United States Court of Appeals
FOR THE DISTRICT OF COLUMBIA CIRCUIT
Argued September 7, 1999 Decided October 1, 1999
No. 98-5254
County of Los Angeles, a political subdivision
of the State of California, owner and operator of
Los Angeles County/USC Medical Center,
Harbor/UCLA Medical Center,
Martin Luther King Jr./Drew Medical Center,
Olive View Medical Center and High Desert Hospital, et al.
Appellees/Cross-Appellants
v.
Donna E. Shalala, Secretary,
U.S. Department of Health and Human Services
Appellant/Cross-Appellee
Consolidated with
Nos. 98-5255, 98-5256, 98-5257, 98-5258, 98-5259,
98-5260, 98-5261, 98-5262, 98-5325, 98-5326,
98-5327, 98-5328, 98-5329, 98-5330, 98-5331,
98-5332, 98-5333
---------
Appeals from the United States District Court
for the District of Columbia
(No. 93cv00146)
(No. 93cv00147)
(No. 93cv00479)
(No. 93cv00692)
(No. 93cv00836)
(No. 93cv00837)
(No. 93cv01188)
(No. 93cv02069)
(No. 94cv01485)
Peter R. Maier, Attorney, United States Department of
Justice, argued the cause for appellant/cross-appellee. With
him on the briefs were Frank W. Hunger, Assistant Attorney
General, Wilma A. Lewis, United States Attorney, and Bar-
bara C. Biddle, Attorney, United States Department of Jus-
tice.
Lloyd A. Bookman argued the cause for appellees/cross-
appellants. With him on the briefs were David H. Eisenstat,
Byron J. Gross, John R. Hellow, Michael G. Hercz, John R.
Jacob, and David B. Palmer.
Before: Wald, Silberman, and Tatel, Circuit Judges.
Opinion for the Court filed by Circuit Judge Wald.
Wald, Circuit Judge: Brought by the owners of Medicare-
provider hospitals ("Hospitals") and the Secretary of Health
and Human Services ("Secretary"), these cross-appeals pres-
ent two issues. First, under the Medicare statute, must the
Secretary provide hospitals with retroactive reimbursements
to ensure that aggregate outlier payments during any given
fiscal year meet minimum statutory targets? And second,
has the Secretary adequately explained why, when calculating
outlier thresholds for fiscal years 1985-1986, she relied on a
1981 database instead of more contemporaneous records from
1984 Medicare discharges? Finding that Congress had spo-
ken directly and unambiguously to the first question, the
district court granted partial summary judgment to the Hos-
pitals. With respect to the second issue, however, the court
perceived nothing unreasonable in the Secretary's choice of
data, and entered judgment accordingly for the Secretary.
Because we disagree with the district court on both points, we
now reverse.
I. Background
Through a "complex statutory and regulatory regime,"
Good Samaritan Hosp. v. Shalala, 508 U.S. 402, 404 (1993),
the Medicare program reimburses qualifying hospitals for the
services that they provide to eligible patients. See Social
Security Act, Pub. L. No. 89-97, tit. XVIII, 79 Stat. 286, 291
(1965) (codified as amended at 42 U.S.C. ss 1395-1395ggg
(1994 & Supp. III 1997)). From its inception in 1965 until
October 1983, Medicare compensated hospitals for the "rea-
sonable costs" of the inpatient services that they furnished.
See 42 U.S.C. s 1395f(b). Experience proved, however, that
this system bred "little incentive for hospitals to keep costs
down" because "[t]he more they spent, the more they were
reimbursed." Tucson Med. Ctr. v. Sullivan, 947 F.2d 971,
974 (D.C. Cir. 1991).
To stem the program's escalating costs and perceived inef-
ficiency, Congress fundamentally overhauled the Medicare
reimbursement methodology in 1983. See Social Security
Amendments of 1983, Pub. L. No. 98-21, s 601, 97 Stat 65,
149. Since then, this new regime, known as the Prospective
Payment System ("PPS"), has reimbursed qualifying hospi-
tals at prospectively fixed rates. By establishing pre-
determined reimbursement rates that remain static regard-
less of the costs incurred by a hospital, Congress sought "to
reform the financial incentives hospitals face, promoting effi-
ciency in the provision of services by rewarding cost/effective
hospital practices." H.R. Rep. No. 98-25, at 132 (1983),
reprinted in 1983 U.S.C.C.A.N. 219, 351.
Calculating prospective-payment rates begins with deter-
mining the "federal rate," a standard nationwide cost rate
based on the average operating costs of inpatient hospital
services. See 42 U.S.C. s 1395ww(d)(2)(A)-(B); 49 Fed. Reg.
234, 251 (1984). To account for regional variations in labor
costs, the Secretary then establishes a wage index that aug-
ments the adjusted standardized payment depending on the
location of a qualifying hospital. s 1395ww(d)(2)(H),
(d)(3)(E). The final variable is an additional weighting factor
that reflects the disparate hospital resources required to treat
major and minor illnesses. s 1395ww(d)(4). For each of 470
medical conditions--known as diagnosis related groups or
"DRGs"--the Secretary assigns particular weights by which
the federal rate is to be multiplied. The more complicated
and costlier the treatment is, the greater the weight assigned
to that particular DRG will be. To calculate the final "DRG
prospective payment rate" for a patient discharge, the Secre-
tary takes the federal rate, adjusts it according to the wage
index, and then multiplies it by the weight assigned to the
patient's DRG. By statutory mandate, the Secretary must
publish the weights and values that she will factor into the
prospective-payment calculus before the start of each fiscal
year. s 1395ww(d)(6).
Despite the anticipated virtues of PPS, Congress recog-
nized that health-care providers would inevitably care for
some patients whose hospitalization would be extraordinarily
costly or lengthy. To insulate hospitals from bearing a
disproportionate share of these atypical costs, Congress au-
thorized the Secretary to make supplemental "outlier pay-
ments." During the years at issue in these cross-appeals, the
outlier-payment provisions were set forth in four clauses of
the Medicare statute. 42 U.S.C. s 1395ww(d)(5)(A)(i)-(iv)
(Supp. IV 1986). With the first two clauses, Congress estab-
lished two classes of outlier payments: day outliers and cost
outliers. s 1395ww(d)(5)(A)(i)-(ii). A hospital could qualify
for a day-outlier payment if the patient's length of stay
exceeded the mean length of stay for that particular DRG by
a fixed number of days or standard deviations.
s 1395ww(d)(5)(A)(i). Along the same lines, the Secretary
would make cost-outlier payments when a hospital's cost-
adjusted charges surpassed either a fixed multiple of the
applicable DRG prospective-payment rate or such other fixed
dollar amount that the Secretary established.
s 1395ww(d)(5)(A)(ii). In the third clause, Congress provided
that outlier payments "shall be determined by the Secretary
and shall approximate the marginal cost of care beyond the
cutoff point applicable" to the day or cost outlier.
s 1395ww(d)(5)(A)(iii).
It is the fourth and final clause, however, that forms the
textual nub of the present controversy.
s 1395ww(d)(5)(A)(iv). During 1985 and 1986, paragraph
(5)(A)(iv) provided:
The total amount of the additional payments made under
this subparagraph for discharges in a fiscal year may not
be less than 5 percent nor more than 6 percent of the
total payments projected or estimated to be made based
on DRG prospective payment rates for discharges in that
year.
