Points of View Innovativeness of New Drugs and Price Premium 1) Pilot study of a new drug for which no similar drug exists

Printable PDF

Junichi Nishimura, Professor, Gakushuin University, Visiting Fellow, Pharmaceuticals and Industrial Policy Research Institute
Sadao Nagaoka, Professor, Tokyo Keizai University, Director, Pharmaceuticals and Industrial Policy Research Institute

1. Introduction

⁽ Policy Research Institute News No. 622 ^{) and} No. 643 ⁾ conducted an international comparative analysis of the relationship between innovativeness and the price premium (the degree to which the price of a new drug is higher than that of a similar drug when evaluated by the price ratio of the new drug and the similar drug) for new drugs that were subject to NHI price calculation under the Japanese comparable drug pricing system, using price data for the similar drug as a control group. An international comparative analysis was conducted. As a result, it was confirmed that in Japan, the U.S., and Germany, new drugs with high innovativeness were set at a higher price premium, and the degree of this premium was highest in the U.S. However, when a similar drug is not available, the price premium is higher in the U.S. than in Germany.

However, how the price of a new drug is set in each country in relation to its innovativeness in the absence of a comparable drug remains an important research question. Theoretically, even in the absence of a similar drug, if we know the extent to which the use of a new drug improves health outcomes (e.g., as measured by Quality Adjusted Life Years (QALYs)) compared to the absence of a similar drug (or the use of some alternative treatment), then there should be If we know how much the health benefits (measured, for example, in Quality Adjusted Life Years (QALYs)) can be improved compared to the cost of using an alternative treatment (or some alternative treatment), it is possible to get an indication for a price that reflects the innovation. In Europe and the U.S., it is thought that in many countries, drug prices for new drugs are calculated based on such a cost-effectiveness approach. On the other hand, in Japan, the cost accounting method has been used in the calculation of drug prices for new drugs when similar drugs are not assumed. However, there is no direct causal relationship between the cost of a drug and its health benefits, and as a result, the cost accounting method may not adequately reflect innovation in Japan compared to Europe and the United States.

In this paper, we conduct an international comparative analysis of the extent to which the prices of new drugs calculated using the Japanese cost accounting method reflect innovativeness, using a matched sample of Japanese, U.S., and European drugs that share the same generic ingredient name, dosage form, and specification unit. Although it is difficult to directly assess the impact of innovativeness on prices in each country because similar drugs are not assumed in the NHI price calculation, it is possible to estimate the extent to which the degree to which innovativeness is assessed differs between two countries (e.g., the United States and Japan) under certain ^{assumptions4)}. Using the data used in Policy Research Institute News No. 64, we will apply the same method as in this paper to new drugs subject to the comparable drug effect method, and conduct a comparative analysis on a trial basis to determine how the degree of evaluation of innovativeness differs between Japan, the United States, and Europe for new drugs subject to the cost accounting method.

2. data and analysis methods

First, the details of the data for new drugs and similar drugs that were subject to the comparable drug method are described in Policy Research Institute News No.62 and No.64, and the details are included in Appendix 1. Next, the price data for the new drugs subject to the cost accounting method are constructed in the same way as for the comparable drug method. In other words, the pricing data for drugs launched in Japan, the U.S., Germany, the U.K., France, and Japan between July 2010 and March 2019 is constructed at the quarterly level using "Pricing Insights" licensed from IQVIA. ⁵⁾ In the case of the cost accounting method, the data are used to calculate the price of the new drug in the same way as in the case of the similar drug effect comparison method. In the case of the cost accounting method, information on similar drugs cannot be obtained from the drug price calculation information of the Chuikyo, so for the new drugs concerned in this paper, existing drugs belonging to the same ATC drug classification (4-digit classification) were used as a control group, and price data for these existing drugs were also ^collected6). The new drugs and existing drugs were selected to have the same generic ingredients, dosage forms, and specification units in common in Japan, the U.S., and Europe. The 36 new drugs were selected for inclusion in the costing system, but only 19 of them could be matched with existing drugs to enable international comparisons.

For new drugs subject to the cost accounting method, the data obtained in the above procedure is used to conduct an international comparative analysis using the price premiums of new drugs compared to existing drugs in the same therapeutic area to determine how they reflect innovativeness. The price premium is calculated from the following equation (1).

