FTC Staff Comment Before the Department of Health and Human Services Concerning Medicaid Reimbursement of Pharmacies

BERC-356-E.
COM.\1ENTS OF THE BUREA US OF ECONOMICS,
CONSUMER PROTECTION AND CO\1PETITION OF THE
FEDERAL TRADE COM\lISSION
Washington, D.C. 20580
October 20, 1986
Introduction
The staff of the Federal Trade Commission ("FTC") is pleased to offer
comments on 	the 	Notice 	of a 	Proposed 	Rule 	published 	by the 	Health 	Care
Financing 	Administration 	("HCFA") 	of 	the 	U.S. 	Department 	of 	Health 	and
Human Services. l 	HCFA requests comments on three alternative systems for
determining the maximum level of federal funding for state reimbursement of
retail pharmacies for drugs dispensed to Medicaid customers. 	In fiscal 	year
1984, Medicaid program expenditures on prescription drugs totalled a bou t $2
billion. 2
The FTC has a long-term interest in the development of competition in
the 	prescription 	drug 	market, 	as 	reflected 	in 	a 	number 	of 	FTC 	studies
assessing competitive conditions in that market. 3 	Further, the FTC and 	the
1 	On 	the 	subject 	of 	Limits 	on 	Payments 	for 	Drugs 	in 	the 	Medica:d
Program, file 	cede BERC-356-P, 51 	F.R. 29,560 (August 	19, 1986) (hereafter
referred 	to 	as 	the 	"Notice"). 	These 	comments 	represent 	the 	views 	of 	the
Bur~aus of 	Economics, 	Consumer 	Protection 	and 	Competition 	and 	do 	not
necessarily 	reflect 	the 	views 	of 	the 	Federal 	Trade 	Commission 	or 	any
individual Commissioner. 	The Commission, however, has authorized the filing
of these 	comments. 	Inquiries 	regarding 	these comments should 	be directed
to John Woodbury, Bureau of Economics.
2 	R. 	Helms, 	"Statement 	by 	the 	Acting 	Assistant 	Secretary 	for
Planning and Evaluation, Department of Health and Human Services, Defore
the House Subcommittee on Health and 	the Environment" (July 	15, 1985), p.
2.
3 	These studies 	include 	R. Bond and 	D. Lean, Promotion and 	Product
Differentiation 	in 	Two 	Prescription 	Drug 	\1arkets 	(FTC', 	February 	1977);
Bureau 	of 	Consumer 	Protection, 	FTC, 	Drug 	Product 	Selection 	(FTC, 1979)
(hereafter, "Bep"); and 	A. 	Masson 	and 	R. Steiner, Generic Subst::ution 	and
Prescription Drug Prices (FTC, October 1985) (hereafter, "Masson and Steiner").

Food and 	Drug 	Administr:ltion ("FDA") jointly develope': :l model state law
th:lt 	would 	permit 	pharm3cists to substitute lower-price: 	··ersions of a drug
for 	more 	expensive 	versions. 4 	It 	is 	against 	the 	:-:;.::kground 	of 	our
experience and 	expertise in examining the prescription 	c~:..:g market that we
offer these comments to HCF A.
In 	section 	I, 	we 	describe 	the 	current 	reimburserne::! 	scheme 	and 	the
proposed 	alternatives. 	Section 	II 	discusses 	the 	goals 	against 	which 	the
alternative 	proposals 	should 	be 	evaluated. 	In 	sections 	III, 	IV 	and 	V, 	we
assess 	the 	extent 	to 	which 	the 	proposed 	alternatives 	a,e 	likely 	to 	attain
these 	goals. 	In 	section 	VI, 	we 	summarize 	our 	_conclusions 	and
recommenda tions.
In 	brief, 	the 	touchstone 	of 	our 	analysis 	is 	econor:::: 	efficiency. 	No
reimbursement scheme should unnecessarily skew marketpJ::.:e outcomes to the
detriment of Medicaid and non-Medicaid consumers alike. 	Unfortunately, the
current 	reimbursement 	scheme 	may 	have 	unduly 	hindered 	competition 	by
limiting 	the 	incentive 	to 	dispense 	the 	lower-cost 	versio~5 of 	drugs. 	This
may 	have 	occurred 	because 	relatively 	few 	of 	the 	dr::~ 	categories 	with
multiple versions of the drug available were included in :he current scheme
that 	encourages 	pharmacists 	to 	dispense 	lower-cost 	versio:ls of 	the 	drug 	to
Medicaid 	consumers. 	We 	therefore 	endorse 	HCFA's 	intention 	under 	each
alternative to expand 	the number of drug categories for 	which 	there is an
incentive 	to 	dispense 	lower-cost 	versions of the drug. 	F~rther, we suggest
that 	HCFA 	consider 	complementing 	the 	pharmacist's 	incentive 	to 	dispense
lower-priced versions in a drug category with an incenti"e (::amely, a
4 	See BCP, pp. 273-288.
2

percentage 	copayment) 	for 	Medicaid 	consumers 	to search 	:lmong 	ph:lrmacies
for lower-cost drugs.
However, 	none 	of 	the 	proposals permit Medicaid 	co::sumers 	to register
their preferences for a particular version of a drug, inclu.::ng a brand name.
As a 	result, 	the 	proposals 	may 	unnecessarily 	discourage 	tr:e development 	of
brand-name 	reputations 	and 	non-price 	differences 	amor.g 	therapeutically-
equivalent drug products. 	Brand names may provide consumers with valuable
information 	regarding 	non-price 	differences 	among 	versions 	in 	a 	particular
drug 	category. 	We 	therefore 	suggest 	that 	in 	whatever 	s:heme 	is 	adopted,
HCFA 	permit 	Medicaid 	consumers 	to 	purchase 	the 	versic::l. of a 	drug 	they
prefer 	if they 	are 	willing 	to 	pay 	the 	difference 	between 	the 	price 	of 	the
preferred version and the reimbursement level determined by HCFA.
For each proposal, we also suggest that 	HCF A carefully consider wha t
the most efficient distribution system for 	the delivery of prescription drugs
to Medicaid consumers might 	be. 	A lower reimbursement level will 	reduce
the 	costs 	of 	the 	program 	to 	the 	government. 	However. 	the 	number 	and
quality (for example, locational convenience) of participating pharmacies may
fall as the reimbursement level is reduced. 	Consequently, the accessibility of
pharmacies to Medicaid consumers may decline. 	In selecting the reimburse-
ment level that yields the most efficient distribution system, we suggest that
HCFA 	balance 	the 	gain 	in 	government 	savings 	from 	lower 	reimbursement
levels 	against 	the 	costs 	of 	reduced 	accessibility 	of 	pharmacies 	to 	Medicaid
consumers.
While 	we 	offer 	a 	number 	of 	suggestions 	concerning 	reimbursement
schemes, 	a 	lack 	of 	available 	data 	prevents 	us 	from 	determining 	which 	of
3

the 	proposed 	alternatives 	is 	best. 	We 	suggest 	the 	kind 	0:" 	further 	analysis
that might be undertaken to eliminate these information ga;:s.
1. THE CURRENT AND PROPOSED :-'1EDICAID REl~1BCRSE\fE0JT SYSTE\1S
FOR DRUGS
While 	the 	states 	are 	directly 	responsible 	for 	imple:::.enting 	Medicaid
programs, 	the 	rules 	governing 	federal 	reimbursement 	of 	state 	Medicaid
expenditures exert a strong influence on the state programs. 	The purpose of
the 	current 	federal-state 	program 	is 	to 	reimburse 	the 	retail 	pharmacist 	for
the 	entire 	cost 	of 	a 	Medicaid 	prescription 	drug 	beyond 	the 	small 	fixed
copayment that may be paid by the consumer. 5 	States set C:1ny rules, which
may 	include 	the 	consumer's 	copayment, 	a 	list 	of 	c:-ugs 	for 	which
reimbursement will be made, mandatory substitution of lowe:- cost versions of
a 	drug 	for 	higher 	cost 	versions, 	and 	the 	amounts 	of 	rei:::bursement. 	The
state 	in 	turn 	is 	reimbursed 	by 	federal 	monies, 	in 	accord::nce 	with 	HCFA
regulations. 	It is these HCFA regulations that are now to be :-evised.
The federal Medicaid drug reimbursement system curre:1tly in place is a
composite .of several 	interlocking elements. 	The 	HCFA 	regulations 	for 	the
reimbursement for 	multisource drug categories (i.e, drug c::.:egories with two
or 	more 	therapeutically-equivalent 	versions 	a v ailable)6 req::ires 	payment 	by
5 	Twenty-two 	states 	currently 	require 	a 	fixed;opayment, 	most
frequently 	$.50 or $1.00. 	Drug 	Topics (March 	24, 	1986), 	p. 29. 	For some
drug 	categories 	with 	multiple 	versions 	of 	the 	drug 	availz:le, 	these 	typical
copayments amount to 9 percent and 	18 percent respectively of the average
price of the 10"'Jer-priced versions. 	The price data used is from Professional
Management 	Associates, 	Inc., 	Interim 	Report 	on 	Medicaj-' 	Reimbursement
Policies 	(September 	8, 	1986), 	Table 	I, 	Column 	I. 	The 	imposition 	and
magnitude of copayments are governed by 42 U.S.c. Section 1396(0).
6 	Multisource 	drug 	categories 	are 	those 	with 	two 	or 	more 	products
containing 	the 	same 	active 	chemical 	ingredients. 	Many 	:nultisource 	drug
categories 	contain 	products 	that 	are 	both 	chemically 	ac: 	therapeutically
equivalent. 	In 	this 	Comment, 	we 	confine 	our 	attention 	to 	that 	class 	of
multisource 	drug 	categories 	that 	contain 	chemically 	an': 	therapeutically
4