Id. Traditionally, the Secretary has read paragraph
(5)(A)(iv) to mean that at the start of each fiscal year, she
must establish the fixed thresholds beyond which hospitals
will qualify for outlier payments at levels likely to result in
outlier payments totaling between five and six percent of
projected DRG payments for that year. In making this
estimation, the Secretary first settles on the per-diem outlier
payment, which pursuant to s 1395ww(d)(5)(A)(iii), must ap-
proximate the marginal cost of care. She then examines
historical Medicare-discharge data to determine which thresh-
olds, when multiplied by the per-diem payment rate, would
probably yield total outlier payments falling within the five-
to-six-percent range in paragraph (5)(A)(iv). As the Secre-
tary observed during the rulemaking, however, "given the
data available, forecasts of probable future outlier payments
are inexact." 50 Fed. Reg. 35,646, 35,710 (1985). If it turns
out that the Secretary overestimated the mean length of stay
for DRGs, the actual total outlier payments at the end of the
year may amount to less than five percent of estimated DRG-
related payments. Conversely, underestimating the mean
length of stay might produce outlier payments in excess of six
percent of estimated total DRG payments.
Whether the Secretary's projections prove to be correct
will depend, in large part, on the predictive value of the
historical data on which she bases her calculations. For fiscal
year 1984, the Secretary relied on data culled from the 1981
Medicare Provider Analysis and Review ("1981 MEDPAR")
file, a database containing 1.6 million Medicare discharges
from 1981. As a product of the old reasonable-cost system,
however, the 1981 MEDPAR file obviously did not reflect one
of "[t]he most commonly accepted expectation[s] about the
PPS at the time of its inception[:] that it would result in
shorter stays for Medicare patients." Office of Research &
Demonstrations, Health Care Fin. Admin., U.S. Dep't of
Health & Human Servs., Pub. No. 03231, Report to Congress:
Impact of the Medicare Hospital Prospective Payment System
6-13 (1984). By 1984, however, preliminary data indicated
that the mean length of stay for virtually all DRGs had, as
anticipated, declined dramatically under PPS. The Secre-
tary, nevertheless, chose to rely again on the 1981 MEDPAR
file in setting outlier thresholds for fiscal years 1985-1986.
During those years, though the Secretary set thresholds at a
level projected to result in outlier payments at or above
paragraph (5)(A)(iv)'s five-percent floor, actual outlier pay-
ments in 1985 constituted only 3.0 percent of estimated DRG-
related payments and in 1986 they amounted to 4.4 percent.1
Given the enormity of the Medicare program, these seemingly
modest percentage differences represent substantial sums of
money: $241 million for fiscal year 1985 and $101 million for
fiscal year 1986.
Because actual outlier payments for fiscal years 1985-1986
amounted to less than five percent of projected DRG-related
payments, the Hospitals petitioned the Secretary for retroac-
tive reimbursements to satisfy the difference. According to
the Hospitals, paragraph (5)(A)(iv) does more than instruct
the Secretary about where she should set outlier thresholds
__________
1 Note that these percentages were derived from the actual total
DRG prospective payments in 1985 and 1986, not the projected
payments as set forth in the statute. Both parties have agreed that
this is the only practical method of calculating shortfalls at this
point. See Br. of Sec'y at 13 & n.4; Br. of Hosps. at 4 & n.3.
at the beginning of each fiscal year; it affirmatively com-
mands her to recalibrate retroactively the outlier thresholds
if, after the fiscal year concludes, actual outlier payments do
not equal at least the five-percent statutory target. More-
over, the Hospitals claimed that the Secretary had acted
arbitrarily and capriciously by relying on the 1981 MEDPAR
file when she forecast outlier thresholds for fiscal years 1985-
1986. The Secretary rejected both claims, and the Provider
Reimbursement Review Board authorized the Hospitals to
seek expedited judicial review pursuant to 42 U.S.C.
s 1395oo(f).
In an opinion and order dated January 20, 1998, the district
court granted in part and denied in part both the Secretary's
and the Hospitals' cross-motions for summary judgment. In
granting a portion of the Hospitals' motion, the court pro-
ceeded no further than the first step of Chevron U.S.A. Inc.
v. Natural Resources Defense Council, Inc., 467 U.S. 837,
842-43 (1984), concluding that paragraph (5)(A)(iv) unambigu-
ously requires the Secretary to adjust outlier payments retro-
actively to ensure that total actual outlier payments fall
within the statute's five-to-six-percent range. County of Los
Angeles v. Shalala, 992 F. Supp. 26, 31-33 (D.D.C. 1998).
Holding, however, that the Secretary's decision to favor the
1981 MEDPAR file over the more recent, though preliminary,
1984 data when determining outlier thresholds was "a rational
choice between two imperfect databases," the district court
granted the Secretary's motion for summary judgment on the
Hospitals' claim of arbitrary and capricious agency action.
Id. at 34-36.
Having determined that paragraph (5)(A)(iv) required the
Secretary to make retroactive outlier payments to the Hospi-
tals, the district court instructed the parties to meet and
confer on how to structure the final remedy. On April 30,
1998, based largely on a joint stipulation filed by the parties,
the district court entered an order granting judgment to the
parties on each of the claims on which they respectively
prevailed. While the April 30 order also instructed the
Secretary to tender retroactive outlier payments to the Hos-
pitals, it did not specify a sum certain to be paid; rather, the
court left it to the Secretary to calculate the amount owed.
Thereafter, the Secretary noted a timely appeal and the
Hospitals noted timely cross-appeals.
II. Analysis
A. Jurisdiction
Before reaching the merits, it is necessary to examine our
jurisdiction to entertain these cross-appeals. Section 1291 of
the Judicial Code provides that the courts of appeals may
review "all final decisions of the district courts of the United
States." 28 U.S.C. s 1291 (1994). In the proceedings below,
the district court, managing this case as it would any garden-
variety civil suit, adjudicated not only the respective legal
rights of the parties but also took steps toward decreeing a
proper remedy. Thus in its January 20, 1998 order, the court
resolved the merits of the Hospitals' claims, and with its April
30, 1998 order, directed the Secretary to calculate the amount
of outlier payments due to the Hospitals and to make pay-
ment accordingly. This latter order has spawned some confu-
sion about our jurisdiction because of the general rule appli-
cable to civil actions that "where assessment of damages or
awarding of other relief remains to be resolved," a district
court's judgment is not " 'final' within the meaning of 28
U.S.C. s 1291." Liberty Mut. Ins. Co. v. Wetzel, 424 U.S.
737, 744 (1976); see also A & S Council Oil Co. v. Lader, 56
F.3d 234, 238 (D.C. Cir. 1995) (holding that an order estab-
lishing liability but referring the issue of damages to arbitra-
tion is not final). For it is clear that neither of the district
court's orders resolved the precise quantum of payments to
be made to the Hospitals.
This rule of finality does not apply here, however, because
this is not an appeal from an ordinary civil judgment ren-
dered by the district court. With both of their claims, the
Hospitals challenged the Secretary's actions under section
10(e) of the Administrative Procedure Act, 5 U.S.C. s 706(2).
As we have often observed, "[w]hen a final agency action is
challenged under the APA in district court, if the relevant
substantive statute does not provide for direct review in the
court of appeals, the district court does not perform its
normal role" but instead "sits as an appellate tribunal." PPG
Indus., Inc. v. United States, 52 F.3d 363, 365 (D.C. Cir.
1995) (quoting Marshall County Health Care Auth. v. Shala-
la, 988 F.2d 1221, 1225 (D.C. Cir. 1993)); accord James
Madison Ltd. v. Ludwig, 82 F.3d 1085, 1096 (D.C. Cir. 1996)
("Generally speaking, district courts reviewing agency action
under the APA's arbitrary and capricious standard do not
resolve factual issues, but operate instead as appellate courts
resolving legal questions."), cert. denied, 519 U.S. 1077 (1997).
