GAB Decision 226
October 30, 1981 Ohio Department of Public Welfare; Docket No.
79-237-OH-HC Ford, Cecilia; Settle, Norval Garrett, Donald
By letter dated December 20, 1979, the Ohio Department of Public
Welfare (Ohio, State) requested reconsideration of $3,304,558 in Federal
financial participation (FFP) disallowed by the Health Care Financing
Administration (HCFA, Agency).The amount disallowed had been claimed by
Ohio under Title XIX of the Social Security Act for services to Medicaid
recipients by nursing homes during the period July 1, 1973 to June 30,
1976. Subsequently, in a March 28, 1980 supplement to its request, Ohio
withdrew its appeal as to $316,849, which represented unallowable
payments for rest home care. This left a balance of $2,987,709 on
appeal.
This dispute is the result of an audit of the Medicaid program in
Ohio by the Audit Agency of the (then) Department of Health, Education
and Welfare. With respect to the issues involved in this appeal, the
auditors recommended that the State return the following amounts of FFP,
for the reasons indicated:
1. $1,898,921 - absence of provider agreements.
2. $1,058,631 - overpayments due to adjustment of retrospective
rates.
3. $30,157 - unallowable payments for life care contracts and
failure to deduct patients' income and resources.
In this decision we uphold in part and reverse in part the $1.9
million disallowance based on the absence of provider agreements. We
uphold the entire amount of the disallowance on the remaining points.
Our decision is based on the appeal, HCFA's response, various
supplemental filings by both parties, an Order to Show Cause and the
responses by both parties, and an Order to Develop the Record and the
responses by both parties.
A. Absence of Provider Agreements
HCFA disallowed $1,898,921 FFP based on its finding that the State
had reimbursed both skilled nursing facilities (SNFs) and intermediate
care facilities (ICFs) for services to Medicaid recipients during
periods when those facilities did not have provider agreements with the
State. To qualify for FFP, a State must have such agreements in effect.
Section 1902(a)(27) of Title XIX of the Social Security Act, 42 U.S.C.
Sec. 1396a(27), and 45 CFR Part 249 (1973-1976).
(2) Ohio does not deny that some of the facilities for which it
claimed FFP did not have provider agreements. Ohio's principal argument
is that the disallowance based on the lack of provider agreements should
be reversed because of the method by which it was calculated. There are
three sub-issues:
1) Where does the burden lie of determining the quantum of
unallowable costs?
2) Has HCFA shown that its method was valid?
3) Assuming the method was not valid, what is the remedy?
1) Responsibility for the quantum
The parties each rely on Board decisions to support their arguments
that the burden rests with the other to produce a figure for the quantum
of unallowable costs. Ohio cites University of California -- General
Purpose Equipment, Decision No. 118, September 30, 1980. HCFA cites
Massachusetts Department of Public Welfare, Decision No. 155, March 20,
1981; and California State Department of Health, Decision No. 55, May
14, 1978. HCFA also cites Georgia v. Califano, 446 F. Supp. 404 (N.D.
Ga. 1977). We find the instant case distinguishable from the situations
prompting those decisions and instead elect to follow a principle of
shared responsibility such as the one under-pinning the result in
California Department of Health Services -- San Joaquin Foundation,
Decision No. 182, May 29, 1981.
In Georgia v. Califano, supra at 410, the court held that the state
was "ultimately charged with the duty of proving the allowability of
deferred claims." The court made that statement in the context of the
state's challenge of the use of sampling per se and went on to hold that
the sample was correctly done. Id. at p. 411. We do not find this
decision relevant in a situation where the Agency has not shown its
sample to be valid.
In General Purpose Equipment, the Board sustained the grantee's
appeal from a disallowance the quantum of which was based on statistical
sampling found to be questionable. Here, because of the State's failure
to monitor the renewal of provider agreements, the Agency should not
bear alone the responsibility for determining the amounts to be
disallowed.
