Innovation Incentives Part 2: Patent Valuation
Understanding the Consequences of Linking Market and Regulatory Incentives for Drug Development: Part 2
Editor's note: This is the second installment of a three-part series.
In new work by our group, we have outlined a tandem of new methodological tools to identify and quantify new and follow-on drugs and patent valuation. The first is a harmonized method to quantify drug approvals, patents and associated chemical components that summarizes and extends our previous work on topic. The second provides a new “innovation index” that incrementally grades the value, not only for patents in the life sciences and other technology-intensive sectors, but also for associated regulatory approvals, chemical components, patent characteristics, etc. The innovation index values are based on evidentiary hurdles and prioritizations for several classes of “new” and “follow-on” drugs disclosed by drug regulators. As indicated by the titles of the articles, one focuses on the quantitative side while the other focuses on the qualitative side of the analysis.
The Boston Article presents a harmonized method to collect, compare, and quantify regulatory approval data from multiple cohorts of new and follow-on drugs. We looked in some detail at about 2,000 regulatory approvals, 5,000 patents, and 130 chemical components. The analysis encompasses all drug classes enumerated, described and prioritized by domestic drug regulators. The drug classes were gleaned from the usual literature reviews, supplemented by several hours of consultation with Health Canada regulators and review of Health Canada Guidance Documents on topic. A second purpose of this work was to go beyond simplified descriptors of new and follow-on drugs found in the literature, to categorize classes of new, line extension and generic approvals according to the nomenclature used by regulators themselves. This latter point is relevant is relevant, as we found different scholars use different approaches and nomenclatures, sometimes very different, and that these approaches were not always the same as those used by regulators themselves.
Read the rest of this post . . . .The innovation index work described in the companion Santa Clara Article was driven by the fact that almost all published patent assessment methods measure innovation using primarily quantitative methods, otherwise referred to as ‘counting methods.’ For reasons discussed in work on topic by Kingston at Trinity, Lemley at Stanford and Polk and Parchomovsky at Penn, and the sources cited therein, while quantitative models are widely considered to be problematic, a model that assesses patent value using qualitative methods that track, or are at least designed to track social benefits, has not yet emerged. A second reason for developing the two methods is that is that even when many scholars and commentators do look at the “innovative” aspect of the data, they simply accept data provided by regulators in their respective annual reports in a per se manner.
While developing a novel scientific method for either obtaining or analyzing legal data is fraught with its own problems, this step nevertheless forms a necessary component of the “trial and error” heuristic typical in the hard sciences. As more individuals with prior experience in medical science enter law and legal scholarship, we will undoubtedly see more and more scientific studies of law, including importing of fundamental mathematical, statistical, curve fitting, modeling, and graphing methods. In the Santa Clara paper, a qualitative innovation index is reported that we hope may fill some of the gaps in patent valuation. One of the figures from this work, relating to regulatory approvals, is shown below.
The strengths and weaknesses of the hybrid “subjective-objective” nature of data transformation, and the similarities to subjective-objective hybrid models that are already widely accepted for use in the fields of drug approval, patent grant, and the adjudication of patent claims by the courts are discussed more fully there.
Data can, of course, be fit to many types of numerical functions, linear or non-linear; increasing or decreasing. Fig. 1c above shows that the data in the bar graph of Fig. 1b fit to a declining exponential function. As can be seen by the close fit of the data to the function, the choice of an exponential relationship was well founded. The data are interesting as they demonstrate an exponential decline in the numbers of drugs in classes with relatively high innovation index values. In other words, the vast majority of drugs approved in Canada have a very low index value, and indeed are primarily follow-on Me Too drugs.
