RS
Drug safety and pharmaceutical regulation in the age of global development
2020 年 2 巻 3 号 p. 115-117
詳細
2020 年 2 巻 3 号 p. 115-117
As the drug development landscape becomes more globalized, the trade-off between global and local development pathways has become an increasingly important issue. Trade-offs are observed in a variety of consequences, including the resources and time required for development and the impact on patient health. From a public health perspective, studies suggest that the choice of global development pathways for new drugs by pharmaceutical companies may have a significant impact on drug efficacy and safety in Japan. In particular, the fact that dose-finding or confirmatory Phase III studies in Japan have led to improved post-marketing safety demonstrates the potential value of collecting local evidence. The global vs. local trade-off has been overlooked because we lack the concepts and tools to discuss it rigorously in the current drug evaluation science. We need to build a framework for drug evaluation science for a new generation.
● With the globalization of drug development, the trade-off between global and local drugs is becoming more important but latent in the efficacy and safety of drugs.
● Studies show that local trials may lead to improved post-marketing safety in national markets.
● We need to build a framework for drug evaluation science that enables an explicit discussion of global vs. local trade-offs.
As the drug development landscape becomes more globalized, the trade-off between global and local development pathways has become an increasingly important issue. This is most evident with the so-called “drug lag”. The strategy of coordinating the timing of drug launches in different countries has been devised in order to maximize the profits of pharmaceutical companies but has not been shown to have a public health impact in Japan [1].
Drug lag is likely a consequence of profit-maximizing behavior under uncertainty, in which the probability of success of new drug development is used as a parameter. [2] By optimally coordinating the progressive stages of development around the world, pharmaceutical companies can maximize their expected profits, which are determined by the timing of the product entry into markets around the world and the probability of success. Studies have shown that companies decide when to launch new drugs in each country (i.e., the length of the drug lag) based on this trade-off.
There is no doubt that these behaviors of the pharmaceutical industry have a significant impact on the timing and extent of access to promising new drugs and, as a result, on public health around the world. However, the impact on local public health at the country level does not occur solely through access to new drugs.
Drug development is the process of exploring—and ultimately establishing—the efficacy and safety of a drug in a target population. In other words, the primary objective of drug development is to suggest the optimal use of the drug in the target population. The problem is, however, that there is no single, self-evident target population for pharmaceutical companies to choose from. A global company may consider all patients worldwide with a disease as a potential target population. A company doing business only in Japan might consider only Japanese patients as its target population. Another company might consider Westerners in a large market as its primary target group and Japanese consumers in a smaller market as a lower priority target group.
Optimization of drug use in low-priority target populations, or in patient populations not identified as target populations in the first place, will not be achieved to the same extent as optimization in high-priority target populations because commercial drug development takes place under various resource constraints, which in itself is one of the factors for profit maximization.
In reality, it is difficult to measure with objective indicators the extent to which drug use optimization is achieved. The main reason for this is that there has probably never been a single commercial clinical trial that has rigorously and unambiguously pre-identified a target population and extracted samples from that population using random sampling techniques. Because we have never attempted to achieve optimization even hypothetically, “optimization in the target population” is treated as only a conceptual ideal rather than an objectively measured metric. Another challenge to optimizing drug use is the fact that pharmaceuticals have numerous, complex characteristics that effect both efficacy and safety. Different people will have different opinions on which characteristics of a drug product to measure the degree of optimization of its use.
Differences in the degree of optimization of drug use between Western countries and other countries (such as Japan) that represent smaller markets will inevitably arise when new drugs are developed by profit-maximizing pharmaceutical companies, and these differences cannot be easily detected without careful analysis. To maximize public health in countries such as Japan, we must assess whether the differences in drug optimization are significant. Researchers involved in pharmaceutical regulation and evaluation have an obligation to conduct such evaluations.
