Department of Health and Human Services
DEPARTMENTAL APPEALS BOARD
Appellate Division
SUBJECT: Colorado Department of Social Services
DATE: August 19, 1991
Docket No. 90-155
Audit Control
No. A-08-89-00211
Decision No. 1272
DECISION
The Colorado Department of Social Services (State) appealed
a
determination by the Health Care Financing Administration
(HCFA)
disallowing $792,966 in federal financial participation (FFP) claimed
by
the State under Title XIX of the Social Security Act (Medicaid)
for
payments made to Medicaid providers from October 1, 1985 to March
31,
1987. The disallowance was based on a report issued by the Office
of
Inspector General, Office of Audit, of an audit of Medicaid
payments
made by the State's fiscal agent, Computer Sciences Corporation
(CSC).
1/ CSC was under contract with the State to operate a
Medicaid
Management Information System (MMIS) that would, among other
things,
process claims submitted by providers for payment of services
rendered
to Medicaid recipients.
For the reasons stated below, we uphold the disallowance on the basis
that
excessive or improper payments to providers are not medical
assistance under
the Act; to the extent the State received FFP in these
payments, the State
has received an overpayment of federal funds, to be
adjusted under section
1903(d)(2) of the Act. However, we agree with
the State that the amount
of the disallowance should be recalculated to
take into account those
payments which the Board finds were not made in
error (3 of the 24 sample
payments at issue) and to reflect the proper
statistical method for
projecting sample payment errors to the universe
of claims.
Background
On December 20, 1989, the Office of Inspector General (OIG), Office
of
Audit, issued the final report of an audit of Medicaid payments made
by
the State during the period October 1, 1985 through March 31,
1987.
State Exhibit (Ex.) 1. The OIG audited a sample of Medicaid
claims paid
during the relevant time period by the State's fiscal agent,
CSC. The
primary objective of the audit was to determine whether the
provider
payments were allowable Medicaid costs. State Ex. 1, p. 3.
The audit sample was divided into 12 strata, each stratum
containing
claims submitted by different types of Medicaid providers, such
as
pharmacies, physicians, and nursing homes. A sample of 100 claims
was
drawn from each stratum and audited to determine whether the
amounts
paid on the claims were allowable. In the final audit report,
the OIG
listed 24 sample claims in which payment errors were found.
State Ex.
1, Ex. I. The OIG had reduced the number from the 44 payment
errors
previously found in the second draft report. This reduction
occurred
because the State was able to document 8 claims which previously
had
been "missing" or inadequately documented, and because 12 other
payments
to providers of nursing home services, or inpatient or
outpatient
hospital services were in three claims strata which the State
had
reprocessed. See State Ex. 1, Exs. I and II. The State
identified
$15,598,510 in payment errors through the reprocessing and
adjusted the
federal share ($7,799,255 in FFP).
HCFA's final determination on the findings contained in the audit
report
was to disallow $792,966 in FFP, based on HCFA's projection of
the
sample results to the universe of claims. HCFA found sample
payment
errors based on its conclusions that--
1. Claims or history documentation was missing
and, therefore, was
insufficient to support the
claims;
2. Copayments were not deducted as required by the State Plan; and
3. Overpayments of the amount due, duplicate
payments, and
payments to the wrong provider are not
considered "medical
assistance" under the Act.
CSC Ex. 2.
The State and the intervenor (CSC) presented legal arguments regarding
the
propriety of the disallowance action here. They also raised
arguments
relating to the findings on the 24 sample payments that were
identified as
errors in the audit, and relating to HCFA's projection of
certain types of
errors to the universe of claims. CSC also challenged
the statistical
validity of HCFA's method of projecting the sample
errors to the universe of
claims.
Below, we first discuss the general legal issues raised. We explain
why
we conclude that no FFP is available generally under the Act
for
excessive or improper payments to providers and that HCFA's
disallowance
here is not arbitrary or contrary to its own regulations.
We then
present our analysis of the individual sample payments cited as
errors
in the audit report. We uphold HCFA's findings with respect to
21
sample payments and reverse HCFA's findings on 3 sample payments.
Next,
we explain why we conclude that HCFA properly projected manual
claims
processing errors and Medicare "crossover" errors to the universe
of
claims. Finally, we discuss the statistical sampling issue and
state
why we conclude that HCFA should recalculate the disallowance
since
HCFA's calculations were not based on a statistically valid method.
Analysis
I. General legal issues
A. Statutory background
Title XIX of the Act establishes the statutory context for
determining
whether provider payments are allowable for federal
reimbursement.
Title XIX of the Act authorizes federal grants to states to
aid in
financing state programs which provide "medical assistance" and
related
to services to needy individuals. A state must have a state
plan that
meets the requirements of section 1902(a) of the Act, including
setting
forth the methods and procedures for payment for care and
services
available under the plan. Section 1902(a)(30). Section
1903(a) of the
Act, the basic provision governing payment of FFP for
Medicaid, requires
the Secretary to pay each state with an approved plan ". .
. an amount
equal to the Federal medical assistance percentage of the total
amount
expended during the quarter as medical assistance under the State plan
.
. . ." (Emphasis supplied.) Section 1905(a) of the Act
defines
"medical assistance" as "payment of part or all of the cost" of
certain
covered services provided to eligible individuals.
States receive payments each quarter, based on estimates of
their
expenditures. Section 1903(d)(2)(A) provides:
The Secretary shall then pay to the State, in
such
installments as he may determine, the amounts
so
estimated, reduced or increased to the extent of
any
overpayment or underpayment which the
Secretary
determines was made under this section to such State
for
any prior quarter and with respect to which
adjustment
has not already been made.
This Board and the courts have held that, under section 1903(d)(2),
HCFA
may require adjustment of the grant award for the federal share
of
firmly established overpayments to providers of services, even if
a
state has not yet recovered these amounts from the providers. 2/
See,
e.g., Massachusetts Dept. of Public Welfare, DAB No. 262 (1982);
New
York State Dept. of Social Services, DAB No. 311 (1982); Ohio Dept.
of
Public Welfare, DAB No. 637 (1985), and cases cited therein. 3/
The
reason is that excess or improper payments to providers are not
"medical
assistance under the State plan" within the meaning of
sections
1903(a)(1) and 1905(a) of the Act and, therefore, FFP in such
payments
constitutes an overpayment to a state to be adjusted under
section
1903(d)(2).
B. The audit was authorized under
the statute and
applicable regulations.
The State argued that it was unaware of any statutes or regulations
that
authorize an audit such as that performed here. State brief, p.
3.
HCFA, however, correctly pointed out that Public Law 95-454, The
Inspector
General Act of 1978, as amended, provides that:
. . . It shall be the duty and responsibility
of each
Inspector General, with respect to the
establishment within
which his Office is established
. . . to provide policy
direction for and to
conduct, supervise, and coordinate
audits and
investigations relating to the programs and
operations of such establishment; . . . to recommend
policies for, and to conduct, supervise, or
coordinate
relationships between such establishment
and other Federal
agencies, State and local
government agencies, and
non-governmental entities
with respect to . . . all matters
relating to the
promotion of economy and efficiency in the
administration of, or the prevention and detection of
fraud
and abuse in, programs and operations
administered or
financed by such establishment . . .
.
Moreover, Medicaid regulations provide at 42 C.F.R. 430.33 (1988):
(a) Purpose. The Department's Office of
Inspector
General (OIG) periodically audits State
operations in
order to determine whether-- (1) The
program is being
operated in a cost-efficient manner; and
(2) Funds
are being properly expended for the
purposes for which
they were appropriated under Federal
and State law and
regulations.
See also 45 C.F.R. 201.12 (1985-1987).
The purpose of the Inspector General Act and the regulatory
provisions
authorizing OIG audit is to ensure that federal funds are
properly
expended and that programs are run in a cost-efficient manner and
to
detect fraud and abuse in federally funded programs. There is no
doubt
that the statute and the regulatory provisions clearly authorize
the
audit of the State here.
C. HCFA's disallowance action
was not arbitrary or
capricious.
Both the State and CSC argued that HCFA's action to take a
disallowance
based on this audit was arbitrary and capricious. The
State contended
that even though its error rate was very low (approximately
.3%
according to the State), HCFA had singled the State out
for
disallowance.
Both the State and CSC pointed out that HCFA had developed
detailed
requirements and criteria for Claim Processing Assessment Systems
(CPAS)
to examine and evaluate the accuracy of claims processed and
paid
through an MMIS. They argued that the disallowance here is
improper
because the claims processing error rate achieved by the Colorado
MMIS
during the audit period met or exceeded the norms for error
tolerance
under the CPAS. Moreover, the State argued that HCFA requires
a
mandatory CPAS only in states that exceed a certain error rate
in
processing claims under their Systems Performance Review (SPR) for
their
MMIS. 4/ The State and CSC argued that the threshold error rate
under
the CPAS is 1% and annual erroneous payments exceeding $1 million
in
federal funds, whereas the State's error rate during the audit
period
was only .3% and the erroneous payments of federal funds for 18
months
was less than $800,000. 5/
The State also contended that HCFA "knows" of similar or larger
claims
processing errors in other states from those states' SPR results,
yet
Colorado was the only state to receive a disallowance for
erroneous
payments made as a result of claims processing errors. The
State
acknowledged, however, that there are differences between the SPR
and
the audit conducted here.