In Massachusetts, the Board sustained the disallowance, even though
the Agency admitted that there were defects in the statistical sampling
methodology used by the auditors. The Board held that in the "narrow
circumstances" where Massachusetts had made a "conscious and deliberate
choice" not to apply a reduction in physicians' fees mandated by the
legislature, the State had to produce its own estimate or be bound to
pay the amount established by HCFA. Decision, pp. 3, 4.
(3) Here, although the Agency does not admit to the defects is the
sampling methodology, those defects can be demonstrated. See page 7,
infra. And although Ohio is responsible for the failure which occasioned
the disallowance, its wrong was not one of the deliberately choosing not
to apply a program requirement, as in Massachusetts.
The California case (Decision No. 55) actually did not involve a
sampling issue like the one in this case. The Board ruled against the
State's use of a ratio derived from a sample to estimate the number of
persons eligible for Medi-Cal in a group for which the State did not
have records. /1/ HCFA had treated the entire group as ineligible. The
Board found that the State has "the ultimate burden of establishing and
documenting eligibility" and, in any event, "beneficiaries for whom
records are not found are more likely to be ineligible than those for
whom records are found." Decision, pp. 4, 5. Here, even though we find
that the State bears at least part of the burden of establishing the
quantum, we also find the figure chosen by HCFA to be of questionable
validity.
San Joaquin (Decision No. 182), like California, involved a question
of payments for persons not eligible for Medicaid. Similar to the
situation in the instant case, the dispute was over the methodology in
determining the amount, not the disallowance itself. In the face of an
admission by a HCFA witness that its sample was defective and the
Board's analysis showing that HCFA's choice of a figure for the
disallowance was arbitrary, the Board decided not to uphold the
disallowance "as it now stands." Decision, p. 5. The Agency auditors,
assisted by the State, were given an opportunity to modify or reaffirm
the disallowance by means of full review or a valid sampling technique.
Id. at p. 6.
Thus, in situations where a disallowance is obviously called for, but
the quantum has not been properly established, the Board has not found
that the State has the burden of proving its claim under all
circumstances. While as a general proposition it is true that the State
bears the ultimate burden of justifying its claims for FFP, the Agency
shares the burden of going forward with evidence, in the sense that the
Agency cannot merely make arbitrary counterclaims; if the Agency
disputes a claim by the State, and does so using statistical sampling
methodology, the Agency must, if asked, show the reasonableness of its
methodology. The State bears a similar burden, in that it has
responsibilities of justifying its claim and facilitating Agency access
to necessary records. These shared responsibilities track the
cooperative nature of Federal-state relations in assistance programs
generally and of the Medicaid program in particular.
2. HCFA's showing of validity
The $3,517,403 ($1,898,921 FFP) which HCFA concluded the State paid
to nursing homes without provider agreements was the total of one actual
finding and two (4) estimates derived from samples. /2/ The principal
question here is whether the sample estimates reasonably can be said to
be valid.
The audit encompassed 880 nursing homes. Five SNFs were examined;
four of these did not have provider agreements and had been paid a total
of $1,400,175. /3/ The remaining 875 facilities consisted of 333 SNFs
and 542 ICFs. /4/ Samples of 58 SNFs and 114 ICFs were taken. /5/
In the sample of 58 SNFs, five were found not to have provider
agreements. /6/ Inexplicably, one of these had not received any
payments, although it was on the list of approved providers. HCFA
August 1981 Response, p. 9. The other four had received payments
totalling $347,407. /7/ From this the Agency estimated that the total
paid to those SNFs without provider agreements (in the universe of 333)
was somewhere between $1,151,847 and $2,837,493. For the disallowance,
it chose a point midway between the two extremes -- $1,994,670
(hereinafter called the "point estimate"). It claimed a "confidence
level" of 90 percent. /8/
(5) In the sample of 114 ICFs, nine were found not to have provider
agreements. The nine received payments totalling $119,868. /9/ From
this the Agency estimated that the total paid to those ICFs without
provider agreements was somewhere between $388,072 and $751,212. For
the disallowance, it chose a point midway between the two expremes --
$569,642 ("point estimate"). It claimed a "confidence level"
of 90
percent.
