Get documents about "Fed Reckoning
Document Sample
REPORT OF THE COUNCIL ON MEDICAL SERVICE
CMS Report 4 - I-00
(December 2000)
Subject: Medicare Fraud Analysis
Presented by: Joseph M. Heyman, MD, Chair
Referred to: Reference Committee A
(Richard L. Stennes, MD, Chair)
1 At the 1999 Interim Meeting, the House of Delegates adopted Resolution 222 which called for the
2 AMA to (1) study and report, with a detailed settlement analysis on the extent of Medicare fraud
3 and abuse: (a) in total, (b) by physicians (differentiating inadvertent coding errors from inadequate
4 documentation and true fraud), and (c) all other health care providers; (2) study and report on what
5 elements comprise the government's statistical estimate of Medicare fraud or abuse involving 10%
6 of “all expenditures,” and provide assurances that the government is not including in that
7 “estimate” the government‟s other findings of wrongdoing, which does not include physicians and
8 providers; and (3) study and report on what elements comprise the government‟s actual settlement
9 of Medicare fraud or abuse, and discourage the government‟s policy of reporting “estimates” of
10 Medicare fraud or abuse.
11
12 The following report summarizes the history of the Office of the Inspector General‟s (OIG) Chief
13 Financial Officer (CFO) Audit; presents the findings from an AMA/American Hospital Association
14 (AHA) analysis of the OIG CFO Audit; reviews a recent U.S. General Accounting Office (GAO)
15 study of Medicare fraud; summarizes current AMA policy and previous policy reports; and
16 discusses the implications of these combined analyses.
17
18 BACKGROUND
19
20 Since 1997, the OIG has issued an annual report summarizing its estimates of the percentage of
21 Medicare payments that were the result of “improper billing.” In its initial June 1997 report, the
22 OIG claimed that 14% of Medicare payments were the result of such “improper billing,” totaling
23 $23.2 billion. The study that yielded this estimate was replicated in each of the succeeding three
24 years, generating “improper billing” estimates that totaled $20.3 billion, $12.6 billion, and $13.5
25 billion, respectively.
26
27 From the first iteration, the OIG CFO Audit has been based on a stratified sample of claims for 50
28 Medicare patients from each of 12 Medicare contractor/quarters. For example, in the first OIG
29 CFO Audit, 50 patients were selected from the carrier Pennsylvania Blue Shield with respect to
30 their claims from the first quarter of 1996. Of the 12 contractor/quarters selected, Pennsylvania
31 Blue Shield was the only contractor that was solely a carrier. Four were fiscal intermediaries only,
32 and the remainder were contractors that were both carriers and intermediaries. Thus, the study was
33 based on the experience of three months each for 600 Medicare patients.
34
CMS Rep. 4 - I-00 -- page 2
(December 2000)
1 It should be noted that in determining whether an action may be characterized as fraud, the Health
2 Care Financing Administration‟s (HCFA) Regional Office Manual provides definitions for key
3 terms. In stating that the terms “fraudulently” and “with intent to defraud” are synonymous, the
4 manual says that these terms:
5
6 refer to an intentional deception or misrepresentation which the individual knows to be
7 false and which the individual makes knowing that his deception could result in some
8 unauthorized benefit to himself or some other person. The requisite intent is present if the
9 representation was either knowingly false or fraudulent, or made in reckless disregard of
10 what the facts might be.
11
12 Despite the fairly careful construction of this definition, the bulk of federal government statements
13 about the extent of fraud in the Medicare program have not been as careful. In fact, the AMA‟s
14 response to the year 2000 publication of the OIG CFO Audit stated, “the AMA is committed to
15 effective use of Medicare funds, but releasing a gross dollar amount of „improper billing‟ -
16 especially a questionable dollar amount - is irresponsible grandstanding. It doesn‟t help physicians,
17 and it certainly doesn't help patients.” In addition, although the OIG ritualistically notes each year
18 that it cannot and does not provide a separate estimate of fraudulent billings, the intense media
19 coverage of this report typically refers to the total as the government‟s estimate of Medicare
20 “fraud.”
