ORIGINAL ARTICLE
Effect of a financial incentive scheme on promoting prescription of biosimilars: a interrupted time-series analyses
2026 Volume 8 Issue 2 Pages 54-61
Details
2026 Volume 8 Issue 2 Pages 54-61
BACKGROUND
In Japan, biosimilars have the potential to reduce drug expenditure because their official price is lower than the price of the original products. To promote biosimilars, the Japanese government introduced a new financial incentive scheme for medical institutions to prescribe biosimilars to outpatients in April 2020. However, the impact of the incentive remains unevaluated. Hence, in this study, we conducted an interrupted time-series analysis to evaluate the impact of the incentive scheme on biosimilar prescription.
METHODS
We used the DeSC database in Japan. From this database, we included 3,348 patients who required self-injection and were prescribed insulin, human growth hormone, or etanercept. Interrupted time-series analyses were conducted by fitting Prais–Winsten linear regression models to assess the association of the financial incentive with the outcomes, namely, monthly proportions of biosimilar prescriptions between April 2019 and March 2021.
RESULTS
No significant changes were observed in the monthly proportion of biosimilar prescriptions immediately after the introduction of the incentive. The sustained effect, representing the effect of the intervention over time, was also not significant.
CONCLUSIONS
Our study suggests that the financial incentive introduced in April 2020 in Japan was not associated with an increase in biosimilar prescriptions.
Biopharmaceuticals are manufactured using biotechnologies, such as genetic recombination and cell culture. They are generally expensive because of the cost of specialized manufacturing, which is different from that of other general pharmaceutical products, ranging from hundreds to millions of dollars for a single use. The market size of all pharmaceutical products in 2020 was reported to be approximately 10 trillion yen in Japan, and biopharmaceuticals were reported to account for approximately 22% of the total market1). Hence, the impact of biopharmaceuticals on healthcare finances is considerable in Japan, because it is necessary to continuously use biopharmaceuticals in cases of chronic autoimmune diseases.
In Japan, 114 biopharmaceuticals (such as human insulin, human growth hormone, interferon, erythropoietin, and granulocyte colony-stimulating factor) were launched as new molecular entities between 2013 and 20212). The number of approved biopharmaceuticals, including new molecular entities and biosimilars, was 30 in 2013, and this number increased to 52 in 20213). Biopharmaceuticals accounted for 133 of the top 300 pharmaceutical products sold, making them an important part of the pharmaceutical market4). In Japan, one biosimilar was approved for the first time in 2009, and 40 biosimilars to 18 original products have been approved as of January 20245). Biosimilars have the potential to reduce drug expenditure because the Ministry of Health, Labour and Welfare (MHLW), Japan, have set their official prices at 70% of those of the original products. In such situations, the government aims to replace at least 80% of the components with biosimilars on a volume basis and that at least 60% of the total number of components will be replaced by their biosimilars by the end of fiscal year 20296). As one of the measures to promote biosimilars, the government introduced a new financial incentive scheme for medical institutions to prescribe biosimilars to outpatients in April 2020. If a doctor explains biosimilars to a patient who requires self-injection and prescribes biosimilars, the medical institution receives a specific reimbursement7). The financial incentive scheme was used to account for approximately 5,300 times in 2020, 10,000 times in 2021, and 11,000 times in 20228). However, the effectiveness of this incentive in increasing the proportion of biosimilar prescriptions remains to be evaluated. Therefore, we conducted an interrupted time-series analysis to evaluate the impact of the financial incentive on the prescription of biosimilars to outpatients who require self-injection.
The DeSC database (DeSC Healthcare Inc.), which comprises commercially available administrative claims and health checkup data, was used to obtain data for the period from April 2014 to November 2022. The database includes health insurance claims data from three health insurers: (i) National Health Insurance, for some municipalities, (ii) Kempo, operated by various health insurance societies, and (iii) the Advanced Elderly Medical Service System, for some prefectures, which covers individuals aged ≥75 years. Individuals covered by the National Health Insurance are typically self-employed individuals, freelancers, part-time workers, unemployed individuals, or retired individuals aged 65–74 years. Those covered by Kempo are mostly under 65 years of age and mainly full-time employees. The Advanced Elderly Medical Service System covers individuals aged ≥75 years. This database includes information on approximately 12,500,000 individuals, and claims data at the individual level are anonymized. It covers data on young, middle-aged, and older individuals and includes the following information: individual identifier, birth month, sex, and drugs dispensed based on drug codes9),10).