We find no basis for concluding that the State was somehow
arbitrarily
singled out for this audit and resulting disallowance, even
assuming
that HCFA is not free to exercise something tantamount to
prosecutorial
discretion in selecting audit targets. First, since the
regulations
authorize such audits, the State could not reasonably rely on
any
failure to actually conduct similar audits in other states as a
basis
for thinking it would not be subject to such an audit. Under
the
regulations, HCFA had the authority to request OIG to perform such
an
audit. The circumstances here suggest that due to media
coverage
regarding the fiscal agent's performance, HCFA decided that an
audit
should be performed to determine whether a reported claims
processing
backlog had been cleared and to assure that only allowable
payments had
been made. The OIG was asked to perform the audit and to
determine
whether the expenditures claimed for FFP were allowable.
Neither the State nor CSC provided any evidence that HCFA had
comparable
information about inadequacies in claims processing in another
state and
did not audit. There is no indication from the record or the
applicable
law that the audit here was arbitrary in its conception or
performance.
Both the State and the fiscal agent referred to the 1% error rate and
$1
million level established for the CPAS, but failed to explain how
the
CPAS requirements are relevant to the issue of whether HCFA
could
properly disallow for those claims which are not allowable under
the
State Plan. HCFA clearly explained in establishing these
threshold
standards that they would be used for determining whether a state
was
required to operate a more comprehensive "mandatory" CPAS. 50 Fed.
Reg.
21839 (May 29, 1985); see State Ex. 20. There is no indication
that
HCFA considered erroneous payments up to these thresholds to
be
allowable; indeed, HCFA referred to such payment errors as
"misspent
Federal funds." 50 Fed. Reg. at 21845; see also 42 C.F.R.
431.800(j).
Moreover, the SPRs pertain to determinations of whether a state's
MMIS
system is operating properly so that the state may receive
enhanced
reimbursement for the administrative costs of operating the
system. See
generally 42 C.F.R. Part 433, Subpart C. The
regulatory provisions on
the CPAS and the SPR do not conflict with or
supersede HCFA's authority
to conduct oversight reviews or to request audits
to identify and
recover FFP for provider payments which do not meet statutory
or
regulatory requirements. To conclude otherwise would render
meaningless
HCFA's oversight function which it is required by statute and
regulation
to perform. 6/
We also reject CSC's contention that, in the context of "error
tolerance
rates," the impossibility of operating an error free claims
processing
system has long been acknowledged. CSC cited several Board
cases for
that proposition. First, California Dept. of Health Services,
DAB No.
170 (1981), was the only case cited which specifically dealt
with
improper payments made as a result of claims processing errors. In
that
case, the Board specifically stated that quality control
provisions,
i.e., tolerance levels for errors, have never been applied to
claims
processing errors; therefore, the Board upheld the disallowance.
The
other three cases are inapplicable here. FY 1981 Medicaid
Quality
Control Disallowances, DAB No. 948 (1988) refers to
eligibility
determination errors to which the Medicaid quality control
provisions
and tolerance levels specifically apply. See section 1903(u)
of the
Act; 42 C.F.R. Part 431, Subpart P. Louisiana Dept. of Health
and Human
Resources, DAB No. 580 (1984), and Maryland Dept. of Human
Resources,
DAB No. 358 (1982), dealt with the Title IV-A program, not
with
Medicaid.
Finally, we reject the State's contention that counsel for HCFA made
a
tacit admission that the disallowance decision here was arbitrary
and
capricious. The State pointed to a statement in a letter from
HCFA
counsel indicating that the CPAS and SPR procedures --
do not "delineate or define any accepted or
permitted claims
processing performance threshold,
accuracy level or error
rate tolerance" for purposes
of determining the availability
of Federal financial
participation (FFP) in funds claimed as
"medical
assistance" (i.e., a State's request for FFP in the
cost of claims for Medicaid covered medical services,
as
distinguished from claims for enhanced FFP in
the
administrative costs of operating an approved
MMIS);
however, they are generally relevant to the
issue of
determination of an error rate for CPAS and
SPR purposes.
State Ex. 23.
The State misinterpreted what HCFA counsel stated to mean that HCFA had
no
guidelines or standards for disallowing claims processing errors
and,
therefore, HCFA was acting arbitrarily. Such guidelines or
standards
might be necessary if HCFA were establishing a tolerance level
for
claims processing errors similar to that for eligibility
determination
errors. No such guidelines or standards were required
here. The
absence of any such established tolerance level clearly means
generally
that HCFA should disallow FFP in all payments which are excessive
or
improper due to claims processing errors since they do not
constitute
medical assistance under a state plan.
Thus, we conclude that HCFA was authorized to disallow FFP in excessive
or
improper payments to providers and that HCFA was not arbitrary or
capricious
in deciding to audit the State here and to base a
disallowance on the
identified claims processing errors.
II. Whether HCFA correctly found that 24 sample payments were
erroneous
The parties presented arguments related to the 24 alleged
payment
errors. The State and the intervenor, CSC, concurred with
HCFA's
finding of errors in the following seven sample payments: 410,
680,
764, 796, 824, 945 and 984. 7/ We discuss the parties'
arguments
regarding the 17 remaining disputed payments in sequential order
by
claim type below.
We note here, however, that the State's arguments were based on
several
erroneous assumptions. With respect to many of the payment
errors, the
State had initially responded to the audit report by
acknowledging that
the payments were erroneous. During Board
proceedings, the State for
the first time challenged HCFA's findings on the
basis that the State
was not required to document the allowability of the
payments or that
HCFA had not articulated a sufficient basis for
disallowance.
The State clearly had the burden, however, to document the allowability
of
its payments. The applicable regulations require states to
ensure
proper and efficient payment of claims by maintaining an
accounting
system and supporting fiscal records (42 C.F.R. 433.32 and 45
C.F.R.
Part 74, Subpart H) and to conduct prepayment claims review to
verify,
among other things, appropriate service, nonduplication of claims,
and
that the payment does not exceed any reimbursement rates or limits
in
the state plan (42 C.F.R. 447.45(f)).
In determining what rates or limits were established by the State
plan,
the auditors relied in part on lists which had been provided by
the
State to CSC. The State challenged these findings here on the
basis
that HCFA had not specifically cited any regulations establishing
what
the correct rates or limits were under the State plan. The State
did
not, however, provide any evidence to show that its lists
were
incorrect, nor, indeed, did the State even allege that they
were
incorrect. In our view, HCFA was entitled to rely on lists
provided by
the State to CSC in the absence of any evidence that these lists
did not
reflect the amounts payable under the State plan.
Moreover, given the findings in an audit conducted according to
generally
accepted auditing principles, the State clearly had the burden
of going
forward to refute those findings.
We also note generally that CSC's arguments in large part went to
the
issue of whether CSC was liable to the State for the errors. That
issue
is irrelevant here. The question before us is whether HCFA
correctly
found that no FFP was available in the payments.
CSC argued that, according to the Board's Practice Manual, once
the
appellant has come forward with evidence and argument sufficient to
call
into question the respondent's determination, the burden shifts to
the
respondent to support its determination and to show that it had
a
reasonable basis. 8/ CSC argued that it and the State had
presented
argument and evidence that was sufficient to question
HCFA's
disallowance and to shift the burden to HCFA. CSC contended that
HCFA
cannot carry its burden and cannot rest on vague speculation
and
unfounded speculation.
The burden described in the Practice Manual is twofold. First,
the
appellant has the obligation of coming forward with evidence
sufficient
to support the arguments it makes and sufficient to question
the
respondent's determination. Moreover, the party with knowledge of
facts
pertinent to the Board's decisionmaking and peculiarly within
that
party's knowledge and control, has the burden of supplying
that
information to the Board. If the appellant has come forward with
this
kind of evidence, the burden will shift to the respondent.
However, the
question of whether a party has met this burden is ultimately
for the
Board to make. The Board weighs the evidence presented and
determines
by the weight of the evidence presented whether a particular
argument
has been established by that evidence.
In this case, the parties presented evidence on the individual
claims
cited as errors. The Board weighed the evidence presented for
each
claim in making its determination. The fact that a party may
have
presented evidence, however, does not mean the evidence was
sufficient
to prove the matter it was intended to prove. Consequently,
whether
HCFA or the appellant met its burden of going forward with evidence
was
for the Board to determine here and the Board did so in making
its
findings on the evidence and argument presented on the 24
individual
claims.
We next turn to the individual sample payments at issue. We
have
divided our discussion into six sections related to the six types
of
claims.
A. Claim Type: Pharmacy
Claim 39
The auditors found that this pharmacy claim was overpaid because
the
payments made were higher than could be computed using the rates on
file
with CSC for the type of drug provided on the date the drug
was
provided. State Ex. 1, p. 7.
The State argued that the finding of error for this claim cannot
stand
because the audit report did not explain how the auditors determined
the
"correct" price nor did the report show how the price actually paid
was
inconsistent with the State Plan or with any federal regulation
or
statute. CSC argued that a review of the pertinent documentation
and
pricing requirements confirms that this claim was paid correctly.
