A PROGRAM FOR RESEARCH ON SOCIAL AND ECONOMIC DIMENSIONS OF AN AGING POPULATION Effects of ‘authorized generics’ on Canadian drug prices Paul Grootendorst SEDAP Research Paper No. 201
For further information about SEDAP and other papers in this series, see our web site:
Requests for further information may be addressed to:
Effects of ‘authorized generics’ on Canadian drug prices Paul Grootendorst SEDAP Research Paper No. 201
The Program for Research on Social and Economic Dimensions of an Aging Population (SEDAP) is aninterdisciplinary research program centred at McMaster University with co-investigators at seventeen otheruniversities in Canada and abroad. The SEDAP Research Paper series provides a vehicle for distributingthe results of studies undertaken by those associated with the program. Authors take full responsibility forall expressions of opinion. SEDAP has been supported by the Social Sciences and Humanities ResearchCouncil since 1999, under the terms of its Major Collaborative Research Initiatives Program. Additionalfinancial or other support is provided by the Canadian Institute for Health Information, the CanadianInstitute of Actuaries, Citizenship and Immigration Canada, Indian and Northern Affairs Canada, ICES:Institute for Clinical Evaluative Sciences, IZA: Forschungsinstitut zur Zukunft der Arbeit GmbH (Institutefor the Study of Labour), SFI: The Danish National Institute of Social Research, Social DevelopmentCanada, Statistics Canada, and participating universities in Canada (McMaster, Calgary, Carleton,Memorial, Montréal, New Brunswick, Queen’s, Regina, Toronto, UBC, Victoria, Waterloo, Western, andYork) and abroad (Copenhagen, New South Wales, University College London).
Effects of ‘authorized-generics’ on Canadian drug prices* Paul Grootendorst Associate Professor Leslie Dan Faculty of Pharmacy University of Toronto [email protected] 416 946 3994 Abstract
This paper examines how the use of ‘authorized-generics’ (AGs) influences Canadian prescription drug prices. An authorized-generic is the actual brand name drug product, manufactured by the brand firm, but sold as a generic by a licensee or subsidiary of the brand, competing with independent generics (IGs), which operate independently from the brand firm. In theory, AGs have offsetting effects on drug prices. On the one hand, AGs compete against IGs and increases in the number of generic competitors should lower prices. On the other hand, the threat of AG entry into a therapeutic market might deter entry by IGs and this would lessen competition. Moreover, brand firms might increase prices of their brand drugs to increase demand for their AG. I find that when AGs are first to enter a drug market, average drug prices drop by about 12%; average prices drop by smaller amounts, the larger the AG share of the generic market. I could not directly assess whether the threat of AG entry into a market might deter entry by IGs. IG executives, however, state that the threat of AG entry has decreased their incentive to challenge ‘marginal’ drug markets. In particular the threat of AG entry has increased from $5m to $10m the threshold market size – the value of brand drug sales in the 10th year that it has been on the market, below which the IG firm will not attempt to enter. IG executives also stated that AGs have seriously reduced IG retained earnings. The reduction in retained earnings has hampered their ability to challenge brand drugs with annual sales well above $10m, but which have particularly high entry costs. Finally, the IG executives claimed that brand firms have attempted to use the threat of AG entry to negotiate agreements with the IG to delay entry (or not enter at all). A comprehensive evaluation of the competitive effects of AGs would need to verify and quantify these costs and compare these to the benefits of AG competition. *Financial support from Industry Canada, the Competition Bureau and the Program for Research on Social and Economic Dimensions of an Aging Population (SEDAP) is gratefully acknowledged. MS Shim, Yuqian Lu and Ivana Todorovic provided capable research assistance. The author thanks Aidan Hollis and participants at the Competition Bureau Intellectual Property Symposium, Ottawa, for helpful comments on an earlier draft. Keywords: authorized-generic drugs, independent generic drugs, drug prices, Canada JEL Classifications: I11, I18, L65 Résumé
Cette étude se penche sur les effets de l’usage de médicaments « génériques autorisés » (GAs) sur le prix des prescriptions canadiennes. Un GA est un produit de marque original vendu comme un médicament générique par un tiers parti autorisé ou une filiale de la marque, rentrant en concurrence avec les produits génériques indépendants (GIs), et qui opère ainsi indépendamment de la compagnie mère. En théorie, les GAs ont des effets compensateurs sur le prix des prescriptions. D’un côté, les GAs rentrent en concurrence avec les GIs et la hausse du nombre de producteurs de médicaments génériques devrait entraîner une diminution des prix. D’un autre côté, la menace liée à l’introduction des GAs sur le marché thérapeutique pourrait décourager l’entrée de nouveaux GIs, et provoquer une baisse de la concurrence. De plus, les marques pourraient augmenter les prix de leurs médicaments originaux afin d’augmenter la demande des GAs. Nous trouvons qu’à la suite de l’introduction des AGs sur le marché, le prix moyens des médicaments diminuent de 12%; mais ces derniers diminuent plus modérément selon l’importance de la part de marché des GAs sur le marché des médicaments génériques. Nous n’avons pas été capable de déterminer directement si la menace posée par l’introduction des GAs sur un marché décourageait l’entrée de nouveaux GIs. Les cadres de compagnies de GIs, cependant, affirment que la menace de voir apparaître des GAs a diminué volonté de rentrer sur des marchés de médicaments plus marginaux. En particulier, l’apparition des GAs a entraîné une hausse de $5m à $10m du seuil critique de taille du marché – la valeur des ventes d’un médicament de marque durant la dixième année de sa mise en marché, seuil en dessous duquel une compagnie de GI ne tentera pas de pénétrer le marché. Les cadres des compagnies de GIs ont aussi souligné que les GAs ont sérieusement réduit leurs bénéfices non distribués. Cette réduction des revenus stables a entravé leur aptitude à compétitionner avec les médicaments de marque dont les ventes sont supérieures à $10m, mais dont les coûts d’introduction sont particulièrement élevés. Enfin, les cadres des compagnies de GIs soutiennent que les compagnies de marque ont tenté d’utiliser la menace d’introduire des GAs pour négocier des délais (ou des retraits) d’introduction avec les GIs. Une évaluation complète des effets compétitifs des GAs devrait vérifier et quantifier ces coûts et les comparer aux bénéfices de la concurrence émanant des GAs.
Introduction
This paper examines how the use of ‘authorized-generics’ influences Canadian prescription drug prices. An authorized-generic (hereafter ‘AG’)1 is “the actual brand name drug product, manufactured by the brand company, but sold as a generic by a licensee or subsidiary of the brand, competing with independent generics.” (Hollis and Liang, 2006) AGs are to be distinguished from independent generics (hereafter ‘IG’s) which operate independently from the brand firm. Theoretically, AGs have offsetting effects on drug prices. On the one hand, AGs compete against IGs and as I discuss below, increases in the number of generic competitors should lower prices. On the other hand, as Hollis (2003) and Kong and Selden (2004) have noted, the threat of AG entry into a therapeutic market might deter entry by IGs and this would lessen competition. Moreover, brand firms might increase prices of their brand drugs to increase demand for their AG if there are no IGs available. Why might the threat of AG entry deter entry by an IG? According to Hollis (2002), the reason is that the AG, should it be the first generic to enter a market, typically captures a significant and durable market share. This reduces the demand for the IG and may make entry into marginally profitable therapeutic markets unprofitable. One would therefore expect that the threat of AG entry might deter IG entry in markets with only modest revenue potential. Generic drug competition in the Canadian drug market
How does generic drug competition lower drug prices? In a conventional market, increases in the number of suppliers reduce the average price paid by consumers – suppliers lower prices in an attempt to gain market share. In the limiting case, known as a perfectly competitive market (one with very many buyers and sellers of a homogenous product), price will decline to the marginal production cost. The Canadian market for generic prescription drugs is different. First, two generic companies, Apotex and Novopharm, control over half of the market (Skinner 2004). Second, although there are many consumers in this market, including private insurance plans and cash paying customers, the market is dominated by several large provincial government drug plans, notably the plans operating in Ontario (the Ontario Drug Benefit program) and Quebec (RAMQ) that cover the drug costs incurred by seniors, those with high drug costs relative to income, and other groups. The ODB plan is the single largest drug plan in Canada; it dictates the amount that it is willing to pay for generic drugs, and these prices become the defacto prices charged to other customers. The amount that is ODB willing to pay for generic drugs is a proportion of brand drug prices. During the period May 1993 – September 1998, ODB invoked the ‘75/90’ regulations. The 75/90 regulations stipulated that the first generic entrants’ price could not exceed 75% of the incumbent branded drug price. When the second generic entered the market, the maximum reimbursement price dropped to 67.5% of the branded drug price (i.e. 75% of the first generic entrants price: 75% of 90%=67.5%). During the period November 1998 – September 2006,