HCFA 	of 	the 	lowest 	of 	a) 	the 	usual 	and 	customary 	re:3.il 	price 	of 	the
indi\'idual pharmacist In question for the exact product dis;;ensed, h) the sum
of 	a 	dispensing 	fee 	and 	the 	Estimated 	Acquisition 	Ccst 	(to 	the 	retail
pharmacy) 	("EAC") 	of 	the 	drug 	product 	or 	c) 	the 	sum 	of 	3. 	dispensing 	fee
and 	the Maximum 	Allowable 	Cost ("MAC") 	of 	the 	drug. 	The 	MAC system
was 	designed 	to 	encourage 	pharmacists 	to 	dispense 	low-:ost 	generic 	drug
products. 	For 	multisource 	drugs 	for 	which 	no 	~1AC 	limit 	has 	been
esta blished, reimbursement is the lower of (a) or (b).7
In 	place of the 	current scheme, HCF A 	proposes to adopt one of three
alternatives for multisource drug reimbursement: 	(a) a re\'ised MAC program;
(b) a Pharmacists Incentive Program ("PhIP"); or (c) a Co~petitive Incentive
Program ("CIP"). 	For all 	three alternatives, the list of drugs to be included
in the special multisource 	reimbursement schemes would encompass all drug
categories that are deemed by the FDA to contain therapeutically equivalent
equivalent 	drug 	products. 	Within 	that 	class, 	we 	often 	refer 	to 	"geDeric"
versions of a drug, products chemically and therapeutically equivalent to the
leading 	or 	pioneer 	versions 	of 	the 	drug. 	Finally, 	the 	discussion 	below
focuses 	only 	on 	schemes 	which 	reimburse 	the 	pharmacist 	for 	dispensing
particular products in a multisource drug category when the pharrr:.1cist has
the 	discretion 	to 	select 	the 	particular 	product 	dispensed. 	The 	FTC's
expertise 	is 	most 	extensive 	in 	the 	assessment 	of 	competition 	in 	the
multisource drug categories.
7 	The 	prescription 	dispensing 	fee 	is 	set 	in 	each 	state 	after
pharmacies' costs of dispensing are surveyed and analyzed. 	Dispensing fees
are updated occasionally. 	The EAC is established by the states by reference
to 	published sources of advertised 	wholesale prices for 	individual 	products.
Finally, 	the 	MAC 	system 	applies 	to 	a 	small 	list 	of 	multisource 	drugs 	for
which 	the 	maximum 	allowable 	cost 	for 	each 	drug 	category 	is 	currently
established 	at 	the 	70th 	percentile 	of 	actual 	wholesale 	invoice 	costs 	of
pharmacists as determined from survey data. 	Thus, pharmacies dlspensing a
drug under the MAC system receive a fixed level of reimbursement regardless
of the particular manufacturer's product actually dispensed, provided that the
EAC of 	the 	drug 	product dispensed 	is 	greater 	than 	the \L\c. 	The list of
multisource drugs for which MAC limits are determined is set by HCFA after
a period 	of 	public comment and 	in 	consultation 	with 	the 	FDA. 	The MAC
limits are also established by HCFA after a period of public comment.
5

products 8 	and 	for 	which 	the 	advertised 	wholesale 	prJces 	of 	at 	least 	three
suppliers are 	included 	in 	the most current edition 	of the 	Red 	Book or Blue
Book. standard industry sources.
The MAC limits 	in 	the 	revised 	MAC proposal 	would 	be determined 	in
streamlined administrative 	proceedings. 	EAC and 	the dispensing fee 	would
continue to be determined as in the current program.
Under the PhIP proposal, reimbursement of multisource drugs would be
:' specific percentage (for example, the HCFA - proposed 	150 percent) of the
lowest 	price 	advertised 	in 	each 	multisource 	drug 	category 	in 	the 	most
current Red Book or Blue Book, plus an estimated reasonable dispensing fee.
However, 	HCFA 	also 	proposes 	to 	limit 	the 	payment 	to 	no 	more 	than 	$4.00
and no less than $1.50 above the lowest price advertised in each multisource
drug category.
Under 	CIP, 	HCFA 	proposes 	to 	use 	as 	the 	basis 	of 	its 	reimbursement
limits 	for 	all 	(not 	just 	multisource) 	drugs 	the 	actual 	price charged 	by 	the
pharmacist to its non-Medicaid customers, less a discount of 5 or 10 percent.
However, if the pharmacist chooses to dispense a higher-priced leading brand
when a lower priced generic version could ha ve been dispensed, the discoun t
rises to 25 percent. 	HCFA is also proposing that in 	no case would the CIP
reimbursement exceed 125 percent of the median retail price in the state for
a particular manufacturer's product.
8 	The list includes all drug products in a specific drug category that
are identical in 	terms of 	their active chemical ingredients, dosage form and
strength and 	that 	have 	been 	classified as "A" in 	the FDA's A::mroved 	Drug
Products with Therapeutic Equivalence Evaluations.
6

II. 	GOALS OF REVISION
In 	assessing 	the 	proposed 	schemes 	as 	well 	as 	the 	r:::!.sons 	for 	HCFA's
dissatisfaction with the current !\-1AC program, we begin t:. 	noting that any
reimbursement 	program 	should 	contribute 	to 	a 	num:er 	of 	sometimes
conflicting goals.
A. 	Economic efficiency.
First, any scheme of prescription drug reimbursement should contribute
to economic efficiency. 	Thus, the scheme should not hinder the development
of 	market 	mechanisms 	that 	assist 	consumers 	(i.e., 	reduce 	their 	costs) 	in
making 	product choices; that facilitate 	the least-costly dis:ribution of goods
to consumers; and 	that 	promote 	price competition or, 10 2. dynamic market
environment, 	promote 	the 	production 	of 	new 	goods 	and 	product 	attributes
desired 	by 	consumers. 	A 	scheme 	that 	unnecessarily 	irr.pedes 	the 	use 	or
development 	of 	such 	mechanisms 	or 	otherwise 	unne;:essarily 	distorts
marketplace 	outcomes 	may 	impose 	costs 	on 	Medicaid 	and 	non-Medicaid
consumers 	that 	must 	be 	weighed 	against 	any 	resulting 	administrative 	or
reimbursement cost savings to the government.
For 	example, 	in 	many 	economic 	markets, 	firms 	haye 	an 	incentive 	to
develop a brand name reputation. 	While consumers may NY a higher price
for 	brand name goods, brand names often 	have value to consumers because
they may indicate that the expected quality of a product is higher than that
for 	non-branded 	products, 	and 	the 	existence 	of 	brand 	D:lmeS 	reduces 	the
costs 	of 	consumer 	search 	for 	products 	that 	best 	satisfy 	consumer demands.
Brand 	names 	provide 	an 	incentive 	for 	firms 	to 	develop 	;:,roduct 	attributes
that 	consumers 	value 	and 	to 	compete 	along 	these 	and 	other 	non-price
dimensions, 	for 	example, 	by 	communicating 	these 	non-price 	differences 	(as
7