FINANCIAL INCENTIVE SCHEME TO PROMOTE BIOSIMILARSA new health insurance reimbursement for outpatients who need self-injection, “Fee for the Initial Introduction of Biosimilars,” was implemented on April 1, 2020. The fee (10 US dollars) is paid only once a month for 3 months starting from the month of the first prescription date of biosimilars if a doctor explains their effectiveness and safety and prescribes it to an outpatient who needs self-injection. The biosimilars subject to the incentive were insulin, somatropin (human growth hormone), etanercept, and teriparatide.
PATIENTS AND OUTCOMESAll patients who required self-injection and were prescribed either insulin, somatropin, or etanercept at the start of the observation were included in our analyses. The observation period was from April 2019 to March 2021. We excluded any product that was newly launched or discontinued during the study period to ensure that the biosimilars analyzed remained the same. Patients who required self-injection were defined as those who paid a fee for self-injection guidance. The measured outcomes were the monthly proportions of prescribed biosimilars. A list of drug codes for the biosimilars used in this study is presented in Supplementary Table 1. For each month of the study period, the outcomes were calculated as the number of prescriptions in the claims data for that month.
VARIABLESThe baseline characteristics of patients who required self-injection and were prescribed either insulin or etanercept included age, sex, insurer, fiscal year, and proportion of biosimilars prescribed.
STATISTICAL ANALYSISFirst, the background characteristics of patients stratified by the type of biosimilar (insulin, etanercept, and somatropin) were analyzed. Second, aggregated monthly data from April 2019 to March 2021 were prepared for interrupted time-series analysis. These data were used to evaluate the effects of the financial incentive scheme. A Prais–Winsten linear regression model was fitted to account for autocorrelation11). and the effect of the financial incentive scheme was analyzed by using the following segmented regression model12):
Yt = β0 + β1Tt + β2Xt + β3TtXt
where Yt is the monthly proportion of outcome at time t; Tt represents the months elapsed from the start of the observation period; X1t is a dummy variable (assuming 0 and 1 for before and after the introduction of the financial incentive, respectively); TtXt is the interaction term, which indicates the months passed from the introduction of the financial incentive; β0 is the intercept monthly proportion of the outcome; β1 estimates the slope of the monthly proportion of the outcome before the introduction of the financial incentive; β2 estimates the effect on the monthly proportion of the outcome immediately after the introduction of the financial incentive; β3 estimates the sustained effect following the introduction of the financial incentive and represents the effect of the intervention over time; β1 + β3 estimates the slope after the introduction of the financial incentive. In addition, the Durbin–Watson test was used to evaluate the autocorrelation between data points in the generalized least-squares regression model13). Third, we analyzed the relationship between the prescription of biosimilars and the “Fee for the Initial Introduction of Biosimilars” from April 2020 to March 2021. The proportion of prescriptions, medical institutions, and incentives reimbursed was assessed. Statistical significance was defined as p < 0.05. All analyses were performed using the Stata/SE software (version 17.0; StataCorp, College Station, TX, USA). This study was approved by the Institutional Review Board of the University of Tokyo (approval number:3501-(5), May 19, 2021). The requirement for written consent was waived due to data anonymity.
SUBGROUP ANALYSISPatients who self-injected insulin or etanercept were identified. Subsequently, the same analyses were repeated as in the main analysis. The total number of prescribed somatropins was small (n = 5), and a subgroup analysis was not conducted for somatropin.
SENSITIVITY ANALYSISThe robustness of the results was tested using two alternative time bandwidths: between October 2019 and September 2020 and between June 2018 and January 2022. The start of the latter alternative time was set in June 2018 because the first biosimilar of etanercept went on sale in May 2018.