21
22 AMA/AHA ANALYSIS OF THE OIG CFO AUDIT
23
24 The AMA and the American Hospital Association (AHA) jointly sponsored a review of the OIG
25 CFO Audit process by an outside team of statistical experts, Michael Intriligator, Ph.D. of UCLA,
26 an econometrician, and Rod Little, Ph.D. of the University of Michigan, a classically trained
27 statistician (hereinafter, I&L). Their findings regarding the OIG CFO Audits from 1998 and 1999
28 were as follows:
29
30 The OIG CFO Audits do not explain how claims were classified as proper or improper.
31 The reports do not show what instructions, if any, were given to reviewers. Because there
32 is no documentation as to what is being measured, it is impossible to evaluate the estimated
33 improper billing. In 1998, the AMA submitted a Freedom of Information Act (FOIA)
34 request to the OIG that asked for any and all reports from contractors used in developing
35 the OIG CFO Audit. A report from the medical reviewers should reveal information about
36 how the medical review decisions were made. The OIG and HHS responded to the FOIA
37 request by asserting that such contractor reports do not exist. Nonetheless, the initial OIG
38 CFO Audit includes the following vignette—one of fifteen included to illustrate improper
39 billings:
40
41 A physician billed Medicare for an electrocardiogram and various laboratory tests.
42 After reviewing the medical records submitted by the provider, the medical
43 reviewer concluded the billed services should be denied because the services were
44 performed as part of the beneficiary‟s routine yearly physical examination, which
45 is not a Medicare-covered service.
CMS Rep. 4 - I-00 -- page 3
(December 2000)
1 If this vignette was not reported to the OIG by the medical reviewers, one might conclude
2 that this vignette and the fourteen others were simply fabrications. Without the underlying
3 documentation of how the medical review process was designed or conducted, it is
4 impossible to evaluate whether the medical reviewers correctly determined that EKG and
5 lab tests were performed solely as part of a routine annual physical. The qualifications of
6 this single “medical reviewer” were not reported, hence are unknown.
7
8 The complete statistical and computational methods used by the OIG aren‟t included in the
9 report. According to I&L, the OIG does not even present, much less adequately explain,
10 the statistical and computational methods used in the CFO Audit. Therefore, if one
11 followed the OIG‟s own audit approach of disallowing all payment when there is missing
12 documentation, one might say that the OIG CFO Audits had never happened.
13
14 The CFO Audit fails to identify the necessary weights used in the audit. The OIG sample
15 is biased toward sicker patients. The claims in the OIG‟s sample have a much higher
16 dollar value than those observed in the Medicare universe as a whole. The OIG‟s sample
17 beneficiaries exhibit more claims, more expensive claims, and higher total charges per year
18 than the Medicare population. This situation makes it imperative that the sample be
19 properly weighted. I&L point out that the OIG has not documented this process. In
20 response to the AMA‟s FOIA request, the OIG has claimed that it neither printed out nor
21 saved the weights used in its extrapolations from the sample to national estimate. In effect,
22 HHS‟s lead audit agency has said that its own audits do not require an audit trail.
23
24 The OIG‟s results don‟t come close to a match with HCFA‟s own audit numbers. With
25 respect to hospital desk audits of 5000 cost reports, HCFA found an error rate of less than
26 1% on hospitals under the Medicare Prospective Payment System, compared to the OIG‟s
27 estimate of 11%. Further, the OIG‟s methods do not meet the statistical requirements that
28 HCFA requires of its own contractors.
29
30 The statistical variability reported by the OIG (i.e., plus or minus 20% or more) is
31 excessive given the magnitude of the numbers and their importance to the policy debate.