CSC
argued that the drug formulary file, which contains the National
Drug
Codes and is updated by the State in accordance with its MMIS
system
design requirements (see CSC Ex. 44, p. 1), shows the State
provided
prices for this particular drug. This document indicated two
different
prices for this particular drug with the same effective date.
The
fiscal agent argued that its selection of the higher of the two
prices
was consistent with the State's MMIS requirements.
Contrary to the State's assertions, the audit report specified why
the
amount paid for this claim was inconsistent with the
applicable
regulations and statute and the State Plan. State Ex. 1, pp.
6-7.
Under section 1902(a)(30) of the Social Security Act (Act), the
State
Plan for medical assistance must --
(A) provide such methods and procedures
relating to the
utilization of, and payment for,
care and services available under
the plan . . . as
may be necessary to safeguard against unnecessary
utilization of such care and services and to assure that
payments
are consistent with efficiency, economy,
and quality of care . . .
.
(Emphasis added.) The audit report also cited the federal
regulations
beginning at 42 C.F.R. 447.200 which prescribe the State
Plan
requirements for setting payment rates to implement, in part,
section
1902(a)(30) of the Act. The general provisions regarding
payment
methods set forth in 42 C.F.R. 447.200 et seq. require that the
State
Plan meet the requirements of this subpart as well as describe
the
policy and methods to be used in setting payment rates for each type
of
service included in the State Plan. 42 C.F.R. 447.201. The
regulations
also provide that the State agency must maintain documentation
of
payment rates and make these available to HHS. 42 C.F.R.
447.203. The
audit report also cites to the regulations in Part 447,
which set forth
specific provisions on the proper computation of payment for
specific
types of services covered under the State's Medicaid State
Plan. The
regulations in 42 C.F.R. Part 447, subpart D, provide for the
proper
computation of payment for drugs. 42 C.F.R. 447.331. The
State
generally may not pay more for prescribed drugs than the lowest of
the
maximum allowable cost of the drug, if any; the estimated
acquisition
cost of the drug; or the provider's usual and customary charge to
the
public. 42 C.F.R. 447.332.
The State did not deny that it created the formulary lists used by CSC
to
pay the claims. Moreover, the State did not deny that the
formulary
list was intended to be used and was used to determine the payment
rate
for drugs pursuant to the statute, the regulations and the State
Plan
provisions.
The documentation indicates that the formulary file listed two unit
prices
effective the same date. The State offered no explanation why
the
higher of the two unit prices should apply. The State did not
allege
and presented no evidence that the amount paid was the amount
properly set
forth under the State Plan and the regulations.
Moreover, CSC presented no evidence, documentation, or substantiation
for
its assertion that paying the drug claim at the higher of the two
listed
prices was consistent with the State's MMIS requirements.
CSC's
documentation merely states a policy (effective 2/06/87) that CSC
would
"[a]ccept all changes and additions to the Drug Master File
through
on-line entry by State staff." CSC Ex. 44, first unnumbered
page. Even
if CSC had shown that paying the higher price was consistent
with what
it was required to do under its MMIS contract with the State,
however,
this fact would go only to the liability of CSC for the overpayment;
it
does not establish the amount of FFP allowable in payment of
the
provider's claim.
Finally, we note that, although a handwritten notation on the bottom
of
the claim printout indicates that the rate was changed to .0935
(the
higher of the two listed per tablet estimated acquisition costs)
on
2/12/86, this does not necessarily indicate that this cost
superseded
the .0891 estimated acquisition cost (EAC) used by the
auditors. The
drug list shows that costs with a later effective date
generally appear
on the list to the left of costs with an earlier effective
date.
Although both the .0891 and .0935 have the same effective date
of
2/03/86, the fact that the .0891 appears to the left of the
.0935
indicates that the .0891 superseded the .0935. CSC Ex. 44.
In the
absence of any evidence to the contrary, HCFA reasonably determined
that
the .0891 EAC applied.
Thus, we conclude that HCFA correctly determined that this claim had
been
overpaid.
Claim 59
This pharmacy claim was denied because neither the original claim
for
services nor a microfilmed copy of the claim was produced for review
by
the auditors.
CSC argued that it had submitted a copy of the microfilmed claim in
its
appeal file at Exhibit 45. That exhibit includes two pages, one
of
which is a computer printout entitled "Division of Medical Assistance
-
MMIS - Glider - OIG Audit." This document has a handwritten
notation
"Sample #59." Close scrutiny of this document reveals that the
name of
the recipient of the services is "Allen," and the amount of the
payment
is $5.43. The second document included in Exhibit 45,
entitled
"Colorado Department of Social Services - Bureau of Medical Services
-
OIG Audit of Colorado Claims Process - Audit Claim Report" indicates
a
recipient name of Cristelli, a claimed amount of $3.97, and a
payment
amount of $0. CSC also submitted a copy of a microfilmed claim
form
(copied backwards) showing the recipient name "Cristelli," with a
claim
amount of $3.97. CSC Ex. 44, p. 5. The audit report shows a
payment
amount of $5.43 for Claim No. 59. State Ex. 1, Ex. I to the
audit
report, p. 1. This supports a finding that sample Claim No. 59
was for
recipient Allen rather than for recipient Cristelli. Thus, we
conclude
that, contrary to what CSC argued, it did not produce a copy of
the
microfilmed claim form for the sample claim.
Therefore, we uphold HCFA's finding on Claim No. 59 since CSC did
not
submit the required source documentation of the provider's
claim
necessary to verify that the claim for Allen was paid correctly.
B. Claim Type: Physician
Claim 402
This physician claim was cited as paid in error on the basis that
a
required copayment was not deducted from the amount allowed.
CSC explained that under the State's MMIS requirements, a
recipient
copayment amount is properly deducted from amounts otherwise
payable to
providers for services rendered to non-exempt Medicaid
beneficiaries
under some circumstances. In the case of a community
mental health
center provider, the State MMIS requirements specify that only
one
copayment deduction for these services is required per date of
service.
CSC Ex. 46, p. 4. CSC argued that it did not deduct a
copayment for the
sample claim because the requisite copayment had already
been deducted
from the same provider for the same recipient for another claim
for the
same date of service for the same provider and recipient.
HCFA did not deny that only one copayment deduction is required per
date
of service. HCFA acknowledged that the claims history provided by
CSC
shows a deduction of a $1 copayment for a service with procedure
code
X0171 provided on 7/16/86, the same date as the sample claim, but
argued
that CSC did not supply the original claim form for this claim.
HCFA
argued that neither the State nor CSC had supplied sufficient
supporting
documentation for their contention that the required copayment
was
deducted for a service on the same date as the sampled claim.
With its reply brief, the State submitted the requisite claim form
showing
a service with procedure code X0171 provided on 7/16/86,
supporting its
contention that the required copayment had been deducted.
State Ex. 21.
Thus, we conclude that this claim was paid correctly and
reverse HCFA's
finding that Claim No. 402 was overpaid by $1.
Claim 407
This physician claim was cited as paid in error on the basis that
an
incorrect rate was used to price the claim, since the payments made
were
higher than could be computed using the rates on file for the
service
provided on the dates the service was provided.
The State argued that the audit report does not show that the
price
actually paid on this claim was inconsistent with the State
Plan,
federal regulation, or the Act. The State argued that the
auditors
selected the "correct price" from documentation that was not
generated
until approximately one month after the claim was actually
paid. State
Ex. 4. CSC argued that this claim was paid correctly,
explaining that
the claim was initially suspended since there was no price on
file for
the procedure code. To remedy this, a joint State/fiscal agent
task
force was formed to determine, among other things, a price for
this
procedure code. In the interim, however, the State instructed CSC
to
pay the affected claims as billed, so long as the price was
not
unreasonable. CSC was instructed on 3/27/86 by the State to load
a
price for this procedure code with an effective date of 7/01/84.
CSC
explained that the claim in question, however, was processed prior
to
3/27/86 and was paid as billed because the prices were not
unreasonable.
HCFA contended that the pricing information for this procedure code
was
first established effective 7/01/83 at $11.75. Consequently, that
was
the effective price on 8/26/85, the date of service. This was
the
pricing figure used by the auditor. HCFA argued that whether the
price
was determined before or after this claim was paid, the State
clearly
intended the price to be retroactive to 7/01/83. HCFA,
therefore,
concluded that since the amount allowed ($30) was not
adjusted
accordingly, the claim was properly cited as an error.
The issue here is not the reasonableness of CSC's actions in paying
the
claim. Rather the central question is whether the claim was paid
in
accordance with the State plan and therefore, FFP is allowed for
the
claim. The evidence presented shows that the correct price under
the
State plan for this procedure code for the date of service (according
to
the information provided by the State to CSC) was $11.75. Absent
any
evidence from the State that the information it provided to its
fiscal
agent was incorrect, we conclude that the amount ($30) paid for
this
claim was in excess of the amount set forth under the State plan.
Therefore, we sustain HCFA's finding that this claim was overpaid
by
$18.25 ($30.00 - $11.75).
Claim 420
This claim was cited as paid in error on the basis that a
required
copayment of $1 per visit was not deducted from the payment on
this
claim.
The State argued that the audit report did not explain how it
was
determined that a copayment should have been deducted. The State
argued
that the audit workpapers indicated that the auditors recognized
that
the documentation in the claim file was insufficient to question
the
nondeduction of the $1 copayment and that the available
documentation
supports a finding that one copayment had already been made and
that a
second copayment would not apply. State Ex. 6.