Price premium = ln (price of new drug/price of existing drug or similar drug) (1)

However, because comparable drugs are not defined for new drugs calculated by the cost accounting method, there is no standard level of treatment (daily dosage) common to new drugs and existing drugs (in the case of the comparable drug method, this could be calculated from the daily drug price of the Chuikyo). Therefore, the price premium expressed as a ratio of the price of a new drug to that of an existing drug may reflect differences in dosage or intensity between the new drug and the existing drug, rather than differences in their health benefits. To eliminate this measurement bias, the estimation uses the U.S. price premium as the basis for the analysis of the difference between the U.S. price premium and the Japanese and European price premiums.

Price premium difference = U.S. price premium - Japanese and European price premiums (2)

Appendix 2 provides the theoretical background for the above analysis. By taking such a price premium differential, measurement bias based on differences in dosage, intensity, etc. between new drugs and existing drugs can be removed.

In this paper, we make two estimates. The first is a model that estimates the extent to which the price premiums in Japan and Europe are below (or above) the U.S. price premium. Although both new drug prices and existing drug prices are higher in the U.S. than in Japan and Europe, it is not known whether the price premium, which is the price ratio of a new drug to an existing drug, is higher in the U.S. than in Japan and Europe. For new drugs subject to the comparable drug pricing system in Japan, we reported in Policy Research Institute News No. 62 that the price premiums in Japan and Europe are significantly lower than those in the U.S. (z in the estimation model below is positive). In this paper, the following estimation model is applied to new drugs subject to the comparable drug price comparison method and the cost accounting method, respectively, to obtain the average value z of the difference between the price premium in each method and the U.S. price premium in each method.

Price premium difference = z + control for change over time + error term (3)

where z is, according to the theoretical model in Corollary 2, the difference between the extent to which the health benefits (degree of innovation) of a drug are reflected in its price in the U.S. and the extent to which such benefits are reflected in its price in Japan and Europe ( ^{β*-β in} Corollary 2) and the ratio of the health benefits of a new drug and an existing drug (or similar drug) (ln( _qi/qo,i )) consists of the ratio of the health effects of the new drug and the existing drug (or similar drugs). The latter ratio of health benefits of new drugs and existing drugs is considered to be common in Japan, the U.S., and Europe. Therefore, z is expected to mainly reflect differences in the extent to which innovativeness is reflected in prices between Japan, the U.S. (or Europe and the U.S.). It should be noted, however, that the ratio of the health benefits of new drugs and existing drugs in the comparable drugs method and the ratio of the health benefits of new drugs and existing drugs in the cost accounting method are not necessarily equal, so the estimated value of z in the comparable drugs method and z in the cost accounting method cannot be simply compared.

In the estimation equation (3), as a control factor, we also included year dummies for the elapsed time since the launch of a new drug in the U.S. and Japan/Europe and for the period used in the estimation. Since the prices of new drugs and existing drugs (or similar drugs) change over time after the launch of a new drug, the inclusion of control factors in the model allows us to account for this effect. In this case, the estimated value is the difference in price premiums in the initial year.

Next, we estimate the relationship between the innovativeness of a new drug and its price premium. The innovativeness of a new drug is constructed as a synthetic index by weighted average of novelty of mechanism of action and science intensity (number of cited papers, number of citations of these papers, and speed of science utilization) of the underlying patents (substance patents, crystal patents, and use patents), as analyzed in Policy Research Institute News No. 64. Their detailed construction procedures are summarized in Appendix 1. In this paper, we also measure the innovativeness of similar drugs or existing drugs using the same procedure as for new drugs. In the costing method, the technological basis for drug discovery may differ significantly between new drugs and the existing drugs we selected, and since we used data from the patent protection period of both new drugs and existing drugs, we assumed that the difference in innovativeness would also affect the price ratio of new drugs and existing drugs. Therefore, in line with the explained variable, the price premium (the logarithm of the ratio of the prices of new drugs to existing drugs), the explanatory variable, innovativeness, was also incorporated into the estimation model by taking the logarithm of the ratio of the innovativeness of new drugs to existing drugs (Equation (1.6) in Appendix 2 also takes the ratio of health benefits of new drugs to existing drugs). In order to maintain comparability, the innovativeness variable is constructed as the logarithm of the ratio of innovativeness of the new drug to that of a similar drug in the same way.

In equation (5) below, γ corresponds to the difference between the degree to which the price reflects the degree of innovativeness of a drug in the U.S. and the degree to which that degree is reflected in the price in Japan and Europe ( ^{β*-β in} Appendix 2).