well 	as 	price 	differences) 	to 	consumers 	via 	advertising. 	The 	information
contained 	in 	advertising 	is 	valuable to consumers not only directly 	but also
indirectly because it may facilitate entry by new suppliers and thus advances
competition.
Thus, the selected reimbursement scheme should not unduly hinder new
product introduction, brand name differences and other non-price diIi:ensions
of competition, just as it should not hinder price competition.
B. 	Incentive to dispense the lowest-cost generic product.
One significant 	reason 	why HCFA is considering repl:1cing the current
reimbursement scheme 	is to reduce the government's reimbursement costs by
encouraging Medicaid consumers to seek out and pharmacists to dispense the
lowest cost generic drug 	products. 	This goal will 	be easier to attain as the
number of multisource drug categories that fall within this incentive scheme
increases. 	If multisource drug products were reimbursed at a fixed level, the
structure 	of 	the 	current 	MAC 	program 	would 	yield 	incentives 	for 	low-cost
generic 	drug 	dispensing 	for 	drugs 	on 	the 	MAC 	list: 	in 	dispensing 	a 	MAC
drug, 	the 	pharmacist 	could 	retain 	the 	difference 	between 	his 	actual
acquisition plus dispensing costs and the MAC plus the estimated dispensing
cost. 	However, 	it 	is 	our 	understanding 	from 	discussions 	with 	HCFA staff
tha t, at least for some MAC drug products, the EAC is less than the MAC.9
Consequently, the incentive to dispense the lowest cost generic drug product
9 	This 	can 	occur 	because 	each 	MAC 	is 	determined 	by 	reference 	to
the range of wholesale prices fer all drug products within a particular drug
category. 	By 	contrast, 	the 	EAC 	is 	established 	for 	each 	specific 	drug
product.
8

may be reduced under the current MAC scheme. lO 	Even if :his were not the
case, the current administrative machinery has proven cost]:. and cumbersome
for 	determining which 	multisource 	drug categories should 	:-::ne 	MAC 	limits
and 	what 	those 	limits 	should 	be. 	Consequently, 	only 	a 	relatively 	sm:l!l
number of multisource drug categories have HCFA - determined MAC limits.
The 	opportunity 	for 	government 	savings 	through 	enhancing 	the
incentive 	for 	low-cost 	generic 	use 	is 	large. 11 	Approximately 	two-thirds 	of
all 	prescriptions 	written 	in 	the 	U.S. 	are 	for 	multisource 	drug 	categories.
The difference between the price of the leading brand in a prescription drug
entity 	and 	the 	price 	of 	alternative 	brands 	in 	the 	same 	entity 	is 	typically
large. 12 	Savings 	depend 	not 	only 	on 	whether 	the 	pharmacist 	dispenses 	a
generic 	drug 	product 	but 	also 	on 	which 	generic 	is selected 	because of 	the
wide variation in the cost to the pharmacist of different generic versions of
10 	For 	example, 	consider 	two 	versions, 	A 	and 	B, 	in 	the 	same 	drug
category. 	The MAC for 	both is $5 and the (correctly estiQated) dispensing
fee for both is $3. 	The EAC for A is $5.50 while that for B is $4.50. 	The
actual wholesale price to the pharmacist is $4.00 for A and S3.75 for B. 	If
the 	pharmacist 	dispenses 	the 	lowest-cost 	version 	(B), 	the 	amount 	of
reimbursement 	is 	$7.50 	(EAC 	plus 	dispensing 	fee), 	leaving 	the 	pharmacist
with a profit of $.75 ($7.50 less dispensing costs less the wholesale cost). 	If
instead 	A 	is 	dispensed, 	the 	reimbursement 	would 	be 	58.00 	(MAC 	plus
dispensing fee), leaving the pharmacist with a S1.00 profit.
11 	One 	indication 	of 	the 	savings 	potential 	from 	restoring 	the
incentive 	to 	use 	generic 	drug 	products 	in 	the 	Medicaid 	program 	is 	an
estimate 	of 	savings 	due 	to 	states' 	allowing 	generic 	substitution. 	An 	FTC
staff 	report 	estimated 	the 	annual 	reduction 	in 	consumer 	expenditures
attributable 	to 	state 	policies 	that 	permit 	generic 	substitution 	to 	be 	$44
million to $80 million in 1980 and perhaps three times that in 1984. 	Masson
and Steiner, p. 183.
12 	A 	1980 	average 	price 	across 	37 	leading 	multi-source 	drugs,
weighted 	by 	number 	of 	prescriptions sold, 	was $8.22 for 	the leading 	brand
and $6.22 for 	the average of other brands, a difference of $2.00, or nearly
25 percent of the leading brand price. 	Masson and Steiner, pp. 5, 356-36.
9

the 	same drug. 13 	Thus government savings can 	be incre:E:::d 	if 	pharm:lcists
choose low-cost rather than high-cost generics.
C. 	Pharmacy particiQ3tion
A third goal 	is the selection of the most efficient SCt of pharmacies to
handle 	Medicaid 	drugs. 14 	The level of Medicaid 	reimbursement 	will 	affect
not 	only 	the 	number 	of 	pharmacies 	that 	opt 	to 	participate 	but 	also 	the
"quality" 	of 	such 	pharmacies. 	Other 	things 	being 	equal 	(for 	example, 	the
number and types of Medicaid prescriptions dispensed, thc incentive for the
pharmacist 	to 	dispense 	the 	lowest-cost 	generic 	product, 	and 	the
administrative 	costs), 	a 	lower 	reimbursement 	level 	will 	not 	only 	reduce
government costs but also reduce pharmacy participation and accessibility of
participating 	pharmacies 	to 	Medicaid 	consumers. 1S 	The 	Notice 	describes
HCF A's 	concern 	that 	in 	fact 	many 	pharmacists 	have 	had 	difficulty 	in
acquiring drugs at or below the MAC levels in the current scheme.
13 	In 	the 	1984 	Red 	Book, 	for 	example, 	the 	wholesale 	price 	per 	100
tablets/capsules 	of 	allopurinol, 	other 	than 	the 	leading 	brand, 	ranged 	from
$14.65 to $33.68; for amoxicillin from $8.95 to $27.25; for metronidazole from
$28.75 to $54.40; for amitriptyline from $2.10 to $10.48. 	The extent of price
dispersion 	among 	generic 	products 	is 	puzzling. 	One 	would 	expect 	that
pharmacists 	would 	be 	knowledgeable 	about 	the 	therapeutic 	equivalence 	of
generic 	products 	and 	therefore 	that 	competition 	among 	generic 	product
manufacturers 	for 	the 	patronage 	of 	pharmacists 	would 	greatly 	reduce 	the
degree 	of price dispersion. 	One 	possible explanation is 	that 	the dispersion
among 	transactions 	prices 	is 	far 	less 	than 	tha t 	for 	ad vertised 	prices.
Alternatively, 	these 	price 	differences 	may 	reflect 	real 	non-therapeutic
differences among generic products.
14 	Our 	understanding from 	discussions 	with 	HCFA 	officials is that
most U.S. pharmacies currently participate in the Medicaid program.
15 	To 	the 	extent 	that 	many 	pharmacies 	depend 	upon 	Medicaid
consumers 	for 	purchases 	of 	drugs 	and 	other 	goods, such 	pharmacies 	migh t
exit 	the 	industry 	if 	the 	reimbursement 	level is 	not sufficient to 	profitably
maintain their participation in the Medicaid program.
10

D. 	AdministrJtivc costs
A 	fourth 	goal 	of 	any 	rcimbursemcnt 	scheme 	IS 	to 	reducc 	the
administrative 	costs. 	including 	any 	costs 	required 	to 	Insurc 	that 	the
reimbursement 	limits 	accurately 	reflect 	changing 	market 	conditions. 	As
notcd 	above, the current administrative process for determining MAC limits
has proven unwicldy.
III. 	EFFICIENCY EVALUA nON
All 	three proposals encourage greater use of generic drug 	products by
expanding the set of multisource drugs for 	which the dispensing of generic
drug 	products 	would 	be encouraged 	from 	the 	present 	rebtively 	short MAC
list 	to 	a 	much 	larger 	FDA 	lisL 16 	The 	FDA 	therapeutic 	equivalence 	list
reflects 	the 	fact 	that 	for 	many 	multi source 	drugs 	there 	is 	now 	substantial
agreemen t that no serious therapeutic ineq ui valence problems exisL 11
Under 	the 	current 	administrative 	scheme, 	neither 	the 	pharmacist 	nor
the customer has the incentive to dispense or use the lowest cost 	version of
16 	Compared 	to 	PhIP 	or 	CIP, 	the 	universe 	for 	possible 	generic
substitutions at any given time might be somewhat smaller under the revised
MAC plan simply because of delays in adding new products or 	new generic
entities. 	The 	revised 	MAC 	system 	requires 	publication 	of 	notices 	of
proposed actions, with a comment period to follow, before a drug is added to
the 	MAC 	list. 	However, 	the 	revised 	MAC 	proposal 	would 	streamline 	the
current administrative process.
11 	However, 	two same-strength 	products 	in 	the 	same 	generic 	entity,
containing the same active chemical ingredients in identical proportions, may
not 	always 	have 	the 	same 	effects 	in 	a 	patient, 	because 	differences 	in
inactive ingredients used for binding or coloring may modify the effects of
the active ingredients or create their own unintended side effects. 	Further,
therapeutically 	equivalent 	products 	may 	differ 	with 	respect 	to 	flavoring,
color, 	shape, 	packaging, 	and 	shelf 	life. 	FDA, 	Approved 	Prescription 	Drug
Products with 	Therapeutic Equivalence Evalua tions (1985), pp. 1-2, 1-3. 	For
a 	more 	extensive 	discussion 	of 	these 	and 	other 	differences, 	see 	M.
Lieberman, The Essential Guide to Generics (Harper and Row, 1986), pp. 4-11.
11