Ohio argues 1) a confidence level of 90 percent is not high enough;
2) it is not possible to substantiate a confidence level of 90 percent
by use of the midpoint; and 3) it could not ascertain how the Agency
decided on the size of the samples, but the extent of variance in the
payments as shown by the $236,633 for one of the five SNFs warranted an
increase in the size of the samples. Submission dated March 28, 1980
(pages not numbered). HCFA has never adequately responded to these
points, even after the Board requested specific information in the Order
to Develop the Record.
In that Order, the Board asked for the following data and attendant
calculations with respect to each sample: the mean, the standard error,
the tolerable error, and the coefficient of variation. /10/ In a
blurred, copier-darkened, handwritten document, /11/ HCFA supplied most
of the data but none of the calculations. August 1981 Response, p. 13
and Exhibit 1. The exhibit states that the tolerable error was not
"established." HCFA reported the mean for the SNF sample to be $5990.
It
supplied no calculations, but $5990 can be obtained by dividing the
$347,407 total for the (6) sampled SNFs lacking provider agreements by
58, the SNF sample size. Multiplying $5990 by 333, the SNF universe,
yields $1,994,670. This may have been how the point estimate
($1,994,670) used for the disallowance was derived. Similar
computations produce a like result for the ICFs.
HCFA reported the coefficient of variation for the SNF sample as
5.35743. /12/ It reported two figures as the standard deviation --
$12,052 allegedly calculated manually and $32,090 generated by use of a
computer. /13/ HCFA does not provide any calculations to show how it
arrived at any of these figures except by a general reference to Dr.
Arkin's textbook, nor does it explain why the manual and computer
results differ. Ohio cites Dr. Arkin's text in support of the State's
contention that the method used by HCFA may not produce a point estimate
sufficiently accurate to present as legal evidence. Ohio September 1981
Response, p. 8.
We do not find that either party has established the validity or
invalidity of the 90 percent confidence level, but using HCFA's figures
in a standard formulation we conclude that the point estimate does not
have a sufficiently valid basis to be used in a disallowance. We will
demonstrate this by using the formula for determining sample size for
variables, because the high coefficient of variation indicates that the
sample was not large enough. /14/ We use the formula for variables
because HCFA's position is that it was sampling to determine the (7)
amount paid (a variable), not the attribute of having a provider
agreement. /15/
The formula is to take the product of the universe (N) multiplied by
the confidence factor (t) squared times the coefficient of variation (V)
squared and divide this product by the sum of N multiplied by the
tolerable error (E) squared plus the product of t squared multiplied by
V squared. /16/
2 2 Nt V Thus, n (sample size) 2 2 2 .
NE + t V
Since the sample size is known and the tolerable error is not, we
convert the formula to
2 2 2 2 E = t V - t V . n N
Using HCFA's figures, we arrive at a tolerable error of 105 percent
for the SNF sample and 77 per cent for the ICF sample. /17/ We find it
unreasonable to base a disallowance on such a high rate of error. As
shown above, the tolerable error rate is a key factor in determining
sample size. By way of contrast, had the Agency chosen a tolerable (8)
error rate of 5 per cent, it would have had to "sample" virtually
the
entire universe in each instance (329 SNFs, 533 ICFs), given the same
high coefficient of variation. The Agency's admission that it did not
establish a tolerable error rate, in addition to our own computations
above, lead us to conclude that the sample size was chosen arbitrarily
and without reference to the problem of variance.
This finding is strengthened by HCFA's response to the Board's
request that the Agency demonstrate how the sample sizes were
appropriate for determining payments to facilities lacking provider
agreements. The Agency response alludes to "problems of cost and time"
and in addition sets out numbers and calculations which at best give
collateral but coincidental support only to the ICF part of the
disallowance. August 1981 Response, p. 11.