32 The CFO Audit has major policy implications for patients, physicians, and the Medicare
33 program. Congress and the Administration make far-reaching policy decisions based on
34 the results of the CFO Audit. In light of the CFO Audits‟ impact on policy, the OIG
35 should collect additional samples when performing the audit in order to narrow the margin
36 of error.
CMS Rep. 4 - I-00 -- page 4
(December 2000)
1 RECENT GAO ANALYSIS
2
3 In the Fall of 1999, the GAO initiated a new study of Medicare fraud. It was believed that this
4 report would attempt to distinguish and measure actual fraudulent activities/billings with respect to
5 the Medicare program as opposed to the “improper billing” estimates from the OIG. The results of
6 this effort were presented in testimony before the U.S. House of Representatives, Committee on the
7 Budget, Task Force on Health on July 12, 2000. Unfortunately, the GAO did not either confirm or
8 refute any existing estimates of Medicare fraud. The report did assert, however, that the OIG CFO
9 Audit methodology was not intended to and would not detect all potentially fraudulent schemes
10 that might be perpetrated against the Medicare program.
11
12 Nevertheless, the GAO report revealed two items of interest. First, the total cost of the OIG CFO
13 Audit for 1999 was reported to be $4.7 million. The actual payments made by HCFA on behalf of
14 the 600 Medicare patients included in the 1999 study was $5.4 million. Accordingly, is appears
15 that the OIG spent 87 cents to review every dollar initially paid on behalf of the sample
16 beneficiaries whose claims were included in the sample. While on the face of it, this may not
17 appear to be particularly efficient, the OIG sample amounts to only less than four ten-thousandths
18 of 1% of the annual Medicare beneficiary universe. As a result, a disinterested observer might
19 object to the OIG sample because either it was particularly expensive to audit, or it was literally
20 infinitesimal, or both.
21
22 The second result of interest involved a display presented in the GAO testimony regarding HCFA‟s
23 Fraud Investigation Database (FID). This database includes all referrals for investigation from
24 1993 through Spring 2000. The GAO testimony presented the percentage distribution of
25 allegations across the ten most common of 60 possible categories (collapsed for the report from ten
26 into six aggregate categories). GAO staff in Atlanta, where the study was conducted, indicated that
27 the top ten categories included some 5700 “duplicated” referrals (i.e., a case alleged to have
28 involved two and only two types of fraud would be entered in the database—and hence counted—
29 twice). Even ignoring the fact that the file counts include duplicates, the distribution presented in
30 the testimony indicates that the maximum number of entries in the FID would be 12,350, or just
31 less than 2000 per year. Given that HCFA contractors process in excess of 900,000,000 claims per
32 year, this yields a fraud allegation rate that at most may be just above two ten-thousandths of 1% of
33 claims per year.
34
35 PREVIOUS AMA POLICY REPORTS AND CURRENT AMA POLICY
36
37 At the 1997 Interim Meeting, the Council on Ethical and Judicial Affairs and the Council on
38 Medical Service reported jointly to the House of Delegates on health care fraud and abuse. With
39 respect to the prevalence of health care fraud, the report concluded that “most of the estimates of
40 the total dollar value associated with fraudulent and abusive health care practices in the United
41 States have been speculative at best, often based on extrapolations from high-profile fraud
42 settlements or fraud demonstration projects.”
43
44 The AMA also has established considerable policy related to fraud and abuse. Policy H-
45 175.981[1] (AMA Policy Database) is firmly committed to eradicating from the Medicare program
46 abuse and fraud—as properly defined, as well as assuring physicians are provided adequate due
47 process and protected from undocumented allegations. Perhaps of most relevance to Resolution
CMS Rep. 4 - I-00 -- page 5
(December 2000)
1 222 (I-99), is Policy H-175.979 which advocates that the AMA seek congressional intervention to
2 halt abusive practices by the federal government and refocus enforcement activities on traditional
3 definitions of fraud, rather than inadvertent billing errors. Similarly, Policy H-445.989 calls for the
4 AMA to monitor statements by government officials relevant to medical care, and to challenge in a
5 highly visible public manner all statements and conclusions not supported by fact or scientifically
6 and statistically based data.