a multisource drug in many multisource drug categorics. IS 	T:'is government-
induced 	indiffcrcnce 	in 	the 	currcnt 	schemc 	rcprding 	·.. hich 	non-\1A.C
multisource 	drug 	IS 	dispensed 	may 	havc 	unduly 	limil~': 	thc 	extent 	of
compctition 	among 	generic 	drug 	products 	and 	bctwec:1 	ge:-.cric 	and 	brand
name drug products, thereby raising the cost of the \1cdicai:: program to the
government 	and 	the 	costs 	for 	non-Medicaid 	consumers 	as 	·"'"ell. 	However,
because 	the 	three 	multisource 	reimbursement 	proposals 	fOC~5 exclusively on
encouraging the dispensing of In"'-cost generic products, the:; appear to veer
too far in discouraging the development of brand names anj non-therapeutic
differences among drug products. 	Between these two extrer::es, we propose a
straightforward middle ground.
A. 	Benefits from 	encouraging pharmacists to dispen5~ 10'wer-cost drug
products
Information 	about 	American 	consumers 	in 	general 	su~gests that 	many
feel ill-informed about generic drug products. 19 	This concluoion is supported
by a 	recent FTC study which points to consumers' lack of ;nformation as a
major cause of reluctance to accept a generic d.rug produc!.:) 	For example,
18 	For 	non-MAC 	multisource 	drugs, 	the 	pharmacist 	is 	currently
reimbursed 	at 	EAC 	plus 	the 	estimated 	dispensing 	fee. 	If 	these 	estimated
costs 	were 	accurate, 	the 	pharmacist 	would 	have 	no 	incentive to 	dispense 	a
lower cost product instead of a higher cost product to lv1edicaid customers.
Because Medicaid customers bear at most a small fixed dollar copayment for
the 	purchase 	of 	the 	drug, 	they 	also 	have 	no 	incentive 	to 	request 	a 	lower
cost product.
19 	According to the CBS Consumer Model, a 	national survey done in
1983, 	consumers 	rate 	themselves 	as 	being 	either 	"not 	very 	informed" 	(29
percent of the sample) or "somewhat informed" (25 percent of the sample) on
the 	effectiveness 	of 	generic 	prescription 	drugs; 	45 	percent 	said 	they 	were
"not 	at 	all 	informed". 	On 	brand 	versus 	generic 	prescri;:'tion 	costs, 	the
typical consumer was less than somewhat informed. 	The CBS survey used a
national probability sample of households. 	The CBS Consu:::.er Model, 	1984,
p. 14.
20 	Masson and Steiner, pp. 5-7.
12

the 	study 	shows 	that 	consumers 	are 	more 	likely 	to 	3.::~~t 	generic 	drug
products if they feel 	their doctor has certified 	their apr~:;:riateness.21 The
most likely explanation of this behavior is that consumC~5 	2nd 	pharmacists)
interpret the 	fact 	that physicians specify a brand as a St~:~g preference on
the 	physician's part for 	that 	particular brand, even 	when 	the physician has
not chosen 	to exercise the 	legal option to prohibit substit>.;:i0n explicitly, an
option available on every prescription. 	The physician rna:; :;ot, in [act, have
a 	strong 	preference, 	but 	the 	consumer's 	uncertainty 	deters 	acceptance of a
substitute brand. n
21 	Consumers 	accept 	generic 	substitutions 	at 	the 	pharmacy 	only 	a
small 	fraction 	of 	the 	time 	when 	the 	physician 	has 	name: 	a 	brand 	(on 	7.3
percent 	of 	prescriptions 	on 	which 	substitution 	was 	permitted 	in 	1980), yet
they 	accept 	generics 	nearly 	all 	the 	time 	(89 	percent 	ir: 	1980) 	when 	the
physician 	prescribes generically. 	Masson and Steiner, pp. :-.116-117. In the
U.S. market as a whole, use 	of generic drug products is i:.:reasing rapidly.
The market share of generic drug products in 1980 was at-out 25 percent of
45 	leading 	multisource 	drugs 	dispensed. 	Masson 	and 	Steiner, 	p. 	117.
Officials at the Pharmaceutical Manufacturers Association report that a drug-
by-drug 	analysis 	of 	the 	market 	share 	of 	generic 	drug 	products 	shows
substantial 	increases 	between 	1980 	and 	1986, 	citing 	:ncreases 	of 	20
percentage 	points in 	some 	instances. 	These 	increases are 	:onsistent 	with 	a
heightened consumer understanding of the therapeutic equivalence of lower-
priced generic products and the higher-priced brand 	names. 	The rise in the
market 	share 	of 	generic 	drug 	products 	may 	alsc 	reflect 	~::cent changes 	in
state 	laws 	that 	now 	permit 	the 	pharmacist 	to 	substit:::e 	generic 	drug
products 	on 	a 	prescription 	written 	for 	a 	brand 	name 	(p~ovided that 	the
prescribing physician does not explicitly prohibit such substitution). 	Masson
and 	Steiner, 	p. 	l. 	As 	a 	result 	of 	these 	changes 	in 	state 	laws, 	the
percentage of brand-written 	multisource prescriptions on wilich substitutions
were made nearly doubled from 5.1 	percent to 9.5 percent cetween 	~ 980 and
1984. 	Stephen 	C. 	Chappell, 	"1 st 	6 	Months 	of 	'84: 	Independents 	& 	Drug
Chains 	Dispensed 	775.7 	Million 	Prescriptions." 	Pharm3~'; Times, 	October
1984, pp. 25-31.
22 	HCFA 	notes 	that 	it 	intends 	to 	retain 	the 	requ:rement 	that 	any
physician veto or override of substitution must be certified :0 the physician's
own 	handwriting. 	We 	endorse 	retention 	of 	this 	requir:ment. 	Previous
research 	has 	consistently 	demonstrated 	that 	the 	format 	0:- 	the 	physician's
prescription 	pad 	has 	a 	substantial 	impact 	on 	the 	incide:.ce 	of 	physician
overrides. 	For example, one study found 	that substitution was 18 percentage
points 	higher 	for 	formats 	requiring 	more 	physician 	ef:'ort 	to 	override
substitution than for other formats. 	See Masson and Steine~. ;Jp. 89-97; 100-
13