Innovativeness differential = ln (innovativeness of new drug / innovativeness of existing drug or similar drug) (4)

Price premium differential = constant term + γ innovativeness differential + control for change over time + error term (5)

In the estimation of equations (3) and (5) above, as mentioned above, the estimation is limited to the price data of new drugs and existing drugs or similar drugs within the patent protection period in both Japan, the U.S. and Europe for a common component in Japan, the U.S. and Europe. This is to exclude from the analysis the effect of the entry of generic drugs after the expiration of patent ^protection7). As a result of this work, the actual number of components in the sample subject to the comparable drug method is approximately 30, while the number of components subject to the cost accounting method is approximately 10.

3. results of estimating the difference in price premiums from the U.S.

（The estimation results of equation (3) are summarized in Tables 1 and 2. Table 1 shows the estimation results of z for the new drugs that were subject to the similar drug effect comparison method. First, looking at the results without including the control factor, the estimated value of z is positive in Japan, Germany, and the U.K. The t-value is very large and can be considered statistically significant. This reflects the extent to which the price premiums in Japan and Europe are below the U.S. price premium, and the large coefficient value indicates that the price premium in Japan is the largest below the U.S. price premium. Since this result does not include the control factors such as year since launch and year dummy, it means that we are looking at the average difference over the estimation period.

Next, when we look at the estimation results including the control factor, the estimated value of z is still positive in Japan, Germany, and the U.K. However, it is not statistically significant. However, only in Japan is the result statistically significant. Since the results with the control factor reflect the initial value of the price premium difference (at the time of market launch), this means that there is no significant difference in the price premium difference between Germany and the U.K. and the U.S. at the time of market launch. On the other hand, in Japan, the price premium was lower than the U.S. price premium from the time of market launch. These results indicate that the price premium in the U.S. tends to increase over time compared to Japan and Europe.

Next, Table 2 shows the z results for the new drugs subject to the cost accounting method. The results, which do not include control factors, show positive coefficient values for Japan and the U.K., and negative coefficient values for Germany. However, only in Japan and the U.K. did the results reach statistical significance. Therefore, it can be said that the average difference between Japan and the U.K. over the estimation period is less than the U.S. price premium. On the other hand, in the case of Germany, there is no significant difference from the U.S. price premium. In the model with control factors over time, the estimated value of z is positive for Japan and negative for Germany and the United Kingdom. However, these t-values are low and not statistically significant. Thus, when looking at the time of market launch, there is no difference in the price premiums in Japan, the U.S., and Europe. As in Table 1, it can be said that the price premium in the U.S. tends to increase over time compared to Japan and Europe.

Since the estimates of z include the ratio of the health benefits of new drugs to existing drugs (or similar drugs), the estimates of z in Table 1 and Table 2 cannot be directly compared. Overall, however, it can be seen that the estimated value of z for the costing method tends to be somewhat lower than the estimated value of z for the comparable drug effect comparison method. This may reflect the tendency that new drugs subject to the cost accounting method are more likely to be evaluated for their innovativeness in Japan and Europe than in the U.S. In Germany, this tendency is more pronounced, with a non-significant negative value of z in the cost accounting method results compared to the similar drug effect comparison method results. This is a major difference in the drug price structure between the similar drug effect comparison method and the cost accounting method. We estimate whether the degree to which innovativeness is evaluated differs between new drugs subject to the similar drug effect comparison method and new drugs subject to the cost accounting method, based on equation (5).

4. estimation results of innovativeness and price premium

（The estimation results of equation (5) are summarized in Tables 3 and ⁴⁸⁾. Table 3 shows the results for new drugs subject to the similar drug effect comparison method. Models (1) and (2) are the results of the price premium difference between Japan and the U.S., (3) and (4) are the results of the price premium difference between the U.S. and Germany, and (5) and (6) are the results of the price premium difference between the U.S. and U.K. Looking at the estimation results including the control factor of the elapsed time since market launch, the coefficient value of the innovativeness difference is positive and significant in both models (1), (3), and (5). As mentioned earlier, the coefficient value of the innovativeness difference in the model that takes into account the elapsed time after the market launch is considered to reflect the effect at the time of the market launch in principle. Comparing the results for Japan, Germany, and the United Kingdom, the coefficient value for Japan is the largest at 0.227, followed by Germany at 0.121, and the United Kingdom at 0.061. This coefficient value is the difference between the degree to which the degree of innovativeness is reflected in prices in the U.S. and the degree to which this degree is reflected in prices in Japan and Europe ( ^β*-β in Appendix 2). Thus, the coefficient indicates that the degree to which innovativeness is reflected in prices is the lowest in Japan.

