Lowering Generic Drug Prices Less Regulation Equals More Competition
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
MEDICAL CARE
Volume 41, Number 1, pp 135–141
©2003 Lippincott Williams & Wilkins, Inc.
Lowering Generic Drug Prices
Less Regulation Equals More Competition
ASLAM H. ANIS, PHD, DAPHNE P. GUH, MSC, AND JOHN WOOLCOTT, MA
BACKGROUND. In Ontario, Canada, the 70/90 as the number of generic firms increased
regulations were instituted in May 1993 to within these periods. However, this decrease
establish provincial government procurement in P was significantly less in the AP (median:
prices for generic drugs. Accordingly, the first 0.75 f 0.68 f 0.67) than in BP (0.71 f 0.61 f
generic entrant’s price could not exceed 70% of 0.53) as the number of generics increased from
the incumbent’s branded price. Subsequent 1 to 2 to 3, respectively. The regression analysis
entrants’ prices could not exceed 90% of the showed that the price ratio in the AP was
first entrant’s price. higher than that in the BP by 0.05, 0.09, and 0.13
OBJECTIVE. These regulations’ impact on ge- for first, second, and third generic entrant
neric market competitiveness are evaluated. respectively.
DESIGN and METHODS. Data on 518 drugs CONCLUSIONS. Our findings show that the
spanning nine therapeutic classifications were 70/90 regulations not only failed to achieve
collected for the period of 04/01/1987 to 12/31/ their goal of lowering the procurement price
1998 from Ontario Drug Benefit formulary and but instead the opposite occurred. The man-
IMS Canada. The period 04/01/1987 to 04/30/ dated procurement price became a focal point
1993 was defined as the before period (BP) and and resulted in a clustering of prices around
05/01/1993 to 12/31/1998 was the after period the maximum allowable levels with little price
(AP). We compared the price ratio (P ⴝ PG/PB) dispersion.
between BP and AP and performed regression Key words: Drug policy; drug price regula-
analysis to assess the determinants of P. tion; generic market competitiveness. (Med
RESULTS. P in both the BP and AP decreased Care 2003;41:135–141)
When available, the use of generic as opposed to natives. This may be because of the lower costs of
brand prescription drugs can significantly reduce production associated with generics who do not
costs for drug insurance plans. Additionally, price have to incur R&D costs and the increased compe-
regulation is often used to further reduce costs. tition which results after the expiry of the branded
Often drug price regulation is tied to drug substitu- products’ patent. This process is facilitated by regu-
tion laws to redirect dispensing toward cheaper lations that attempt to lower generic prices while
generics from more expensive branded drugs. His- increasing generic competition.
torically the prices of generic drugs are significantly Although well intentioned, price regulations are
lower than the prices of the relevant branded alter- often poorly conceived. Several studies have doc-
From the Department of Health Care and Epidemi- Address correspondence and reprint requests to:
ology, University of British Columbia, Centre for Health Aslam H. Anis, PhD, CHEOS, 620-1081 Burrard St.,
Evaluation & Outcome Sciences, St. Paul’s Hospital, Vancouver, B.C., Canada, V6Z 1Y6. E-mail:
Vancouver, British Columbia. anis@cheos.ubc.ca
An earlier version of this paper was presented at the
3rd World Conference of International Health Econom- Received October 12, 2001; initial review December
ics Association, July 22 to 25, 2001, York, UK. 14, 2001; accepted July 12, 2002.