HCFA 	is of course fully aware of the therapeutic e'::;'~~valence between
lower-cost 	glneric 	products 	and 	higher-priced 	brand 	;::,mes. 	'Yet, 	by
reimbursing 	non-MAC 	multisouree 	drugs 	dispensed 	at 	L-\C. 	HCFA's scheme
provides no incentive for 	ph:Hmacists to dispense lower-cc-s: drug products. 23
This may 	have adverse consequences for 	consumers (inclu'::ng non-Medicaid
consumers) 	by 	reducing 	the 	availability 	of 	generic 	proc·~:ts, by 	reducing
downward pressure on prices, and by slowing the diffusior::C information to
consumers about therapeutic equivalence.
By 	expanding 	the 	number 	of 	multisource 	drug 	categories 	for 	which
there 	are 	incentives 	for 	pharmacists 	to 	dispense 	lower-cc~t drug 	prod ucts,
the three proposals ameliorate these inefficiencies. 	The res·.;lting increase in
competition may cause generic drug prices to decline.
In 	each of the proposed alternatives, these benefits cnld be magnified
if 	the 	pharmacist's 	incentive 	to 	dispense 	lower 	cost 	generi: 	products 	were
complemented 	by 	incentives 	for 	the 	Medicaid 	consumer 	to 	request 	these
products and to shop among pharmacies for the lowest-price: version in each
drug 	category. 	We 	would 	therefore 	suggest 	that 	HCFA 	consider 	a
requirement 	that 	Medicaid 	consumers 	pay 	a 	percentage 	:opayment. 	We
recognize 	that 	HCFA 	must 	weigh 	the 	efficiency 	benefits 	of 	a 	copayment
101; and 	106-107. 	See also, C. DeVito, W. Dickson, and J. Gbel, "Evaluating
Kentucky's Generic Substitution Law" in Generic Drug La":s: 	A Decade of
Trial--A 	Prescription 	for 	Pr02ress 	(U.S. 	Department 	of 	He::lth 	and 	Human
Services, June 1986), pp. 401-402.
23 	Indeed, 	the 	pharmacist may 	in 	fact 	be inclined :0 dispense 	brand
names rather than generic drug products under EAC. 	For example, consider
a Medicaid 	prescription 	written 	generically or for a 	brand name but which
permits 	the 	pharmacist 	to 	substitute 	a 	generic 	drug 	p~oduct. 	If 	the
pharmacist were to dispense the generic product, the pharffi:1cist might 	have
to 	bear 	the 	cost 	of 	explaining 	to 	the 	Medicaid 	patient 	that 	the 	generic
product dispensed is therapeutically equivalent to the brand C1me product.
14

against 	any 	income 	distributional 	considerations. 	par!i,:ujJ~ly the 	possibility
that for very expensive drugs, a copayment may be oner:::,;.:s for the \1edic3id
consumer. 24
We 	note, 	ho\.... ever. 	that 	the 	three 	proposals 	He 	not 	identiol 	with
respect 	to 	the 	pharmacist's 	incentives 	to 	dispense 	the 	lowest-cost 	products
in 	any 	particular drug 	category. 	PhIP clearly creates 	a;; 	incentive 	for 	the
lowest cost generic drug product to be dispensed because the pharmacist can
retain 	the 	difference 	between 	the PhIP 	limit 	plus 	the 	estimated 	dispensing
fee 	and 	the 	pharmacist's 	true 	wholesale 	and 	dispensing costs. 	Because 	the
EAC of many products within each multisource drug category may be below
the MAC, 	the 	revised 	MAC proposal 	may not create 	as 	i!luch 	incentive for
the pharmacist to dispense the lowest-cost generic product.
The proposed 25 percent brand-name discount in 	the CIP proposal will
certainly 	encourage 	the 	pharmacist 	to 	dispense 	a 	lower-priced 	generic
product 	instead 	of 	a 	brand 	name 	product. 25 	However. 	we 	can 	detect 	no
24 	One 	way 	of 	ameliorating 	this 	particular 	con:ern 	would 	be 	a
graduated 	copayment 	schedule. 	For 	example, 	on 	any 	prescription 	the
Medicaid consumer might provide a copayment of 20 percer:.t on the first $20
and 	5 percent on anything over $20. 	We recognize 	that implementation of
this recommendation would require statutory changes.
25 	We 	note 	that 	under 	CIP, 	prices 	for 	muItisource 	products 	other
than the leading brand would be discounted by 5 or 10 percent. 	There seem
to be three reasons for this particular discount. 	First, data shov.· that EAC
levels 	are 	approximately 	10 	percent 	below 	retail 	prices 	to 	private 	pay
customers. 	(New results from HHS contractor's analysis, as reported by Walt
Francis 	of 	HCFA.) 	This 	suggests 	that, 	in 	the 	absence 	of 	a 	Medicaid
discount, reimbursement levels 	would 	be higher than Current reimbursement
levels. 	Second, 	pharmacies 	often 	provide 	"senior 	citizens· discounts 	of 	10
percent. 	Third, 	through 	Medicaid 	the 	government 	is 	a 	volume 	buyer.
Volume 	discounts 	appropriat:::ly 	reflect 	real 	savings 	when 	it 	is 	cheaper 	to
deal 	with 	one 	large 	buyer 	than 	with 	several 	smaller 	ones. 	However,
Medicaid 	transactions 	occur 	individually. 	It 	is 	difficult 	to argue 	that 	the
real 	costs 	to 	the 	pharmacy 	are 	lower 	because 	the 	prescriptions 	are 	for
Medicaid customers. 	On 	the contrary, 	pharmacies' costs 	m:?y 	be 	higher for
Medicaid prescriptions than for cash-paying customers bec3use of paperwork
15

mechanism 	in 	CIP 	that 	would 	encourage 	the 	pharma~i:;: to 	dispense 	the
lowest-priced gener ic. Z6
B. 	Costs of excessivelv discouraging use of br:lnd-n3:":':~ drug products
While 	the 	current 	multisource 	reimbursement 	s~~eme 	m3.Y 	h3.\'e
unnecessarily hindered competition between brand name 3.nC generic products,
at 	least 	two 	of 	the 	three 	alternatives (PhIP 	and 	CIP) ap;;ear 	to 	discourage
the 	potentially 	efficient 	development 	of 	brand-name 	rep'Jtations 	and 	non-
price differences among therapeutically-equivalent drug products. 27 A policy
actively discouraging brand-name dispensing is predicated on the assumption
that informed consumers will be indifferent between two drug products that
and delays in reimbursement. 	Therefore, if volume purch2.ocs under Medicaid
were 	the proferred 	basis for 	the discount, we would recommend elimination
of this discount.
26 	In 	a 	recent 	supplementary 	Notice, 	HCFA 	expressed 	concern 	about
an 	apparent 	anomaly 	in 	the 	CIP 	proposal. 	HCFA 	noted 	that 	CIP 	might
generate 	a 	reimbursement 	level 	for 	the 	highest-priced 	generic 	product
dispensed that is greater than the brand-name reimbursem~nt level (after the
25 percent discount is applied to the brand-name product'j. 	Federal Register
(September 	18, 	1986), pp. 33086-33087. 	HCFA considers 	two alternatives 	to
reduce 	the 	incidence 	of 	this 	anomaly. 	The 	first 	is 	to 	supplement 	the 	25
percent 	brand-name 	discount 	with 	a 	limit 	on 	the 	reimtursement 	for 	any
generic 	product, 	the 	limit 	being 	no 	more 	than 	75 	percent 	of 	the 	median
leading brand price. 	While this proposal, if adopted, may have the advantage
of 	both 	discouraging 	the 	unilateral 	dispensing 	of 	higher-priced 	generic
products by the pharmacist and of limiting the extent of the anomaly, CIP as
modified would still not provide incentives for dispensing the lowest-priced
generic 	product. 	HCFA's 	second 	alternative 	would 	be 	to 	raise 	the
reimbursement 	level 	for 	brand 	name 	products. 	Raising 	the 	reimbursement
level 	for 	brand-name products would 	not reduce 	the pharmacist's incentives
to 	dispense 	brand-name 	products. 	Other 	things 	equal, 	the 	first 	alternative
seems preferable.
27 	Because 	it 	sets 	a 	fixed 	reimbursement 	lever, 	PhlP encourages 	the
dispensing of lowest cost generic products. 	Under CIP, the pharmacist would
receive 	only 	75 	percent 	of 	the 	retail 	price 	charged 	if 	a 	brand 	name 	were
dispensed. 	The incenti\'Cs in the revised MAC are 	less cle:.H if the EAC of
many drug 	products falls 	below 	the MAC level. 	The marginal 	profit from
dispensing 	a 	brand-name 	product 	could 	be 	greater 	or 	less 	than 	that 	from
dispensing a generic product.
16