3. The remedy
We find that HCFA's failure to validate the potential inaccuracies in
the point estimates for both SNFs and ICFs as demonstrated above makes
those estimates highly questionable as bases for this disallowance. On
the other hand, the State admittedly failed to monitor the renewal of
provider agreements and a disallowance of some amount is justified.
Accordingly, we sustain this disallowance in part based on what we find
to be the best available figures, as follows:
a. In its attempt to justify the ICF part of the disallowance, HCFA
alleged that the auditors found a total of $524,078 paid to ICFs lacking
provider agreements. This was not based on a sample, but a 100 percent
audit of only those ICFs giving rest home care. /18/ August 1980
Response, p. 8. The Board noted this finding in the Order to Show Cause
and Ohio has not refuted it. Therefore, we uphold the ICF part of the
disallowance to the extent of $524,078, FFP to be determined at the
applicable rate.
b. We also uphold the actual SNF finding to the extent of
$1,400,175, FFP to be determined at the applicable rate. Ohio objected
to the HCFA's combining of this actual finding with the estimates
derived (9) from the two samples, but did not refute the finding itself.
The Board made a tentative finding in the Order to Show Cause that
mixing actual and estimated findings could be valid. Ohio has continued
its objection but has not produced any support for its conclusion.
September 1981 Response, p. 6. Accordingly, we uphold this part of the
SNF disallowance and conclude that it may properly be added to estimated
findings.
c. We uphold the remainder of the SNF disallowance to the extent of
$1,151,847, which is the lower end of the range established by HCFA.
Ohio has argued throughout that the use of the point estimate was not
valid, but it noted the auditors' initial decision to use the lower
boundary and characterized it as "more prudent" than the point estimate.
March 28, 1980 submission (not paginated). The lower boundary may be
subject to some of the same infirmities as the point estimate; but
consistent with our holding that the State shares responsibility with
HCFA for this estimate, we sustain this amount. /19/
d. We also give both parties an alternative to our holdings on the
two estimated findings, "a" and "c", totalling $1,675,925.
With respect
to either or both of the holdings, the parties may conduct a 100 percent
audit or a valid sample. If the State intends to do this, it must
notify the Board and the Agency within 10 days from receipt of this
decision. The effect of our decision is suspended for 25 days to permit
the State time to give notice and will be suspended for an additional 35
days if the State elects to audit or sample. Similarly, HCFA may issue
a new disallowance within 60 days from the date of our decision. If
HCFA does not agree to the State's figures, or the State does not agree
to HCFA's, the State may appeal again to this Board, within 30 days of
receiving HCFA's determination.
4. Other issues
Ohio also has four other points with respect to the provider
agreement issue. We find against the State on each of these.
a. Ohio contended that Adams Manor, one of the sampled ICFs
allegedly lacking a provider agreement, was certified and did have a
provider agreement for December 1975, the month at issue. In support,
Ohio submitted a Certification and Transmittal form signed June 16, 1976
(10) and a provider agreement signed July 12, 1976. Exhibit 2bi, July
31, 1981 submission; Exhibit 2b, July 28, 1980 submission. We hold, in
keeping with prior Board decisions, that the provider agreement may not
be effective earlier than the date of certification -- in this case,
June 16, 1976.Washington Department of Social and Health Services,
Decision No. 176, May 26, 1981; Maryland Department of Health and
Mental Hygiene, Decision No. 107, July 2, 1980. Thus, Ohio has not
shown that there was a valid provider agreement in effect for Adams
Manor in December 1975. Even if it had, our decision on the ICF part of
the disallowance does not depend on the ICF sample and we would still
uphold that part to the extent indicated.
b. Ohio argued that a disallowance by HCFA dated January 28, 1977,
overlapped this disallowance. February 28, 1980 submission. The
disallowance on appeal here relates to claims for services during the
periods January 1, 1974 - February 29, 1976 (SNFs) and May 10 - December
31, 1975 (ICFs). The 1977 disallowance covered claims submitted in
reports for the quarters ended March 31, 1976; June 30, 1976; and
September 30, 1976.