Several studies have shown a complex relationship between the global development pathways and the safety of new drugs at the national level in Japan. For example, drugs with longer drug lags, which indicate a pathway of Western precedence, have generally been shown to be safer in the Japanese market (Table 1) [3]. These results show that the experience of using the product in large foreign markets has helped to improve safety in subsequent Japanese markets.
| Characteristics (explanatory variables) | IRRa,b | P value |
|---|---|---|
| Patient population estimate (in millions) | 1.15–1.19 | <0.05 |
| Launch delay (>6 years)c | 0.24–0.26 | <0.05 |
| Similar mode of actionc | 0.20–0.34 | <0.01 |
| Bridging strategyc | 1.72–3.08 | <0.05 |
| Orphan drugc | 1.09–1.53 | NS |
| Sample size (in thousand) | 1.84 | NS |
| Review time (in thousand days) | 0.80–1.12 | NS |
| All case surveillancec | 0.24–0.25 | NS |
Data summarized from Yamada et al [3]. IRR, incident rate ratio; NS, not significant. a) IRRs, which are estimated by the negative binomial distribution model in [3], indicate the likelihood of safety related regulatory actions for drugs with each of the characteristics. b) IRRs are shown in the range of values obtained in the models in [3]. c) These are dummy variables.
A more detailed categorization of the development pathway reveals an interesting relationship between specific strategies for global development and drug safety that has not been identified in previous studies (Table 2) [4]. When development pathways were categorized by whether a dose-finding study and/or a Phase III confirmatory study was conducted in Japan, the results showed that the drugs with Japanese dose-finding and/or Phase III studies had a higher level of safety than those without such local studies. In addition, some of the drugs that were tested in a dose-finding study in Japan and for which the same dose was selected in Japan and the US were found to be relatively more dangerous than other drugs.
| Characteristics (explanatory variables) | IRRa | P value |
|---|---|---|
| Dose ratio (Japanese dose/US dose) | 1.34 | NS |
| Same dose in US and Japanb | 0.47 | <0.05 |
| Same dose and dose finding study in Japanb | 2.18 | <0.1 |
| Dose finding study in Japanb | 0.39 | <0.05 |
| Confirmatory Phase III study in Japanb | 0.62 | <0.1 |
| Bridging study in developmentb | 0.50 | <0.05 |
| Global study in developmentb | 1.09 | NS |
Data summarized from Kawamura et al [4]. IRR, incident rate ratio; NS, not significant. a) IRRs, which are estimated by the negative binomial distribution model in [4], indicate the likelihood of post marketing drug-related deaths for drugs with each of the characteristics. b) These are dummy variables.
These relationships indicate that for new drugs developed in the current global market, there is room to improve their use and labeling by conducting local trials in Japan. The current rules and guidelines for drug development and approval in Japan have accepted the industry-proposed approach to global development. While this seems to be a reasonable strategy for Japan, a country with a smaller market than the US and Europe, in terms of allowing the fastest possible access to drugs developed by global companies, it should also be noted that such a stopgap strategy has side effects.
The science required to establish the safety and efficacy of pharmaceuticals must enable us to approach the more fundamental issues of drug evaluation and to solve the practical problems, including the industrial and geodemographic challenges we have described. Fundamental issues include, for example, the definition of the target population (as discussed in this review) and the semantic interpretation of the concept of “the drug is safe” (or “the drug is effective”), which is inextricably linked to the definition of the target population. The semantic interpretation would require the aid of predicate logic.
There also appears to be some confusion in the current pharmaceutical evaluation in terms of the semantic interpretation of the statistics usually applied. For example, the debate about whether “the drug is safe” should be defined with a hypothetical randomized controlled trial in the population in mind or with a counterfactual hypothetical model of each individual is only just beginning as some easy-to-use causal inference models become popular.
There is an inevitable trade-off between the efficiency associated with the global development of new drugs and the optimization of local usage. We must not turn a blind eye to the potential flaws in the existing body of drug evaluation science, including the various guidelines, but rather face the flaws and make suggestions for improvement. We should not only focus on developing drugs to satisfy the patients of today but also strive to build a foundation of drug evaluation science that we can hand to the next generation.
The authors have nothing to disclose.
