TITLE:  SmithKline Beecham Corporation, B-283939, January 27, 2000
BNUMBER:  B-283939
DATE:  January 27, 2000
**********************************************************************
SmithKline Beecham Corporation, B-283939, January 27, 2000

Decision

Matter of: SmithKline Beecham Corporation

File: B-283939

Date: January 27, 2000

John E. Daniel, Esq., Morgan, Lewis & Bockius, for the protester.

Thomas L. McGovern III, Esq., and Joy E. Sturm, Esq., Hogan & Hartson, for
Glaxo Wellcome Inc.; and David C. Hammond, Esq., Powell, Goldstein, Frazer &
Murphy, for Hoechst Marion Roussel, the intervenors.

Barbara J. Stuetzer, Esq., and Phillipa L. Anderson, Esq., Department of
Veterans Affairs, for the agency.

Ralph O. White, Esq., and Christine S. Melody, Esq., Office of the General
Counsel, GAO, participated in the preparation of the decision.

DIGEST

Protester's challenge to the price evaluation scheme included in a
solicitation for prescription drugs that anticipates evaluation of a
per-dose price based on the only use for which all three of the competing
drugs are approved by the Food and Drug Administration is denied, even
though the evaluation does not consider certain uses of the solicited drugs
that will have a different cost profile, where the per-dose price requested
provides a common basis for evaluating prices, the agency has no basis for
providing estimates for the other uses of these drugs, and the protester has
not established that the solicitation's approach will produce a materially
misleading result.

DECISION

SmithKline Beecham Corporation (SKB) protests the terms of request for
proposals (RFP) No. RFP-797-NC-99-0020, issued by the Department of Veterans
Affairs (VA) for 5-Hydroxtryptamine (5HT3) receptor antagonists, used for
the treatment of nausea and vomiting (emesis) resulting from chemotherapy.
SKB argues that the RFP's price evaluation methodology will not identify the
proposal that will provide the actual lowest price to the government.

We deny the protest.

This procurement is part of the VA's program to standardize the availability
of pharmaceuticals and medical/surgical items, through the use of its
national formulary. A formulary is a list of prescription drugs, grouped by
therapeutic class, that a health care organization prefers that its
physicians prescribe. Drugs are chosen for a formulary on the basis of their
medical value and price. Defense Health Care: Fully Integrated Pharmacy
System Would Improve Service and Cost-Effectiveness, (GAO/HEHS-98-176, June
12, 1998) at 2 n.1. In addition to standardizing drug availability, the VA
is using its national formulary to increase the continuity of VA care,
standardize the processes for evaluating the safety and efficacy of drugs,
and manage cost growth. VA Health Care: VA's Management of Drugs on its
National Formulary, (GAO/HEHS-00-34, Dec. 14, 1999) at 4.

Prior to issuing this solicitation, the VA's Pharmacy Benefits Management
(PBM) section, together with its Medical Advisory Panel (MAP), conducted a
drug class review of the three known 5HT3 receptor antagonists, [1] which
are: Dolasetron, manufactured by Hoechst Marion Roussel, Inc., and marketed
under the name Anzemetï¿½; Granisetron, manufactured by SKB, and marketed
under the name Kytrilï¿½; and Ondansetron, manufactured by Glaxo Wellcome,
Inc., and marketed under the name of Zofronï¿½. The review concluded that the
three drugs were equally efficacious in managing CINV and PONV; that they
presented similar, and acceptable, levels of side effects; and that a single
5HT3 drug should be selected for the national formulary based on cost.
Agency Report (AR), Tab 7, Drug Class Review, Oct. 1999, at 13.

The RFP, issued on July 21, 1999, sought fixed-price proposals to provide an
indefinite quantity of the three 5HT3 receptor antagonists, and advised that
award would be made to the offeror whose proposal was deemed most
advantageous to the government considering price and past performance. RFP
amend. 0003, at 2, 5. Of these two factors, offerors were advised that price
would be more important than past performance, and that the agency's
objective was to award to the offeror with the lowest per-dose price (PDP).
Id. at 5.