We identified 3,348 eligible patients, including 3,253 (97.2%), 90 (2.7%), and 5 (0.1%) who were prescribed insulin, etanercept, and somatropin, respectively. The baseline patient characteristics are presented in Table 1. The proportions of patients prescribed insulin and etanercept were 50.2% and 66.7%, respectively, in the fiscal year 2019; the corresponding proportions in the fiscal year 2020 were 49.8% and 33.3%. The proportion of patients covered by the National Health Insurance, Kempo, and the Advanced Elderly Medical Service System who were prescribed any biosimilars was 40.6%, 2.6%, and 56.8%, respectively. The proportion of biosimilar prescriptions was approximately 56% and 6% of the overall insulin and etanercept prescriptions, respectively. Table 2 shows the monthly numbers and proportions of prescribed biosimilars. The number of monthly insulin prescriptions was above a certain level during the study period; however, etanercept was prescribed only for 5 months during the study period.
| Characteristics | Total n = 3,348 |
Insulin n = 3,253 |
Etanercept n = 90 |
Somatropin n = 5 |
|---|---|---|---|---|
| Age (years), n(%) | ||||
| 0–20 | 76 (2.3) | 71 (2.2) | 0 (0) | 5 (100) |
| 20–40 | 108 (3.2) | 108 (3.3) | 0 (0) | 0 (0) |
| 40–60 | 421 (12.6) | 408 (12.5) | 13 (14.4) | 0 (0) |
| 60–80 | 1,808 (54.0) | 1,748 (53.7) | 60 (66.7) | 0 (0) |
| 80–100 | 935 (27.9) | 918 (28.2) | 17 (18.9) | 0 (0) |
| Sex, n (%) | ||||
| Male | 1,755 (52.4) | 1,743 (53.6) | 9 (10.0) | 3 (60.0) |
| Female | 1,593 (47.6) | 1,510 (46.4) | 81 (90.0) | 2 (40.0) |
| Insurer | ||||
| National Health Insurance | 1,359 (40.6) | 1,327 (40.8) | 27 (30.0) | 5 (100) |
| Kempo | 86 (2.6) | 81 (2.5) | 5 (5.6) | 0 (0) |
| Advanced Elderly Medical Service System | 1,903 (56.8) | 1,845 (56.7) | 58 (64.4) | 0 (0) |
| Proportion of prescription of biosimilars, n (%) | 1,830 (54.7) | 1,825 (56.1) | 5 (5.6) | 0 (0) |
| Time (Month, Year) | Total population | Insulin | Etanercept | Somatropin | ||||
|---|---|---|---|---|---|---|---|---|
| n | % | n | % | n | % | n | % | |
| April, 2019 | 66 | 47.1 | 66 | 49.3 | 0 | 0 | 0 | (0) |
| May, 2019 | 59 | 47.2 | 59 | 48.0 | 0 | 0 | 0 | (0) |
| June, 2019 | 89 | 64.5 | 89 | 66.9 | 0 | 0 | 0 | (0) |
| July, 2019 | 67 | 45.6 | 67 | 47.9 | 0 | 0 | 0 | (0) |
| August, 2019 | 70 | 49.0 | 70 | 50.4 | 0 | 0 | 0 | (0) |
| September, 2019 | 75 | 55.6 | 74 | 57.4 | 1 | 16.7 | 0 | (0) |
| October, 2019 | 67 | 51.9 | 67 | 54.9 | 0 | 0 | 0 | (0) |
| November, 2019 | 74 | 53.2 | 74 | 54.8 | 0 | 0 | 0 | (0) |
| December, 2019 | 100 | 56.8 | 100 | 58.5 | 0 | 0 | 0 | (0) |
| January, 2020 | 65 | 52.0 | 65 | 54.2 | 0 | 0 | 0 | (0) |
| February, 2020 | 80 | 60.6 | 80 | 61.5 | 0 | 0 | 0 | (0) |
| March, 2020 | 84 | 50.9 | 83 | 53.2 | 1 | 12.5 | 0 | (0) |
| April, 2020 | 82 | 56.9 | 82 | 59.9 | 0 | 0 | 0 | (0) |
| May, 2020 | 76 | 55.5 | 76 | 57.6 | 0 | 0 | 0 | (0) |
| June, 2020 | 79 | 54.5 | 79 | 54.9 | 0 | 0 | 0 | (0) |
| July, 2020 | 91 | 55.2 | 91 | 55.8 | 0 | 0 | 0 | (0) |
| August, 2020 | 95 | 65.1 | 95 | 66.0 | 0 | 0 | 0 | (0) |
| September, 2020 | 68 | 51.1 | 68 | 51.9 | 0 | 0 | 0 | (0) |
| October, 2020 | 74 | 53.6 | 73 | 54.1 | 1 | 33.3 | 0 | (0) |
| November, 2020 | 56 | 52.8 | 55 | 53.4 | 1 | 33.3 | 0 | (0) |
| December, 2020 | 91 | 58.7 | 90 | 58.8 | 1 | 50.0 | 0 | (0) |
| January, 2021 | 67 | 57.8 | 67 | 58.3 | 0 | 0 | 0 | (0) |
| February, 2021 | 76 | 61.3 | 76 | 61.8 | 0 | 0 | 0 | (0) |