Fig. 1d represents the normalized cumulative data fit to a sigmoid (S-shaped log) function, which is a numerical approximation of “how fast” the innovation index data rise to their maximal peak. A fast rise, as we see here, suggests that most of the drugs approved over nearly a decade are in the low index bins and that the data in the low index bins accumulate much more rapidly than do the data in the higher index bins. Similar, though not identical, results were obtained with several indicator Cohorts studied, including a wide Cohort of 2,087 drugs, a narrower Cohort of 95 of the most profitable drugs, and a similar Cohort of associated patents and chemical components.
The innovation index provides a means of weighing legitimate patent protection against perceived societal benefit. As such, it affords a qualitative measure of the innovative nature of drug patents that, when compared to counting methods, may more adequately reveal the outcome of development incentives for firms and regulating bodies insofar as these parties have conflicting interests.
The results from our analysis indicate that it is not the most innovative or even strongly innovative drugs that are attracting the greatest firm patenting effort. Rather, when gauged against development priorities publicly disclosed by regulators and governments, including specifically in the United States and Canada where linkage first came into force, it is the least innovative drugs of all classes investigated that display the strongest regulatory approval and patenting efforts. This issue is touched on in more detail in Part 3 of the series.
In this manner, our data are contrary to the established dogma that the strength of patent protection is proportional to the "strength" of innovation of a given product. As discussed more fully in Part 3, the data obtained also support the conclusion that cluster-based, or portfolio-based, drug development has become the dominant innovation strategy for both brand and generic firms. Indeed, data from our Boston study demonstrates conclusively that generic firms are accruing more patents than their brand counter-parts, especially in the new drug approval category.
Finally, the data suggest that the perception on the part of governments and the public to the effect that societal benefit comes as a kind of “natural consequence” of patenting may need to be reconsidered.
Editor's note: This is the second installment of a three-part series.
In new work by our group, we have outlined a tandem of new methodological tools to identify and quantify new and follow-on drugs and patent valuation. The first is a harmonized method to quantify drug approvals, patents and associated chemical components that summarizes and extends our previous work on topic. The second provides a new “innovation index” that incrementally grades the value, not only for patents in the life sciences and other technology-intensive sectors, but also for associated regulatory approvals, chemical components, patent characteristics, etc. The innovation index values are based on evidentiary hurdles and prioritizations for several classes of “new” and “follow-on” drugs disclosed by drug regulators. As indicated by the titles of the articles, one focuses on the quantitative side while the other focuses on the qualitative side of the analysis.
The Boston Article presents a harmonized method to collect, compare, and quantify regulatory approval data from multiple cohorts of new and follow-on drugs. We looked in some detail at about 2,000 regulatory approvals, 5,000 patents, and 130 chemical components. The analysis encompasses all drug classes enumerated, described and prioritized by domestic drug regulators. The drug classes were gleaned from the usual literature reviews, supplemented by several hours of consultation with Health Canada regulators and review of Health Canada Guidance Documents on topic. A second purpose of this work was to go beyond simplified descriptors of new and follow-on drugs found in the literature, to categorize classes of new, line extension and generic approvals according to the nomenclature used by regulators themselves. This latter point is relevant is relevant, as we found different scholars use different approaches and nomenclatures, sometimes very different, and that these approaches were not always the same as those used by regulators themselves.
Read the rest of this post . . . .The innovation index work described in the companion Santa Clara Article was driven by the fact that almost all published patent assessment methods measure innovation using primarily quantitative methods, otherwise referred to as ‘counting methods.’ For reasons discussed in work on topic by Kingston at Trinity, Lemley at Stanford and Polk and Parchomovsky at Penn, and the sources cited therein, while quantitative models are widely considered to be problematic, a model that assesses patent value using qualitative methods that track, or are at least designed to track social benefits, has not yet emerged. A second reason for developing the two methods is that is that even when many scholars and commentators do look at the “innovative” aspect of the data, they simply accept data provided by regulators in their respective annual reports in a per se manner.