To perform its PDP comparison of the three drugs, the RFP established an
oral dosing level for each drug, as set forth below:

 PRODUCT         STRENGTH       PATIENT        NO. OF
                                DOSAGE         TABLETS

 Dolasetron      100 mg         100 mg         1

 Granisetron     1 mg           2 mg           2

 Ondansetron     8 mg           16 mg          2

RFP amend. 0003, at 2. The VA explains that it selected these dosing levels
for its price evaluation because they are the FDA-approved dosing levels for
these three drugs when treating patients with moderate CINV, and because the
treatment of moderate CINV is the only condition for which there are
FDA-approved dosing levels for the oral version of all three drugs. [2] AR,
Legal Memorandum, Nov. 18, 1999, at 5; see also AR, Tab 7, Drug Class
Review, Oct. 1999, at 2. Thus, put simply, the PDP will be based on the
price for 1 tablet of Dolasetron, and 2 tablets each of Granisetron and
Ondansetron.

SKB protests that the solicitation's price evaluation scheme is defective
because it does not give reasonable assurances that an award to the offeror
with the lowest evaluated price will result in the lowest cost to the
government during actual performance. This argument is based on the premise
that some portion of the VA population will need treatment for high CINV,
rather than moderate CINV, and that the treatment of patients with high CINV
using these drugs will have a relative cost different from the relative cost
of the dosages used to treat moderate CINV.

The key to SKB's argument is the fact that the same dosage level (2 mg) of
SKB's product, Granisetron, is approved by the FDA for the treatment of high
and moderate CINV. In contrast, Ondansetron has been approved for the
treatment of high CINV at a 24 mg dose; and there is no FDA-approved oral
dose of Dolasetron for the treatment of high CINV. AR, Tab 7, Drug Class
Review, Oct. 1999, at 11-12. To illustrate the potential difference in
pricing, the table below includes: (1) the currently published Federal
Supply Schedule (FSS) prices for each of the three 5HT3 drugs; (2) the
dosage levels (and resulting FSS prices) for treating moderate CINV; and (3)
the dosage levels (and prices) for treating high CINV.

 PRODUCT       FSS PRICE             MOD. CINV       HIGH CINV
                                     DOSE/PRICE      DOSE/PRICE

 Dolasetron    $33.83 per 100 mg     1               -- [3]
               tablet                tablet/$33.83

 Granisetron   $21.80 per 1 mg       2               2
               tablet                tablets/$43.60  tablets/$43.60

 Ondansetron   $17.00 per 8 mg       2               3
               tablet                tablets/$34.00  tablets/$51.00

Id.; SKB's Comments, Dec. 3, 1999, at 5 n.5. Thus, as shown above, SKB's
Granisetron is the most expensive of the 5HT3 drugs (using current FSS
prices) for treating moderate CINV, and the least expensive for treating
high CINV. This is the basis for SKB's argument that considering only the
dosing levels for moderate CINV will not provide the lowest cost to the
government in actuality.

The VA defends its price evaluation scheme as a reasonable attempt to
objectively compare the price of these three drugs. As explained above, the
VA notes that the treatment of moderate CINV is the only situation where
there are FDA-approved oral dosage levels for all three drugs. It argues
that if it attempted to calculate a PDP for treatment of high CINV, it would
need to extrapolate an oral dosage level for Dolasetron, which would lead to
further disputes. The VA also explains that while it can provide the
offerors with the number of tablets of each drug used in prior years, it
does not have historical data on, or a prospective estimate of, the portion
of its patient population receiving chemotherapy who present high CINV
versus moderate CINV. Finally, the VA explains that the protester's approach
to focusing only on the distinction between high and moderate CINV is overly
simplistic because the treatment of high CINV often involves combining a
moderate dose of one of the 5HTC drugs together with other antiemetic
agents. [4]