A computer printout submitted by CSC indicates that for the recipient
in
question certain services were received on the same date at the
same
facility. CSC Ex. 48, p. 2. 9/ This document indicates that
a $1
copayment was deducted for this recipient for service procedure
X0171
provided on 6/6/86 at Jeffco Mental Health Center. A computer
printout
for the sampled claim indicates no copayment was deducted for
service
procedure X0147 provided to the same recipient on the same date by
the
same provider. HCFA Ex. 2. HCFA, however, contended that the
claim
form lists only procedure code X0147 and that the State did not submit
a
claim form referencing procedure X0171 for that same date. CSC Ex.
48,
p. 2. Consequently, HCFA argued that neither the State nor CSC
had
provided sufficient documentation either to support the copayment or
to
support a conclusion that the recipient was exempt from copayment.
We agree with HCFA generally that the original claim form (or
a
microfilmed copy) is necessary to document that a provider submitted
a
claim for a service provided on a particular date. In accordance
with
the statute, the regulations, and the State Plan, the State was
required
to retain adequate source documentation to support the
Medicaid
payments. See section 1902(a)(4) of the Act and 42 C.F.R.
431.17; see
also 42 C.F.R. 74.61 and 42 C.F.R. 433.32(a) (requiring
source
documentation to support accounting records) and 74.20 and 42
C.F.R.
433.32(b) and (c) (retention period for records). The absence of
any
claim form here is significant because HCFA's copy of the
computer
printout showing the copayment deduction of $1 has the
auditor's
notation indicating that this printout had an error either in
the
invoice number (which was paid in a different transaction) or in
the
date of service. HCFA Ex. 1.
Thus, we conclude that the State has failed to document that a
copayment
was deducted for another service rendered on the same date as the
sample
claim. Therefore, we sustain HCFA's finding of error for this
claim.
Claim 445
The payment on this claim was cited as an error on the basis that
the
provider, who assisted in certain surgery, was not paid at the
assistant
surgeon's rate. The audit report stated: "Although the
claim did not
have the assistant surgeon modifier on it, subsequent
processing of the
surgeon and anesthesiologist claims should have resulted
in
identification and correction of the overpayment to the
assistant
surgeon." State Ex. 1, p. 7.
The sample claim was for payment to Dr. Pyle, and the
documentation
indicates that Dr. Pyle was the assistant surgeon on the
procedure. See
HCFA Ex. 3; cf. CSC Ex. 49. Furthermore, the
sampled claim was the
first of the claims from three different physicians
processed for this
same procedure. CSC Ex. 49. When the surgeon's
claim was processed
later, apparently the claim did not hit a computer edit
that would have
identified the sample claim as a duplicate.
The State argued that there is insufficient evidence in the
documentation
to show that this claim should have been denied as a
duplicate or paid at a
different price. CSC argued that this claim was
in fact processed
correctly according to the State's MMIS requirements
(which identified a
physician claim as a duplicate only if the provider
was the same) and that
the characterization of this claim as an
overpayment is a recommendation that
future functions of the State's
MMIS should be changed.
Again, the issue here is not the reasonableness of the fiscal
agent's
payment of the claim under the MMIS. The question is whether
this claim
for an assistant surgeon was paid in accordance with the State
Plan and
whether FFP is allowable for this claim. CSC's documentation
indicates
that the State pricing of physician claims differs depending on the
type
of physician (including the price of physician claims for
assistant
surgeon services). CSC Ex. 49, p. 1. Neither CSC nor
the State
specifically disputed the audit finding that the provider listed on
the
sample claim was reimbursed at a price greater than that to which he
was
entitled under the State Plan. Nor did they deny that the claim did
not
have the assistant surgeon modifier on it. As we explained above,
HCFA
said it was determining the proper payment amounts through
lists
provided by the State to CSC. Absent any showing by the State
that a
different amount applied under the State Plan, we find that the
auditors
correctly found that the claim was overpaid. Furthermore,
there is no
indication that this overpayment was corrected.
Therefore, we sustain HCFA's finding that this claim was overpaid
by
$90.20.
C. Claim Type: Medicare Crossover
Claims 534, 539, 564, and 599
The auditors determined that there was no documentation to support
the
payments made on these claims. These claims (referred to as
"crossover
claims") involved Medicaid recipients who were also eligible
for
Medicare.
The State contended that these claims were unlike any others in the
audit
because the original claims were never submitted to the State or
CSC.
Rather, the claims were submitted to the Title XVIII (Medicare)
fiscal
intermediary. For these Medicare crossover claims, the claim
is
transmitted tape-to-tape to Medicaid. The State argued that it
was
unfair to penalize the State because the Medicare fiscal
intermediary
failed to keep needed records. Colorado indicated that it
recognized it
has a duty to maintain necessary records pursuant to 42
C.F.R.
431.17(b). The State asserted, however, that it had no duty to
create
facsimiles of records which should have been kept by others and
over
which the State has no control, such as the Medicare
fiscal
intermediary. Also, the State contended that the facsimiles are
not
"source documentation" pursuant to 45 C.F.R. 74.61.
CSC contended that the auditors did not conclude that these claims
were
processed incorrectly but only that the differences in media employed
in
the different systems (tape-to-tape versus hard copy) frustrate
the
effort to verify the nature and extent of services rendered to
Medicaid
beneficiaries. CSC also argued that the recommended
disallowance for
these claims is insupportable based on statements in the
audit
workpapers that indicated there was not a sufficient basis
for
classifying these claims as errors for FFP disallowance purposes.
We agree with HCFA that 42 C.F.R. 431.17 requires that the Medicaid
agency
maintain records necessary for the proper and efficient operation
of the plan
and the records must include individual records on each
recipient that
contain information on the provision of medical
assistance. See also 42
C.F.R. 433.32 (requiring that the State Plan
must provide that the Medicaid
agency will maintain an accounting system
and supporting fiscal records to
assure that claims for federal funds
are in accord with federal
requirements). Furthermore, on crossover
claims from Medicare the tape
or the facsimile created from the tape are
the only "source documentation"
available to the State. The State would
clearly have the responsibility
under the regulations either to retain
copies of the tapes or to produce
facsimiles of the claims on which it
relied in making Medicaid
payments. Absent any assurance from HCFA that
it was requiring the
Medicare intermediary to maintain records of
Medicare crossover claims
transmitted to the State, the State could not
reasonably rely on the Medicare
fiscal intermediary to fulfill this
State responsibility. Contrary to
what CSC argued, the issue here is
not the form of the records, but the
failure to retain any record to
document the Medicaid pricing and payment of
the crossover claims.
Failure to have any supporting records for the payments
is a sufficient
basis for disallowing the claims for federal reimbursement.
10/
Furthermore, we find no merit in the allegation that certain statements
in
the audit workpapers confirm there was not a sufficient basis for
a
disallowance. First, statements in audit workpapers may represent
only
the preliminary findings of the auditors. Second, it is not clear
the
cited statements are those of the auditors, rather than of State or
CSC
commenters. See CSC Ex. 50, pp. 3, 6, 9, 13. Finally, the
auditors
merely recommend findings to HCFA. It is HCFA that determines
whether
it will go forward with any particular action. 42 C.F.R.
430.33(b)(3).
Thus, we conclude that HCFA properly determined that these claims were
not
allowable for federal reimbursement.
Claim 552
This claim was cited as paid in error on the basis that a
required
copayment was not deducted from the amount allowed.
The State argued that the audit report does not explain how it
was
determined that a copayment should have been deducted. The
State
pointed to a statement in the audit workpapers that it
believed
indicated that the auditors found the documentation in the claim
file
was insufficient to question the nondeduction of the $1 copayment
and
that the auditors were unable to determine whether the medical
condition
to which Claim No. 552 relates is a copay-exempt emergency
service.
CSC argued that no copay deduction was required for this claim because
the
place of service was an inpatient hospital. CSC indicated that
the
claim line detail in the Audit Claim Report indicates a physician
visit
with an inpatient hospital place of service. CSC Ex. 51, p.
3.
Although CSC argued that this claim was exempt from copay because
the
place of service was an inpatient hospital, CSC did not point to
any
specific information on the claim history form supporting such
a
finding. The claim history form indicates that the provider was
a
physician. No hospital or inpatient provider is listed on the
claim
history form. CSC Ex. 51, pp. 2-3.
HCFA pointed out that this was a Medicare crossover claim and that
under
Medicare the place of service code 3, which appears on the audit
claim
report, is for a doctor's office place of service, as opposed to a
place
of service code 3 in Medicaid which represents inpatient place
of
service. CSC Ex. 51, pp. 2 and 3. CSC did not deny that this
was a
Medicare crossover claim or that the Medicare place of service code
3
relates to physician service in the physician's office. According
to
the information CSC provided, this type of visit requires a
copayment.
CSC Ex. 51, p. 1.