arc therapeutically equivalent. 	This m3Y not in fact be the ':3se. 	To use an
extreme 	example, 	a 	Yugo 	and 	a 	Rolls 	Royce 	may 	be 	"tr3nsport3tionally
equivalent" but they are not homogeneous in the view of consu;ners.
Therapeutically-irrelevant 	product 	differences 	mig~~ 	include. 	for
example, 	the 	incidence 	of 	side 	effects 	that 	are 	annoying 	to 	consumers
without 	posing 	health 	hazards, 	the 	size 	and 	shape 	of 	the 	13 bIer. 	and 	taste.
A brand name or the development of a reputation for qualit:: may signify to
the 	consumer 	that 	the 	probability 	of 	experiencing 	annoying 	side 	effects 	is
lessened.
By 	focusing 	on 	price 	as 	the 	only 	important 	differer.:e 	among 	drug
products, HCFA may discourage the development of therapeu:i:ally equivalent
drugs 	that 	differ 	in 	important 	dimensions 	to 	consumers. 	Similarly, HCFA
may discourage advertising to doctors and 	pharmacists and the development
of 	brand 	names 	in 	prescription 	drugs 	even 	if 	these 	brand 	names 	convey
useful information to consumers. 	Further, if the current virtual FDA ban on
direct consumer advertising of individual prescription drugs ""ere relaxed, the
incentive of drug manufacturers to provide information via su.:h advertising
may be diminished under any of the HCFA schemes. 28
We 	are 	not 	aware 	of 	any 	evidence 	that 	would 	indicate 	the 	extent 	to
which 	consumer 	perceptions of 	non-price differences 	among 	therapeutically
equivalent 	drugs 	are 	important. 	Experience 	in 	other 	markets, 	however,
clearly 	suggests 	the 	importance 	of 	non-price 	differences 	to 	consumers. 	For
28 	Evaluating 	the 	role 	of 	brand 	names and 	thus 	the source of price
differences between 	brand name and generic drug products i:l a multisource
drug 	environment 	is 	difficult 	in 	part 	bc:ause 	of 	the 	virtu::i1 	FDA 	ban 	on
direct consumer advertising of prescription drugs and 	therefore 	the 	lack of
this particular source of consumer information on the products of individual
drug 	manufacturers. 	See 	A. 	Masson 	and 	P. 	Rubin, 	"Plugs 	for 	Drugs,"
Regula tion (September jOctober 1986), pp. 37-43 ff.
17

eXJmpJc, following airline deregulation, a hitherto unknow;; variety of price-
sen'lce options available to consumcrs bcgJn 	to flourish. 	\\::h the 3dvcnt of
the 	Jdvertising 	of 	legal 	services. 	J 	numbcr 	of 	"no-frilis' 	3nd 	low 	priced
services were offered to consumers. 	Brand 	nJmcs ha\'c pI:;: cd an importJnt
role 	in 	efficiently 	guiding 	consumer 	choice 	In 	produc:s 	rJnging 	from
appliances to fJst-food chJins.
We 	recommend 	that 	in 	the 	reimbursement 	scheme 	ac::;pted 	by 	HCFA.
Medicaid 	consumers 	have 	the 	option 	of 	purchasing 	any 	:!1ultisource 	drug
provided they pay the pharmacist the difference betwcen the market price of
the purchased drug and HCFA's reimbursement leve1. 29
C. 	Conclusion.
In sum, the current HCFA reimbursement scheme for multisource drugs
may artificially inhibit competition between generic drug products and brar.d
name 	products 	as 	well 	as 	among 	generic 	drug 	products. 	The 	multisource
coverage of the current MAC scheme is limited, and neither pharmacists nor
Medicaid 	consumers 	have 	an 	incentive 	to 	select 	the 	]C',"'er-cost 	generic
29 	For 	this 	reason, 	we 	recommend 	that 	the 	25 	pe~cent 	brand-name
discount 	in 	the 	CIr 	plan 	be 	replaced 	with 	a 	less 	costly 	alternative. 	For
Medicaid consumers that wish to purchase a higher-cost version of a drug in
the 	CIr 	plan, 	we 	suggest 	that 	such 	consumers 	pay 	the 	diiference 	between
the 	price 	of 	the 	lowest 	cost 	product 	dispensed 	by 	the 	ph:irmacist 	and 	the
higher cost 	product. 	The pharmacist, in 	turn, 	would 	be 	reimbursed 	by the
government for the price of the lowest-cost generic dispensed. 	Should HCF A
choose 	to 	retain 	the 	25 	percent 	discount, 	then 	we 	would 	suggest 	that 	the
Medicaid consumer reqLesting a 	brand 	name 	product pay the 	pharmacist 25
percent 	of 	the 	brand 	name 	price. 	In 	either 	case, 	the 	pharmacist 	will 	be
di~couraged from choosing to dispense a 	brand-name product to a 	Medicaid
consumer unless the consumer values the higher priced drug sufficiently.
Because the CIr reimbursement rules as originally proposed distinguish
between leading brand and generic products, we note that !he leading brand
in each drug category must be unambiguous. 	Yet HCFA offers no definition
of 	a 	leading 	brand. 	We 	suggest 	that 	HCFA 	consider 	defining 	the 	leading
brand 	product 	as 	the 	pioneer 	product. 	HCFA 	might 	also 	require 	that 	a
leading brand be defined as any product named by physici3::s on at ';;ast, for
example, 30 percent of all prescriptions in the particular drug ca tego.)'.
18

products 	in 	many 	cases. 	Consequently, 	fewer 	generic 	alternatives 	may 	be
available. the 	prices for available generic drug products ;:1:1Y be higher, and
the 	price 	of 	the 	brand 	name 	drug 	product 	may 	be 	higher 	than 	would 	be
efficient. 	Because 	of 	the 	greater 	rate 	of 	generic 	prvduct 	introduction
expected 	as 	a 	result of 	the streamlined 	FDA 	approval 	pro.::ess 	mandated 	in
the Hatch-Waxman Act, the costs of limits on the dispensing of generic drug
products 	in 	the 	future 	would 	be 	enormous. 	Because 	all 	three 	proposed
alternatives will expand the number of multisource drug c::negories included
under 	a 	reimbursement 	scheme 	designed 	to 	promote 	greater 	use 	of 	generic
drug products, all are improvements over the current scheme. 	However, only
PhIP appears to create unambiguous incentives for the pharmacist to dispense
the lowest-cost generic product.
However, 	we 	note 	that 	the 	assumption 	that 	therapeutic 	equivalence
implies 	product 	homogeneity in 	the view of consumers 	rna y 	not be correct.
We 	would 	therefore 	urge 	HCFA 	to 	modify 	whatever 	scheme 	is 	adopted 	In
order to avoid unnecessarily discouraging the use of brand name products by
consumers willing to pay the higher price and thereby avoid discouraging the
development of non-therapeutic differences across drugs, investment in brand
names, and non-price competition.
Within 	the 	framework 	of 	the 	efficiency 	goal, 	we 	would 	also 	suggest
that 	HCFA 	compare its 	current and 	proposed 	schemes of 	prescription drug
reimbursement 	and 	distribution 	to 	those 	developed 	by 	private 	insurance
companies. 	Those firms have profit incentives to pick the efficient structure
of distribution and reimbursement levels, ceilings on total reimbursement, and
19

incentives 	to 	encour3ge 	both 	ph3rmacists 	and 	consumers 	to 	select 	the
lowest-cost drug products.3o
IV. 	PHARMACY PARTICIPATION
Each 	of 	the 	proposed 	multisource 	reimbursement 	proposals 	will 	likely
have 	different 	effects 	on 	pharmacy 	participation 	rates 	unless 	the 	average
reimbursement 	level 	for 	~vledicaid prescriptions 	dispensed 	were 	identical 	in
each. 31 	As 	previously 	observed, 	HCFA 	has 	noted 	the 	complaints 	of 	some
pharmacists that they are unable to acquire MAC drugs at or below the MAC
limit. 	In 	a 	recently-released 	study conducted for 	HCFA, _the 	revised 	MAC
proposal 	is 	estimated 	to 	result 	in 	the 	highest 	average 	per-prescription
reimbursement level, followed by PhIP and CIP. 3o 	Yet, there is no evidence
that 	HCFA 	has 	considered 	the 	effect 	of 	the 	reimbursement 	level 	on 	the
pharmacy participation rate in setting the rates at which pharmacies will be
paid under each of the three proposals.
30 	However, subsidies 	to 	employer-financed 	schemes 	provided 	by 	the
tax system may encourage excessive insurance coverage.
31 	This 	assumes 	that 	each 	proposal 	provides 	an 	equal 	incentive 	for
pharmacists to dispense lower-cost generic products.
30 	Professional 	Management 	Associates, 	Interim 	Analvsis 	of 	the
Medicaid 	Reimbursement 	Policies 	(September 	4, 1986), Tables 	2, 	3, 	and 	12.
We 	note 	that 	among 	other 	difficulties, 	the 	definition 	of 	multisource 	drug
categories 	in 	the 	report 	is 	different 	from 	that 	described 	in 	the 	Notice.
Further, while the reported reimbursement levels for revised MAC and PhIP
may 	be 	reliable 	indicators 	for 	analysis of pharmacy 	participation, 	that 	for
CIP 	may 	not. 	Unlike 	PhIP 	or 	MAC 	that 	establish a 	single 	reimbursement
level 	for 	all 	pharmacies, the reported reimbursement levels for CIP reflects
pharmacy-spc::;ific prices as well as the effect of the 125 percent screen that
determines 	the 	maximum 	reimbursement 	level. 	Thus, 	the 	average 	per-
prescription reimbursement level for CIP includes drug products whose retail
prices 	are 	less 	than 	the 	screen 	as 	well 	as 	those 	whose 	prices 	are 	at 	or
above 	the 	screen. 	For 	assessing 	pharmacy 	participation, 	it 	is 	only 	this
maximum CIP level (the screen level) tha: is relevant.
20