The Competition in Contracting Act requires that agencies consider the cost
to the government in evaluating competitive proposals. 41 U.S.C. sect.
253a(b)(1)(A), (c)(1)(B) (1994); Health Servs. Int'l, Inc.; Apex Envtl.,
Inc., B-247433, B-247433.2, June 5, 1992, 92-1 CPD para. 493 at 3-4. While it is
up to the agency to decide upon an appropriate and reasonable method for
proposal evaluation, it may not use an evaluation method that produces a
misleading result. Id. at 4. Such method must include some reasonable basis
for evaluating or comparing the relative costs of proposals, so as to
establish whether one offeror's proposal would be more or less costly than
another's. See Health Servs. Int'l, Inc.; Apex Envtl., Inc., supra; Penn,
Ferrara, Adler & Eichel, B-224224, Feb. 9, 1987, 87-1 CPD para. 134 at 3-4.

As a preliminary matter, we note that SKB's premise that some portion of the
VA patient population will need treatment for high CINV, rather than
moderate CINV, is unassailable. We also note that the VA's price evaluation,
by definition, does not consider the dosing levels required to treat high
CINV. In addition, we agree that the current FSS prices for these three
drugs suggest that award to the offeror whose PDP presents the lowest cost
for the treatment of moderate CINV may not, in every instance, result in the
lowest cost to the government during actual performance, depending on how
many VA patients require treatment for high CINV. The question for our
Office is not, however, whether the VA's approach to evaluating prices is
perfect, but whether, despite any limitations, it is reasonable under the
circumstances here. For the reasons set forth below, we conclude that it is.

SKB argues that, as a legal matter, the VA's approach to evaluating prices
here is analogous to the solicitation we reviewed in Globe Air, Inc.,
B-188611, June 6, 1977, 77-1 CPD para. 395, wherein we sustained a protester's
challenge to an agency's approach to evaluating bids for helicopter services
on the basis that the solicitation violated the statutory mandate to
consider the cost of performance in making a contract award. In sustaining
that protest, we noted that the price evaluation considered only one
function (and the associated costs) of the helicopter services that, upon
review, represented only 25 percent of the helicopter time required by the
agency. Globe Air, Inc., supra, at 4. The record showed that for the
remaining 75 percent of the helicopter time required by the agency, a
different set of costs would apply, and also showed that a different offeror
would likely prevail if those costs were considered. Id. at 5.

In our view, Globe Air is distinguishable from the facts presented here. In
Globe Air, the agency had data showing that its price evaluation methodology
accounted for only 25 percent of the helicopter time required. Thus, the
methodology failed to consider the costs for the remaining 75 percent of the
time, and did not provide an accurate determination of the lowest overall
cost to the government. Here, the VA explains that while it has data on the
quantity of 5HT3 drugs it has used in the past, it has no data on the
portion of its patient population that will need treatment for high CINV
versus moderate CINV. In this situation, the VA contends that its
solicitation provides a reasonable and objective basis for comparing the
costs of these drugs, while also providing information about the amount of
drugs that will be needed.

We agree. Under the circumstances here, we think the more appropriate
analogy is to our decision in Aalco Forwarding, Inc., et al., B-277241.15,
Mar. 11, 1998, 98-1 CPD para. 87. There, in a procurement for moving services
where the agency had no historical data regarding the need for different
accessorial services associated with international moves, the solicitation
set forth a notional shipment, similar to a sample task, for which each
offeror was asked to submit a price. The decision there held that where
estimates for various types of required services are not reasonably
available, an agency may establish a reasonable hypothetical, consistent
with the RFP requirements, to provide a common basis for comparing the
relative costs of the proposals. [5] Aalco Forwarding, Inc., et al., supra,
at 11.