Whether it is a court of appeals or a district court, "[u]nder
settled principles of administrative law, when a court review-
ing agency action determines that an agency made an error of
law, the court's inquiry is at an end: the case must be
remanded to the agency for further action consistent with the
corrected legal standards." PPG Indus., 52 F.3d at 365; see
also South Prairie Constr. Co. v. Local No. 627, Int'l Union
of Operating Eng'rs, 425 U.S. 800, 806 (1976); SEC v. Chen-
ery Corp., 318 U.S. 80, 94-95 (1943). Once, therefore, the
district court held that the Secretary had misinterpreted
s 1395ww(d)(5)(A)(iv), it should have remanded to the Secre-
tary for further proceedings consistent with its conception of
the statute. Not only was it unnecessary for the court to
retain jurisdiction to devise a specific remedy for the Secre-
tary to follow, but it was error to do so. See Ommaya v.
National Insts. of Health, 726 F.2d 827, 830 (D.C. Cir. 1984)
("If MSPB relied on incorrect legal grounds, it would be error
for this court to enforce without first remanding for agency
examination of the evidence and proper fact-finding.") (quot-
ing White v. United States Dep't of the Army, 720 F.2d 209,
210 (D.C. Cir. 1983)). Accordingly, because that was all that
the district court had the power to do, we construe its
January 20, 1998 order as a remand to the Secretary, and
ignore, for jurisdictional purposes, its later order on specific
relief.
Of course, properly characterizing the district court's order
as a remand does not, without more, resolve our jurisdictional
quandary, for a "remand order usually is not a final decision."
NAACP v. United States Sugar Corp., 84 F.3d 1432, 1436
(D.C. Cir. 1996). We have recognized "an exception to this
general rule, however, where the agency to which the case is
remanded seeks to appeal and it would have no opportunity to
appeal after the proceedings on remand." Occidental Petro-
leum Corp. v. SEC, 873 F.2d 325, 330 (D.C. Cir. 1989).
Animating this principle is a pragmatic concern. Because an
agency must conduct its proceedings and render its decision
pursuant to the legal standard that the district court articu-
lates in its remand order, "[u]nless another party appeals [the
agency's subsequent] decision, the correctness of the district
court's legal ruling will never be reviewed by the court of
appeals, notwithstanding the agency's conviction that the
ruling is erroneous." Id. Here, were the Secretary unable
to appeal the district court's decision at this point, on remand
she would have to interpret paragraph (5)(A)(iv) as the dis-
trict court has construed it, and disburse millions of dollars in
retroactive outlier payments to various Medicare-provider
hospitals. Absent an appeal from that decision by one of the
Hospitals, the Secretary would have no opportunity to chal-
lenge the legal basis for the disbursements. Our jurisdiction
is therefore proper because the Secretary's appeal falls within
the exception recognized in Occidental Petroleum. Addition-
ally, vested with jurisdiction to review the Secretary's appeal
under s 1291, we may also consider the Hospitals' cross-
appeal of the district court's grant of summary judgment to
the Secretary on their arbitrary and capricious agency-action
claim. See United States Sugar Corp., 84 F.3d at 1436.
B. The Secretary's Appeal
Turning first to the Secretary's challenge, we face a ques-
tion of statutory interpretation that the district court resolved
in the affirmative: whether, under s 1395ww(d)(5)(A)(iv), the
Secretary must make retroactive reimbursements to ensure
that aggregate outlier payments during any given fiscal year
constitute at least five percent of estimated or projected
DRG-related payments. Because we may set aside agency
action only if it is "arbitrary, capricious, an abuse of discre-
tion, or otherwise not in accordance with law," 5 U.S.C.
s 706(2)(A), we accord no particular deference to the judg-
ment of the district court when conducting our review. See
HCA Health Servs. v. Shalala, 27 F.3d 614, 616 (D.C. Cir.
1994); Hennepin County v. Sullivan, 883 F.2d 85, 91 (D.C.
Cir. 1989) ("Our review proceeds as if this case were an
immediate appeal from a decision reached after an adminis-
trative hearing on the record."); Biloxi Reg'l Med. Ctr. v.
Bowen, 835 F.2d 345, 348-49 (D.C. Cir. 1987).
We initiate statutory analyses of the sort presented here by
first asking whether "Congress has directly spoken to the
precise question at issue." Chevron U.S.A. Inc. v. Natural
Resources Defense Council, Inc., 467 U.S. 837, 842 (1984).
For if after "exhaust[ing] the traditional tools of statutory
construction," Natural Resources Defense Council, Inc. v.
Browner, 57 F.3d 1122, 1125 (D.C. Cir. 1995) (quoting Chev-
ron, 467 U.S. at 843 n.9), we ascertain that Congress's intent
is clear, "that is the end of the matter." Chevron, 467 U.S. at
842. But "if the statute is silent or ambiguous with respect to
the specific issue, the question for the court is whether the
agency's answer is based on a permissible construction of the
statute." Id. at 843. Where judicial review is refracted
through this analytic prism, the "view of the agency charged
with administering the statute is entitled to considerable
deference; and to sustain it, we need not find that it is the
only permissible construction that [the agency] might have
adopted." Chemical Mfrs. Ass'n v. Natural Resources De-
fense Council, Inc., 470 U.S. 116, 125 (1985).
Contending that the statutory language boasts an ambigui-
ty, the Secretary maintains that she has reasonably construed
paragraph (5)(A)(iv) as prescribing a specific methodology
that she must follow when setting outlier thresholds at the
beginning of each fiscal year. Under the Secretary's inter-
pretation, paragraph (5)(A)(iv) mandates that she must select
outlier thresholds which, when tested against historical data,
will likely produce aggregate outlier payments totaling be-
tween five and six percent of projected or estimated DRG-
related payments. Because, under the Secretary's construc-
tion, paragraph (5)(A)(iv) speaks only to how she is to calcu-
late outlier thresholds for the forthcoming year, the Secretary
posits that she has no obligation to ensure that actual outlier
payments for the year total five percent of projected DRG-
related payments.
Advocating a results-oriented approach, however, the Hos-
pitals argue, and the district court agreed, that the Secre-
tary's interpretation contradicts the unambiguous meaning of
paragraph (5)(A)(iv). The textual lodestar guiding their
plain-meaning critique is the single phrase "payments made."
According to the Hospitals, by indicating in paragraph
(5)(A)(iv) that "the total amount of the additional payments
made ... may not be less than 5 percent" of total DRG-
related payments "estimated or projected to be made," 42
U.S.C. s 1395ww(d)(5)(A)(iv) (Supp. IV 1986) (emphasis add-
ed), Congress unmistakably meant that the total amount of
additional payments actually made during a fiscal year must
meet the five-percent target. During years in which the
Secretary's chosen thresholds yield outlier payments that fall
short of the statutory mark, the Hospitals' interpretation
would require the Secretary to bridge the difference by
recalibrating outlier variables and making retroactive pay-
ments accordingly.
Standing alone, however, the phrase "payments made"
hardly conveys a single meaning, much less the one that the
Hospitals advance. As it is employed in paragraph (5)(A)(iv),
"payments made" is "simply an adjectival phrase, not a
verbial phrase indicating the past tense, and hence allows
alternative temporal readings." United States Dep't of the
Treasury v. FLRA, 960 F.2d 1068, 1072 (D.C. Cir. 1992). It
is not unlike the phrase "recognized as reasonable," which the
Supreme Court, quoting favorably from our decision in Ad-
ministrators of the Tulane Educational Fund v. Shalala, 987
F.2d 790 (D.C. Cir. 1993), held "does not tell us whether
Congress means to refer the Secretary to action already
taken or to give directions on actions about to be taken."
Regions Hosp. v. Shalala, 522 U.S. 448, ---, 118 S. Ct. 909,
916 (1998) (quoting Tulane Educ. Fund, 987 F.2d at 796).