1 Authorized-generics are sometimes referred to as ‘pseudo-generics’.
ODB limited the first generic entrants’ price to 70% of the incumbent branded drug price. When the second generic entered the market, the maximum reimbursement price dropped to 63% of the branded drug price (i.e. 70% of the first generic entrants price: 70% of 90%=63%). Since October 2006, the ODB reimbursement of generic drugs is limited to 50% of the brand drug price. The ODB, however, will consider a increasing the reimbursement of the first generic entrant by up to 70%, on a case-by-case basis. Anis et al (2003) present evidence that generic prices in Ontario are set close to the ODB maximum generic drug reimbursement prices. The prices charged for generic drugs in other provinces tend to match ODB prices. Generic firms that attempt to charge more would soon be undercut by wholesalers who sell drugs at ODB prices. How then do generic firms compete for market share? Once they become available for sale, generic drugs rapidly replace branded drugs owing to the reimbursement policies of almost all drug plans, public and private alike. Most plans limit reimbursement of multi-sourced drugs (i.e. drugs for which there are both brand and generic forms available for sale) to the price of the generic drug.2 Drug plan beneficiaries do have the option of paying extra to get the higher priced brand version of the drug, but relatively few are willing to pay extra. The first generic therefore gains a large market share and can charge 70% of the branded drug price. Once there are several generics available, generic firms compete for sales to large pharmacy chains such as the Shoppers Drug Mart chain and independent pharmacies by lowering prices charged to these pharmacies. Competition occurs through the generosity of cash and in-kind rebates; these rebates drive down the effective price paid by the pharmacy for generic drugs. The difference between the ODB reimbursement price and the pharmacy’s actual acquisition cost becomes part of the pharmacy’s revenues. Anecdotal evidence suggests that this margin can be as high as 60% of the formulary price of generic drugs. Anis (1992) reports that when the Nova Scotia public drug plan replaced formulary pricing with a system in which only the actual acquisition costs were reimbursed, generic drug prices fell by about 40%. The rebate system was disrupted by regulations enacted by ODB in October 2006. In an attempt to manage program costs, ODB appropriated some of the rebate amounts that were accruing to pharmacy chains. It did so by reducing the maximum that it would reimburse generic drugs to 50% of the brand drug price and limiting the rebate amounts generic firms paid to pharmacies for ODB-reimbursed drugs to 20% of the ODB reimbursement price.3 Although ODB did not regulate rebates paid for drugs that it did not reimburse, it applied a code of conduct that required that all generic drug rebates that Ontario pharmacies received (from sales of generic drugs to both ODB and private payers) be used for patient care. Rebates could no longer be taken as
2 The Quebec RAMQ plan is notable for its willingness to pay for the brand version of the drug for 15 years (from the date of formulary inclusion of the brand drug) even if generic versions are available for sale. 3 To offset these revenue losses, the Ontario government increased the dispensing fee (from $6.46 to $7.00 per prescription) and, in recognition of professional services traditionally rendered by community pharmacists without payment, established new funding (cited to be in the area
of $50 mil ion for 2006-07) as compensation for these services. Further details on the legislation
are available at: http://www.health.gov.on.ca/english/public/legislation/drugs/hu_drugsact.html
personal income. These regulations appear to be actively enforced: ODB just recently announced that each Ontario pharmacy is required to submit quarterly reports on the amount of rebate revenue received and how these revenues are used. It is unclear how these regulations will affect the prices received by generic firms. On the one hand, the maximum rebate that can be paid on ODB prescriptions is limited to 20% of the ODB price – limiting the rebates should increase prices received by generics. On the other hand, this limit does not apply to non-ODB prescriptions, so that generics might still use rebates to compete for this segment of the Ontario market, as well as the rest of the domestic market. Moreover, the maximum generic price is now 50% of the brand price (although as stated earlier, it is possible that the first generic entrant might be able to negotiate a higher amount). It seems likely that that this new, lower generic price will be available to the rest of the domestic market, although pharmacy chains are actively resisting this at present. Why might a brand firm introduce an AG?
Under the terms of the ‘NOC-link’ section of the Patent Act, IG firms must declare to the brand firm their intention of introducing a generic copy onto the market.4 This declaration, and any resulting litigation over the validity of brand firm’s patents, gives the brand firm some indication of the timing of IGs onto the market. If no IG is expected to enter, then neither will the AG. If it appears that IG entry is likely, then the brand firm may launch an AG, for two reasons. First, launching an AG might be profitable. Relative to IG firms, the brand firm faces low fixed market entry costs. Because the AG is the brand drug, the brand firm does not face the costs of conducting bioequivalence studies,5 nor does it face the costs of challenging the validity of any outstanding patents. Also, the costs of establishing manufacturing capacity have already been incurred at the time of the brand drug launch. The brand firm must pay a fee for the licensee to sell the AG drug, but these costs would likely be less than the fixed costs incurred by IG firms. Should it decide to enter a market, the brand firm is faced with the decision of launching its AG before the IG drug, or not. By launching before the IG, the AG initially captures the entire share of the generics market. Moreover, if it is the only generic available on the market, it need not pay pharmacies any rebate (since the pharmacy has no choice but to carry it) and this will increase net reimbursement. These benefits may be short lived however – as the number of generic competitors increase, the AG’s market share will decrease and the amount of rebate needed to gain access to pharmacy shelves increases.6 More importantly, by launching before 4 IG firms must declare using a ‘Notice of Allegation’ or NOA that their generic drug will not infringe any of the brand firm’s patents but this has the effect of signaling their intentions to enter
a therapeutic market. A brand firm can delay the regulatory approval of the generic drug
(technically the issuance by Health Canada of a Notice of Compliance, or NOC, indicating that the drug is bioequivalent to the brand drug) by alleging patent infringement. 5 Health Canada requires evidence that IGs are bio-equivalent with the brand drug and also have acceptable potency, purity, and stability. Hollis (2005) cites a newspaper article that indicates that these development and clinical testing costs typically range from $0.5 m to $3 m
per drug. 6 One way that a major IG firm is reported to have gained AG’s market share is to purchase from pharmacies the AG’s inventory, replacing it with their own product.