135ANIS ET AL MEDICAL CARE
umented the association between price regula- may be charged by manufacturers. Under the
tions and a rise in the average price of generic regulations, a generic drug manufacturer wanting
drugs and a reduced dispersion of prices that their product to be eligible for reimbursement
normally existed between generics.1–3 under the ODB plan must first have their product
In a typical marketplace, free and unfettered listed as interchangeable under the Drug Inter-
exchange between suppliers and demanders of changeability and Dispensing Fee Act (DIDFA)
goods and services occurs on a regular basis. No and listed on the ODB Formulary. To be listed, the
rules or regulation are necessary for the smooth price of the generic drug cannot exceed a specified
functioning of such markets. The equilibrium price (maximum) level with respect to the branded
of the competitive market ensures that there is manufacturer’s price. The 70/90 regulations stipu-
neither “excess”demand nor supply in the market. lated that the first generic entrants’ price could not
With numerous buyers and sellers these markets exceed 70% of the incumbent branded product
are assumed to be competitive, and any attempts price. Subsequent entrants could not price their
to raise prices above the competitive equilibrium product at greater than 90% of the first entrants’
level would be unsustainable because the compe- price, which essentially limits the subsequent en-
tition would attract all the business. The market trants’ price to be at most 63% of the relevant
for prescription pharmaceuticals does not, how- branded product price.
ever, resemble the competitive market scenario. The 70/90 regulations were first introduced as
First, for the most part, there are few, rather than the 75/90 regulations in May 1993 and were
numerous sellers. Second, the demanders (physi- subsequently amended and became the 70/90
cians) are not consumers, and furthermore often regulations in November 1998 (see Appendix A.
the demanders and consumers are not the pur- Table 1). The only difference between the 75/90
chasers. This creates a market imperfection where regulations and the 70/90 regulations is that under
the ramifications of price changes are not directly the former, the first entrant had to offer a 25%
felt by the consumers and therefore have negligi- discount on the incumbents’ branded product,
ble effects on demand. Thus, a rationale for im- which under the latter was increased to 30%. The
plementing price regulation exists. objective of this study was to evaluate the impact
In the province of Ontario, Canada, for the past of these regulations, and given that the current
three decades, Drug Product Selection/Substitu- regulation is known as the 70/90 regulation, we
tion (DPS) laws, have been enforced regarding refer to these regulations as the 70/90 regulations
price, interchangeability, legal liability, etc. for all on generic market competitiveness throughout the
eligible prescription drugs.4 The Ontario Drug remainder of the text.
Benefit (ODB) publishes a formulary listing of all Parenthetically it should be noted that in 1991
eligible drug products and the price for each pharmacoeconomic guidelines were introduced in
dosage form that will be paid directly to pharma- Ontario.6 However, these guidelines applied spe-
cists when the drug is dispensed to anyone cov- cifically to new drugs seeking ODB formulary
ered by the provincial plan. The drug cost for inclusion. Branded products already on the ODB
multiple-source drugs was determined as the low- formulary and generic products (existing or new
est price submitted to the formulary by any eligible entrants) were exempt.
manufacturer and was designated as the Best
Available Price (BAP). For single-source drugs, the
BAP referred to the only price submitted to the Materials and Methods
formulary. The philosophy of the BAP-type system
was to force suppliers of multiple-source drugs to To assess the impact of the 70/90 regulations,
compete among each other by submitting low- we compared the price ratio between the relevant
price bids. By designating the lowest-price bid as generic and branded drug products launched be-
the BAP and additionally requiring pharmacists fore and after the regulations. A sample of drug
via substitution legislation to substitute all pre- products listed under the ODB Formulary between
scriptions for the cheaper BAP products, it was April 1, 1987 to December 31, 1998 was selected
hoped that a substantial cost savings could be from the following therapeutic classifications: car-
achieved.5 diovascular drugs, antiarthritics, psychotherapeu-
The 70/90 regulations specifically govern the tics, antiinfectives, analgesics, neurologic disor-
pricing of generic drugs and limit the prices that ders, cholesterol reducing/lipotropic drugs,
136Vol. 41, No. 1 DRUG PRICE REGULATIONS
proprietary analgesics, and diuretics. These thera- first generic product over the branded product, the
peutic classifications together accounted for more ratio of the second generic entrant over the
than 50.5% of the prescriptions written in Canada branded product and the ratio of the third generic
for 1999.7 The ODB Formulary, updated and pub- entrant over the branded product.