Evincing the same syntactical equivalence, the phrase "pay-
ments made" in paragraph (5)(A)(iv), though certainly capable
of accommodating the Hospitals' interpretation, can just as
easily be read to reflect Congress's intent to "give directions
on actions about to be taken." Id. In other words, instead of
embodying a retrospective inquiry into the amount of outlier
payments that have been made, the phrase "payments made
under this subparagraph" might just as plausibly reflect a
prospective command to the Secretary about how to structure
outlier thresholds for payments to be made in advance of each
fiscal year. Cf. Regions Hosp., 522 U.S. at ---, 118 S. Ct. at
916 ("[T]he phrase 'recognized as reasonable' might mean
costs the Secretary (1) has recognized as reasonable ..., or
(2) will recognize as reasonable...."). Ultimately, whether
the phrase is "recognized as reasonable," "adversely affect-
ed," or "payments made," it is difficult to divine with much
confidence the pellucid intent of Congress because "[t]he
language, in short, is ambiguous." United States Dep't of the
Treasury, 960 F.2d at 1072 (describing as ambiguous the
phrase "adversely affected"); accord Tulane Educ. Fund, 987
F.2d at 796.
Hoping to stave off judicial review under Chevron's defer-
ential second step, the Hospitals attempt to resuscitate their
plain-meaning interpretation by contrasting the two ways in
which Congress modified the word "made" in paragraph
(5)(A)(iv). When it first appears, "made" is used without
modifiers to describe the "total amount of the additional
payments made under this subparagraph"; later, the word
materializes to indicate that the total amount of outlier pay-
ments just described may not be less than five percent "of the
total payments projected or estimated to be made" for DRG-
related payments. 42 U.S.C. s 1395ww(d)(5)(A)(iv). Be-
cause, the Hospitals reason, Congress employed words of
estimation and projection to modify the total amount of DRG-
related payments "to be made" but neglected to do so when
describing the total amount of outlier payments "made," it
must have intended that total outlier payments actually made
during a fiscal year would equal at least five percent of
estimated or projected DRG-related payments.2
__________
2 The Hospitals cite only to this court's decision in Washington
Hospital Center v. Bowen, 795 F.2d 139 (D.C. Cir. 1986), to buttress
their argument that the plain meaning of "made" can be inferred
Whatever logic this internal construction of paragraph
(5)(A)(iv) enjoys, to prevent statutory interpretation from
degenerating into an exercise in solipsism, "we must not be
guided by a single sentence or member of a sentence, but
look to the provisions of the whole law." United States Nat'l
Bank v. Independent Ins. Agents of Am., Inc., 508 U.S. 439,
455 (1993) (quoting United States v. Heirs of Boisdore, 49
U.S. (8 How.) 113, 122 (1849)). Under Chevron step one, "we
consider not only the language of the particular statutory
provision under scrutiny, but also the structure and context of
the statutory scheme of which it is a part." Illinois Pub.
Tele. Ass'n v. FCC, 117 F.3d 555, 568 (D.C. Cir. ), modified,
123 F.3d 693 (1997), cert. denied, --- U.S. ----, 118 S. Ct.
1361 (1998); accord Conroy v. Aniskoff, 507 U.S. 511, 515
(1993) ("[T]he meaning of statutory language, plain or not,
depends on context."); Davis v. Michigan Dep't of Treasury,
489 U.S. 803, 809 (1989) ("[W]ords of a statute must be read
in their context and with a view to their place in the overall
statutory scheme."). By examining paragraph (5)(A)(iv) in its
immediate statutory context, any putative clarity that that
__________
from the different language that Congress used to modify that word
in paragraph (5)(A)(iv). That decision misses the mark, however.
In Washington Hospital Center, we presumed that when Congress
amended a pre-existing section of the Medicare statute by adding
and deleting certain words, it must have intended the amended
provision to have a different meaning from its predecessor provi-
sion. Id. at 146. More on point for the Hospitals, though they did
not cite it, would be the canon of construction that posits that
"where Congress includes particular language in one section of a
statute but omits it in another section of the same Act, it is
generally presumed that Congress acts intentionally and purposely
in the disparate inclusion or exclusion." Russello v. United States,
464 U.S. 16, 23 (1983); accord Independent Bankers Ass'n of Am. v.
Farm Credit Admin., 164 F.3d 661, 667 (D.C. Cir. 1999). Of
course, "[c]anons of construction are, after all, only aids in the
process of statutory construction, nothing more, nothing less."
Eagle-Picher Indus. v. United States EPA, 759 F.2d 922, 927 n.6
(D.C. Cir. 1985). As we demonstrate, infra, this canon does not
resolve the ambiguity in paragraph (5)(A)(iv).
provision might arguably have quickly recedes to ambiguity
once again.
Preceding paragraph (5)(A)(iv) by two paragraphs,
s 1395ww(d)(3)(B) provided during the time relevant to this
litigation:
The Secretary shall reduce each of the average standard-
ized amounts under subparagraph (A) ... by a propor-
tion equal to the proportion (estimated by the Secretary)
of the amount of payments under this subsection based
on DRG prospective payment amounts which are addi-
tional payments described in paragraph (5)(A) (relating
to outlier payments)....
42 U.S.C. s 1395ww(d)(3)(B) (Supp. IV 1986) (emphasis add-
ed). Not only does this provision expressly indicate that total
outlier payments are to be estimated by the Secretary, but it
also employs language that closely parallels the language that
later appears in paragraph (5)(A)(iv). In our endeavor to
determine whether the "total amount of the additional pay-
ments made under this subparagraph" contemplates outlier
payments actually made or those estimated to be made, we
find it significant that in paragraph (3)(B) Congress provided
that the "amount of payments ... which are additional pay-
ments described in paragraph (5)(A)" are to be "estimated by
the Secretary." s 1395ww(d)(3)(B). Given that in paragraph
(3)(B) it had already indicated that the Secretary would
estimate the amount of outlier payments described in subpar-
agraph (5)(A), Congress could have reasonably concluded that
there was no need to provide expressly in paragraph
(5)(A)(iv) that the phrase "payments made" referred to pay-
ments estimated to be made. Thus, whatever can be said for
Congress's disparate modification of the word "made" in
paragraph (5)(A)(iv), when we open the analytic aperture to
examine that clause in its proper statutory context, the
inherently ambiguous phrase "payments made" becomes no
clearer.
Nor does a passing observation in the Conference Report
that "the Secretary would be required to provide additional
payments for outlier cases amounting to not less than 5
percent, and not more than 6 percent, of total projected or
estimated DRG related payments," compel us to adopt the
Hospitals' construction of the statute under Chevron step one.
H.R. Conf. Rep. No. 98-47, at 189 (1983), reprinted in 1983
U.S.C.C.A.N. 404, 479 (emphasis added). Ambiguous in its
own right, this passage, if given the gloss that the Hospitals
advance, would chafe against the commentary in the Senate
Report. Suggesting that paragraph (5)(A)(iv) reflects the
prospective inquiry that the Secretary advocates, the Senate
Report provides that the "[t]otal expected payments resulting
from this policy will be not less than 5 percent, nor more than
6 percent, of total medicare payments to hospitals." S. Rep.
No. 98-23, at 51, reprinted in 1983 U.S.C.C.A.N. 143, 191
(emphasis added). The only conclusion that we can safely
draw from these seemingly contradictory passages is that
"the little legislative history that exists for [paragraph
(5)(A)(iv)] is as ambiguous as the statute itself." Deaf Smith
County Grain Processors, Inc. v. Glickman, 162 F.3d 1206,
1212 (D.C. Cir. 1998).