the IG, it will cannibalize its brand drug sales. Specifically, during the period November 1998 – September 2006, the AG realized only 70% of the price paid for the brand version of the drug. Accepting a 30% price cut (and after the recent changes to ODB reimbursement, up to a 50% price cut) makes sense only if the AG realizes a compensatory increase in market share relative to the share it would realize if it launched alongside (or after) the IG. Launching an AG before the IG could increase profits for another reason. If the AG is the only generic on the market then the brand firm could potentially coordinate the prices charged for the brand and AG forms of the drug in a way that maximizes profits. In particular, as Hollis (2005) notes, by pricing the brand version relatively high, it can increase the amount that it charges for the AG version of the drug, subject to price regulations imposed by the federal government and provincial drug plans. The second reason that the brand firm might find it in their interest to launch an AG is to establish a reputation among IGs that it will launch an AG in ‘marginally profitable’ markets – markets in which the presence of the AG can tip the balance between economic profitability and loss. This might deter future IG entry in such markets. Existing Evidence
In theory, the impact of AGs on drug prices is ambiguous. What does the evidence indicate? There is no empirical evidence on the effects of AGs on the decision by an IG to attempt to enter a drug market. The evidence on the effect of AGs on drug markets in which IGs do enter is mixed. The one Canadian study (Hollis, 2005) found AGs to be mildly anti-competitive. Hollis analyzed sales data for the 31 drugs in Canada for which the first generic competitor entered during the years 1994-1997. He concludes: “This paper shows that pseudogenerics appear to increase both brand and generic drug prices, in line with a simple theoretical model. … The econometric analysis shows that a 10% increase in the pseudogeneric share of generic sales leads to brand prices about 1% higher, controlling for the number of generics and the years since the first generic entered. Since the typical pseudo- generic controls about 40% of generic sales in the first few years of generic entry, this implies about a 4% increase in brand prices, compared to the case of no pseudo-generic. The direct impact of pseudo-generics on generic drug prices appears to be similar in scale.” Page 348. The US evidence is mixed. The Pharmaceutical Research and Manufacturers of America (PhRMA), the association of US brand drug firms, commissioned IMS Consulting to assess the effects of AGs on US generic prices. The study7 found that AGs were pro-competitive. Specifically the study compared markets with and without AGs present and found that average
7 The study is available at: http://www.phrma.org/files/IMS%20Authorized%20Generics%20Report_6-22-06.pdf [accessed
http://www.phrma.org/news_room/press_releases/phrma_statement_on_authorized_generics/
generic prices (measured as a fraction of the brand drug price) were lower– on average 15.8 percentage points lower – in markets with AGs present. In response, the Generic Pharmaceutical Research and Manufacturers of America (GPhRMA), the association of US independent generic firms, commissioned Hollis and Liang to evaluate the PhRMA study methods. The study, which will appear in Hollis and Liang (2006), challenged both the measurement of the price variables in the PhRMA study as well as the methods used to identify the impact of AGs on price. For instance, Hollis and Liang suggested that the PhRMA study finding of greater percentage discounts to brand prices in markets with AGs is due in part to higher brand prices in markets with AGs, rather than lower generic prices. In other words, during the period immediately following IG entry, prices of the brand drug increased more in markets with AGs than in markets without AGs. They also question whether the influence of AGs on prices can be captured with a simple comparison of markets with and without AGs present. In particular, they note that these market likely also differ with respect to sales volume, number of anticipated entrants, potential for licensing and other factors that could independently affect price discounts. Finally, Hollis and Liang note that the PhRMA study does not address the effects of potential AG entry on the likelihood that an IG will attempt to enter a market. When they reanalyzed the data used in the PhRMA study, Hollis and Liang found some evidence of a very modest anti-competitive effect. Specifically, in markets with AGs present, the average discount of generics off of branded prices was less than in markets without an AG (17.7% vs. 17.1%). Research Objectives
Industry Canada and the Competition Bureau asked me to investigate the following questions about the prevalence and effects of AGs in Canadian pharmaceutical drug markets:
In what proportion of drug markets are authorized generics present?
How important is the invocation of the automatic 24 month stay under the PM(NOC) Regulations in determining the entry of authorized generics?
What are the characteristics of the drug markets that authorized generics enter? In particular, how important is market size in explaining authorized generic entry?
What is the impact of authorized generics on drug prices and the entry of independent generics?
How important are market opportunities in the U.S. at explaining the decisions of independent generic firms to enter a particular market?
Questions (ii) and (v) were addressed using qualitative methods – I interviewed executives of major brand and generic firms as well as representatives of the IG trade association, the Canadian Generic Pharmaceutical Association (CGPA). The remaining questions were addressed by analysing IMS Health Canada data on the national, monthly sales of the generic and brand versions of all dosage forms and strengths of the molecules in which the first generic drug (authorized or independent) launched over the period 1998-2004. I identified the AG drugs using the NOC database produced by Health Canada.8 Specifically AG drugs were identified as those generic drugs which have received New Drug Submission (NDS) status; independent generics have Abbreviated New Drug Submission (ANDS) status.9 Not all NDS status generic drugs are AGs, however. I consulted with IMS Health Canada, Jack Kay of Apotex and Brogan Inc to help identify the NDS drugs that were not AGs. Using these data, I determined the proportion of drug markets in which an AG has entered (question (i)); estimated regression models of the probability of AG entry in the molecule market (question (iii)); and finally estimated regression models of the average price for the molecule as a function of the presence of an AG (question (iv)). The probability of AG entry was modelled to be a function of inter alia an estimate of the expected total sales revenue for all dosage forms and strengths of the molecule in the year before the launch of the first generic; indicators of year of first generic launch; and finally indicators of the company producing the brand drug. I describe the methods to estimate the influence of AGs on drug prices below. Results
1. In what proportion of drug markets are authorized generics present?
I defined a drug market as a unique combination of molecule and dosage form; hence I distinguished, for instance, the oral solid version of cyclosporine from its oral liquid and its injectable forms. I distinguished between the different dosage forms of the same molecule because in some cases the pattern of generic drug entry varied by dosage form. In the case of cyclosporine, for instance, an AG (but not an IG) launched an oral solid, while an IG (but not an AG) launched an oral liquid. Neither an AG nor an IG launched an injectable form. Of the 83 drugs in the sample, however, most were oral solids (tablets and capsules) that were not available in other dosage forms. Table 1 provides a cross-tabulation of IG and AG entry into the 83 drug markets. An AG entered 32 (39%) of the 83 drug markets in the sample, while an IG entered 74 of the 83 markets.
8 http://www.nocdatabase.ca/ 9 According to Jack Kay, COO of Apotex, an AG drug would not cite any bioequivalence data on its product monograph, whereas an IG would. Unfortunately, I could not locate the product monographs of the generic drugs in this study.
Both an IG and an AG entered in 31 markets. The 39% AG entry rate over this time period is consistent with Hollis’ (2005) finding: “In 1999, the total pharmacy sales of these generics drugs totaled approximately $500m, of which the pseudo-generic share was 34.6%.” (p. 332) Table 1: Cross tabulation of AG and IG entry in 83 drug markets =1 if | authorized | =1 if independent generic | generic launched launched | 0 1 | Total -----------+----------------------+---------- 0 | 8 43 | 51 | 15.69 84.31 | 100.00 | 88.89 58.11 | 61.45 -----------+----------------------+---------- 1 | 1 31 | 32 | 3.13 96.88 | 100.00 | 11.11 41.89 | 38.55 -----------+----------------------+---------- Total | 9 74 | 83 | 10.84 89.16 | 100.00 | 100.00 100.00 | 100.00
Note: first number in each cell is the frequency, second number is the row percentage and the third number is the column percentage. Table 2 provides information on the timing of the AG and first IG in the same market. Of the 31 markets in which both an AG and an IG entered, the IG entered first in 16, the AG entered first in 10 and they both entered in the same month in 5 markets. Table 2: Number of Months by which AG Precedes IG Number of Months by which AG Precedes IG Percent Cum. Number of Months by which AG Precedes IG Percent Cum. Note: a positive number means that the AG entered before the IG. Figure 1 plots data on the number of months by which AG entry precedes IG entry, by the total sales of the drug in 1998. There is some evidence that in small to medium markets (markets with sales of $50 M or less), the AG is no more likely than the IG to enter first. In the 4 largest markets, however, the AG comes in no later than the IG. Figure 1: Number of Months by which AG Precedes IG, by 1998 market sales
Table 3 presents summary data on the fraction of markets with AGs and IGs present, by the total sales of the drug in 1998. The fraction of markets with an AG or IG tends to increase with the value of 1998 sales of the drug.