Next, looking at models (2), (4), and (6), which exclude the effect of the year since market launch, the coefficient value of the innovativeness difference is positive and strongly significant in all cases. Since the effect of the post-launch year is not taken into account, these coefficient values indicate the average effect of the innovativeness gap over the estimation period. These coefficient values are all larger than the coefficient values of the model including the post-launch year. Thus, they reflect a relative increase in the price premium in the U.S. over time relative to that in Japan and Europe. This result is consistent with Table ¹⁹⁾. The above estimated results also indicate that the greater degree to which the innovativeness of new drugs is evaluated in the U.S. is an important cause of the larger price premium for new drugs in the U.S.

Next, Table 4 shows the results for new drugs subject to the costing method. Looking at the estimation results including the control factor, elapsed time since launch, the innovativeness differential in model (1) is positive and significant, while it is negative and significant in model (3) and negative and non-significant in model (5). Thus, as in Table 3, the lowest degree to which innovativeness is reflected in prices is common for Japan, but rather more innovativeness is reflected in prices in Germany than in the U.S., and in the U.K. there is no difference in the degree to which innovativeness is reflected in prices between Germany and the U.S. These results are the result of the price premium difference at the time of market launch, which differs significantly from the results of the similar drug effect comparison method in Table 3.

Looking at models (2), (4), and (6), which exclude the elapsed year after market launch, model (2) is positive and significant, model (4) is negative but non-significant, and model (6) is positive and significant. As in Table 3, these coefficient values are all larger than the coefficient values of the models that include the elapsed year after the market launch. Thus, the costing method also reflects that the price premium in the U.S. increases relative to that in Japan and Europe over time. For Japan, the estimated coefficient values in Models (1) and (2) are almost equal to those in Table 3, which means that the degree to which innovativeness is reflected in prices is about the same in Japan as in the U.S., regardless of whether the drug is subject to the comparable drug method or the cost accounting method. On the other hand, the coefficient values for Germany and the U.K. are significantly lower than those estimated in Table 3, suggesting that in Europe, innovativeness is highly reflected in the prices of new drugs that are subject to the cost accounting method in Japan. These results are consistent with those in Table 2.

5. summary of analysis results

In this paper, we analyzed the relationship between price premiums and innovativeness by constructing matched data for Japan, the U.S., and Europe for new drugs subject to the comparable drug method and for new drugs subject to the cost accounting method for which a comparable drug is not assumed. The results of the estimation revealed the following. First, the similar drug effect comparison method revealed that the price premiums in Japan and Europe are lower than those in the U.S., and the extent of the price premium increases over time. Furthermore, the greater degree to which the innovativeness of new drugs is evaluated in the U.S. was an important cause of the larger price premium for new drugs in the U.S. ¹⁰⁾. In particular, we also found that the degree to which the innovativeness of new drugs is evaluated is lower in Japan than in the U.S. and Europe, while it is about the same in Germany and the U.K.

Next, analysis of the costing method showed that the price premiums in Japan and the U.K. were lower than those in the U.S., but this was not as pronounced as in the comparable drug method, and such a trend was not observed in Germany. Furthermore, the degree to which the innovativeness of new drugs was evaluated was lower in Japan than in the U.S. and similar to that in the comparable drug price method. However, in Germany and the U.K., there was no statistically significant difference in the degree of innovation compared to the U.S. In fact, at the time of market launch, Germany rated innovativeness higher than the U.S.

The above results indicate that the degree to which innovativeness is reflected in price is lower in Japan than in the U.S., regardless of whether the similar drug effect comparison method or cost accounting method is used, while a significant difference is observed in Europe. One of the motivations for this paper was to point out the possibility that the innovativeness of new drugs subject to the cost accounting method, in which cost is the basis for NHI prices in Japan, may not be easily reflected in the price of such drugs in Japan. Such new drugs may be inherently highly innovative, considering that similar drugs are not expected. In Germany and the U.K., such innovativeness may have been a result of the evaluation.

Although the number of drugs covered by the cost accounting method is quite limited in this analysis, and the field of drug efficacy is quite narrow, it is still necessary to verify whether the selection of existing drugs is appropriate for evaluating the innovativeness of a new drug. One possible reason for the large difference in the results of the costing method in Europe is that in the absence of similar drugs, the resulting price in the U.S. has a large impact, and we would like to further analyze this issue, including its mechanism.

