken action to implement activities to address the recommendations. DHS supports the implementation activities being undertaken.
The ANAO report states that Medicare Australia has established a number of activities that aim to ensure, monitor, maintain and improve the quality of processing of Medicare claims. DHS acknowledges that these are particularly important where manual processing of claims is involved.
DHS recognises that the challenges for Medicare Australia are already supported by system controls and guidance material to mitigate the risk of error, and appreciates the ANAO's assistance in improving the rigor and robustness relating to the configuration of business rules.
37. The ANAO made four recommendations aimed at achieving further improvements in both system and manual processing of Medicare claims and the associated quality systems. Medicare Australia agreed with all four recommendations.
1 Before 1 October 2005, MA operated as the Health Insurance Commission under the Health Insurance Commission Act 1973 (HIC Act). On that date, the Human Services Amendment Act 2005 commenced and the HIC Act became the Medicare Australia Act 1973.
2 In addition, the agency is responsible for the administration of the Pharmaceutical Benefits Scheme, the Australian Organ Donor Register and as part of the virtual agency, the Family Assistance Office, delivers Family Assistance programmes. Across the various programmes that MA administers, the agency processes more than 500 million transactions each year and is responsible for providing over $30 billion in benefits to the Australian public and to health care providers. Medicare Australia Annual Report 2006–07, p. 11.
3 ibid, p. 25. The number of persons enrolled in Medicare includes non-residents (long-term visitors – more than 6 months – and eligible short-term visitors).
4 The Medicare programme provides health insurance that covers the delivery of certain health services. These services are referred to as Medicare Benefit Schedule (MBS) Items, or Items. For Medicare benefits to be payable, the professional services and the fees for the service must be included in the MBS, which requires them to be provided for under Part II of the Health Insurance Act 1973 (as amended)(the Act) and listed in supporting Regulations. The Act provides that Medicare benefits are payable for medical expenses incurred in respect of professional services. The MBS is updated each November to include changes flowing from the changes to the relevant regulations. The majority of changes in the November update are to allow for the annual adjustment of the fees payable for the professional services.
5 The ANAO's sample was drawn from claims processed in a two week period in October/November 2006, which spanned the annual update of the MBS which occurs on 1 November each year. These claims were for services that were delivered by service providers and processed by MA within the two week period selected. Claims processed by MA during that two week period but which related to services delivered outside the period were not included. In addition some classes of claims were also not included. These were claims which had been adjusted after processing (Latter Day Adjusted), rejected claims, and where claims spanned more than one claim record (required where more than 12 Items are claimed together).
6 The error was due to an Item fee in Medicare Australia's Item database not being updated in the 1 November 2006 annual fee update process, and not as a result of a processing logic error (discussed in paragraphs 2.20 to 2.22 in the audit report).
7 The benefit paid to a claimant for a particular Item does not necessarily match the schedule fee. This occurs for a number of reasons including: whether the service is provided by a GP; whether or not the service is provided to a patient in hospital; the concession status of the patient; and whether or not the patient has reached the relevant Medicare Safety Net Threshold. The interaction of the various rules affecting the benefit amount payable can be complex in some cases. In addition, a benefit payable can vary due to the nature of other items claimed at the same time or within a given period. See paragraphs 4.11 to 4.16 in the audit report for further information. As a result, the testing of the actual benefits payable for a single Item could require the analysis of a patient's prior Medicare transactions for a much greater period, even years. This was impractical in the context of this audit. During the processing of a Medicare claim the actual benefit payable and the nominal MBS Item benefit amount used as the basis for the benefit payable calculation is recorded. It was this nominal MBS Item amount that was tested by this audit.
8 Some 98.7 per cent of claims processed in the two week sample tested by this audit were automatically assessed without any assessing intervention by a MA Service Officer (SO).
9 Manual processing can include the manual entry of a claim by a SO and/or manual assessing intervention by a SO that can be required even for claims entered using systems. System entry of claims includes claims originating from electronic sources and bulk bill vouchers scanned using automatic IT based systems.
10 For example, assessing rulings. Assessing rulings are sections of text associated with some business rules that can be referred to by SOs to provide guidance about how and when to override the specific business rule and pay the claim. See paragraphs 16 to 17.
11 For specific details of the scope of the sample and nature of tests performed see footnotes 5 and 7 respectively.
12 Not all claims can be assessed solely through the application of system based business rules, and may require that a patient's specific circumstances be considered through manual assessing intervention by a MA SO.
13 The principal legislation for the Medicare programme is the Health Insurance Act 1973, which is supported by a number of Regulations that are updated annually to implement benefit increases.
14 System entry of claims includes claims originating from electronic sources and bulk bill vouchers scanned using automatic IT based systems.
15 6 499 841.