Table 3: Fraction of drug markets with authorized generics and independent
generics present, by quintiles of total sales of the drug in 1998 ($millions)
Quintile Mean Median Min Appendix 1 presents detailed information on IG and AG entry into the 83 markets, as well as year specific sales in these markets.
2. How important is the invocation of the automatic 24 month stay under the PM(NOC) Regulations in determining the entry of authorized generics?
My sense from the literature and discussions with industry representatives is that the PM(NOC) (hereafter the “NOC link”) regulations facilitate the entry of AGs, although they are not absolutely necessary. A brand firm will normally not launch an AG unless it is confident that an IG will indeed enter. Under the NOC link regulations, the brand firm can more accurately ascertain the date of expected entry of an IG into a therapeutic market than it could without the NOC link regulations. The reason is that under the NOC link regulations, an IG firm is obliged to notify the brand firm of its intention to enter. The NOC link regulations also give the brand firm the right to delay the issuance of an NOC by alleging patent infringement. The ensuing litigation provides the brand firm with much more information on the entry date than it would otherwise have. For instance, the brand firm can monitor the progress of the trial and hence forecast its outcome, providing it with information with which it can determine whether and when it should ready an AG.
Without the NOC link regulations, conversely, the only information that the brand firm would have about IG entry is the expiration date of its patents, and Health Canada NOC approval of an IG drug.
It is important, however, not to overstate the importance of the NOC link regulations in informing the brand firm of the expected market entry date of an IG. First, my data indicates that brand firms do not always attempt to beat IG firms to the market with an AG, especially in markets with only modest revenue potential. In larger markets however, the AG enters at the same time or some months before the IG. Second, my brand firm contacts mentioned that even if they were notified of the imminent entry of an IG by some other means (say the issuance by Health Canada of an NOC) they could still launch an AG about the same time as an IG. The reason is that an IG will usually achieve significant sales volumes only some time after it receives its NOC. In particular, it takes some time for the provincial drug plans to list the IG on its formulary (which is the point in time at which demand for the drug increases markedly). Moreover, AGs might also be able to launch faster than IGs. According to the CGPA, several
provincial government drug plans – including the plans operating in Alberta and Newfoundland – require generics to submit evidence of bioequivalence before they will reimburse the generic drug. (The other plans accept Health Canada certification of bioequivalence.) Vetting the bioequivalence data for an IG takes longer than that for an AG (since the AG is the brand drug, bioequivalence is a non-issue). Depending on the frequency with which the assessment committees meet and make their recommendations re: bioequivalence, and the frequency with which these recommendations are acted upon by formulary committees, the AG might get listed on a provincial formulary before an IG.10
3. What are the characteristics of the drug markets that authorized generics enter? In particular, how important is market size in explaining authorized generic entry?
I estimated a probit regression model of the probability of AG entry into the drug markets. The probability was modelled as a function of: (i) indicators of the brand drug firm (in the event that different brand firms are more or less likely to use AGs); (ii) indicators of the year of entry of the first generic (authorized or independent); as well as (iii) indicators of the quintiles of market sales in the 12 month period prior to generic entry. These quintiles divide the drug markets into 5 evenly sized groups, ranked by sales from smallest to largest. The markets in the bottom quintile had annual sales of between $0.01 and $1.73 million; the markets in the top quintile had annual sales of between $50.8 and $485 million (Table 4). Table 4: Mean, median, min and max total sales ($million) in drug market in the
year before generic entry, by sales quintile
Quintile Mean Median Min
The model estimates indicate that of the three factors considered above, factor (iii), market size, is the strongest predictor of AG entry. Holding constant the influence of factors (i) and (ii), the
10 An IG executive I interviewed recounted how their launch of ramipril was pre-empted by an AG: “We got our NOC on Dec 12, after years of litigation, and sent our submission to Alberta by
courier on Dec.13. It arrived Dec 14. Sanofi-Aventis arranged for a NOC for authorized generic
to be issued to ratiopharm on Dec 13. Because they did not have to file a substantive submission in Alberta but are automatically accepted because same as brand, Alberta considered their
submission complete on Dec 13. Our submission had to include data and was very large, and
thus could not be transmitted by fax. It arrived at Alberta Health on Dec 14 as aforesaid. Alberta fast-tracks the first generic. Hence they proceed to immediately list the ratio product
and not ours, and ratio thus took the whole Alberta market. …”
bigger is the market size, the greater is the likelihood of AG entry. Compared to markets in the lowest sales quintile, AGs were: 33 percentage points more likely to enter markets in the second lowest sales quintile (p=0.196), 45 percentage points more likely to enter markets in the middle sales quintile (p=0.059), 43 percentage points more likely to enter markets in the second highest sales quintile (p=0.081), and 64 percentage points more likely to enter markets in the top sales quintile (p=0.004). A useful summary statistic is the R-squared; this statistic reflects the amount of variation in the outcome variable that is explained by predictor variables. I note that my AG entry probability model pseudo-R-squared is decent (29%) for a cross sectional data set. This likely reflects the strong predictive performance of pre-IG market sales; the remaining explanatory variables had apparently little explanatory power. Specifically, there were no noticeably different probabilities of AG entry among the different brand firms. (Novartis Pharma and Pfizer had slightly higher probabilities of issuing AGs than the other brand firms, but it is not clear if this is a chance result.) Similarly, there were no discernable trends in the likelihood of AG entry by year.