Appendix 1: Procedure for Constructing Basic Data for New Drugs in the Comparison of Similar Drugs Method

Construction of price data

In this paper, we constructed an analytical dataset using various data ^sources11). The construction procedure is as follows. First, "Pricing Insights," which we licensed from IQVIA, contains monthly price information on marketed drugs from July 2010 to March 2019 for five countries: Japan, the U.S., Germany, the U.K., France, and Germany (although price information for Japan is missing for 2010). Pricing Insights contains price information at each stage of distribution, but this analysis uses pharmacy purchase prices (PPP), which are assumed to be the closest to actual prices in each country.

From this basic database, we created a database by matching the price information of each country for each drug unit that shares the same generic ingredient name, international trade name, dosage form, and strength, and then simply averaged the price data on a quarterly basis for each country. We obtained information on GDP deflators and purchasing power parity as of 2015 for each country from the OECD, realized the data at 2015 prices, and converted the data to dollar-based prices using purchasing power parity.

Next, a list of similar drugs corresponding to the top 300 drugs identified from the IQVIA World Review Analyst 2018 list of top-selling drugs and the NHI drug price calculation information from the Japanese Central Medical Association was created and connected to the database constructed from Pricing Insights. The list of similar drugs was obtained only for those drugs that were subject to the comparable drug pricing system.

Finally, we constructed a list of 69 new drug components that were launched in Japan, for which a new drug price was determined by the comparable drug price method, and that were also launched in the United States or one of three European countries (Germany, the United Kingdom, or France). The drugs in these analyses must have a common generic name, international trade name, dosage form, and strength in each country.

In order to calculate price ratios for new drugs and similar drugs at the standard level of treatment (daily dosage), information on the daily drug price of each drug was obtained from the NHI drug price calculation data of the Central Medical Council to determine the price premium (logarithmic value of price ratio, the rate of price increase of new drugs relative to similar drugs) for new drugs relative to similar drugs in Japan, the United States, and Europe. Therefore, the analysis was based on the assumption that this standard level of treatment is the same in each country.

Construction of innovativeness data

Four new drug innovativeness indicators were developed. The first indicator was developed based on the Chuikyo's "Drug Classification for Selection of Similar Drugs".

Next, a group of patents (substance patents, crystal patents, and use patents) protecting each pharmaceutical ingredient was identified from the San-Ei Report, and a group of scientific papers that were cited in the bibliographic information (front page) of the corresponding U.S. patent group was collected from Web of Science to create the following innovativeness indexes. Since citation information inherently has truncation problems and the propensity of pharmaceutical patents to cite scientific and technical papers changes over time, the year when the new drug was launched was controlled as a fixed effect when the following indicators were used for estimation.

Appendix 2 Background of the Price Premium Model

Assume that the health effects of the new drug i and the corresponding existing drug are measured by the same criterion, qi _andq0,i respectively. These can be thought of as the degree of improvement in QALYs relative to the absence of the drug. Suppose that the price of an existing drug is proportional to the country's willingness to pay for health improvement, w, and also reflects the disease-specific circumstances (e.g., demand for treatment), _αi, and the effect of the drug, as follows.

where β ≥ 0. If β = 1 and _αi = 1, then the price of an existing drug is set proportionally equal to its health benefit, and if β = 0, the price does not reflect its health benefit.

New drugs use the same formula (equation) as existing drugs and at the same time are determined to reflect their high effectiveness, but if a new drug requires _{only θi} prescription volume to achieve _qi effect, the real price is _θipi and can be written as follows.

(1.1) and (1.2), taking the price ratio in equations (1.1) and (1.2),

Suppose that the drug price is determined in the same way in the reference country (the U.S.),

To solve the measurement bias caused by the fact that information on the daily drug price is not available due to the unavailability of similar drugs, as is the case for new drugs subject to the cost accounting method, a further ratio of the price ratio to that of the U.S. can be used. That is, taking the ratio of equations (1.3) and (1.4),

and the effect of _θi can be eliminated.
（Taking the logarithm on both sides of equation (1.5),

However, price premium = ln (price of new drug) - ln (price of existing drug)

（The price premium difference in equation (1.6) can eliminate the effect of θ even if the price of a new drug is θ times the daily drug price of an existing drug in each country (caused by the difference in prescription volume, etc.).

Return to News List

TOP