lished periodically, contained information regard- Occasionally, the market in the before and after
ing the interchangeable drug categories and pric- periods changed from a situation of one branded
ing data for all drug products. Additional data firm to a situation of one branded and two generic
regarding launch dates of individual generic prod- firms without the intermediate case of one
ucts was obtained from the Drugstore and Hospi- branded and one generic firm. Similarly, there also
tal Database constructed by IMS Canada. were instances where the market changed from
The study sample was primarily selected based the situation of one branded and one generic firm,
on the availability of the data. A pharmacist re- to one branded and three or more generic firms
viewed the database and some inconsistencies directly. When this occurred, the price ratios of first
with respect to launch dates were found. These generic product, second generic product and/or
inconsistencies related primarily to situations second generic product over branded product
where: (1) for the same drug product, multiple would not be available. Furthermore, because we
drug manufacturers were named because of li- did not have pricing data before April 1987, no
censing arrangements between parties; or (2) a price ratios could be calculated when more than
new drug information number was subsequently three generic products within the same inter-
issued for an existing drug product as the new changeable drug category were already present in
drug information number may have been issued the market before April 1987.
because of a change in the manufacturing site or a The precise expiry date of each patent during
change in the source of raw materials for the drug the study period was not available to us. Patent
product. In these circumstances, these observa- expiry was inferred from generic entry. The num-
tions were excluded from the study. Also excluded ber of first, second, or third generic entrants is
were generic products owned by branded firms governed entirely by patent expires, market size
known as “fighting generics” or “pseudo- and entry conditions. To observe that there were
generics.” They were excluded from the analysis more first entrants in either period is a function of
because they do not typically compete with the both the number of patent expires, and the profit
incumbent branded firm. On the contrary, these potential of generic entrants, which, among other
“pseudo-generics” allow the owner of the patent factors would be a function of market size in the
(branded firm) to segment the market in a manner relevant therapeutic category.
that maximizes their revenues from both the brand Univariate analysis using nonparametric
version and generic version. (distribution-free) Wilcoxon rank sum test was
As the regulations were first introduced in May performed to compare PG/PB in the before period
1993, the period between April 1, 1987 and April versus the after period.4 We chose the Wilcoxon
30, 1993 was defined as the before period while rank sum test, which tests the hypothesis that the
the period between May 1, 1993 and December 31, distributions of the two populations are the same
1998 was defined as the after period. Each generic because our data are skewed.
drug product was classified as being launched in Multiple regression analysis was additionally
the before period or in the after period. The order performed to identify the determinants of the ratio
at which each generic drug product launched of price. In the final regression model, the vari-
within each interchangeable drug category was ables included were the period at which the
also determined from its launched date. generic product was launched (before period or
For each generic product, the price ratio, which after period), the number of generic product exist-
was defined as the price of generic product over ing in the market the generic product (0, 1, 2), their
the price of branded product (PG/PB) in the same interaction term, and the therapeutic classification.
interchangeable drug category, was calculated. We The regression model was estimated using Gen-
used the first available pricing data for both the eralized Estimating Equations approach because
generic and branded products in the same edition when generic products were from the same inter-
of the Formulary after the launch date of the changeable drug category, the price of these
generic product to calculate PG/PB. Three types of branded products would contribute to more than
price ratio were examined in our study: the ratio of one price ratio and these ratios may not be con-
137ANIS ET AL MEDICAL CARE
sidered as independent observations.8 A simple TABLE 1. Summary of Price Ratio by Generic
exploratory analysis was also performed to check if Entrant Before and After Price Regulations
there is an increasing trend of price ratios over
Estimated Price Ratio
time for first, second, and third entrants (PG/PB)
respectively.