4. What is the impact of authorized generics on drug prices and the entry of independent generics?
Recall that, theoretically, AGs have offsetting effects on drug prices. On the one hand, AGs compete against IGs and increases in the number of generic competitors should lower prices. On the other hand, the brand firm might raise prices on its branded drug to increase the sales revenue accruing to its AG drug. Moreover, the threat of AG entry into a therapeutic market might deter entry by IGs and this would lessen competition. One therefore needs to assess the net effect of AGs on drug prices. One can do so by considering the effects of AGs on drug prices in drug markets with and without generic drugs present. The impact of AGs on markets with a generic drug (IG or AG) present can inform the impact of AG competition on drug prices, after accounting for the fact that brand firms might increase branded drug prices when an AG is present. I develop and estimate a drug price model that addresses this question. This model uses IMS data on the average prices charged for the drug molecules in which a generic drug entered between 1998 and 2004. This model, however, does not inform the question as to how the threat of AG entry deters IG entry in marginally profitable markets and how the attendant reduction in competition translates into higher prices. The impact of AGs on IG entry is the difference between the likelihood of IG entry with the threat of AG entry present and the likelihood of IG entry without the threat of AG entry. It is difficult to estimate this impact. While data on the drug markets in which IGs have entered is readily available, it is difficult to determine which drug markets IGs would have attempted to enter had it not been for the threat of AG entry. In particular, while we can directly observe the drug markets in which all relevant patents have expired but in which there are no IGs, it is not obvious which patent-protected drug markets would have been successfully challenged by IG firms (perhaps because the patents were sufficiently weak) but were not due to the threat of AG entry. Yet these data are required to address the question. I explored various approaches to addressing this question. For instance, I investigated whether it was possible to analyze annual data on the threshold market size, above which an IG enters. If
the advent of AG competition in the Canadian market occurred recently then it might be possible to examine time series data on these market size thresholds in period before and after the advent of AG competition. It turns out however that AGs have competed in the Canadian market since at least 1987, making it difficult to collect data in the period before AG competition began. Instead of conducting a quantitative analysis of AGs on the likelihood of IG entry, I used qualitative methods. Specifically I interviewed executives of a major domestic IG firm about the effect of the threat of AG entry on the threshold market size above which the IG will attempt to enter. I describe the results of these interviews below. Quantitative analyses
I developed a linear regression model of average prices charged to consumers (i.e. drug insurers, and patients paying out of pocket) each month, in each of the 83 drug markets described in Appendix 1. The average monthly price in each of these markets is simply total monthly sales revenues across all strengths and manufacturers of the drug (including brand and generic firms) divided by the total number of milligrams (mg) of the drug dispensed. The total mgs of the drug dispensed, in turn, is the number of units dispensed times the strength in mgs per unit. I deemed expenditure per mg of drug dispensed to be the best measure of average price charged – it automatically standardizes prices charged consumers for differences in the market shares of the brand and generic products. Importantly, it is sensitive to any increases in brand price that might occur after AG entry. And unlike a price measure based on expenditure per prescription, my measure is sensitive to the number of units dispensed per prescription and the strength per unit. The IMS sales data are derived from a sample of total prescription drug sales and volumes at over 2000 representative retail pharmacies. The sales data include all wholesale and retail markups and dispensing fees, so that some of my consumer price variable represents payments to wholesalers and pharmacies. Below, I explain how my model will account for the influence of variation in such payments. I modelled the average drug and month specific price as a function of the following covariates:
• Indicators of various levels of the authorized generic share of the generic prescription
volume, measured as the total number of mgs dispensed. As Hollis (2005) notes, this variable was thought to better capture the degree of AG market penetration than a simple binary indicator of AG entry. The use of multiple indicators, each indicating the degree of AG penetration (0%, 1-20%, 21-40%, 41-60%, 61-80%, 81-100%) allows for non-linear effects of AG entry on price.11 12
11 Hollis (2005) presents evidence that the AG share of the generic market is endogenous in the price model; i.e. AG share is correlated with an unobserved component of price. He therefore
developed an instrumental variables estimator using as an instrument the number of months by
which the AG beats the first IG to market. I deal with the possible endogeneity of AG share in my price model by using the one period lagged values of AG share and number of IGs instead
of their contemporaneous values. 12 In some versions of the empirical price models, I distinguished between AGs that are licensed to a subsidiary of the brand firm from the AGs that are licensed to an IG firm, in order to assess
• Indicators of various levels of the number of independent generic competitors
• Drug market specific indicators (included to pick up time-invariant differences in the
price per mg of the different molecules; this will also pick up the influence of market size and other time-invariant factors that might affect entry by IG and AG firms)
• Quarter of year indicators (included to pick up time period-specific changes in the price
common to all the different drug markets, such as changes in pharmacy and wholesaler mark-ups and fees)
I modelled the effects of AG generic market shares and other predictor variables on the log average price per mg. Focussing on log price allowed me to assess the effects of changes in the AG generic market share on relative (percentage) price changes, which is intuitively more appealing that absolute changes in price per mg.13 I used all available data, except that I removed observations on nefazadone after it was pulled from the market due to safety concerns in January 2004. Although there were a few price observations available after this date, the average price was very poorly estimated. I also estimated the price model using the subset of the data which remained after dropping observations on drug markets in which the price and generic market share data were measured with considerable error.14 Measurement error can decrease the precision of one’s model estimates and I wanted to ensure that my results were robust to measurement error.15 The reader is invited to inspect drug market specific data on monthly average price, average brand drug price, generic drug share of the market, and authorized generic drug share of the market, presented at http://individual.utoronto.ca/grootendorst/pdf/drugpricegraphs.pdf to assess the degree of measurement error in these drug markets. The price models fit the data quite well – R squared values all exceed 95%. (See Appendix 2 for details of the model results.) Moreover, the parameter estimates are plausible. The models suggest, for instance, that increases in the number of independent generic drugs reduce average prices.
whether this distinction makes a difference to the estimated impact of AG generic drug share on average price. The data were not sufficiently informative to detect a difference. 13 The log transformed price model is also more consistent with the data than the linear price model: I subjected the average price model to the boxcox transform to determine whether a linear or log specification was more appropriate – the linear specification was handily rejected
in favour of the log specification. 14 Recall that the IMS sales data are estimates based on a representative sample of pharmacies. For some drugs with low prescribing volumes, the price estimates are highly variable. This could
be because of the limitations of survey sampling. It could also reflect a limitation with my data. I
had to enter manually into my database sales data for the years 1998-2000. (IMS data older than 72 months are available in hardcopy only.) The dol ar sales and quantities data in these
hardcopies are rounded to the nearest 1000 dol ars or units, which can introduce measurement
error when the underlying sales volumes are small. 15 Specifically, I removed the follow drugs: Amiodarone Inj, Cefuroxime Axetil Liquid, Ciprofloxacin Liquid, Clavulanic Acid Liquid, Clindamycin Inj, Clindamycin Liquid, Cyclosporine
Inj, Cyclosporine Liquid, Fenofibrate, Fluconazole Inj, Fluconazole Liquid, Ketorolac Inj, Labetalol Inj, Lamotrigine Chewable, Piperacillin, Propofol, Tizanidine, Vancomycin Inj, Warfarin Inj.
The models suggest that AGs are mildly pro-competitive. Focusing on the results in model 1.3, when AGs capture 80-100% of the generic market, that is, when they are the first or among the first to enter the market, they unambiguously reduce prices by about 11%. When the AG generic market share is in the range of 21-80%, prices drop by a smaller degree – between 5 and 7%. When the AG generic market share is in the range of 1-20%, prices drop by an even smaller degree – about 3%. This suggests that the average per mg price of the brand drug increases as IGs take market share away from the AG (otherwise the additional competition provided by IGs would lower average price). To investigate further, I estimated models of the impact on AG generic market share on the average brand drug price per mg. These models suggest that average brand prices are slightly higher (between 2-8%) when AGs control between 1-40% of the generic market compared to when they are not sold at all. (Brand prices are essentially unchanged when AGs control 41- 100% of the market.) Qualitative analyses
The generic firm executive indicated that AGs have their primary anticompetitive effect on the decision of IGs to enter a drug market. He indicated that with AGs present, and under the current set of ODB drug reimbursement regulations, he will not bother challenging a drug market with less than $10 million in annual sales. He reasoned as follows. Generics can charge at most 50% of the brand price so that potential revenues drop to $5 million. Even if his IG firm is the first on the market, he can get at most 75% of the entire market (recall that RAMQ and some private plans will continue to cover brand drugs even with generics available). This reduces revenues to 0.75×$5m = $3.75m. Under the new ODB regulations, the cost of rebates to pharmacies is about 40% (down from over 50%). Hence revenues net of rebates are 0.6×$3.75m= $2.25m. Production costs (raw materials, labour and indirect overheads, including depreciation) are roughly 40% of revenues; revenues net of production costs are therefore 0.6×$2.25m=$1.35m. Selling expenses, including administration, distribution, and the legal costs of challenging patents reduce net revenues by a further 15%; hence 0.85×$1.35m=$1.15m remains after deducting selling expenses. R&D costs (including the costs of obtaining regulatory approval to sell the generic) represent a further 15% of sales hence 0.85×$1.15m=$0.98m remains. Hence net revenues are about $980,000 for each year in which they control the generic market. The anticipated flow of net revenues over time, accounting for the usual patterns of IG competition and gradual obsolescence of the drug market, is sufficiently lucrative for the firm to challenge a drug market. However, if an AG does enter, the executive I interviewed stated that he assumes that he will lose about 50% of the market so that net revenues fall to 0.5×$980,000 =$490,000 annually which, the executive claims, is the point at which he would not bother challenging. If he was certain that there would be no AG entry, he would challenge drug markets with $5+ million in annual sales.