The State also argued that HCFA had not shown that the service was not
an
emergency service exempt from the copayment requirement pursuant to
42 C.F.R.
447.53(b). However, as discussed above, the burden was on the
State to
show why a particular finding is incorrect. Here, it is a
requirement
of federal regulation for the State to specify in its State
Plan the
procedures for implementing the exclusions from the
copayment
requirement. The State did not show how the State Plan
determines
whether medical services constitute "emergency services" so as to
be
exempt from the copay requirement. There is no reason why the
State
could not have submitted the relevant portions of its State Plan
that
explain what constitutes an emergency service. In this instance,
one of
the computer printouts submitted by CSC indicates that the diagnosis
was
a urinary tract infection. CSC Ex. 51, p. 7. Under
Medicaid
regulations, in order to be exempt from copay as emergency
services,
services must be provided after the sudden onset manifesting itself
by
acute symptoms of sufficient severity, including pain, such that
the
absence of immediate medical attention could be expected to result
in
placing the patient's health in jeopardy, serious impairment to
the
bodily function or serious dysfunction of any bodily organ or part.
42
C.F.R. 447.53(b)(4). A urinary tract infection would not
necessarily
fit this description. Consequently, the record does not
support a
finding that this claim was exempt from the copay requirement as
an
emergency service.
Thus, we conclude that HCFA correctly determined that a $1 copayment
is
necessary, and therefore Claim No. 552 was overpaid by $1.
D. Claim Type: Supply
Claim 620
This claim was cited as an error on the basis that it was a
duplicate
payment of an earlier submitted and paid claim for the same
service
provided by the same provider to the same recipient.
The State said that "the earlier claim appears to have been paid
on
September 16, 1985, before CSC became Colorado's fiscal agent."
State's
brief, p. 10. The State contended that the sample claim is
not
identical to the earlier paid claim because the procedure code
is
different (although the State conceded that the written description
of
the service is the same). State Exs. 9.and 10. The State
argued that
there is nothing in the State plan, federal regulations, or
statutes
that requires an MMIS to determine if a claim has been paid under
old
procedure codes used by a former fiscal agent. The State contended
that
the MMIS can read only the procedure codes that are keypunched into
the
system and it cannot react to written descriptions of the services
on
the claim forms. The State, thus, argued that this claim was
processed
correctly and should be removed as an error.
CSC, however, never mentioned the previous claim that was paid
in
September 1985 but instead argued that reimbursement was properly
paid
for two different eyeglass lenses (right and left) that were
provided.
Regarding the sample claim, CSC argued that the claim line
(02)
initially was suspended as an apparent duplicate. CSC Ex.
52. CSC
asserted that on further review the examiner properly overrode
the
duplicate edit since one claim line was for a left lens and the
other
for a right lens. CSC further asserted that it had verified
through a
recipient profile that the history did not indicate any other
claims
which would be in conflict.
The documentation includes two claim forms submitted by the same
provider
(Rainbow Optical) for the same recipient for the same date of
service.
The written descriptions of the procedures are identical (two
lenses and a
frame), but the sample claim uses the procedure code V2100
for each lens and
the other claim uses the code 09231 for each lens.
The sample claim form is
marked "Rebill." State Exs. 9, 10. HCFA
argued that it is clear
from this that the sample claim, adjudicated
6/21/86, is a rebilling of the
first claim and that both claims were
paid, the second one in error.
Again, the question before us is not whether the sample claim
was
processed correctly under the MMIS but, rather, whether the payment
is
allowable. Clearly, if the provider received reimbursement twice
for
the same service, the provider has been overpaid. The State cannot
be
excused from its responsibility for such an overpayment merely
because
it changed fiscal agents. The State had a responsibility for
ensuring
that such a change would not result in duplicate payments.
CSC's theory about what happened is not totally implausible, but is
not
sufficiently supported, particularly in view of the State's
admission
that the earlier "claim" was paid on 9/16/85 by the other fiscal
agent.
Although the State said payment was .under a different procedure
code,
presumably this means under code 09231 rather than code V2100 and
does
not distinguish one lens from the other. Thus, the implication is
that
payment was made on 9/16/85 for both lenses, and no evidence
was
submitted to show that only one lens was paid for at that time. Nor
is
there any evidence which supports a finding that any claim for
the
second lens (item 02 on the claims forms) was suspended. While
the
claims history printout for the sample payment indicates there was
a
previous transaction, the TCN for that transaction has the identifier
00
at the end, rather than the 02 identifier one would expect if it
related
to the second lens. CSC Ex. 52, p. 2.
Finally, CSC failed to present any evidence from the recipient
history
which would support a finding that payment was made only once for
the
second lens.
Therefore, we uphold HCFA's finding for Claim No. 620, involving a
payment
of $11.75.
E. Claim Type: Dental
Claim 768
This claim was cited as an overpayment on the basis that the
third-party
insurer paid the full amount allowed for this service for the
Medicaid
program.
Both the State and CSC argued that this claim was processed correctly.
CSC
contended that this claim was initially suspended with error code
51, which
means that the billed claim indicates Third Party Liability
but there is no
indication that the third party carrier has been billed.
CSC Ex. 53, p.
4. CSC argued that under the Colorado MMIS requirements
for this error
code, it was to pay the claim, overriding the error code,
and forward a copy
to the State to research and update the recipient
eligibility file.
The State contended that a copy of the third party insurer's
form
indicated that the insurer paid $2 and Medicaid paid the remaining
$2.
State Ex. 12. The State contended that Medicaid properly paid
the
amount not paid by the private insurer.
HCFA contended that at the time this claim was submitted for payment
to
Medicaid, the claim history form indicated no third party
liability
payment. HCFA said that the claim was suspended, but CSC then
paid the
Medicaid scheduled amount of $2 for this type of procedure and .sent
the
claim to the State to research and update the recipient file. See
CSC
Ex. 53, p. 4. In its reply brief, the State argued that HCFA did
not
reference any State Plan provision or regulation indicating that $2
is
the maximum allowable charge for this service. The State argued
that
HCFA relied on the claims history form without attempting to
determine
whether the amount paid by CSC was the maximum allowable
charge.
As we have stated previously, the State's records should support
the
Medicaid payments made. See section 1902(a)(4) of the Act and 42
C.F.R.
431.17; see also 45 C.F.R. 74.61 and 42 C.F.R. 433.32(a)
(requiring
source documentation to support accounting records) and 45 C.F.R.
74.20
and 42 C.F.R. 433.32(b) and (c) (retention period for records).
Here,
the third party insurer's form indicates only that the provider
charged
$4, that the insurer paid $2, and that the reason the remaining $2
was
not paid by the insurer was that the charge exceeded the
allowable
benefit. State Ex. 12. Contrary to what the State
alleged, the
insurer's form does not indicate in any way that the Medicaid
payment of
$2 was intended to cover the part of the charge the third party
insurer
would not pay.
Moreover, the remaining documentation is consistent with CSC's and
HCFA's
explanation that CSC made a payment based on an override of the
error code
(leaving it to the State to recover from the third party
insurer), rather
than because CSC determined that Medicaid was liable
for the $2 the insurer
did not pay. The claim form submitted to CSC by
the provider does not
specifically indicate that $2 had been paid by the
insurer for the line item
at issue. CSC Ex. 53, p. 2. The claims
history form shows no
third party payment for this particular service
and indicates that the
provider billed $4 and CSC paid $2. Under the
circumstances of the
override, the most reasonable inference is that $2
is the maximum amount
allowed by Medicaid. While the State argued that
HCFA has not shown
that $2 was the maximum amount payable by Medicaid
for this service, the
burden is on the State to show that HCFA's
findings were wrong. This
mere allegation by the State without any
documentary support is insufficient,
particularly in light of the claims
history documentation here.
Therefore, we sustain HCFA's finding that there was a Medicaid
overpayment
of $2 for this claim.
F. Claim Type: Early and
Periodic Screening,
Diagnosis and Treatment
Claim 837
This payment for EPSDT (Early and Periodic Screening, Diagnosis
and
Treatment) services was cited as an error.
The claim history forms show there were three claims for the
same
recipient for EPSDT services on the same date and for the
same
procedure. Each of the claims was reimbursed at $4.15. The
original
claim, TCN 6013-440-001-40-05, was adjudicated on 3/10/86 and paid
to
Dr. Dunn on 3/12/86. HCFA Ex. 7. The next claim,
TCN
6072-052-0008059-02, was adjudicated on 4/6/86 and paid to
Greeley
Medical Clinic, where apparently Dr. Dunn worked, on 4/7/86.
The third
claim, which was the sample claim, TCN 6160-099-0017-03, was
adjudicated
on 6/21/86 and paid to Greeley Medical Clinic on 6/23/86.
State Exs.
13, 14. CSC also submitted a computer printout which shows a
retraction
of a $4.15 payment, TCN 6245-693-1008022-02. CSC Ex. 54, pp.
1 and 3.
The State argued that the sample claim, TCN 6160-099-0017-03, was not
a
duplicate payment because the payment made under this claim
was
retracted under TCN 6245-693-1008022-02. The State contended that
any
error in payment of TCN 6072-052-8059-02 is beyond the scope of
the
audit because that claim was not selected as a sample claim and
the
subsequently paid sample claim was cured by the retraction.
HCFA argued that the retraction of payment did not retract the
sample
claim but rather retracted the payment made under
TCN
6072-052-0008059-02. HCFA claimed that the 02 at the end of the TCN
for
the retraction identifies a line item on the claim form and
therefore
the retraction must relate to the payment with that same 02
identifier.