7
8 DISCUSSION
9
10 The AMA has zero tolerance for actual fraud and abuse. In particular, the AMA does not condone
11 any use of the Medicare payment system to either bill for services not provided or provided
12 unnecessarily, or to seek excess reimbursement by exaggerating the nature of services that were
13 provided to Medicare patients. At the same time, the AMA has been dismayed over the continued
14 governmental publication of virtual “guestimates” of the extent of fraud and abuse in the Medicare
15 program, and has so indicated in its responses to the publication of the OIG CFO Audit results that
16 have been released to the press.
17
18 Wherever possible, the AMA has attempted to provide additional scrutiny of the government‟s
19 estimates of fraud and abuse. Internal analyses of the OIG CFO Audit methods have focused on
20 both the methodological and theoretical approaches repeated each year. As previously discussed,
21 because the OIG CFO Audit sampling frame involved all Medicare contractors, no effort was made
22 to separately audit Medicare‟s Part A and Part B experience. As a result, the “patient histories” in
23 the sample did not necessarily include all of those patients‟ claims for the quarter in question. In
24 fact, of the 600 patient quarters selected for the study for fiscal year 1996, only 288 had any
25 physician claims that had been processed by a Medicare carrier. That is, less than half of the
26 sample had any physician claims. On the other hand, 60% of sampled beneficiaries had either
27 hospital claims or hospital outpatient claims or both. Given that more than 90% of Medicare
28 beneficiaries receive one or more physician services during a year while only about 20% are
29 hospitalized, this again calls into question the representativeness of the sample.
30
31 Materials supplied in response to AMA‟s FOIA request indicate that the OIG tabulated two
32 different totals with respect to improper billing. The first involved the total value of individual
33 items that were suspected to involve improper billing. The second involved the total value of all
34 claims that may have involved one or more, but not necessarily all, individual line items suspected
35 to involve improper billing. (To illustrate, one might consider a $100 claim with two line items-on
36 for $40 and a second for $60. If only the first line item was alleged to involve improper billing, the
37 total value of suspected line items would be $40, while the total value of claims with one or more
38 suspected line items would be $100.) In the unweighted sample data provided by the Department
39 of Health and Human Services (HHS), the total value of questioned claims was roughly double that
40 of the total value of questioned individual items. Accordingly, it cannot be determined which of
41 these two measures was included in the OIG CFO Audit.
CMS Rep. 4 - I-00 -- page 6
(December 2000)
1 AMA staff did meet with representatives of the OIG, in May 1998, to discuss the statistical
2 methodology used in the CFO Audit. Also attending that meeting on behalf of the AMA, were
3 statistical and economic research experts from the consulting firm of PricewatershouseCoopers. At
4 this meeting, the OIG staff presented materials on the statistical basis of their sampling frame that
5 had not been included in any of their other reports. However, in response to the suggestions from
6 AMA staff and the PricewatershouseCoopers consultants for a simple approach to gauge the
7 representativeness of the sample, the OIG responded as follows:
8
9 While the additional information you requested may have provided an interesting academic
10 exercise, it would not change the results of the audit nor would it affect the CFO audits that
11 are currently underway. Additionally, developing and assembling the requested
12 information would be an inordinately labor-intensive undertaking, requiring resources
13 beyond those available. Therefore, we will not be providing the requested information.