Hence, taking the statements of the IG executive at face value, the effect of the threat of AG entry is to increase the minimum market size that justifies a challenge from the IG from $5 to $10 million. Assuming that generic prices would be about 50% of brand prices if an IG did enter, and assuming that no other generics enter markets with less than $10m in annual sales, the threat of AG competition raises consumer costs by about 50% for every market not challenged. Assuming that the average size of the markets that go unchallenged on account of the threat of AG entry is $7.5m, the threat of AG entry increases drug payer costs by $3.75m annually (over the remaining life of the drug) for every market that goes unchallenged. The IG executive stated that AG competition discourages them from challenging brand drugs with sales well above $10m for which developing and producing a generic drug is particularly costly, for two reasons.16 First, as before, the threat of AG would take away IG market share and hence revenues. Second, AG competition in markets the IG has entered has reduced their retained earnings. This reduction in retained earnings has hampered their ability to finance development of specialized production capacity, and the cost of clinical studies. (According to the IG executive, many drugs are not amenable to bio-equivalence studies and require comparative clinical trials that can cost many millions or even tens of millions of dollars per study.) The CGPA shared with me a survey of eight participating generic firms conducted by IBM Global Business Services that, at face value, corroborates the claims of the IG executive. (The study was based on firms’ responses that were not audited or otherwise validated by IBM.) The IBM study concludes: “Canadian generic pharmaceutical manufacturers continue to experience a financial loss on their sales in Canada. The net loss on Canadian sales was -2.7% in 2002 and the net loss for 2005 remained at -2.7%” In order to validate these claims, one would need to analyse data on the minimum market size below which no generics will enter. According to the executive at the IG firm, the decision by an IG firm to challenge a brand drug is typically based on the sales of the brand drug in the 10th year that it has been on the market. To verify the claims by the IG executive, one would need to construct a time series of the year 10 sales of brand drugs and determine, for each of these drugs, whether an IG was observed to eventually enter. Further deterrent effects of AGs on IG entry
Finally, the IG executive claimed that brand firms have attempted to use the threat of AG entry to negotiate agreements with their IG firm to delay entry until entry by another IG firm is imminent, at which time their IG firm will launch an AG. The proposed agreements involve
16 The IG executive cited as examples the fol owing brand drugs (with annual sales) which wil likely not be challenged: Eprex $200 m, Remicade $170 m, Enbrel $160 m, Herceptin $160 m,
Flovent HFA $130 m, Advair $125 m, Duragesic $115 m, Aranesp $115 m, Rituxin $110 m, Rebif $ 85 m, Neupogen $ 80 m, Prevnar $ 75 m, Symbiocort $ 60 m, Spiriva $ 60 m, Advair MDI $ 60 m,
Copaxone $ 60 m, Zoladex LA $ 55 m, Betaseron $ 50 m, Lupron Depot $ 50 m. A brand firm
executive countered that he would be very surprised if at least some of these drugs were not challenged by a generic.
sharing with the IG firm some of the consumers’ savings that would have been realized by IG competition.17
5. How important are market opportunities in the U.S. at explaining the decisions of independent generic firms to enter a particular market?
IG firm representatives stated that market opportunities in the U.S. do not affect their decisions to enter domestic drug markets. The reason is that they incur fixed entry costs each time they challenge a brand drug in one of their three primary markets (Canada, US and the EU). These costs include the legal and related costs of challenging outstanding patents, and the costs of obtaining regulatory approval, the requirements of which typically vary by market. The different jurisdictions do not recognize court decisions made in other jurisdictions so that, for example, if an IG earns the right to sell a generic drug in the US that does not automatically allow them to sell in Canada. Moreover, regulations concerning the size, shape and color, and chemical properties of both the active ingredients and binding agents vary by jurisdiction so that the costs of obtaining regulatory approval must be incurred separately for each market. Discussion
My analyses suggest that when AGs are among the first to enter a market, prices drop by about 12%. This seems plausible: the first generic is regulated to launch at a price no greater than70% of the brand price and this should depress average prices paid for the drug. When AGs control between 1-40% of the generic market, the average price paid per mg of the brand drug increases slightly. It is unclear why this is the case. A brand firm contact did not believe that actual prices increase because both federal regulators (i.e. the PMPRB) and provincial drug plans limit drug price growth. One possibility is that the brand sales that do accrue after substantial entry by IGs are for rarely-used, non-genericized dosage forms and strengths that carry a higher per mg price than the forms/strengths that have been genericized. Irrespective of the impact of AG market share on brand prices, however, I find that AGs are mildly pro-competitive in drug markets that face some generic competition. But IG firm executives state that the threat of AG entry has decreased their incentive to challenge ‘marginal’ drug markets. In particular the threat of AG entry has increased from $5m to $10m 17 One IG executive I interviewed stated: “Just as is now happening systematically in the US, brand companies in Canada are also using the threat to launch an authorized generic to extort deals with generic firms. i.e. they say: "if you proceed with an independent generic, we will
launch an authorized generic and sell at low prices (subsidized by the continuing high priced
sales under the brand label) and you wil make nothing. So let's make a deal (at the expense of consumers) to avoid or delay competition and somehow share the benefits." The deals involve
1. Generic agrees not to proceed with its own product, and to sel the authorized generic, but not until another independent generic is approved or about to be approved
2. Generic agrees not to launch its own product, but instead takes over manufacture of the
product for the brand, and gets paid more for doing so than it would earn if it competed.”
the threshold market size – the value of brand drug sales in the 10th year that it has been on the market, below which the IG firm will not attempt to enter. IG executives also stated that AGs reduce IG earnings in markets which IGs do enter. The attendant loss of retained earnings hampers their ability to challenge brand drugs with sales well above $10m, but which have particularly high entry costs. Finally, the IG executives claimed that brand firms have attempted to use the threat of AG entry to negotiate agreements with their IG firm to delay entry until entry by another IG firm is imminent, at which time their IG firm will launch an AG. A comprehensive evaluation of the competitive effects of AGs would need to verify and quantify these costs and compare these to the benefits of AG competition.