04/01/87- 05/01/93-
Comparison 04/30/93 12/31/98 P
st
Results Branded vs. 1 generic
N 52 19
A total of 125 branded drug products (ie, 125 Mean 0.70 0.73
interchangeable drug categories) and 510 generic SD 0.13 0.03
products were selected. Of those, 19 branded drug Median 0.71 0.75 0.42
products had only one generic product competing Q1–Q3 0.64–0.75 0.71–0.75
in the market during the study period, 23 with Range 0.32–1.00 0.67–0.75
Branded vs. 2nd generic
two, 37 with three, 32 with four, and 14 with more
N 56 22
than four competing generic products. Because of Mean 0.55 0.71
the availability of the pricing data as explained SD 0.21 0.13
earlier, the number of price ratios we were able to Median 0.61 0.68 0.01
obtain was 71, 78, and 69 for first, second, and Q1–Q3 0.47–0.71 0.67–0.75
third entrant, respectively. Out of a total of 125, we Range 0.05–0.90 0.46–1.11
only had 71 first entrant prices because 47 of the Branded vs. 3rd generic
remaining 54 cases resulted in a market where two N 33 36
generics entered simultaneously and for the re- Mean 0.50 0.65
maining seven cases, the branded drug was un- SD 0.21 0.08
Median 0.53 0.67 0.0001
listed and its price was no longer available. In total,
Q1–Q3 0.38–0.60 0.60–0.67
23 branded drugs were excluded for the reasons Range 0.04–0.95 0.46–0.90
listed in the previous section; 21 were launched in
the before period whereas two were launched in
the after period. Among the excluded branded
products, 11 (all in the before period) were before period by 0.05, 0.09, and 0.13 for first,
“pseudo-generics”. second, and third generic entrant, respectively.
Table 1 and Figure 1 present the summary Moreover, as the number of generic product in-
statistics and box plots of the price ratio between a creases from 1 to 2 and 3, the price ratio decreases
new generic product and the respective branded in the before period by 0.08 and 0.16 whereas it
product according to the before period and after only decreases in the after period by 0.04 and 0.08,
period. It shows that the price ratio in both the respectively. The results of the multivariate analy-
before period and the after period decreased as the sis resolve that a significant degree of change in
number of existing generic products increases. the price ratio (PG/PB) can be attributed to the
However, the data clearly show that this decrease period in which the generic drug was introduced.
in price ratio is substantially less in the after period These findings, like those of the univariate analy-
than in the before period. Additionally, the sub- sis, are consistent with the hypothesis that the
stantially smaller spread of the price ratio in the pricing regulations have the intended effect of
after period suggests that the generic products in raising generic prices.
the after period were priced at the level close to
what the regulation set. However, in the before
period, many generic products were priced at a Discussion
much lower level. Additionally, plots of price ratios
over time by generic entry did not reveal signifi- Our findings suggest that Ontario’s 70/90 Reg-
cant trends. ulations failed to achieve their goal of lowering the
The results from the regression analysis are cost of generic drugs for the ODB plan. The
presented in Table 2. The results suggested that generic marketplace appears to be an inherently
after the therapeutic class being adjusted, the price competitive marketplace in which there is a natu-
ratio in the after period was higher than that in the ral, market-imposed pressure on the first generic
138Vol. 41, No. 1 DRUG PRICE REGULATIONS
FIG. 1. Price ratio by order
of generic entry.
entrant to not set prices too high and as the such that all generics are treated as a homogenous
number of competitors increases with entry of good. That is generic firm specific prices do not
additional generic firms, prices are driven down- exist. All generics are paid the same reimburse-
ward. The data also suggests that in the after ment amount, the BAP. Hence, cutting prices does
period, the decrease in price ratio caused by in- not necessarily generate market to the firm cutting
creased number of generic firms is much lower, the price but that all generics in the specific market
which is supportive of the premise that price now receive a lower reimbursement.