Ultimately, neither the text, structure, nor legislative histo-
ry of paragraph (5)(A)(iv) illuminates Congress's unambigu-
ous intent. Although the Hospitals' interpretation, and the
one that the district court adopted, is plausible, it is not the
"only possible interpretation," Sullivan v. Everhart, 494 U.S.
83, 89 (1990), and it certainly is not "an inevitable one."
Regions Hosp., 522 U.S. at ----, 118 S. Ct. at 917. Because
we find the statute ambiguous, we proceed to assess whether
the Secretary's interpretation of paragraph (5)(A)(iv) is "rea-
sonable and consistent with the statutory scheme and legisla-
tive history." Cleveland v. United States Nuclear Regulatory
Comm'n, 68 F.3d 1361, 1367 (D.C. Cir. 1995).
In marking off the metes and bounds of our review under
the second step of Chevron, we accord particular deference to
the Secretary's interpretation of paragraph (5)(A)(iv) "given
the tremendous complexity of the Medicare statute." Appa-
lachian Reg'l Healthcare, Inc. v. Shalala, 131 F.3d 1050, 1054
(D.C. Cir. 1997); accord Methodist Hosp. v. Shalala, 38 F.3d
1225, 1229 (D.C. Cir. 1994). The Hospitals, however, urge us
not to defer in any way to the Secretary's interpretation of
paragraph (5)(A)(iv) because, they contend, that provision
does not delegate interpretative authority to the Secretary
but explicitly limits her discretion. The problem with this
argument, of course, is that it assumes the truth of the
proposition that we just rejected. Were paragraph (5)(A)(iv)
truly "an explicit limitation on the Secretary's discretion," Br.
of Hosps. at 40, there would be no need to analyze the
provision under Chevron step two. While paragraph
(5)(A)(iv) is certainly designed to regulate the Secretary's
discretion to some extent, as we have already concluded, the
precise contours of that provision are hardly explicit but are
instead ambiguous. Nor is it problematic, as the Hospitals
suggest, that Congress did not expressly delegate interpreta-
tive authority to the Secretary in paragraph (5)(A)(iv). Def-
erence to agency interpretation is warranted "when Congress
has left a gap for the agency to fill pursuant to an express or
implied delegation of authority to the agency." Chevron, 467
U.S. at 843-44 (internal quotations omitted). Where, as here,
Congress enacts an ambiguous provision within a statute
entrusted to the agency's expertise, it has "implicitly delegat-
ed to the agency the power to fill those gaps." National Fuel
Gas Supply Corp. v. FERC, 811 F.2d 1563, 1569 (D.C. Cir.
1987); see also Chevron, 467 U.S. at 843-44.
Equally untenable is the Hospitals' argument that the
Secretary's interpretation is not entitled to deference because
she did not adopt it through either formal rulemaking or
adjudication. But as our precedents make clear, "an agency
need not promulgate a legislative rule setting forth its inter-
pretation of a statutory term for that term to be entitled to
deference." Association of Bituminous Contractors, Inc. v.
Apfel, 156 F.3d 1246, 1251-52 (D.C. Cir. 1998). In fact,
"[e]ven if the legal briefs contained the first expression of the
agency's views, under the appropriate circumstances we
would still accord them deference so long as they represented
the agency's 'fair and considered judgment on the matter.' "
United Seniors Ass'n v. Shalala, 182 F.3d 965, 971 (D.C. Cir.
1999) (quoting Auer v. Robbins, 519 U.S. 452, ----, 117 S. Ct.
905, 912 (1997)); see National Wildlife Fed'n v. Browner, 127
F.3d 1126, 1129 (D.C. Cir. 1997) ("The mere fact that an
agency offers its interpretation in the course of litigation does
not automatically preclude deference to the agency."); Tax
Analysts v. IRS, 117 F.3d 607, 613 (D.C. Cir. 1997).
There is no reason to suspect that the Secretary's interpre-
tation of paragraph (5)(A)(iv) embodies anything other than
her fair and considered opinion. In a final rule published on
January 3, 1984, the Secretary articulated the same interpre-
tation of paragraph (5)(A)(iv) that she has pressed before
both us and the district court. See 49 Fed. Reg. 234, 265
(1984). With that rule, she not only observed that under her
interpretation "there is no necessary connection between the
amount of estimated outlier payments and the actual pay-
ments made to hospitals for cases that actually meet the
outlier criteria," id., but she also admonished Medicare pro-
viders that, in the event she overestimated the amount of
outlier payments, she would "not adjust the DRG rates to
compensate hospitals for funds that were not actually paid for
outlier cases." Id. at 266. Even if, as the Hospitals com-
plain, the final rule failed to provide a "cogent explanation" of
the policies undergirding the Secretary's interpretation, the
fact remains that for the past fifteen years, the Secretary has
never wavered from that interpretation. We are confident
that the interpretation of paragraph (5)(A)(iv) under review
embodies the Secretary's "fair and considered judgment on
the matter." Auer, 519 U.S. at ----, 117 S. Ct. at 912. It,
accordingly, demands deference from the judiciary.
Having settled on the scope of our review, we have no
difficulty concluding that the Secretary has advanced a rea-
sonable interpretation of paragraph (5)(A)(iv). Congress es-
tablished outlier payments because it recognized that "there
will be cases within each [DRG] that will be extraordinarily
costly to treat ... because of severity of illness or complicat-
ing conditions, and [will] not [be] adequately compensated for
under the DRG payment methodology." S. Rep. No. 98-23, at
51 (1983), reprinted in 1983 U.S.C.C.A.N. 143, 191. But as
the term "outlier" suggests, these payments were not intend-
ed to subsidize hospitals simply for treatments, which in
absolute terms, were extraordinarily costly or lengthy. Rath-
er, Congress directed the Secretary to provide "additional
payments for cases which are extraordinarily costly to treat,
relative to other cases within the DRG." Id. (emphasis
added); accord H.R. Rep. No. 98-25, at 134-35 (1983), reprint-
ed in 1983 U.S.C.C.A.N. 219, 353-54 ("Your Committee is
concerned that under the prospective payment system, there
will be cases within each [DRG] that will be extraordinarily
costly to treat, relative to other cases within that DRG....").
Thus the House version would have required the Secretary to
tender outlier payments only when a patient's length of stay
exceeded by thirty days the average length of stay for cases
in that same DRG, see H.R. Rep. No. 98-25, at 135, reprinted
in 1983 U.S.C.C.A.N. at 354, while the Senate bill, which the
Conference Committee adopted, authorized outlier payments
for cases in which a patient's length of stay eclipsed the mean
length of stay for discharges within that same DRG by a fixed
number of days or standard deviations. See 42 U.S.C.
s 1395ww(d)(5)(A)(i) (Supp. IV 1986); S. Rep. No. 98-23, at
51, reprinted in 1983 U.S.C.C.A.N. at 191; H.R. Conf. Rep.
No. 98-47, at 188 (1983), reprinted in 1983 U.S.C.C.A.N. 404,
478.
The Secretary's interpretation of paragraph (5)(A)(iv)
evinces far greater fidelity to Congress's conception of outlier
payments than does the view espoused by the Hospitals.
Under the Secretary's reading of the statute, if it turns out
that actual outlier payments do not meet the five-percent
target at the end of the fiscal year, it is because the lengths of
stay for DRGs in that year proved to be shorter than the
historical averages reflected in the data on which the Secre-
tary based her threshold calculations. Whether it is because
hospitals became more efficient or a miracle drug was intro-
duced during the year, the shorter lengths of stay mean that
there were fewer extraordinarily costly cases during the year.
In other words, there were fewer outliers--and therefore,
fewer outlier payments needed to be made.