References
Anis A. Pharmaceutical prices with insurance coverage and formularies. Canadian Journal of Economics 1992; 25(2):420-437. Anis A, Guh DP, Woolcott J. Lowering generic drug prices: less regulation equals more competition. Medical Care 2003; 41:135-141. Hollis A. How do brands’ ‘own generics’ affect pharmaceutical prices? Review of Industrial Organization 2005; 27:329–350. Hollis A. The anti-competitive effects of brand-controlled ‘pseudo-generics’ in the Canadian pharmaceutical market. Canadian Public Policy 2003; 29(1):21–31. Hollis A. The importance of being first: Evidence from Canadian generic pharmaceuticals. Health Economics 2002; 11:723–734. Hollis A, Liang BA. Assessing the effects of authorized generics on consumer prices. Journal of Biolaw & Business, forthcoming 2006. Kong Y, Seldon J. Pseudo-generic products and barriers to entry in pharmaceutical markets. Review of Industrial Organization 2004; 25:71–86. Skinner B. Generic drugopoly: Why non-patented prescription drugs cost more in Canada than in the United States and Europe. Fraser Institute Public Policy Sources, no. 82, 2004. http://www.fraserinstitute.ca/admin/books/files/GenericDrugopoly.pdf
Appendix 1: Characteristics of the drug markets in which a generic launched between 1998-2004
Appendix 2 Models of log average price in markets with generic entry Models of log average price Model 1.1 Model specification: All drug markets used Indicators of one period lagged AG generic market share Indicators of total number of generic drugs introduced Drug market specific indicators Quarter of year indicator variables dropped Robust standard error estimator used Linear regression Number of obs = 6021 F( 97, 5922) = . Prob > F = . R-squared = 0.9913 Root MSE = .24311 ------------------------------------------------------------------------------ | Robust lap | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- laggshare20 | -.0145032 .0134825 -1.08 0.282 -.0409338 .0119274 laggshare40 | -.0586768 .0113652 -5.16 0.000 -.0809568 -.0363968 laggshare60 | -.0604383 .0123816 -4.88 0.000 -.0847108 -.0361657 laggshare80 | -.0479509 .014568 -3.29 0.001 -.0765094 -.0193924 laggshare100 | -.1086092 .0119177 -9.11 0.000 -.1319722 -.0852462 ldtotgen1 | -.1157372 .0135166 -8.56 0.000 -.1422345 -.0892398 ldtotgen2 | -.1955292 .0128613 -15.20 0.000 -.220742 -.1703164 ldtotgen3 | -.2491113 .0159344 -15.63 0.000 -.2803485 -.2178741 ldtotgen4 | -.2650327 .0121951 -21.73 0.000 -.2889397 -.2411258 ldtotgen5 | -.3045973 .0131185 -23.22 0.000 -.3303143 -.2788804 ldtotgen6 | -.2723459 .0191917 -14.19 0.000 -.3099686 -.2347233 ldtotgen7 | -.3274495 .0146875 -22.29 0.000 -.3562424 -.2986566 ldtotgen8 | -.3444357 .0169227 -20.35 0.000 -.3776103 -.311261 ldtotgen9 | -.4054365 .0179062 -22.64 0.000 -.4405392 -.3703339 ldtotgen10 | -.5595747 .014723 -38.01 0.000 -.5884372 -.5307122 ldtotgen11 | -.5645092 .0149968 -37.64 0.000 -.5939084 -.5351101 drug2 | 3.428896 .0088876 385.81 0.000 3.411473 3.446319 drug3 | -2.8618 .0104877 -272.87 0.000 -2.882359 -2.84124 drug4 | 1.195959 .0904644 13.22 0.000 1.018616 1.373302 drug5 | 4.191443 .0106382 394.00 0.000 4.170588 4.212298 drug6 | -.6420946 .0084044 -76.40 0.000 -.6585702 -.625619 drug7 | 4.755167 .0097976 485.34 0.000 4.73596 4.774373 drug8 | 2.575338 .0230683 111.64 0.000 2.530116 2.620561 drug9 | -2.639401 .0288492 -91.49 0.000 -2.695956 -2.582846 drug10 | -.475864 .0107014 -44.47 0.000 -.4968427 -.4548853 drug11 | -4.09638 .0116516 -351.57 0.000 -4.119222 -4.073539 drug12 | -3.324511 .0101459 -327.67 0.000 -3.3444 -3.304621 drug13 | -4.622968 .021668 -213.35 0.000 -4.665445 -4.580491 drug14 | -3.289993 .0104862 -313.74 0.000 -3.31055 -3.269437 drug15 | -5.053125 .0084384 -598.82 0.000 -5.069667 -5.036582 drug16 | -1.00875 .0104285 -96.73 0.000 -1.029194 -.9883065 drug17 | -3.977562 .0121738 -326.73 0.000 -4.001427 -3.953697 drug18 | -5.18226 .0090506 -572.59 0.000 -5.200002 -5.164517 drug19 | -3.392844 .0106829 -317.60 0.000 -3.413787 -3.371902 drug20 | -2.768377 .045196 -61.25 0.000 -2.856977 -2.679776 drug21 | -4.541323 .0122875 -369.59 0.000 -4.56541 -4.517235 drug22 | -1.417252 .0116229 -121.94 0.000 -1.440037 -1.394467 drug23 | -1.513434 .010407 -145.42 0.000 -1.533835 -1.493032 drug24 | .0904029 .0192183 4.70 0.000 .052728 .1280777 drug25 | -1.257087 .0219797 -57.19 0.000 -1.300175 -1.213999 drug26 | -1.015056 .0311714 -32.56 0.000 -1.076164 -.9539489 drug27 | -1.565534 .0989576 -15.82 0.000 -1.759527 -1.371541 drug28 | -1.946037 .0150807 -129.04 0.000 -1.975601 -1.916473 drug29 | -4.584397 .0140168 -327.06 0.000 -4.611875 -4.556918 drug30 | .5060018 .0094763 53.40 0.000 .4874248 .5245789 drug31 | -3.169093 .0096308 -329.06 0.000 -3.187973 -3.150213 drug32 | -3.514334 .0207674 -169.22 0.000 -3.555045 -3.473622 drug33 | -2.973235 .0105299 -282.36 0.000 -2.993878 -2.952593 drug34 | -4.164655 .008904 -467.73 0.000 -4.18211 -4.1472 drug35 | -4.587265 .0123362 -371.86 0.000 -4.611449 -4.563082 drug36 | -.790091 .0140769 -56.13 0.000 -.8176868 -.7624951 drug37 | -.3180133 .0092303 -34.45 0.000 -.336108 -.2999185 drug38 | 1.876656 .3634283 5.16 0.000 1.164204 2.589108 drug39 | -1.949321 .0287602 -67.78 0.000 -2.005702 -1.892941 drug40 | -.0539281 .0161459 -3.34 0.001 -.0855798 -.0222763 drug41 | -.9879428 .0095992 -102.92 0.000 -1.006761 -.9691248 drug42 | -3.983419 .0098794 -403.20 0.000 -4.002786 -3.964051 drug43 | -3.374512 .0116816 -288.87 0.000 -3.397412 -3.351612 drug44 | -4.066363 .0085229 -477.11 0.000 -4.083071 -4.049655 drug45 | -4.063714 .0133109 -305.29 0.000 -4.089808 -4.037619 drug46 | -.645858 .0129643 -49.82 0.000 -.6712727 -.6204432 drug47 | .2043898 .0173437 11.78 0.000 .1703898 .2383899 drug48 | -4.27634 .0108578 -393.85 0.000 -4.297625 -4.255054 drug49 | 1.067576 .1796458 5.94 0.000 .7154051 1.419748 drug50 | -2.562369 .0118583 -216.08 0.000 -2.585616 -2.539123 drug51 | -1.551582 .0121523 -127.68 0.000 -1.575404 -1.527759 drug52 | 1.086123 .0083284 130.41 0.000 1.069796 1.102449 drug53 | -.4997754 .0099135 -50.41 