controls have interfered with natural competitive More generally speaking, the lowest submitted
forces. The question that must then be addressed price was designated as the BAP, yet the submit-
by policy makers is, whether there is a role for ting bidder did not get any special rights as the
regulation in the market of post patent pharma- sole supplier at this bid. All generics were allowed
ceuticals? The answer which this paper along with to supply at this price. This led to a flawed process
others show that to date the stated goals of for generating competitive bids and the addition of
legislation are not being met and therefore differ- the 70/90 regulations did not alter this basic flaw,
ent approaches should be explored.1–3,8,9 ie, lowest bidder did not get the sole supplier
It is our contention that the 70/90 regulations contract.5
created a focal point for setting the initial price of As illustrated by Caves et al,11 in unregulated
any generic product. A potential generic entrant markets, generics were found to be priced approx-
has no incentive to submit a price lower than the imately 60% less than branded products facing
mandated ceiling price because, given inelastic only one generic competitor. In addition, the paper
demand and substitution legislation, it is the profit showed that the greater the number of generics in
maximizing price except for instances where entry the market the lower the generic prices. Using
is unprofitable. The preceding is similar to the Canadian data within the context of formulary
effect of implementing price caps or cost ceilings.10 pricing similar to the ODB framework, Anis5 dem-
Furthermore, the nature of the generic market is onstrated a similar result. Using six countries (US,
139ANIS ET AL MEDICAL CARE
TABLE 2. Estimated Coefficients of the this regulation led the price of branded products
Regression Model* on Price Ratio (PG/PB) facing generic competition to increase and generic
Estimated
firms also raising their prices with fewer firms
Variable Coefficient (SE) P entering the market. When faced with entry price
regulation, Abbot3 noted that branded firms, given
After May, 1993 0.05 (0.03) 0.11 their monopoly power, set launch prices 50%
No. of existing generic product ⫺0.04 (0.01) 0.003 higher than they would in an unregulated market.
Interaction 0.04 (0.02) 0.04 A limitation of our study is that because the
Therapeutic classification ODB Formulary was normally updated and pub-
Cardiovascular 0.52 (0.03) 0.0001 lished semi-annually, the drug prices listed could
Antiarthritics 0.47 (0.03) 0.0001 be different from the launch price of those phar-
Psychotherapeutic 0.37 (0.05) 0.0001 maceuticals. Also, we did not look at the usage
Antiinfectives 0.50 (0.03) 0.0001 levels of the generic products to analyze whether
Analgesics 0.54 (0.04) 0.0001 the 70/90 regulations had an effect on usage levels.
Neurological disorders 0.47 (0.05) 0.0001 Another issue that was beyond the scope of this
Cholesterol/lipotropic 0.65 (0.03) 0.0001 paper was the impact of the 70/90 regulations in
Proprietary analgesics 0.44 (0.01) 0.0001 preventing generic entry. This would occur if
branded firms were to price their products below
N ⫽ 218.
the normal levels or if the mandatory discount
*Pearson Goodness-of-Fit statistic over its degrees
of freedom is 1.06. prevented potential entrants from being able to
feasibly making a profit. A lack of a control group
was another limitation of this study and we real-
UK, Denmark, Holland, Germany, and South Af- ized that there might be some confounding factor
rica) with varying levels of pricing freedom with that caused the differences in the price ratios
respect to new product pricing, Reekie2 compared between periods. Because Ontario is the largest
pricing of the top five products. He showed that in market in Canada, Ontario generic prices were
markets with greater pricing freedom, competition likely to be applicable to the rest of the country as
created by rival products lowered the price of well. Therefore, even though the 70/90 regulations
medicines. were specific to Ontario, other provinces in Can-
The Gordon Commission report),13 which in- ada could not be used as controls. Because the
vestigated drug pricing in Ontario was particularly regulation originally restricted the first generic at
critical of the failure of the ODB formulary prices 75% of the branded product, we have observed
to reflect the true acquisition cost of drugs and price ratios of first generic entrant to the branded
referred to this practice as “spread pricing.” Ac- product over 0.70 and close to 0.75. However, we
cording to this practice, manufacturers included a believed that subsequent amendments to the pric-
margin over cost in their price submission to the
ing restrictions that set the maximum price ceiling
formulary in attempts to prompt pharmacists to
for the first generic at 70% instead of 75% of the
price select in their favor ie, choose to dispense
incumbent’s price, were merely incidental to the
their particular drug product.13 Lexchin9 found
original change in paradigm in 1993.