HCFA alleged that its examination of the work papers for the audits
on which the two disallowances were based indicated no overlap. HCFA
also pointed out that the audit in the January 1977 disallowance did not
begin until five months after the completion of the earlier audit.
August 1980 Response, p. 9.
In the Order to Show Cause, the Board in effect called upon Ohio to
demonstrate where any overlap had occurred, but Ohio did not respond on
this point. Accordingly, we find that an overlap has not been shown to
exist with the other disallowance.
c. Ohio alleged that it continued payments to some of the
facilities, despite the absence of provider agreements, because of court
orders. The Board has held that under certain circumstances FFP is
available in payments to nursing homes appealing the State's termination
of or refusal to execute a renewal of a provider agreement. Ohio
Department of Public Welfare, Decision No. 173, April 30, 1981.
The Board repeatedly requested Ohio to provide copies of the court
orders, but Ohio has produced court orders dealing with only one
facility -- Convalescent Care, Inc. Although similar in name, that
facility is not identified as one of those involved in this
disallowance. /20/ Even if it were, the orders relied on by Ohio do not
direct the State to continue payments; to the contrary, the Court of
Appeals of Franklin County held that it was not a denial (11) of due
process to discontinue payments pending a hearing on the renewal of the
provider agreement. Ohio Response to Order to Show Cause, Exhibit A-3,
p. 1104.
Accordingly, we find that Ohio has failed to make the necessary
showing.
d. Ohio contended that various policy and equitable considerations
justified its continued payments to facilities lacking provider
agreements, and that the 30-day limit for transferring Medicaid
recipients from such facilities abridged the rights of those recipients.
February 1980 submission. With respect to the first point, this Board
has held that such arguments do not outweigh the regulatory
requirements. Ohio, supra, p. 5. As for the second, even if this were
the proper forum for attacking a regulation on constitutional grounds
(HCFA contended it was not), the parties at interest would be the
recipients, not the State. We find these two points without merit and
hold against the State on them.
B. Overpayments due to adjusted retrospective rates
HCFA disallowed $1,974,676 ($1,058,631 FFP) identified as the
equivalent of amounts owed to the State by nursing homes as a result of
adjustments in per diem rates. The nursing homes had been paid in
advance based on interim rates. The homes submitted cost reports at the
end of the period and on the basis of these reports the rates were
adjusted downward. The nursing homes allegedly are required under Ohio
law to refund the difference to the State. Agency Audit Report, p. 7.
Ohio's principal argument is that "good logic and reason" compel
a
delay by the Agency in seeking reimbursement from the State until
completion of administrative hearings on State audits which were
conducted in September 1980. February 1981 Response to Order to Show
Cause, p. 10. /21/ In addition, Ohio relied on a 1977 Agency "Action
Transmittal" (AT 77-85) as a basis for the State not accounting for
overpayments until the nursing homes have had the opportunity to exhaust
all administrative remedies. Id. at p. 11; and Exhibit B-2.
HCFA cited section 1903(d) of Title XIX of the Social Security Act,
42 U.S.C. Sec. 1396b(d), as authority for its right to disallow,
contending that the Secretary lacked the discretion to delay (12)
effecting a disallowance based on an overpayment. /22/ HCFA argued that
AT 77-85 did not apply to this situation because the disallowance was
not based on a State audit, much less one almost a year after the
disallowance. AT 77-85 interprets a similarly worded predecessor to 42
C.F.R. Sec. 447.296, a regulatory provision on which HCFA initially
relied. HCFA later agreed with Ohio that $447.296 was inapplicable to
this case because the disallowance was not based on a State audit. /23/
HCFA submission of August 1980, p. 4; Ohio submission of February 1981,
p. 8; HCFA submission of June 1981, p. 7.
We find that the State has pointed to no authority to justify
delaying the State's accounting for this overpayment. The auditor's
findings were based on the State's review of "cost proposals" and
"cost
reports" supplied by the nursing homes themselves. Audit Report, p. 8.