Under the Hospitals' interpretation, however, regardless of
actual costs or inpatient lengths of stay during a fiscal year,
Medicare providers are guaranteed a substantial and fixed
sum of outlier payments. As they read the statute, even
during a fiscal year in which the length of stay for every
inpatient discharge in every DRG in every hospital equaled or
exceeded by just a day the mean length of stay for each
respective DRG, the Secretary would nonetheless have to
reward hospitals with additional "outlier" payments totaling
five percent of the entire DRG budget. One need not be well
versed in the discipline of statistics to recognize that such
insignificant deviations from the mean do not constitute outli-
ers. To sanction the Hospitals' interpretation of paragraph
(5)(A)(iv) would not only require us to assume that Congress
did not appreciate the meaning of outlier--a term, it should
be noted, that appears throughout both the legislative history
and the text of the Medicare statute--but it would also
transform the character of the outlier-payment regime from a
system intended to insulate hospitals from aberrational and
extraordinary costs into nothing more than an entitlement
program for Medicare providers. Such was hardly Con-
gress's intent, for if anything, Congress indicated that it was
"equally concerned that adjustments may be required for
cases which have an unusually short length of stay or which
are significantly less costly than the DRG payment." H.R.
Conf. Rep. No. 98-47, at 478, reprinted in 1983 U.S.C.C.A.N.
at 478 (emphasis added); see also H.R. Rep. No. 98-25, at 135,
reprinted in 1983 U.S.C.C.A.N. at 354 ("The Secretary would
be required to study ... the appropriateness of, and necessi-
ty for, adjustments in payment rates for extremely short
lengths of stay within a DRG...."). Proposals like these
reflect a reluctance to reimburse Medicare providers at rates
grossly disproportionate to the cost of treatment. We find it
unlikely that Congress nevertheless would have wanted hospi-
tals to reap additional compensation over and above the
standard DRG payment where treatment costs for a particu-
lar discharge were not extraordinarily costly relative to the
mean costs for that DRG.
Moreover, compared to the Hospitals' interpretation of
paragraph (5)(A)(iv), the Secretary's reading better harmon-
izes each of the four clauses in paragraph (5)(A). As did the
district court, the Hospitals struggle to reconcile their con-
ception of the fourth clause with the language of the third,
which provides that the amount of each outlier payment "shall
approximate the marginal cost of care beyond the cutoff point
applicable" for each DRG. 42 U.S.C. s 1395ww(d)(5)(A)(iii)
(Supp. IV 1986). Under the Hospitals' construction of the
statute, outlier payments might bear little relationship to the
marginal cost of care. At the end of each fiscal year, if actual
outlier payments fell short of the five-percent target, the
Secretary would be required retroactively to supplement
those payments to satisfy the difference. Depending on how
great that initial disparity was, by the time that these cura-
tive outlays were made, the newly computed outlier payments
might not approximate anything close to the marginal cost of
care as paragraph (5)(A)(iii) mandates. By contrast, outlier
payments under the Secretary's interpretation will always
approximate the marginal cost of care because when deter-
mining where to set outlier thresholds for DRGs at the
beginning of each fiscal year, she directly factors the margin-
al cost of care into her calculus.
Echoing the district court's holding, the Hospitals discount
paragraph (5)(A)(iii) as merely a "guideline" while contending
that paragraph (5)(A)(iv) operates "as a limitation on the
Secretary's discretion." County of Los Angeles, 992 F. Supp.
at 32. Based on this view, the Secretary is supposed to set
outlier thresholds at the beginning of each year "at marginal
cost and then, when actual outlier data is known, adjust[ ] the
final payments to ensure that the Secretary has met her
statutory obligation to the providers." Id. at 31. Why this
parsing of the statutory language is more reasonable than
that of the Secretary's--much less compelled as an unambigu-
ously plain reading of the provision, as the Hospitals have
urged--is not at all clear. After all, paragraph (5)(A)(iii)
employs mandatory language of the sort not normally used if
all that Congress intended to do was to offer a discretionary
guideline for the Secretary to follow. See
s 1395ww(d)(5)(A)(iii) ("The amount of such additional pay-
ment under clauses (i) and (ii) shall be determined by the
Secretary and shall approximate the marginal cost of
care....") (emphasis added). Nevertheless, even were the
Hospitals' synthesis of the third clause into the remainder of
paragraph (5)(A) plausible, it would not be enough to impugn
the otherwise reasonable interpretation that the Secretary
has advanced since "we need not find that it is the only
permissible construction that [the Secretary] might have
adopted but only that [her] understanding of this very 'com-
plex statute' is a sufficiently rational one to preclude a court
from substituting its judgment for that of [the Secretary]."
Young v. Community Nutrition Inst., 476 U.S. 974, 981
(1986) (internal quotation omitted).
Moreover, the Secretary's interpretation avoids the sub-
stantial administrative burden attendant with the Hospitals'
vision of paragraph (5)(A)(iv). It strains credulity to assume
that Congress would have directed the Secretary to establish
outlier thresholds in advance of each fiscal year, see
s 1395ww(d)(3)(B), (d)(6), and process millions of bills based
on those figures, only to have her at the end of the year
recalibrate those calculations, reevaluate anew each of the
millions of inpatient discharges under the revised figures, and
disburse a second round of payments. As we have held in an
analogous context, "[u]nder these circumstances, retroactive
corrections would cause a significant, if not debilitating, dis-
ruption to the Secretary's administration of the already-
complex Medicare program." Methodist Hosp., 38 F.3d at
1233. Nor is this administrative process rendered less awk-
ward and unwieldy if, as the Hospitals suggest, the Secretary
actively monitors outlier payments and adjusts the thresholds
as the fiscal year unfolds. Apart from the tremendous re-
sources that would be required to maintain such a vigilant
watch over a program as expansive as Medicare, intermittent-
ly modifying outlier thresholds at various times during the
year would mean that different hospitals would likely receive
different outlier reimbursements for the same treatments
based on nothing more than the fortuity of when they treated
a patient.
By the same token, this uncertainty and fluidity in outlier-
payment amounts under the Hospitals' interpretation lead us
to the final virtue of the Secretary's construction. One of the
touchstones of the Prospective Payment System, as its name
suggests, is prospectively determined reimbursement rates
that remain constant during the fiscal year. In setting, prior
to each fiscal year, fixed outlier thresholds and per-diem
reimbursement rates that are not later subject to retroactive
correction, the Secretary promotes certainty and predictabili-
ty of payment for not only hospitals but the federal govern-
ment--concerns that played a prominent role in Congress's
decision to adopt PPS. See H.R. Rep. No. 98-25, at 132,
reprinted in 1983 U.S.C.C.A.N. at 351 ("The bill is intended
to improve the medicare program's ability to act as a prudent
purchaser of services, and to provide predictibility [sic] re-
garding payment amounts for both the Government and
hospitals."). To be sure, we have previously speculated that
"the real linchpin of the [PPS] system may not be that the
exact reimbursement figure is known in advance, but rather
may be that the hospital knows that nothing it does in
providing services will lead to a higher reimbursement level."
Georgetown Univ. Hosp. v. Bowen, 862 F.2d 323, 330 (D.C.
Cir. 1988) (Georgetown II). Yet while we, therefore, have
recognized that retroactive corrections may not ultimately
undermine PPS, we have emphasized that that "does not
establish that a prospective-only policy is unreasonable."
Methodist Hosp., 38 F.3d at 1232.