While developing a novel scientific method for either obtaining or analyzing legal data is fraught with its own problems, this step nevertheless forms a necessary component of the “trial and error” heuristic typical in the hard sciences. As more individuals with prior experience in medical science enter law and legal scholarship, we will undoubtedly see more and more scientific studies of law, including importing of fundamental mathematical, statistical, curve fitting, modeling, and graphing methods. In the Santa Clara paper, a qualitative innovation index is reported that we hope may fill some of the gaps in patent valuation. One of the figures from this work, relating to regulatory approvals, is shown below.
Fig. 1. Innovation Index Data for Total Approval Cohort. Bar graphs showing the number of total approvals expressed as a function of the level of innovation (LOI) before (a) and after (b) of generic approval data. c Brand approvals expressed as a function of LOI. Solid line is a fit of the data to a single exponential function. d Cumulative normalized brand approvals expressed as a function of LOI. Solid line is fit using a sigmoidal function.The figure presents data for many classes of new and follow-on drugs and categorizes these classes using a linear scheme. Raw data values are given in the Y axis of Fig. 1a and 1b, the difference being generic data were subtracted in Fig. 1b to isolate data only from ‘innovator’ firms. The X axis in both panels represents the innovation index data. The innovation index data are referred to as transformed data, because the raw data pertaining to drug approvals, drug patents, and chemical components are transformed into qualitative values on a linear scale (0-15) using the methods outlined in the Santa Clara paper.
The strengths and weaknesses of the hybrid “subjective-objective” nature of data transformation, and the similarities to subjective-objective hybrid models that are already widely accepted for use in the fields of drug approval, patent grant, and the adjudication of patent claims by the courts are discussed more fully there.
Data can, of course, be fit to many types of numerical functions, linear or non-linear; increasing or decreasing. Fig. 1c above shows that the data in the bar graph of Fig. 1b fit to a declining exponential function. As can be seen by the close fit of the data to the function, the choice of an exponential relationship was well founded. The data are interesting as they demonstrate an exponential decline in the numbers of drugs in classes with relatively high innovation index values. In other words, the vast majority of drugs approved in Canada have a very low index value, and indeed are primarily follow-on Me Too drugs.
Fig. 1d represents the normalized cumulative data fit to a sigmoid (S-shaped log) function, which is a numerical approximation of “how fast” the innovation index data rise to their maximal peak. A fast rise, as we see here, suggests that most of the drugs approved over nearly a decade are in the low index bins and that the data in the low index bins accumulate much more rapidly than do the data in the higher index bins. Similar, though not identical, results were obtained with several indicator Cohorts studied, including a wide Cohort of 2,087 drugs, a narrower Cohort of 95 of the most profitable drugs, and a similar Cohort of associated patents and chemical components.
The innovation index provides a means of weighing legitimate patent protection against perceived societal benefit. As such, it affords a qualitative measure of the innovative nature of drug patents that, when compared to counting methods, may more adequately reveal the outcome of development incentives for firms and regulating bodies insofar as these parties have conflicting interests.
The results from our analysis indicate that it is not the most innovative or even strongly innovative drugs that are attracting the greatest firm patenting effort. Rather, when gauged against development priorities publicly disclosed by regulators and governments, including specifically in the United States and Canada where linkage first came into force, it is the least innovative drugs of all classes investigated that display the strongest regulatory approval and patenting efforts. This issue is touched on in more detail in Part 3 of the series.
In this manner, our data are contrary to the established dogma that the strength of patent protection is proportional to the "strength" of innovation of a given product. As discussed more fully in Part 3, the data obtained also support the conclusion that cluster-based, or portfolio-based, drug development has become the dominant innovation strategy for both brand and generic firms. Indeed, data from our Boston study demonstrates conclusively that generic firms are accruing more patents than their brand counter-parts, especially in the new drug approval category.
Finally, the data suggest that the perception on the part of governments and the public to the effect that societal benefit comes as a kind of “natural consequence” of patenting may need to be reconsidered.
0 Comments:
Post a Comment
<< Home