Ohio did not allege that any of the homes had appealed the State's
efforts to collect. Whatever bearing AT 77-85 may have on the September
1980 audits, it is of no relevance to this disallowance. Accordingly,
we uphold the disallowance.
(13) C. Other overpayments
HCFA disallowed $55,988 ($30,157 FFP) because the State had included
in its payments to nursing homes the cost of "life care" contracts
and
also had failed to subtract applicable patient income and resources.
/24/ Prior to its response to the Order to Show Cause, Ohio did not
contest this disallowance, asking only that repayment be deferred until
the State had collected from the nursing homes. Since that Response,
Ohio has by reference applied its argument on the other overpayment
issue to this one.
We find that the State has not shown any basis for overturning or
even delaying this disallowance. Accordingly, we uphold it.
Conclusion
We uphold the disallowance in this case for $5,106,764 and overturn
it in the amount of $888,387. These amounts include both the federal
and State shares, because we could not determine the FFP on the reduced
provider agreement amount. The effect of our decision in the amount of
$1,675,925 of the disallowed amount is suspended for 25 days to give the
State time to decide whether to pursue the 100 percent audit or sample
alternatives and then for an additional 35 days if the State elects to
do so. Page 9, supra. If HCFA is going to issue a new or modified
disallowance to recoup more than $5,106,764, it must do so within 60
days of the date of this decision or the matter of the quantum will be
res judicata. /1/ Medi-Cal is the term for the Medicaid program in
California. /2/ A total of $3,964,487 was identified, but HCFA
subtracted an amount totalling $447,084 which was subject to
disallowance also for other reasons. /3/ In its August 21, 1980
response to the appeal, HCFA indicated that there were four SNFs which
had been paid $1,427,175. Page 6. In its August 19, 1981 response to
the Order to Develop the Record HCFA stated that the four SNFs were part
of a "judgment sample" (not explained) of five SNFs and that the
payments to the four totalled $1,400,175. Page 8. /4/ The audit
report states that there were "about" 541 ICFs, 6 SNFs, and 333 ICF/SNF
facilities. Page 1. In its response to the Order to Develop the
Record, HCFA states that there were as of September 15, 1975, 542 ICFs
and 338 SNFs (333 remaining after the five in the "judgment sample"
were
subtracted). Page 9. HCFA explained that the figures at page one of
the audit report were current at the time the report was prepared (1978)
but change from month to month and were included only to "provide
perspective." Id. at 8. /5/ HCFA does not explain why the
original sample of 107 ICFs was expanded to 114. August 1980 Response,
p. 7. /6/ Exhibit 1 to HCFA's August 1981 Response indicates
there were only four, but in the Response itself HCFA states there were
five. Page 9. /7/ The payments were $47,665; $40,198; $12,911;
and $236,633. HCFA August 1981 Response, p. 13. /8/ Another way of
expressing this is that there is 1 chance in 10 that the correct
amount does not lie within the alleged range. /9/ The payments were:
$53,673 $9,400 6,180 6,895 252 1,260 5,927 4,245 32,036 Id. at p.
13. /10/ These are standard concepts used in statistical
sampling. See, for example, Statistics for Management (1977) by B. J.
Mandel, Professor Emeritus of Statistics and Former Chairman of the
Statistics Department, University of Baltimore. HCFA relied on the
published work of Herbert Arkin, Professor of Business Statistics at the
City University of New York. /11/ One explanation for this might
be that the exhibit is a copy of the audit work papers, but HCFA did not
say so. It would have been better if HCFA had typed the information, or
at least submitted a more legible copy, and identified the source of the
information. See Standards for Audit of Governmental Organizations,
Programs, Activities, and Functions by the Comptroller General of the
United States (1981), pp. 25-26 "Working Papers." /12/ The
coefficient of variation is a measure of the extent of variation of
units (here, payments) from their mean value. Thus, the presence of one
value of $12,911 and another of $236,633 in a set of four SNFs presages
a very high coefficient. Expressed as a percentage (multiplying it by
100), it would be 536 per cent (rounded). We do not know how HCFA
calculated its coefficient, but dividing the standard deviation
($32,090) by the mean ($5990), we obtain 5.36. /13/ The standard
deviation is a measure of the average difference of the units (payments)
from their mean value; it is the square root of the mean of squared
differences of the values from their mean. We were not able to
ascertain how HCFA arrived at either the manual or computer generated
figures. /14/ We realize that our analysis may not be the only
approach for demonstrating the weakness of this sampling methodology.