In addition to our view that the agency's approach appears reasonable given
the limited information available to it, we also note that the protester has
offered little evidence that it will be harmed by the solicitation's
approach to evaluating prices. First, while we have no way of knowing what
percentage of the VA patient population will need treatment for high CINV,
the only evidence in the record suggests that it is a relatively small
percentage of the VA patients receiving chemotherapy. Specifically, one of
the intervenors, Glaxo Wellcome, submitted evidence with its comments which
suggests that only 10 percent or less of the VA patient population receiving
chemotherapy will need treatment for high CINV. Glaxo Wellcome Comments,
Dec. 3, 1999, exh. 4. In contrast, SKB's comments included a table showing
the cost per 100 VA patients (using FSS prices) with different percentages
of patients needing treatment for moderate and high CINV. SKB Comments, Dec.
3, 1999, at 5. This table shows that even if the patient mix is 50-50
moderate/high CINV, SKB's drug would not be the lowest-priced option for
treating the VA patient population. Instead, SKB's own materials do not show
SKB to be the offeror with the lowest price until 75 percent of the patient
population needs treatment for high CINV. [6]

Finally, we note that the national formulary is a listing of drugs, not a
set of guidelines for their use. The drug that prevails in this competition
will be listed on the formulary for VA physicians and pharmacies dispensing
a 5HT3 receptor antagonist. How the drug is prescribed, how it is dosed, and
which drug is prescribed for which indication, remains within the medical
judgment of the VA's treating physician. This is not a requirements
contract; the two other 5HT3 drugs will remain on the FSS, and will continue
to be available to VA physicians if the contracted item is not the
appropriate drug treatment therapy. Further, the record here shows that
simply increasing the dosage of a 5HT3 drug--which is the foundation of
SKB's argument that the solicitation's PDP evaluation method may not yield
the lowest price in practice--may not be the most desirable medical approach
to treating high CINV. These facts, unrebutted during this protest, further
dilute the possibility that the VA's price evaluation method will fail to
reflect the most likely actual cost to the government of using these drugs.

Accordingly, without some reason to conclude that the agency's approach will
lead to an unreasonable evaluation of the likely cost to the government, and
given the fact that the VA lacks data on the percentage of its patients that
will need treatment for high, versus moderate, CINV, we have no basis to
conclude that the solicitation's intended approach to evaluating prices is
improper.

The protest is denied.

Comptroller General
of the United States

Notes

1. The VA's drug class review for the 5HT3 drugs analyzed efficacy of the
drugs in the prevention and treatment of postoperative nausea and vomiting
(PONV), the prevention of radiotherapy-induced nausea and vomiting (RINV),
and the prevention of highly and moderately emetogenic chemotherapy-induced
nausea and vomiting (CINV). AR, Tab 7, Drug Class Review, Oct. 1999, at 2.
The protester's arguments focus on differences in the drugs in combating
highly and moderately emetogenic CINV.

2. As discussed further below, only two of the drugs, Granisetron and
Ondansetron, have FDA-approved dosing levels for treating patients with high
CINV.

3. While there is not yet an FDA-approved dosage of oral Dolasetron for the
treatment of high CINV, the protester points out that the American Society
of Health-System Pharmacists Therapeutic Guidelines (included in the
agency's report) state that "a dose of 200 mg [of Dolasetron] may be
required in patients receiving highly emetogenic chemotherapy." AR, Tab 21,
at 742. Using this guidance, the price for using Dolasetron to treat high
CINV would be $67.66 per dose.

4. Moving beyond the protester's focus on differences in the dosage levels
of these drugs to treat moderate and high levels of CINV, we note that there
are other uses of these drugs--such as the prevention of PONV and RINV--that
are also not captured by the PDP comparison set forth in the solicitation.

5. An issue in Aalco that is not present here was a contention that the
composition of the agency's so-called "notional shipment" was unreasonable.
Here, the protester does not dispute the solicitation's dosage levels for
the treatment of moderate CINV. SKB's Comments, Dec. 3, 1999, at 3 n.1.

6. To be fair to SKB, we note that it offered these figures to demonstrate
its point that the VA's price evaluation methodology could result, under
some circumstances, in award to an offeror whose approach might actually
yield a higher overall cost to the government. SKB points out, and we agree,
that there is no reason to conclude that the prices received in response to
this solicitation will be the same as the currently available FSS prices. On
the other hand, these materials suggest that without a dramatic shift in the
relative prices of these drugs, SKB will not be harmed by the VA's price
evaluation scheme unless a sizable percentage of the VA patients receiving
chemotherapy--probably more than a majority--need treatment for high CINV.