14
15 The AMA also has attempted to cooperate with the federal agencies involved in operating or
16 monitoring the Medicare program. Repeated efforts have been made to provide professional input
17 into ongoing federal efforts regarding detecting and eliminating fraud and abuse in federal health
18 care programs. Most notably, in February 2000, a joint meeting was held with the AHA to present
19 the results of the review of the OIG CFO Audit to representatives of GAO, the Congressional
20 Budget Office, and the Medicare Payment Assessment Commission. While the consensus of those
21 in attendance was to concur with the critical review of the CFO Audit methods and results, none of
22 the government representatives thought that their respective agencies had any role in suggesting the
23 OIG should respond to those criticisms and improve its approach.
24
25 It has been suggested that there are forces in society that are pushing hard to deprofessionalize
26 physicians and re-make them as just “workers” in the health care system. The notion that all
27 physicians should be monitored regarding possible fraud and abuse because a few physicians put
28 self-interest above patient interest strikes directly at the ideal of professionalism maintained by the
29 AMA since 1847. This “big brother” approach is illustrated in the inflexibility of the evaluation
30 and management (E&M) guidelines; the “Who pays? You Pay” campaign initiated by HHS to
31 recruit seniors to report any and all suspicions about their doctors; and the replication of the OIG
32 CFO Audit without regard to possible refinements, or to opening its subjective process of medical
33 care review to greater professional scrutiny.
34
35 The Council on Medical Service believes that any amount of health care fraud is too much fraud.
36 In particular, fraudulent behavior by physicians is clearly an ethical violation for which there can
37 be no excuse. However, while the existence of health care fraud is a problem, the Council does not
38 believe that this justifies any and all efforts at eliminating fraud. The Council has found that
39 existing national estimates of fraud in the Medicare program are unreliable. Yet, those estimates
40 are cited to justify additional government expenditures “to correct the problem,” potentially taking
41 funds away from additional patient care services. Further, the environment of fear of fraud now
42 overly impinges on the practice of medicine. Physicians are now being advised to implement anti-
43 fraud compliance programs with no evidence of either their effectiveness nor of the costs in
44 additional practice expense and time taken away from direct patient care.
45
CMS Rep. 4 - I-00 -- page 7
(December 2000)
1 RECOMMENDATIONS
2
3 The Council on Medical Service recommends that the following be adopted and the remainder of
4 this report be filed:
5
6 1. That the AMA publicize to physicians, Congress, and the general public the shortcomings and
7 bureaucratic arrogance of the methods employed by the Office of the Inspector General Chief
8 Financial Officer Audit. (Directive to Take Action)
9
10 2. That the AMA strongly urge the U.S. Congress to establish a public technical advisory panel to
11 examine the methodology and results of the Office of the Inspector General Chief Financial
12 Officer Audit. (Directive to Take Action)
13
14 3. That the AMA strongly urge the U.S. Congress to direct the Department of Health and Human
15 Services to measure the effects on physician practice costs of introducing compliance programs
16 and to adjust all federal payment levels upward to reflect those costs. Such study should also
17 verify or refute the alleged deterrent effects of compliance programs as proposed by the Office
18 of the Inspector General. (Directive to Take Action)
19
20 4. That the AMA continue to work with federal agencies to improve both their efforts to identify
21 and deter health care fraud and abuse and their estimating and reporting techniques with
22 respect to the extent of health care fraud in federal programs. (Directive to Take Action)
23
24 5. That the AMA reaffirm Policy H-445.989, which calls for the AMA to increase its efforts to
25 monitor statements by government officials relevant to medical care and to challenge in a
26 highly visible public manner all statements and conclusions not supported by fact or
27 scientifically and statistically based data. (Reaffirm HOD Policy)
28
29 6. That AMA staff review all possible legal actions that may be taken to challenge the continued
30 release of inaccurate statistics in the Office of the Inspector General (OIG) Chief Financial
31 Officer (CFO) Audit. Subsequent to such review, the AMA shall undertake the most expedient
32 and effective, if any, legal means to challenge release of the OIG CFO Audit statistics, and
33 report its efforts to the House of Delegates at the 2001 Annual Meeting. (Directive to Take
34 Action)
Related docs