However, we were forced to perform our own evaluation because neither
party made an adequate showing of the validity or invalidity of the
sampling methodology. /15/ In its Response to the Order to
Develop the Record, HCFA stated the number of SNFs estimated to lack
provider agreements -- 29 -- had no bearing on the dollar point
estimate. Page 10. 16 The confidence factor is used to apply the
confidence level. It can be found in a table referred to as the Normal
Distribution Curve (bell-shaped) and is probably a standard feature of
all statistics textbooks. See Appendix C in Dr. Mandel's text, supra
(1977 edition). For a confidence level of 90 percent, the factor is
1.65. Thus, 90 percent of the true values lie within plus or minus 1.65
times the standard deviation from the mean. /17/ The confidence
factor squared is 2.7, based on the confidence level of 90 percent. The
coefficients of variation squared are 28.7 (SNFs) and 32 (ICFs).
Multiplied by the confidence factor, they are 77.6 (SNFs) and 86.5
(ICFs). These products, divided by the respective samples (58-SNFs, 114
- ICFs) and universes (333 - SNFs, 542 - ICFs), the universe quotients
subtracted from the sample quotients, and the square root taken of the
results, gives us 1.05 for the SNFs and.77 for the ICFs. /18/ The total
includes payments to ICFs selected in the sample. FFP was not
available for rest home care and HCFA disallowed $584,077 ($316,849 FFP)
which Ohio is no longer appealing. There is no overlap with the amount
disallowed for lack of provider agreements because the Agency subtracted
$447,084 in arriving at the latter, specifically to eliminate
duplication with disallowances for other reasons cited in the audit
report. Notification of Disallowance, p. 3. /19/ In its response to
the Order to Develop the Record, Ohio designated the method by
which HCFA calculated the range as the "average range method" and
cited
a general statement in the Arkin text that "it would be well" to use
an
unspecified "more exact method" if the results were to be presented
as
legal evidence. We think this falls short of a clear repudiation of the
lower boundary. /20/ Convalescent Center is one of the SNFs
found not to have had a provider agreement. /21/ Ohio initially
contended it had collected and credited $1,138,141 to the Agency, but
subsequently abandoned that contention. See Ohio submission dated March
1980; HCFA submission dated August 25, 1980, Exhibit 1; Order to Show
Cause. /22/ That section states, in pertinent part: (d)(1)
Prior to the beginning of each quarter, the Secretary shall estimate the
amount to which a State will be entitled... (2) 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 determined was made under this section
to such State for any prior quarter and with respect to which adjustment
has not already been made under this subsection. HCFA conceded, however,
that under certain circumstances (not shown to be present here) Agency
regulations permit a State to pay by installments. August 1980
submission, p. 4. /23/ 42 CFR Sec. 447.296 (1979) states: The
agency must account for overpayments found in audits on the quarterly
statement of expenditures no later than the second quarter following the
quarter in which the overpayment was found. AT 77-85 interprets "the
quarter in which found" as "the quarter during which the administrative
hearing procedures of the State have been exhausted and a determination
of overpayment has been sustained." /24/ Life care contracts are
arrangements whereby patients agree to turn over to nursing home
operators the patients' property in return for care for the rest of
their lives. The cost of such contracts may be eligible for FFP under
certain conditions, but those conditions were not met here. Nursing
homes are required to deduct, for each patient, Social Security benefits
and other income in excess of $25 per month and any liquid assets in
excess of $300. Audit Report, pp. 2, 27.
SEPTEMBER 22, 1983