0.000 -.5192095 -.4803414 drug54 | -.5312374 .0124582 -42.64 0.000 -.55566 -.5068148 drug55 | -2.922354 .0118087 -247.47 0.000 -2.945503 -2.899204 drug56 | 1.37354 .0214483 64.04 0.000 1.331493 1.415586 drug57 | -1.262043 .0103086 -122.43 0.000 -1.282251 -1.241834 drug58 | 2.67766 .0099918 267.99 0.000 2.658072 2.697247 drug59 | 3.719948 .0105767 351.71 0.000 3.699214 3.740682 drug60 | -4.641267 .0132655 -349.88 0.000 -4.667272 -4.615262 drug61 | -3.315048 .0095249 -348.04 0.000 -3.333721 -3.296376 drug62 | -3.339985 .0118625 -281.56 0.000 -3.36324 -3.316731 drug63 | -.4396876 .0081908 -53.68 0.000 -.4557447 -.4236306 drug64 | -4.836603 .0126927 -381.05 0.000 -4.861486 -4.811721 drug65 | 2.316317 .0862683 26.85 0.000 2.1472 2.485435 drug66 | -.7146585 .0137181 -52.10 0.000 -.741551 -.687766 drug67 | 3.314119 .0727712 45.54 0.000 3.171461 3.456777 drug68 | -.7801151 .0111608 -69.90 0.000 -.8019942 -.7582359 drug69 | -3.564098 .0143805 -247.84 0.000 -3.592289 -3.535907 drug70 | -.4325102 .0687193 -6.29 0.000 -.5672252 -.2977953 drug71 | -1.866399 .0097092 -192.23 0.000 -1.885433 -1.847366 drug72 | -.5040115 .0163011 -30.92 0.000 -.5359676 -.4720555 drug73 | 6.605436 .0276082 239.26 0.000 6.551314 6.659558 drug74 | -2.419718 .0124577 -194.23 0.000 -2.44414 -2.395297 drug75 | -2.084197 .009074 -229.69 0.000 -2.101985 -2.066408 drug76 | -3.694298 .0146621 -251.96 0.000 -3.723041 -3.665555 drug77 | .1037097 .0152011 6.82 0.000 .07391 .1335093 drug78 | 3.304044 .018963 174.24 0.000 3.26687 3.341219 drug79 | -3.893479 .0096407 -403.86 0.000 -3.912379 -3.87458 drug80 | -1.210858 .007939 -152.52 0.000 -1.226421 -1.195295 drug81 | -2.370689 .1018381 -23.28 0.000 -2.570329 -2.17105 drug82 | -.2474393 .0126145 -19.62 0.000 -.2721684 -.2227103 drug83 | 2.228096 .1489024 14.96 0.000 1.936193 2.519999 _cons | -1.604222 .0076982 -208.39 0.000 -1.619313 -1.58913 ------------------------------------------------------------------------------ Variable definitions variable name description ----------------------------------------------------- lap log average brand price dag =1 if AG on market ldag =1 if dag[t-1]==1 aggshare20 =1 if 0% < aggshare <= 20% aggshare40 =1 if 20% < aggshare <= 40% aggshare60 =1 if 40% < aggshare <= 60% aggshare80 =1 if 60% < aggshare <= 80% aggshare100 =1 if 80% < aggshare <= 100% laggshare20 =1 if 0% < aggshare[t-1] <= 20% laggshare40 =1 if 20% < aggshare[t-1] <= 40% laggshare60 =1 if 40% < aggshare[t-1] <= 60% laggshare80 =1 if 60% < aggshare[t-1] <= 80% laggshare100 =1 if 80% < aggshare[t-1] <= 100% ldtotgen1 =1 if number of IGs[t-1] = 1 ldtotgen2 =1 if number of IGs[t-1] = 2 ldtotgen3 =1 if number of IGs[t-1] = 3 ldtotgen4 =1 if number of IGs[t-1] = 4 ldtotgen5 =1 if number of IGs[t-1] = 5 ldtotgen6 =1 if number of IGs[t-1] = 6 ldtotgen7 =1 if number of IGs[t-1] = 7 ldtotgen8 =1 if number of IGs[t-1] = 8 ldtotgen9 =1 if number of IGs[t-1] = 9 ldtotgen10 =1 if number of IGs[t-1] = 10 ldtotgen11 =1 if number of IGs[t-1] = 11 drug2 drugid==AMCINONIDE drug3 drugid==AMIODARONE drug4 drugid==AMIODARONE INJ drug5 drugid==BETAXOLOL drug6 drugid==BISOPROLOL drug7 drugid==BRIMONIDINE drug8 drugid==BUTORPHANOL drug9 drugid==CALCITONIN (SALMON) drug10 drugid==CARVEDILOL drug11 drugid==CEFADROXIL drug12 drugid==CEFUROXIME AXETIL drug13 drugid==CEFUROXIME AXETIL LIQUID drug14 drugid==CIPROFLOXACIN drug15 drugid==CIPROFLOXACIN LIQUID drug16 drugid==CITALOPRAM drug17 drugid==CLAVULANIC ACID drug18 drugid==CLAVULANIC ACID LIQUID drug19 drugid==CLINDAMYCIN drug20 drugid==CLINDAMYCIN INJ drug21 drugid==CLINDAMYCIN LIQUID drug22 drugid==CLOBAZAM drug23 drugid==CLOZAPINE drug24 drugid==COLISTIN drug25 drugid==CYCLOSPORINE drug26 drugid==CYCLOSPORINE INJ drug27 drugid==CYCLOSPORINE LIQUID drug28 drugid==DEFEROXAMINE drug29 drugid==DIVALPROEX drug30 drugid==DOXAZOSIN drug31 drugid==ETIDRONIC ACID drug32 drugid==FENOFIBRATE drug33 drugid==FENOFIBRATE MICRO drug34 drugid==FLAVOXATE drug35 drugid==FLOCTAFENINE drug36 drugid==FLUCONAZOLE drug37 drugid==FLUCONAZOLE 150MG drug38 drugid==FLUCONAZOLE INJ drug39 drugid==FLUCONAZOLE LIQUID drug40 drugid==FLUNARIZINE drug41 drugid==FOSINOPRIL drug42 drugid==GABAPENTIN drug43 drugid==GLICLAZIDE drug44 drugid==HYDROXYCHLOROQUINE drug45 drugid==HYDROXYUREA drug46 drugid==KETOROLAC drug47 drugid==KETOROLAC INJ drug48 drugid==LABETALOL drug49 drugid==LABETALOL INJ drug50 drugid==LAMOTRIGINE drug51 drugid==LAMOTRIGINE CHEWABLE drug52 drugid==LEFLUNOMIDE drug53 drugid==LORATADINE drug54 drugid==MELOXICAM drug55 drugid==METHAZOLAMIDE drug56 drugid==MIDAZOLAM drug57 drugid==MIRTAZAPINE drug58 drugid==MISOPROSTOL drug59 drugid==MOMETASONE drug60 drugid==NABUMETONE drug61 drugid==NEFAZODONE drug62 drugid==NORFLOXACIN drug63 drugid==OMEPRAZOLE drug64 drugid==OXAPROZIN drug65 drugid==PAMIDRONIC ACID drug66 drugid==PAROXETINE drug67 drugid==PIPERACILLIN drug68 drugid==PRAVASTATIN drug69 drugid==PROPAFENONE drug70 drugid==PROPOFOL drug71 drugid==SERTRALINE drug72 drugid==SIMVASTATIN drug73 drugid==SUFENTANIL drug74 drugid==TERBINAFINE drug75 drugid==TERCONAZOLE drug76 drugid==TICLOPIDINE drug77 drugid==TIZANIDINE drug78 drugid==TRIFLURIDINE drug79 drugid==TRIMEBUTINE drug80 drugid==VANCOMYCIN drug81 drugid==VANCOMYCIN INJ drug82 drugid==WARFARIN drug83 drugid==WARFARIN INJ
Doubled Haploids in Plant Breeding What is a doubled haploid plant? Cells are the building blocks of plants. A doubled haploid plant has cells containing 2 gene sets which are exactly identical. Each cell contains 2 sets of genetic information which are almost (but not exactly) identical. For example, one gene set may carry a gene for disease resistance when Ho
Green-top Guideline LONG-TERM CONSEQUENCES OF POLYCYSTIC OVARY SYNDROME This is the second edition of this guideline, which was previously published in May 2003 under the same title. Purpose and scope This guideline has been produced to provide information, based on clinical evidence, to assist clinicians witha special interest and for updating the generalist who manages women with pol