Thus, HCFA argued, that the sample claim was still a duplicate
payment.
CSC Exhibit 54 establishes that the claim retracted was more likely
the
sample payment than the payment HCFA said was retracted. The
printout
for the retraction specifically refers to TCN 6160-099-0017 (with
only
the line item identifier being different), and to an adjudication
date
of 6/21/86, which is consistent with the sample payment, rather
than
with the payment to which HCFA said the retraction referred.
Moreover,
even if the retraction listed an 02 line item identifier, rather
than
the 03 line item identifier like the sample payment, the
retraction
refers to a procedure code (for a skin test) which is consistent
with
the sample payment. It is highly unlikely that the provider
submitted
two line item claims for this same EPSDT procedure provided to the
same
recipient on the same date. The documentation the State and
CSC
provided is at least sufficient to shift the burden to HCFA to show
that
it was not the sample claim that was retracted. We find no
substantial
evidence that this was so.
Therefore, we reverse HCFA's finding on Claim No. 837 for $4.15.
Thus,
HCFA may not project this payment as an error to the universe of
claims.
11/
Claim 851
This claim for an injection was cited as an error on the basis that it
was
manually priced at a higher rate than that allowed under the State's
Medicaid
program.
The State argued that the audit report does not explain how the
alleged
correct price was determined nor does it show how the price
paid
violated the State Plan or any federal regulation or statute.
CSC
argued that procedure code 90749 is the code for an unlisted
therapeutic
injection. See also CSC Ex. 55. Therefore, CSC said,
the claim was
suspended during the claim processing in order to be priced
in
accordance with error code 405. The claim was manually priced at
$12.
HCFA indicated that the claim form for this claim shows procedure
code
90749, and then the written description, "H-Flu (Hemophilus)," as
the
service provided. CSC Ex. 55, p. 1. HCFA contended that the
correct
procedure code for Hemophilus Influenza Vaccine was 90737.
HCFA
contended that because the initial claim listed the wrong
procedure
code, the claim was paid at the amount billed of $12, an
incorrect
price. HCFA stated that the correct pricing factor for the
time at
which the service was provided was $4.53, which is then multiplied
by
the medical conversion of 2.35 to give the correct payment amount
of
$10.65.
HCFA presented a computer printout of a procedure inquiry from
the
reference file which, contrary to the State and CSC's
assertions,
indicates a procedure code and pricing factor for Hemophilus
Influenza
Vaccine. HCFA Ex. 9. While the doctor filing the claim
may have used
the procedure code for a therapeutic injection of a type not on
the
procedure code list, it is apparent that a procedure code and
pricing
data for this injection did exist. Apparently, CSC overlooked
the fact
that an appropriate procedure code and pricing factor existed, and
paid
the claim as billed. Absent any evidence that the pricing
information
for this procedure provided by HCFA did not apply to the sample
claim,
we find that HCFA correctly used that information to determine
the
correct price of the claim. The State did not show that this
pricing
information was incorrect or was inapplicable to the service
actually
provided. Indeed, the State did not even allege that the
amount CSC
paid was correct.
Accordingly, we conclude that the auditors correctly determined that
this
claim was overpaid by $2.23.
Claim 862
This claim for payment of administration of oral polio vaccine was
cited
as erroneously paid on the basis that it was a duplicate payment.
The State argued that this claim was not a duplicate of an earlier
payment
because the earlier payment was made to a different physician.
In its reply
brief, the State contended that the issue is not whether it
is more or less
likely that this claim is a duplicate but whether
federal regulation requires
a state to treat similar but non-identical
claims as duplicates. The
State argued that in the absence of a
specific citation to a federal
regulation, there is no basis to question
FFP in the payment of this
claim.
CSC argued that according to the Colorado MMIS Detail System Design,
a
duplicate claim requires:
o same provider o same recipient o same date of
service o
same procedure code o same billed amount
CSC argued that since the claims in question were submitted by
two
different providers they do not constitute duplicates and
therefore
there is no evidence that any improper or unauthorized services
were
rendered. CSC Ex. 56.
HCFA reviewed the two claims and found that the first transaction,
TCN
06230-097-0004027-04, for a Polio Virus Vaccine, Live, Oral resulted
in
a payment of $12 to Dr. Derrington on 9/15/86. HCFA Ex. 10.
The second
payment, TCN 06279-501-0009011-04, for the same procedure on the
same
date to the same recipient was made to Dr. Mahan on 10/28/86. HCFA
Ex.
11. The claim for the second payment was marked "Rebill" at the
top
left of the form. HCFA Ex. 12. HCFA contended that the second
payment
was priced correctly at $13.30 but that $12 of that amount
duplicated
the first payment. HCFA also submitted evidence from the
Pueblo,
Colorado telephone book indicating that Drs. Derrington and Mahan
are
listed at the same address and with the same telephone number. HCFA
Ex.
13. HCFA contended that although the providers listed are
different, it
is unlikely that an oral polio vaccine would be administered to
the same
one-year old recipient on the same day. HCFA further argued
that given
the fact that the second claim for payment was marked "Rebill,"
the most
likely inference to be drawn is that a duplicate payment was
made.
One of the requirements under 42 C.F.R. 447.45(f) for reviewing claims
is
to verify whether the claim duplicates or conflicts with one
previously
reviewed and to check whether the services rendered are
logically consistent
with the recipient's characteristics and
circumstances, such as type of
illness, age, sex, and service location.
The question here is not whether CSC correctly processed the claim
under
the requirements of its contract with the State. The question
is
whether this claim is allowable for FFP. A review of the
claims
documents in the process of prepayment review should have alerted
a
reviewer to the fact that this payment was for the same
service
previously paid. The fact that the providers may be different
is not
enough. The regulations specifically require that claims be
reviewed
for logical consistency with the recipient's characteristics.
Here, it
is illogical that a one-year old child would receive two oral
polio
vaccines on the same day. Given this and the fact that the second
claim
was marked "Rebill," the reviewer should have realized
this
inconsistency if the prepayment review had been properly performed
in
accordance with the regulations. The fact that the second payment
may
not have fit the State's definition of a "duplicate" may explain why
CSC
did not identify the problem, but it does not mean that FFP is
available
in two payments for the same service.
The documentation in the record establishes that the sample claim
resulted
in a payment of $12 more than should have been paid for this
service.
As a result, we sustain HCFA's finding of a $12 overpayment.
Claim 866
This claim was cited as a duplicate payment.
The State argued that the sample claim was paid correctly. The
State
indicated that the auditors had noted that the error in payment, if
any,
occurred only in later claims that were not part of the sample
chosen
for the audit. According to the State, the error alleged here
was in
non-sampled claims, and the sample claim should be removed as an
error.
CSC also argued that the claim was paid correctly.
HCFA argued that the sample claim erroneously showed the age of
the
recipient as being over 21 and that this would indicate that
the
recipient was ineligible for EPSDT services because these services
are
available only to recipients under 21 years of age. CSC Ex. 57, p.
2;
42 C.F.R. 440.40(b). HCFA indicated that this was an error that
should
have been projected whether or not any subsequent payments were
made.
HCFA indicated that subsequent payments were made for the same
recipient,
who was identified as a one-year old. The second payment was
paid on
the same date as the first payment. CSC Ex. 57, p. 3. HCFA
then
contended that a third payment on a physician claim was made on
11/4/86 for
the same procedure to the same recipient on the same date by
the same
provider.
The record presented indicates that the sampled claim was not a
duplicate
payment as HCFA asserted. First, the error in the sample
claim was
apparently a typographical error in the birthdate. CSC Ex.
57, p.
2. This was not the error cited by the auditors or by HCFA in
its
disallowance. Moreover, the State could properly make a
payment,
irrespective of the typographical error in the birthdate, so long as
the
recipient was in fact under 21. HCFA implied that CSC had
erroneously
failed to stop payment for a claim for an individual identified
as being
over 21 and that this type of error could have recurred in the
universe.
If the sample claim is truly representative of the universe,
however,
then the inference is that other claims were properly paid in spite
of
an erroneously listed birthdate.
We also reject HCFA's argument that an error should be found because
the
second claim for payment paid on the same date as the first was
a
duplicate payment. CSC Ex. 57, p. 3. The claims history
document
indicates that the warrant number of the sample claim and the
alleged
second duplicate payment are the same and indicate the same
payment
date. See CSC Ex. 57, pp. 2 and 3. Consequently, it
appears that there
was no duplicate payment between these claims but rather
the second
claim was in fact a correction and or retraction of the sample
claim.
While the record indicates that a later physician claim was apparently
a
duplicate of the sample claim, the State is correct here that the
sample
claim and its correction was a proper payment. 12/
Consequently, we overturn HCFA's finding of error for sample Claim No.
866
for $4.15.