that indeed the greater the level of competition
Finally, with previous studies showing that
and consequently the greater the number of pro-
ducers the smaller is the spread and consequently prices resulting from Ontario’s BAP-type of for-
the lower the cost of drugs. mulary not being competitive5,8,13 and the anti-
Similar to the 70/90 regulations, the Omnibus competitive effects of the 70/90 regulations dem-
Budget Reconciliation Act of 1990 (OBRA 1990) onstrated in this study, the spill-over effects of
allowed Medicaid (US federally funded health these regulations on other Canadian provinces
insurance program) to adopt a most-favored- needs to be considered. Given that Ontario is the
customer rule with respect to pharmaceutical pur- largest market in Canada, the impact of the 70/90
chases.12 The OBRA 1990 allowed Medicaid to regulations could result in increased generic drug
purchase pharmaceuticals at a given percentage prices across the country. This is a concern that
below average price if the best price was not low should not be ignored by policy makers both
enough. It has been shown by Scott-Morton1 that within and outside of the Ontario market.
140Vol. 41, No. 1 DRUG PRICE REGULATIONS
Acknowledgments Ontario and Canada. Pharmacoeconomics 1993;
3:354 –361.
The authors acknowledge the assistance of Dianne 7. IMS Health. Canadian Pharmaceutical Industry
Calbick, and IMS Canada in researching and preparation Review. Pointe Claire, Quebec; 2000.
of this manuscript. 8. Liang KY, and Zeger SL. Longitudinal data
analysis using generalized linear models. Biometrika
1986;73:13–22.
References 9. Lexchin J. Effect of generic drug competition on
the price of prescription drugs in Ontario. CMAJ
1. Scott-Morton F. The strategic response by phar-
1993;148:35–39.
maceutical firms to the Medicaid most-favored-
customer rules. Rand J Econ 1997;28:269–290. 10. Laffont J, Tirole J. A theory of incentives in
procurement and regulation. Cambridge, MA: MIT Press;
2. Reekie WD. How competition lowers the costs
1993:75–76.
of medicines. Pharmacoeconomics 1998;14(Supp 1):107–
113. 11. Caves RE, Whinston MD, Hurwitz MA.
Patent, expiration, entry, and competition in the US
3. Abbott TA. Price regulation in the pharmaceu-
pharmaceutical industry. Brookings Papers on Economic
tical industry: Prescription or placebo? J Health Econ Activity, Microeconomics 1991;1– 48.
1995;14:551–565.
12. US Congress, Office of Technology Assess-
4. Anis AH. Substitution law, insurance coverage, ment. Pharmaceutical R&D: Costs, risks and rewards.
and generic drug use. Med Care 1994;32:240 –256. Washington DC: US Government Printing Office;
5. Anis AH. Pharmaceutical prices with insurance OTA-H- 1993:522.
coverage and formularies. Can J Econ 1992;25:420 – 437. 13. Gordon JRM. Report of the Commission on the
6. Detsky A. Guidelines for economic analysis of pricing of multiple-source drug products in Ontario.
pharmaceutical products: a draft document for Kingston, Ontario: Mimeo; 1984.
Appendix
APPENDIX A. TABLE 1. 70/90 Regulations: Legislative Framework
Time Regulations
May 1993 75/90 program was imposed under the Prescription Drug Cost Regulation Act (PDCRA) and the
Ontario Drug Benefit Act (ODBA). According to the regulations, new first generic products
were not designated and listed unless they were priced at or lower 75% of the equivalent brand
name product. New second and subsequent generic products were not listed and designated
unless priced at or below 90% of the lowest priced listed comparable product.
Nov 1998 The new 70/90 Regulations were embraced under section 11 of Regulation 201/96 under the ODBA.
And under Subsection 7(2) of Regulation 935 under the Drug Interchangeability and
Dispensing Fee ACT (DIDFA) which, prior to May 27, 1996 was known as PDCRA. According
to the amended regulations, the drug benefit price of new first generic products must be equal
to or less than 70% of the price of the original products. Restrictions on the pricing of new
second and subsequent generic products were maintained.
141You can also read
