E cdip/15/inf/2 original: english date: january 8, 2015 Committee on Development and Intellectual Property (cdip) Fifteenth Session Geneva, April 20 to 24, 2015

Therapeutic classes

Nearly all ISP registrations include an indicator of the therapeutic class for which the drug is intended. After cleaning and standardizing their names as described in Appendix A, there are 183 distinct classes in 19 broad therapeutic groups. There may be up to 6 of these classes per product, although the majority (two-thirds) of the drugs indicate only a single class. Columns 1 and 3 of Table 4 below show that 19 therapeutic classes account for over half the products between them, with the remainder accounted for by the other 164 classes. The remaining columns (2 and 4) of Table 4 show the number of active ingredients associated with each therapeutic class. The class with the largest number of active ingredients is vitamins, which includes various homeopathic remedies that tend to be mixtures containing a number of ingredients.

Table 4

Table 5 shows the number of primary and secondary patents associated with each therapeutic class.³³ The shares of primary patents vary considerably: recall that product patents were not available in Chile before October 1991. This means that classes like anti-ulcer which had important patents prior to that date are covered only by secondary patents. In contrast, newer areas like anti-virals and anti-neoplastics (anti-cancer) have a large share of primary patents.

Table 5

Activities of Chilean Firms

As we report in the appendix, ISP registrations include the function of the various firms attached to the production and distribution of the relevant product. Chilean firms are rarely the source or licensor for a product, which is consistent with their lack of patenting in pharmaceuticals. The only two drugs for which a Chilean firm is listed as the source are the anitbiotic Meropenem Trihydrate (M Moll Quality Control) and the anticoagulant Warfarin Sodium (Volta Lab), neither of which has associated patents. There are approximately 14 distinct drugs for which a Chilean firm is listed as a licensor: half of these belong to D&M Lab and the remainder to a range of firms including Pfizer Chile. They are largely not patented, as they are predominantly older antibiotics, analgesics, and vitamins.
About 100 Chilean firms are engaged in various kinds of pharmaceutical manufacturing and they are associated with 6,930 ISP registrations. 2,152 of these registrations (31 per cent) are associated with a Chilean patent application and 109 (1.6 per cent) with a primary Chilean patent application. Thus the Chilean firms do not seem to be manufacturing products that are covered by a primary patent, not surprisingly, although they do manufacture products covered only by secondary patents. As we describe in the data appendix, when drugs are registered at the ISP, a number of registrants are supplied, and their functions given. The first four of these functions are those associated with manufacturing of the drug, with 11,718 registrations that include Chilean manufacturing firms and 14,334 registrations that include foreign manufacturing firms. For each of the 381 patent-protected active ingredients, we constructed the share of associated manufacturing firms that were Chilean. This allows investigation of the extent to which Chilean firms are involved with manufacturing patent-protected drugs. Figure 8 shows the distribution of these shares: a large number of active ingredients have no associated Chilean manufacturing firms and a slightly smaller number have only Chilean firm.
Figure 8: Share of Chilean manufacturing firms by active ingredient

In order to investigate the role of Chilean manufacturing further, we regressed the share of Chilean companies on some characteristics of active ingredient patenting: a dummy for whether there was any associated primary patent, the log of the number of associated ISP registrations, and the log of the number of associated patents. The results are shown in Table 6; because of the large numbers of zeros and ones for the share, our preferred estimates are those using the two-limit Tobit model, shown in the second set of columns. They show that active ingredients covered by a primary patent are far less likely to be manufactured by a Chilean firm, which is consistent with the lack of patent ownership. In addition, those active ingredients with a large number of ISP registrations are more likely to be manufactured by Chilean firms, as one might have expected. These active ingredients tend to be generics that have some patent coverage via secondary patents on formulation, delivery type, etc., but the coverage is not really exclusive.

Table 6

Figure 9 combines information on companies that are engaged in pharmaceutical manufacturing, their corresponding ISP registrations (products), and the patents that protect the active ingredients contained by these ISP registrations. The figure distinguishes between active ingredients with at least one primary patent and others. It shows that most active ingredients that are patent-protected are protected only by secondary patents. The figure also shows that Chilean manufacturers account for a substantial fraction of manufacturers of drugs associated with a given active ingredient protected only by secondary patents. This is less the case for products that contain active ingredients that are protected either only by primary patents or by a combination of primary and secondary patents. In both cases, most drugs that contain such active ingredients are exclusively manufactured abroad.
Figure 9: Primary vs secondary patents and products vs share of Chilean manufacturers