III. Whether HCFA properly projected to the
universe errors in
claims processed
manually and in Medicare crossover claims
The State argued that the audit was conducted to determine if a
large
number of errors had been made by the State's MMIS during the
audit
period. The State contended that MMIS is an almost completely
automated
system, and it is reasonable to assume that, if an error is made by
the
automated system with a particular kind of claim, the system will
repeat
the error when similar claims are processed. The State contended
that
some of the claims cited as errors have nothing to do with the
automated
MMIS because, for example, the claim was priced or processed
manually
and the examiner incorrectly determined the price. See Claim
Nos. 410,
680 and 824. The State claimed there is no reason to assume
that
similar errors would be made in other cases. Consequently, while
the
State agreed that these claims were paid in error, it contended
that
these claims cannot be the basis of projecting errors onto the
entire
universe of claims. The State argued that these claims should be
a
separate stratum so that similar claims could be compared to each
other,
rather than to project errors from these claims onto a
universe
containing nearly all automatically priced claims. Similarly,
the State
argued the Medicare crossover claims, if found to be paid in
error,
should not have been projected onto any universe of claims.
We disagree with the State's contention that the purpose of this audit
was
to examine the operation of the State's MMIS and therefore, there is
no valid
reason to project manual processing "errors" to the universe of
paid Medicaid
claims. The audit report states:
The primary objective of this audit was to determine
if the amount
claimed by the CDSS [the State] for
medical services provided to
recipients was
allowable.
CSC Ex. 13, p. 3.
As HCFA indicated, the stated purpose of the audit was not to examine
the
automated MMIS. In fact, there is no dispute that the State's MMIS
had
not been approved and certified as operational during the time
period being
reviewed by the audit. The universe here consisted of all
Medicaid
claims which were paid by the State during the 18-month period
of the audit,
rather than only those claims which were automatically
processed during this
period.
Moreover, the regulations providing for timely claims payment
recognize
that regardless of whether a claim is processed by an automated
system,
the State is still required to conduct prepayment and postpayment
claims
reviews of all claims. 42 C.F.R. 447.45(f). Furthermore,
it is
understood that even if an automated system is used, the system may
have
edits that will suspend a claim and require that the claim be
processed
manually. It is logical and consistent with the intent of the
audit,
the sample chosen, and the nature of the claims payment process
to
project manual errors to the universe of all the Medicaid
claims
processed and paid during this period. This universe would
include
manually processed claims as well as automatically processed
claims.
We also cannot agree with the State that there is no reason to project
the
errors in Medicare crossover claims to the universe. The State
argued
that, even if these claims were paid in error, the error is the
fault of the
Medicare intermediary and not the State Medicaid agency.
The State seems to
reason that since the Medicare intermediary purged
its files and the State
Medicaid agency did not have either facsimiles
or other documentation to
support the payment of this claim, there is no
reason to believe that any of
these claims were paid in error. The
State, however, overlooks
applicable requirements that the State
Medicaid agency must have source
documentation to support its claims for
medical services for which federal
reimbursement is being sought. 42
C.F.R. 433.32; 45 C.F.R. Part 74,
Subpart H. The requirement for
documentation of claims is simple;
without source documentation, there
is no record that a payment was made as
claimed.
Furthermore, the State has not convinced us that either the
manually
processed claims or the Medicare crossover claims in the sample were
not
representative of the universe of claims. The State did not allege
that
the sample was not drawn in a statistically valid manner.
The
assumption underlying a validly drawn sample is that it
is
representative of the universe from which it is drawn.
Thus, we conclude that HCFA correctly projected the errors in
manually
processed claims and in the Medicare crossover claims to the
universe of
paid Medicaid claims for the audit period.
IV. Whether HCFA's method of projecting the
errors to the universe
of claims was statistically valid
As indicated above, HCFA's sample of the State's claims was a
stratified
sample, with 100 claims drawn from each stratum. (The strata
chosen
were based on the type of provider submitting the claim.) After
the
second draft audit report, the State reprocessed three strata:
nursing
home, inpatient hospital, and outpatient hospital. HCFA had
originally
calculated a point estimate and lower limit based on the 44 errors
found
in all strata. The State then reprocessed the three strata,
however
(and returned FFP in approximately $15.6 million of excess or
improper
payments identified as a result of the reprocessing). These
three
strata had included 12 sample errors. Also, for 8 sample payments
in
which the auditors had found missing or inadequate documentation,
the
State provided documentation which the auditors found satisfactory.
In recalculating the disallowance based on the final report, HCFA
obtained
a point estimate of $3,129,924 by subtracting from the original
point
estimate the point estimates for the 8 sample payments for
which
documentation was later found and for the 12 payments in the
reprocessed
strata. HCFA also calculated a lower limit of the two-sided
90%
confidence interval, which HCFA determined was $1,585,932 ($792,966
in
FFP). 13/ HCFA's calculation was based on use of the precision
figure
(50.67%) calculated for the original audit results.
Neither the State nor CSC challenged HCFA's use of sampling in general
as
a basis for disallowance, nor HCFA's calculation of the point
estimate.
Indeed, the courts and this Board have upheld the use of
statistically valid
sampling methods as a basis for determining a
disallowance amount. See,
e.g., Rosado v. Wyman, 437 F.2d 619, 627-628
(2nd Cir. 1970), aff'd, 402 U.S.
991 (1970); Georgia v. Califano, 446 F.
Supp. 404, 409-410 (N.D. Ga. 1977);
California Dept. of Social Services,
DAB No. 816 (1986), pp. 4-5.
CSC presented an affidavit from a statistical sampling expert
which
challenged HCFA's calculation of the lower limit of the
confidence
interval. He asserted that HCFA had erred in how it
calculated the
lower limit. Specifically, he said that HCFA's use of
the 50.67%
precision figure related to the original point estimate
(calculated
based on 44 sample errors) was inappropriate and inconsistent
with
standard statistical procedures. He stated his opinion that
(assuming
24 errors) the statistically correct projection of the lower limit
for
overpayment would be $87,258 ($43,629 in FFP). CSC Ex. 58.
CSC also pointed out that its expert's projection was similar
to
calculations performed by OIG in August 1989, but apparently
discarded.
CSC Br., p. 35, citing CSC Exs. 34 and 35.
In response, HCFA argued that the conclusions drawn by CSC's
statistician
were erroneous because they were based on 24 projected
errors rather than 44
projected errors. HCFA Br., pp. 40-41.
The Board then issued a number of questions to the parties, noting in
a
preliminary analysis that it appeared that there was merit to
CSC's
expert's position.
CSC's expert provided a detailed response to the Board's questions and
a
supplementary affidavit with citation to the statistical treatises
on
which he relied. CSC submission of 1/12/91, CSC Ex. 66. HCFA
then
submitted an affidavit by the Regional Statistical Specialist, OIG,
who
had advised HCFA on how to evaluate the sample results. HCFA
submission
of 5/3/91, enclosure (Samworth Declaration). In its cover
letter, HCFA
asserted:
[T]here is no material dispute as to the results of
using the
statistical formulae employed by the
parties in this case; rather,
the dispute comes down
to whether it is appropriate in making a
statistical
projection of the total overpayment to include as
errors all the claims cited as overpayments by the agency.
Stated
in its most basic terms, CSC's statistical
expert has made his
projections from a
different--and in the agency's view, an
inappropriate--set of assumptions as to which claims should
form
the basis for the projection and, not
surprisingly, has arrived at
a total overpayment
amount that is lower than that reached by the
agency.
CSC then submitted a point-by-point response by its statistical expert
to
the Regional Statistical Specialist's affidavit.
Based on our analysis of the record, we conclude that CSC's
expert
correctly pointed out a flaw in HCFA's method for calculating the
lower
limit.
HCFA provided absolutely no support for its position that the 44
errors
should be projected to the universe, rather than the 24 errors.
As the
Board indicated in its preliminary analysis:
o With respect to the 8 audit findings which related to missing
or
inadequate documentation where the adequate documentation was
later
found, we see no reason to project those findings to the
universe. If
those sample cases are in fact representative, then it is
reasonable to
infer that adequate documentation would ultimately be provided
for the
corresponding cases in the universe.
o The mere fact that HCFA used the variance percentage
originally
calculated for the 44 errors does not mean that the 44 errors were
being
projected to the universe of claims. HCFA itself subtracted the
best
estimates of the dollar amount of the errors associated with the
8
sample cases where adequate documentation was ultimately found and
with
the 12 sample cases in the three strata which were reprocessed.
14/
Thus, HCFA applied its variance calculation to the figure
representing
the best estimate of the amount of errors associated with only
the 24
remaining sample cases.
HCFA's position here is internally inconsistent in that HCFA
acknowledged
that there should be no disallowance for the individual
sample payments in
which the errors have been corrected, yet argued that
somehow those errors
could be a basis for a projected disallowance.
This uses sampling--which is
simply a means of determining a
disallowance amount--as a justification for a
disallowance which is not
otherwise supportable. If there is no longer
any basis for disallowing
the sample payments, there is no basis for
disallowing corresponding
payments in the universe.
We also base our conclusion that HCFA's calculation of the
disallowance
amount was flawed on the following:
o The Regional Statistical Specialist did not disagree with
CSC's
expert's assertion that the revised point estimate was adjusted
downward
so that it is in effect the point estimate for projecting only
the
remaining 24 errors. Indeed, this is obvious from comparing
the
relevant figures and from the fact that the audit report indicated
it
was removing as errors the reprocessed claims and claims for
which
documentation had ultimately been located. See State Ex. 1
(audit
report); Exs. I and II to the audit report.
o The Regional Statistical Specialist, while stating conclusorily
that
"the 44 errors found were properly projected to the universe
sampled,"
also asserts that the State "was given credit for corrections made
and
documents found. That credit does not negate the significance of
the
remaining errors." Samworth Declaration, pp. 2-3. It is
inconsistent,
however, to use a point estimate based on 24 errors and then
to
determine the confidence interval for that point estimate using
a
precision figure calculated based on projection from 44 errors.
o The only rationale offered by the Regional Statistical Specialist
for
using the 50.67% precision figure was that the sample was designed
and
selected based on a projection of all 12 strata to the universe.