In Methodist Hospital, we found the Secretary's decision
not to recalculate retroactively the DRG wage index to be
reasonable, in part, because the Secretary's prospective policy
advanced the principles of PPS.3 With language applicable to
the present case, we held:
__________
3 The Hospitals cannot successfully distinguish Methodist Hospi-
tal. Admittedly, unlike the DRG wage index at issue in Methodist
Hospital, outlier payments do not factor directly into every inpa-
tient discharge. But outlier payments do influence indirectly the
overall DRG payment rate that governs all discharges. As already
discussed, pursuant to s 1395ww(d)(3)(B), the Secretary must re-
duce the standard DRG payment rate by a factor equal to the
outlier payments that she predicts she will have to disburse during
the forthcoming year. Nor is it accurate to claim, as the Hospitals
do, that outlier payments are entirely divorced from PPS. As an
initial matter, the provisions relating to outliers are contained in the
same subsection of s 1395ww as those establishing the PPS regime.
See s 1395ww(d). Moreover, Congress established outlier pay-
ments not as a distinct reimbursement methodology but as a
While retroactive adjustments might leave the "linchpin"
of PPS intact, that does not mean that a prospective-only
policy would not further, to some degree, the overall
goals of PPS.... [H]ospitals, like other businesses, do
make projections about future costs and service levels
based on their experience and historical patterns. To
the extent that the Secretary's prospectivity policy per-
mits hospitals to rely with certainty on one additional
element in the PPS calculation rate ... the Secretary
could reasonably conclude that it will promote efficient
and realistic cost-saving targets.
Id. The same, quite reasonably, can be said of the Secre-
tary's interpretation of paragraph (5)(A)(iv). Under her con-
struction of the statute, at the outset of each fiscal year,
hospitals know the point beyond which a patient's length of
stay will trigger outlier payments and the corresponding rate
at which they will be reimbursed for each day beyond the
threshold. A less determinate policy would not only deprive
hospitals of the ability to make accurate projections about
outlier payments for the forthcoming year but also threaten
them at the end of each year with the prospect of actually
having to forfeit a portion of those payments to the Secretary;
for as the Hospitals concede, under their interpretation,
Medicare providers collectively would be bound to repay any
portion of outlier payments that exceeded six percent of
estimated DRG-related payments. See Br. of Hosps. at 31-
32. We conclude that the Secretary has advanced a reason-
able interpretation of an ambiguous statutory provision, and,
therefore, reverse the judgment of the district court with
respect to the Secretary's appeal.
C. The Hospitals' Cross-Appeal
Because outlier thresholds are measured by their distance
from the mean length of stay, accurately forecasting outlier
__________
carefully crafted supplement to PPS. For that reason, Georgetown
II, which concerned retroactive adjustments under the pre-PPS
"reasonable cost" system--clearly a payment methodology lacking
any relationship to PPS--is inapposite.
payments depends, in large part, on how closely the mean
length of stay reflected in the Secretary's historical data
reflects the actual average length of stay for that particular
DRG. When setting outlier thresholds for fiscal years 1985-
1986, the Secretary relied on the 1981 MEDPAR file, a
database containing patient-specific data for a random sample
of 20 percent of all Medicare-hospital discharges during 1981.
Compiled during an era when Medicare still reimbursed
hospitals under the reasonable-cost system, the 1981 MED-
PAR file could not have predicted how, under PPS, the
average length of stay for virtually all DRGs would eventually
decline dramatically. The Hospitals observe, however, that
by July 27, 1984, the Secretary had already collected data
from 2.5 million discharges under PPS that indicated that the
average length of stay for all DRGs had declined from 9.5
days to 7.5 days under the new payment methodology. Pursu-
ant to section 10(e) of the APA, 5 U.S.C. s 706(2)(A), the
Hospitals claim that the Secretary acted arbitrarily and capri-
ciously when she calculated outlier thresholds for 1985-1986
based on the 1981 MEDPAR file instead of the preliminary
1984 data and failed to explain adequately her decision.
Rejecting the Hospitals' claim, the district court agreed with
the Secretary that her decision to use the 1981 MEDPAR file
over the more contemporaneous but preliminary 1984 data
"was a rational choice between two imperfect databases."
County of Los Angeles, 992 F. Supp. at 36.
Under the APA, we may set aside agency action found to
be "arbitrary, capricious, an abuse of discretion, or otherwise
not in accordance with law." 5 U.S.C. s 706(2)(A). Foreclos-
ed from substituting our judgment for that of the agency, we
do not set aside agency action lightly. See Motor Vehicles
Mfrs. Ass'n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29,
43 (1983); Petroleum Communications, Inc. v. FCC, 22 F.3d
1164, 1172 (D.C. Cir. 1994). Nevertheless, we intervene to
ensure that the agency has "examine[d] the relevant data and
articulate[d] a satisfactory explanation for its action." State
Farm, 463 U.S. at 43. "Where the agency has failed to
provide a reasoned explanation, or where the record belies
the agency's conclusion, we must undo its action." BellSouth
Corp. v. FCC, 162 F.3d 1215, 1222 (D.C. Cir. 1999) (citation
and quotation omitted).
The only contemporaneous explanation that the Secretary
offered for using the 1981 MEDPAR file consisted of two
sentences in the Federal Register: "Based upon outlier and
DRG payment data received through July 27, 1984, there is
no evidence to suggest that total outlier payments are below
the levels intended. Therefore, as discussed above, we are
continuing to set the outlier thresholds on the basis of the
1981 MEDPAR data." 49 Fed. Reg. 34,728, 34,769 (1984)
(emphasis added). We agree with the Ninth Circuit, which
recently considered this same issue, that the Secretary's
"explanation that there was no evidence of an outlier shortfall
was simply not supported by the record before her and did
not explain her failure to use the more recent data." Alvara-
do Community Hosp. v. Shalala, 155 F.3d 1115, 1122 (9th
Cir. 1998). Data that the Secretary possessed as late as July
27, 1984, indicated that the average length of stay for prac-
tically all DRGs had declined considerably under the nascent
PPS program. More concretely, at that point during the
fiscal year, outliers constituted only 1.9 percent of total PPS
discharges instead of 5.0 percent as predicted. And while
these conclusions were drawn from preliminary data, that
data reflected 2.5 million patient discharges under PPS; the
1981 MEDPAR file, by contrast, contained 1.6 million dis-
charge records. Failure to account for this trend is all the
more perplexing insofar as the Secretary herself had antici-
pated that the average length of stay for DRGs would decline
under PPS. In 1984 she observed that "[t]he most commonly
accepted expectation about the PPS at the time of its incep-
tion was that it would result in shorter stays for Medicare
patients.... [R]educed length of stay was to be one of the
major vehicles through which hospital costs were to be con-
trolled under the PPS." Office of Research & Demonstra-
tions, Health Care Fin. Admin., U.S. Dep't of Health &
Human Servs., Pub. No. 03231, Report to Congress: Impact of
the Medicare Hospital Prospective Payment System 6-13
(1984). At bottom, for the Secretary to say that she had "no
evidence to suggest that total outlier payments [were] below
the levels intended," 49 Fed. Reg. at 34,769, runs "counter to
the evidence before the agency" and "is so implausible that it
could not be ascribed to a difference in view or the product of
agency expertise." State Farm, 463 U.S. at 43.