Because, in 	principle, any desired 	reimbursement le\:::; can 	be attained
under 	any 	of 	the 	three 	alternatives,31 	the 	reimburseme:;, 	level. 	and 	the
participation 	rate 	implied 	by 	that 	level, should 	not 	be 	used 	as 	a 	basis 	for
choosing 	one 	proposal 	rather 	than 	another. 	Rather. 	the 	choi:e 	of
reimbursement 	mechanism 	should 	be 	made 	on 	other 	grounds, 	such 	as 	the
plan's effect on economic efficiency and 	its administrative costs. 	After the
plan 	is 	selected, 	attention 	should 	then 	be 	directed 	to 	selecting 	the
reimbursement level that will induce the most efficient number and types of
pharmacies to participate in the program. 32
In 	considering 	the 	efficient 	level 	of 	pharmacy 	parti~ipation, we 	note
that 	designing 	a 	system 	of 	Medicaid 	drug 	distribution 	that 	includes 	most
operating 	pharmacies 	may 	result 	in 	excessively 	high 	reimbursement 	costs.
31 	The 	choice 	of 	the 	MAC 	limit 	for 	each 	drug 	category 	(plus 	the
estimated 	dispensing 	fee) 	in 	the 	revised 	MAC program 	will 	determine 	that
scheme's 	reimbursement 	level. 	The 	percentage and 	dollar 	multiples 	of 	the
Red Book's or Blue Book's lowest cost generic product in each drug category
(plus the estimated dispensing fee) will determine PhIP's reimbursement level.
The 125 percent screen will similarly establish CIP's maximum reimbursement
level.
We also note that the Notice proposes that the PhIP ceiling (excluding
dispensing fee) 	be set at 	150 percent of 	the lowest published 	wholesale list
price, 	but 	that 	the 	ceiling 	reimbursement 	(excluding 	dispensing 	fee) 	be 	no
less 	than 	$1.50 	and 	no 	more 	than 	$4.00 	above 	that 	lowest 	price. 	This
approach 	seems 	more 	arbitrary 	than 	necessary. 	HCFA 	might 	consider 	a
somewhat 	less 	arbitrary 	mechanism 	that 	preserves 	the 	central 	principle 	of
the plan -- tying 	the reimbursement 	to a 	low-ranked 	wholesale 	price -- but
is simpler: choosing the third lowest (or the median) price as the PhIP level.
An advantage would be the elimination of need for minimum and maximum
dollar 	incentives 	and 	the 	percentage 	multiple. 	Their 	elimination 	would
remove two problems. 	First, since prescription prices will change along with
the 	level 	of 	prices 	in 	general, 	any 	minimum 	and 	maximum 	dol~ar figures
would 	have to be inflation-adjusted over time. 	Second, the specification of
any dollar minimum or maximum incentive or a percentage multiple might for
some drugs be either insufficient to assure availability of low cost products
or much more than necessary to assure their availability.
32 	Thus HCFA can adjust the details in each of the three schemes to
attain any desired level of savings in government reimbursement costs.
21

For 	the same drug dispensed 	in 	any giycn dosage 	form. 	the 	price 	nriation
among 	drug 	stores 	In 	any 	given 	geographic 	area 	is 	usuJ.lly 	substantia1. 33
Further. for each pharmacy participating In the !",ledicaid program. there may
be 	a 	fixed 	administrative 	cost 	borne 	by 	HCFA. 	Thus. 	the 	higher 	HCFA's
reimbursement 	level 	for 	prescription 	drugs 	dispensed 	to 	\1edicaid 	patients.
the 	greater 	will 	be 	the 	number 	of 	p:lrticipating 	pharmacies 	of 	any 	given
quality 	(cost) 	and 	the 	higher 	will 	be 	the 	government's 	reimbursement 	and
ad mi n istra ti ve costs.
In choosing the number and quality of participating pharmacies, HCFA
must 	also 	consider 	the 	accessibility 	of 	these 	pharmacies 	to 	Medicaid
customers. 	The 	lower 	the 	number 	of 	participating 	pharmacies, 	the 	less
accessible 	participating 	pharmacies 	will 	be 	to 	Medicaid 	customers. 	The
optimum 	distribution 	system 	is 	one 	that 	just 	balances 	the 	additional
government 	reimbursement 	costs 	of 	increasing 	the 	number 	and 	type 	of
participating pharmacies with the additional benefits of increasing Medicaid
customers' access to participating pharmacies.
We 	do 	not 	have 	sufficient 	information 	to 	conduct 	this 	balancing.
Without 	knowing 	how 	various 	classes 	of 	pharmacies 	will 	decrease 	their
participation 	in 	response 	to 	a 	lower 	reimbursement 	level 	and 	how 	those
decreases 	affect 	accessibility 	to 	Medicaid 	consumers, 	we 	cannot 	provide
HCFA with even a cursory discussion of each proposal's eff~cts on pharmacy
participation.
33
Formula
(Austin),
See 	D. 	Kreling, 	Developing 	A 	Prescription 	DruQ 	R::imbursemcnt
For 	Texas, 	unpublished 	Ph.D. 	Dissertation 	(University 	of 	Texas
1984).
22

V. 	ADMI?,:ISTRATIVE COSTS
All 	thrce 	propos::J.Is 	cnt3iJ 	subst3nti31 	costs 	of 	3c::1inistr3tion. 	:\1]
require 	pharm3cists to 	file 	c13ims and 	states 	to 	review 	a:'.j 	p3y 	thcm. 	All
require an31ysis of d3t3 to set 3ccept3ble reimbursement Ic\ ~ls.
As 	compared 	to 	the 	current MAC 	program, 	the 	re .. ioed 	MAC 	propos31
envisions 	a 	streamlined 	"notice 	and 	comment" 	procedure 	by 	which 	HCFA
would 	determine 	MAC 	limits 	for 	multisource 	drugs 	covered 	by 	the
reimbursement scheme. 	Nonetheless, HCFA staff would sti;1 have to monitor
FDA 	therapeutic 	evaluations, 	prepare 	notices 	for 	the 	Fec~ral Register. 	and
evaluate 	comments. 	Private 	firms 	would 	incur 	costs 	in 	responding 	to 	the
request 	for 	comments for 	setting 	the 	initial 	MAC limit and 	for 	cha nges 	in
those 	limi ts. 	HCFA's 	availability 	survey 	and 	the 	states' 	dispensing 	fee
surveys would 	be retained. 	HCFA 	would continue to pur:hase and analyze
invoice 	cost 	data 	in 	order 	to 	determine 	the 	MAC 	reimbusement 	level 	for
each drug category. 	Further, HCFA proposes to set regional MAC limits to
insure 	availability 	of 	multisource 	drugs 	to 	pharrn:?:ists, 	an 	added
administrative burden. 	Finally, the states would still be required 	to set the
EAC for each 	product and the dispensing fee. 	This rrocess promises to be
expensive, 	but 	we 	have 	no 	data 	on 	what 	the 	actual 	administrative 	costs
might be or how they might compare with current costs.
As in 	the 	current and 	revised 	MAC, 	the 	PhIP propcsal 	would 	require
periodic surveys 	by 	the states of pharmacists' dispensing costs,34 as well 	as
34 	Actually, the 	Notice seems to suggest 	that 	these surveys would 	be
continued 	even 	under 	CIP. 	We 	can 	see 	no 	usefulness 	of 	dispensing 	cost
surveys if CIP is adopted.
23