He
expressed a concern that CSC was manipulating the policy of using
the
lower limit through "selective correcting of problems for
different
combinations of strata." Samworth Declaration, p. 3.
While CSC's
expert showed that his result could also be obtained by
projecting the
errors in the 9 strata which were not reprocessed, his
original result
was based on a projection from all 12 strata. Moreover,
we do not see
what the State and CSC did as "manipulating" the data. To
the extent
errors of this type were corrected, there is no need to project
them to
the universe of claims.
o The issue here is properly viewed as a question of the
statistical
validity of the method of projecting the remaining 24 errors to
the
universe of claims. CSC's expert supported his opinion with a
detailed
explanation that makes sense mathematically and that was supported
by
citation to statistical treatises. The Regional Statistical
Specialist
cited no support for his opinion that use of the method proposed
by CSC
was inappropriate, relying instead on assertions about CSC's
motives.
15/
o The Regional Statistical Specialist did
not express any
disagreement with CSC's expert's opinion that one factor
which affects
the amount of sampling precision for a sample is the
"variability," that
is, the range of dollar values of the sample
errors. The analysis of
the sample data presented with the CSC's
expert's original affidavit
clearly shows a difference in variability in the
dollar values when all
44 errors are considered and when only the 24
remaining errors are
considered. CSC Ex. 58, Ex. B. Since the
lesser variability in the
reprocessed strata would no longer balance out the
greater variability
in the other strata, it makes sense that there would be
less precision
overall to the revised point estimate than to the original
point
estimate for the 12 strata.
In sum, we conclude that HCFA's method used for calculating
the
disallowance amount was flawed. HCFA may not use this method
in
recalculating the disallowance based on the 21 errors for which
we
uphold the sample findings. HCFA may use the method explained in
CSC's
expert's affidavits, or such other method it determines is
a
statistically valid method appropriate to the circumstances.
Conclusion
For the reasons stated above, we find that HCFA properly disallowed
for
excess or improper payments resulting from the claims processing
errors.
However, the disallowance should be recalculated to reflect
the
following:
o Since we reversed HCFA's findings on Claim Nos. 402
($1),
837 ($4.15), and 866 ($4.15), HCFA should remove those
amounts
from the sample errors in the relevant strata; and
o HCFA should recalculate the disallowance consistent
with
the Board's discussion above regarding the appropriate
sampling
methodology.
HCFA should notify the State of how HCFA recalculated the
disallowance
amount. If the State disagrees with how HCFA recalculated
the
disallowance amount, the State may return to the Board, within 30
days
after receiving HCFA's notice, on the limited question of
the
recalculation.
_____________________________
Donald
F.
Garrett
_____________________________
Norval
D. (John)
Settle
_____________________________
Judith
A.
Ballard
Presiding
Board
Member. 1. CSC
moved
to intervene in these proceedings and the Board granted this
request.
2. Since 1985, section 1903(d)(2)(D) of the Act has provided that if
a
state is unable to recover a debt or a portion of a debt representing
an
overpayment made to a person or entity because that debt has
been
discharged in bankruptcy or is otherwise uncollectible, no adjustment
in
the federal payment to the state shall be made on account of
such
overpayment. The State did not allege here that it was unable
to
recover the overpayments from the providers.
3. The Board's prior holdings on overpayments issues have been
upheld
in three decisions by United States Court of Appeals:
Massachusetts v.
Secretary, 749 F.2d 89 (1st Cir. 1984), cert. denied, 472
U.S. 1017
(1985); Perales v. Heckler, 762 F.2d 226 (2d Cir. 1985); and
Missouri
Department of Social Services v. Bowen, 804 F.2d 1035 (8th Cir.
1986).
4. Under the Act and the regulations, HCFA will review yearly a
state's
MMIS operation initially approved as meeting certain conditions
and
reapprove the system for 75% FFP for operational expenditures
providing
certain conditions are met. Section 1903(r) of the Act; 42
C.F.R.
433.119. This review is called the System Performance Review
(SPR).
Part of the SPR tests the claims processing accuracy of the
system. If
a state's system does not meet the conditions for
reapproval, HCFA will
reduce the FFP in expenditures for system operations
from 75% to no more
than 70% and no less than 50%. Section 1903(r) of
the Act; 42 C.F.R.
433.120(b).
5. While the State contended its error rate was only .3% and
erroneous
payments under the audit amounted to less than $800,000 in
federal
funds, this is not accurate. The State's computation of the
error rate
does not take into account the substantial errors found in the
three
strata that were reprocessed ($15,598,510 in unallowable payments
to
nursing home, inpatient hospital and outpatient hospital services).
It
was only after the audit that the State reprocessed these claims
and
attempted to recoup the erroneous payments.
6. The preamble to the final regulations for the CPAS
specifically
states that HCFA may initiate an audit to identify and recover
misspent
federal funds based on an SPR or a state assessment claims
processing
review or any other information. 50 Fed. Reg. 21850 (May 29,
1985).
7. For Claim Nos. 410, 680, and 824, the State and fiscal agent
agreed
that the claims were paid in error, but argued that there should be
no
extrapolation of the errors to the universe of paid claims. We
discuss
this issue below.
8. In presenting the facts of this dispute, CSC attempted to call
into
question the audit results by suggesting that the audit, performed by
an
OIG contractor, was flawed because OIG auditors, reviewing
the
contractor's work, had disagreed with some of the contractor's
findings.
These alleged disagreements, however, are not relevant here.
There is
no dispute that OIG may hire auditing firms under contract to
perform
work. There is no dispute that since the contractor is working
for OIG,
the OIG has the right to review the contractor's work and
suggest
changes before it approves a draft report as OIG's final
report.
Regardless of what differences of opinion may have arisen during
the
process of producing a final report, it is only the final report that
is
the "official" recommendation of the OIG to HCFA. CSC did not point
to
any way in which the audit, as corrected by the OIG, was
inconsistent
with generally accepted auditing principles or otherwise
flawed. In any
event, such flaws would be relevant only to the extent
that they caused
an erroneous finding, and the State and CSC have had ample
opportunity
to challenge the audit findings.
9. CSC also submitted certain documentation pertaining to a
recipient
other than the recipient in question. See CSC Ex. 48, pp. 1
and 3.
This documentation was submitted to show that the recipient was
exempt
from copayment, but it is not relevant to the questioned claim.
10. CSC argued that it had claims history forms for the claims, but
did
not provide them, and the audit workpapers suggest that
basic
information, such as procedure codes, was missing. CSC Ex.
50. If the
State knew the provider numbers, the Medicare fiscal
intermediary would
not have been the only potential source of records for the
claims. The
providers might have had documentation to show what was
claimed and what
had been paid by Medicare. We also note that the
auditors found that
the State's Medicaid system would ordinarily create a
facsimile of a
crossover claim, but that the State had said that CSC's
ability to do
this was not operational at the time these claims were
processed. State
Ex. 1, pp. 5-6; CSC Ex. 50, p. 1.
11. HCFA may take an individual disallowance of $4.15 for the
second
claim since it duplicated the first claim, but to project the error
in
the second claim to the universe of claims (even though the second
claim
was not selected as a sample claim) would destroy the validity of
the
sampling process. Only the third claim was selected according
to
statistical principles.
12. HCFA may disallow the amount of the third payment made on
11/4/86
but may not consider it a sample error to be projected to the
universe
of claims since this physician claim was not selected as a sample
claim.
See note 11 above.
13. Use of the lower limit of the confidence interval, rather than
the
point estimate, is consistent with HHS audit policy. See, e.g.,
HCFA
submission of 5/3/91, enclosure, p. 3.
14. In the three strata in which 12 sample errors were
originally
identified, all of the claims in the strata were reprocessed, and
an
adjustment of FFP was already made for errors identified through
the
reprocessing. Thus, a disallowance based on projection of these
sample
errors to the universe would be duplicative.
15. The Regional Statistical Specialist acknowledged that the
lower
limit found by CSC's expert approximates the lower limit "our
sampling
software produces with the difference estimator" if "the
same
assumptions" are applied. Samworth Declaration, p. 2. The
only
assumption pointed out, however, was of 24 errors versus 44
errors.
Moreover, as CSC's expert explained, the slight difference between
the
lower limit CSC's expert calculated and that produced by the
"difference
estimator" software appears to be attributable to the fact that
CSC's
expert deducted a $6 credit for sample claim no. 796 (for which
the
State already adjusted FFP) from the lower limit, rather than from
the
point estimate. We agree with CSC's expert that the
appropriate
treatment of the $6 credit is to deduct it from the lower limit,
after
that figure is calculated according to statistical sampling
principles,
since the $6 should be treated as a credit to the amount of
the
erroneous payments in which FFP is being