In her brief, the Secretary now contends that what she
meant by "no evidence" was "no reliable evidence." To
bolster this specific claim and her broader argument that the
1984 data were too suspect and incomplete to make accurate
outlier projections, the Secretary appended to her summary
judgment motion in the district court an affidavit from Rose
Connerton, an official with the Health Care Financing Admin-
istration ("HCFA") who helped develop the outlier thresholds
for 1985-1986. Essentially, the Connerton affidavit claims
that the 1984 data were not complete and did not represent a
random sample of cases, that because they were based on a
partial year they would not reflect seasonal and regional
variances, and that any analysis drawn from them would be
skewed.4 See J.A. 90 (Aff. of Connerton pp 10, 12, 15). The
Hospitals contend that the Connerton affidavit, having sur-
faced for the first time during litigation, is an impermissible
post-hoc rationalization that the district court should have
stricken. See SEC v. Chenery Corp., 318 U.S. 80, 87-88
__________
4 Through the Connerton affidavit, the Secretary attempts to
dramatize the unreliability of the partial 1984 data. As of April 27,
1984, reported outliers constituted only 1.9 percent of total PPS
discharges. See J.A. 71. By the end of fiscal year 1984, however,
actual outlier payments ended up totaling 5.3 percent of total PPS
payments, suggesting, Connerton avers, that the preliminary data
were in fact unreliable. Although Connerton's calculations are
accurate, the conclusions that she draws from them are subject to
debate. During a portion of fiscal year 1984, the Secretary errone-
ously provided hospitals with additional outlier payments for non-
PPS-covered treatments, but never sought to recoup these surplus
amounts. That outlier payments amounted to 5.3 percent that year
thus may say less about the reliability of the 1984 data and more
about the scope of the Secretary's clerical error. Whatever the
reason, this dispute underscores the wisdom of Benjamin Disraeli's
sardonic quip (attributed to him by Mark Twain) about the three
great falsehoods: "lies, damn lies, and statistics."
(1943); Reeve Aleutian Airways, Inc. v. United States, 889
F.2d 1139, 1144 (D.C. Cir. 1989). Indeed, faced with a similar
affidavit from Connerton, the Ninth Circuit held that the
district court erred in considering it. See Alvarado Commu-
nity Hosp., 155 F.3d at 1124. Ultimately, we need not reach
this question, for even were we inclined to accept everything
in the Connerton affidavit, we would still remand to the
Secretary for a more adequate justification for her database
selection.
"A long line of precedent has established that an agency
action is arbitrary when the agency offer[s] insufficient rea-
sons for treating similar situations differently." Transactive
Corp. v. United States, 91 F.3d 232, 237 (D.C. Cir. 1996); see
also State Farm, 463 U.S. 29, 57 (1983); Airmark Corp. v.
FAA, 758 F.2d 685, 691-92 (D.C. Cir. 1985); Local 777,
Democratic Union Org. Comm. v. NLRB, 603 F.2d 862, 872
(D.C. Cir. 1978). Although maligning the 1984 data as too
unreliable to calculate outlier thresholds for fiscal years 1985-
1986, the Secretary nonetheless used those same data on
August 31, 1984, to reduce across-the-board all 470 DRG
weighting factors by 1.05 percent. See 49 Fed. Reg. 34,728,
34,770-71 (1984). Such an adjustment was necessary, the
Secretary noted at the time, because "[t]he emerging experi-
ence under the prospective payment system"--an experience
gleaned from the preliminary 1984 data--revealed that the
different incentives that hospitals faced under PPS were
producing unexpected distortions. See id. In making this
correction, the Secretary expressly endorsed the reliability of
the 1984 data: "To date, we have now analyzed 2.5 million
discharges under the prospective payment system, which fully
reflect the effect of those incentives, and we believe this
affords us a better measure of the effect of coding improve-
ments in the average case mix." Id. at 34,771. Moreover, in
responding to a question about the legitimacy of the prelimi-
nary data during a 1984 congressional oversight hearing, the
HCFA Administrator responded that "[o]ur sample now is
based on approximately 50 percent of all of the claims or the
admissions that we had projected for this year. We think
that's a fairly representative sample." Adjustments in Medi-
care's Prospective Payment System: Hearing Before the
Subcomm. on Health of the Comm. on Fin., 98th Cong. 62
(statement of Mr. Davis, Administrator, HCFA). In sum, the
Secretary has inadequately explained why the 1984 data were
suitable for one significant calculation but unreliable for
another. Her sole justification is that preliminary data may
be used to make across-the-board adjustments, as was done
to reduce all DRG weighting factors by 1.05 percent, but that
they may not be used for setting outlier thresholds because a
unique standard deviation must be calculated for each of the
470 DRGs. What renders this explanation inadequate is that
DRG weighting factors, like outlier thresholds, are ordinarily
determined on a DRG-by-DRG basis. Indeed, the very pur-
pose of a DRG weighting factor is to reflect the different
costs of treating minor and major illnesses; to do so, each
DRG must be assigned its own unique weight based on the
cost and complexity of treatment peculiar to that DRG. The
Secretary's proffered distinction is thus not reasonable. She
may in her discretion, of course, rely on preliminary data to
make an across-the-board adjustment to variables that ordi-
narily are determined on a case-by-case basis. But when she
does so, she must be prepared to explain why she cannot also
use that data to make a similar adjustment to variables that
are also typically calculated on an individual basis. As broad
as her discretion is, it "is not a license to ... treat like cases
differently." Airmark Corp., 758 F.2d at 691; accord Teva
Pharms., USA, Inc. v. FDA, 182 F.3d 1003, 1010-11 (D.C.
Cir. 1999); Transactive Corp., 91 F.3d at 237; Local 777, 603
F.2d at 872.
This case must therefore be remanded to the Secretary to
allow her either to recalculate outlier thresholds for fiscal
years 1985-1986 or to offer a reasonable explanation for
refusing to use the 1984 data in setting outlier thresholds
during those years. In reaching this conclusion, we necessar-
ily part ways with the Ninth Circuit, which, in Alvarado
Community Hospital, chose not to remand to the Secretary,
but instead ordered her to adjust outlier thresholds for fiscal
year 1985 based on final 1984 data. Alvarado Community
Hosp., 155 F.3d at 1125. As the Supreme Court has instruct-
ed, however, where "the record before the agency does not
support the agency action, ... the proper course, except in
rare circumstances, is to remand to the agency for additional
investigation or explanation." Florida Power & Light Co. v.
Lorion, 470 U.S. 729, 744 (1985); see also Dunlop v. Bachow-
ski, 421 U.S. 560, 574-75 (1975) ("Where the statement inade-
quately discloses his reasons, the Secretary may be afforded
opportunity to supplement his statement."), overruled on
other grounds by Furniture & Piano Moving v. Crowley, 467
U.S. 526, 549-50 n.22 (1984). We find no reason to depart
from that course here. While we have identified significant
inconsistencies and gaps in the Secretary's rationale for using
the 1981 MEDPAR file, bedrock principles of administrative
law preclude us from declaring definitively that her decision
was arbitrary and capricious without first affording her an
opportunity to articulate, if possible, a better explanation.
See Bechtel v. FCC, 10 F.3d 875, 887 (D.C. Cir. 1993);
Philadelphia Gas Works v. FERC, 989 F.2d 1246, 1251 (D.C.
Cir. 1993); Sullivan Indus. v. NLRB, 957 F.2d 890, 905 n.12
(D.C. Cir. 1992); Tex Tin Corp. v. EPA, 935 F.2d 1321, 1324
(D.C. Cir. 1991); see also Checkosky v. SEC, 23 F.3d 452, 463
(D.C. Cir. 1994) (Silberman, J., concurring) (citing some of the
"many instances where we have remanded to an agency for a
better explanation before finally deciding that the agency's
action was arbitrary and capricious"). Because we fail to
perceive any "rare circumstances" that would warrant a
break with established administrative practice, we adhere to
the proper course of remanding this matter to the Secretary.
III. Conclusion
For the foregoing reasons, we reverse the judgment of the
district court with respect to the Secretary's appeal, and
remand with instructions to enter judgment in the Secretary's
favor. As for the Hospitals' cross-appeal, we reverse the
judgment of the district court, and instruct it to remand the
case to the Secretary for further proceedings consistent with
this opinion.
So ordered.