the 	EAC 	for 	each 	product. 35 	PhIP 	would 	also 	require 	::.:: 	analysis 	of 	list
price 	data 	for 	multisource 	drugs. 	Because 	the 	data 	ar~ :::.w 	computerized
and 	updated 	monthly 	by 	the 	publishers. and 	the 	prograr:::::ing required 	for
determining 	any 	reimbursement 	level 	tied 	to 	a 	low-rank::: 	price as 	well 	as
changes 	in 	that 	level 	is 	simple, 	the 	costs 	of 	this 	pr::edure 	would 	be
moderate. 	The administrative costs of PhIP may be subsr::.::tiaIIy lower than
those of the revised and current MAC.
For a 	number 	of reasons, 	the administrative costs c:- the CIP proposal
are 	more 	difficult 	to 	evaluate 	than 	those 	of 	the 	other 	~roposed schemes.
Unlike 	the 	current 	or 	revised 	MAC 	or 	PhIP, 	CIP 	doc:. 	not 	require 	an
estimate 	of 	the 	EAC 	or 	the 	dispensing 	fee. 	Nor 	is 	an 	::.vailability survey
required. 	However, 	in 	order 	for 	the 	125 	percent 	screen 	to 	reflect 	actual
retail prices, we presume that periodic surveys by states "'-ould be required,
an added administrative burden.
We also suggest that an additional monitoring and enf:rcement cost may
have to be incurred with CIP. 	Any scheme that bases rei=:-ursement on the
prices 	charged 	by 	individual 	pharmacists 	may 	create 	::.n 	incentive 	for
pharmacists to charge the government (i.e., Medicaid cusroc~rs) higher prices
than 	other customers. 	To 	the 	extent 	that 	the 	125 	percer:: screen 	typically
exceeds 	the 	price 	charged 	to 	retail 	customers, 	pharmaci::s 	will 	have 	an
incentive to charge Medicaid consumers the screen price w::'ile charging non-
Medicaid consumers a lower competitive price. 	First, pharcacists know who
their Medicaid customers are. 	Second, Medicaid customers do not care what
price 	is charged 	for 	the 	prescription since 	their copaymen:. when 	required,
35
not 	fall
product's
In the PhIP as in the current and revised MAC, ~~oducts which do
into 	a 	multisource 	drug 	category 	""ould 	be 	r:::mbursed 	at 	the
EAC plus estimated dispensing costs.
24

is a fixed 	dollar amount. 	Thus, phJrmacists can charge \tcdicaid customers
infbted 	prices and 	the \kdiC3,id customers will 	haH 	no ir::cnti\'e to seck a
new 	pha rmacist. 36 	Further, 	unless 	the 	screen 	\Vere 	defined 	In 	terms 	of
priv3te pay prices only, the screen itself might also rise.
The 	invitation 	to 	eng3ge 	In 	dual 	pricing 	would 	o.ifect 	all 	Medicaid
prescriptions. 	Were 	dual 	pricing 	to 	be 	unchecked, 	the 	added 	cost 	to 	the
Medicaid program could 	be substantial. 	Although HCFA would 	require that
the 	screen 	be 	determined 	by 	the 	prices paid 	by 	non-Medicaid 	customers in
each st3te, HCFA might find it cost-effective to audit the \fedicaid and non-
Medicaid prices of at least some individual pharmacists. 37
In 	sum, 	while 	PhIP 	clearly 	has 	lower 	administrati\-e 	costs 	than 	the
revised MAC, a lack of data 	does not permit us to comp:l:e either PhIP or
the revised MAC to CIP.
VI. 	CONCLUSION.
Y,'c 	would 	urge 	HCF A 	to 	adopt 	a 	scheme 	that 	is 	designed 	to 	be
economically 	efficient. 	In 	this 	regard, 	we 	endorse 	HCFA's 	intention 	to
expand 	the 	number 	of 	drug 	categories 	subject 	to 	rnu]tisource 	drug
reimbursement. 	This will reduce or eliminate any inefficient use of brand-
name 	drugs 	that 	results 	from 	the 	limited 	coverage 	of 	the 	current 	MAC
36 	A percentage copayment by Medicaid customers would restore some
of 	the consumer's incentive 	to search for 	lower prices and 	thus 	reduce 	the
likely extent of dual pricing.
31 	Some 	pharmacies ("Medicaid 	mills") 	may 	evade 	the intetlt of such
an 	audit 	by 	restricting 	their 	clientele 	to 	Medicaid 	customers. 	These
pharmacies 	would 	then 	have 	no 	prices 	charged 	non-\fedicaid 	customers
against 	which 	the 	prices 	charged 	lvledicaid 	consumers 	could 	be 	compared.
One 	possible 	solution 	is 	the 	proposed 	125 	percent 	screen. 	While 	such 	a
screen 	will 	reduce 	the 	profitability and 	possibly 	the 	incidence of Medicaid
mills, 	we 	are 	unable 	to 	evaluate 	how 	effective 	a 	deterrent 	such 	a 	screen
might 	be. 	We do 	note 	that a 	screen will 	also likely 	redu:e 	the 	number of
pharmacies participating in Medicaid.
25

program. 	However, because each of the proposed alternatj'.es would provide
strong disincentives to dispensing brand-names when a ge::eric drug 	product
is available, 	and 	because of some 	uncertainty 	regarding t~e significance of
current or potential non-therapeutic differences among drugs, we would urge
HCFA to permit IV1edicaid consumers to purchase the brane-name products if
they choose to pay the difference between 	the price of the 	brand name and
the maximum reimbursement level for a generic drug produ:t. 	We would also
suggest 	consideration 	of 	a 	percentage 	copayment 	by 	Medi:aid 	customers 	to
provide them with an incentive to seek the lowest-cost drug ~vailable.
We are unable to completely rank the proposals against the other goals
of 	a 	reimbursement 	scheme. 	We 	would 	suggest 	that 	to 	facilitate 	such 	a
ranking HCFA consider drawing a random sample of multisource and single-
source 	drugs, 	and 	compare 	the 	estimated 	likely 	reimbursement 	levels 	and
administrative costs, given 	the efficient selection of pharmacy participation.
But 	we 	would also 	suggest 	that 	HCFA 	examine a 	variety 	of 	market-tested
private insurance schemes to assess the means by which reimbursement levels
are 	determined. 	There 	may 	be 	information available 	in 	the 	private 	sector
regarding 	reimbursement 	program 	design 	that 	HCFA 	may 	find 	particularly
useful in developing its Medicaid scheme.
With 	respect 	to 	the 	efficient 	distribution 	of 	prescription 	drugs 	to
Medicaid 	consumers, 	we 	would 	similarly 	suggest 	that 	an 	examination 	of
market-tested private reimbursement schemes would be useful. 	One approach
that could be employed with any of the three proposed schemes would 	rely
on 	market 	responses 	to 	eliminate 	the 	arbitrary 	nature 	of 	the 	proposed
reimbursement levels and simultaneously result in the lowest-cost distribution
systems consistent with HCFA's mandate. 	HCFA could in effect auction off
26

the 	right 	for 	pharmacies 	to 	participate 	in 	the 	:-'1edicaid 	;:>rogram. 38 	First,
HeFA (or the states) would 	have to determine the efficie::t number and the
efficient 	quality 	of 	the 	Medicaid-drug 	distribution 	syste;n. 	Sv.:h 	qualit:-'
aspects 	might 	include 	the 	distances 	of 	the 	ph:l.rmacies 	from 	!\1edicaid
consumers, hours of operation, average waiting time. and the like. 	HCFA (or
the 	states) 	might 	then 	conduct 	regional 	auctions 	based 	upon 	the 	selected
reimbursement scheme (for 	example, PhIP).39 	In 	each 	region, 	HCFA 	would
then announce that Medic:lid drugs under PhIP would be reimbursed at the
lowest-cost generic drug product in 	the Red 	Book or 	Blue Book 	minus, say,
10 percent. If the number and mix of pharmacies agreeing ro participate falls
short 	of 	HCFA's 	goals, 	a 	second 	auction 	would 	be 	undertaken, 	by
announcing, 	for 	example, 	that 	Medicaid 	reimbursement 	would 	be 	at 	the
lowest-cost 	generic 	drug 	product 	minus, 	say, 	5 	percent. 	HCFA 	would
continue to revise the announced reimbursement limits until the number and
mix of participating pharmacies was efficient.
This type of auction scheme is attractive because it eiiminates the need
for arbitrarily establishing the upper limits of reimbursement. 	While we are
unsure 	of 	the 	magnitude 	of 	the 	administrative 	costs 	ent::liled, 	the 	propos::l1
might 	require 	some 	monitoring 	to 	in~ure that 	in 	fact 	the 	quality 	of 	the
services 	provided 	by 	the 	participating 	pharmacies 	is 	maintained. 	Further,
under 	this scheme 	there 	would 	be 	no incentive 	to 	develop 	either Medicaid
"mills" or dual pricing schemes.
38 	It 	is 	our 	understanding 	that 	the 	Department 	of 	Justice 	will 	be
proposing a 	variant of the auction scheme. 	While we are not aware of the
precise details of the proposal and therefore cannot endorse the proposal, we
believe this novel proposal merits consideration.
39 	We use 	PhIP 	for 	illustrative purposes only. 	The described 	::luction
could be used in conjunction with any of the proposals.
27