| March, 2021 | 79 | 54.5 | 79 | 56.0 | 0 | 0 | 0 | (0) |
The results of the interrupted time-series analysis are presented in Table 3 and Fig. 1. The monthly proportions showed no significant trends throughout the study period. The slope changes after the introduction of the financial incentive were also not significant. In addition, the immediate effect of the introduction was not significant. The results of the sensitivity analyses were similar to those of the primary analyses. (Supplementary Tables 2 and 3 and Supplementary Figs. 1 and 2).
| Monthly proportion | Coefficient | 95% Confidence interval | P-value |
|---|---|---|---|
| Slope before the introduction | 0.52 | −0.21 to 1.25 | 0.153 |
| Immediate effect after the introduction | −0.95 | −5.68 to 3.78 | 0.678 |
| Sustained effect | −0.36 | −1.14 to 0.42 | 0.353 |
| Slope after the introduction | 0.17 | −0.18 to 0.51 | 0.336 |
An interrupted time-series analysis was performed to analyze monthly data of patients who were prescribed biosimilars (between April 2019 and March 2021) to evaluate the effect of the financial incentive scheme introduced in April 2020 in Japan. A Prais–Winsten linear regression model was fitted to account for autocorrelation. The horizontal and vertical axes represent time and monthly proportion, respectively. Black dots and lines represent the actual and predicted monthly proportions, respectively. The dashed line indicates when the incentive scheme was introduced.
Table 4 shows the data related to the prescription of biosimilars and the assessment of “Fee for the Initial Introduction of Biosimilars” from April 2020 to March 2021. The proportion of biosimilar prescriptions of the overall biopharmaceutical prescriptions was 58%. The proportion of medical institutions that prescribed biosimilars to the overall number of medical institutions was 69%. The proportion of fee assessments to the number of initial prescriptions of biosimilars was 6%. The proportion of medical institutions that assessed fees to the number of medical institutions that prescribed biosimilars was 9%. The proportion of medical institutions that assessed fees between April 2020 and March 2021 to the number of medical institutions that assessed fees between April 2020 and January 2022 was 45%.
| The proportion | % |
|---|---|
| The proportion of prescription of biosimilars relative to the prescription of overall biopharmaceuticals | 58.0 |
| The proportion of medical institutions that prescribed biosimilars relative to overall medical institutions | 69.3 |
| The proportion of the incentives reimbursed relative to the number of the initial biosimilar prescriptions | 5.6 |
| The proportion of medical institutions that reimbursed incentives relative to those that prescribed biosimilars | 8.8 |
| The proportion of the incentives reimbursed from the first prescription date of biosimilars to 3 months relative to the initial prescription of biosimilars | 1.1 |
| The proportion of the incentives reimbursed in the period between April 2020 and March 2021 relative to those in the period between April 2020 and January 2022 | 40.7 |
| The proportion of medical institutions that reimbursed incentives in the period between April 2020 and March 2021 relative to those in the period between April 2020 and January 2022 | 44.6 |
We conducted the interrupted time-series analyses stratified by biosimilars. The results were similar to those of the main analyses (Supplementary Tables 4 and 5 and Supplementary Figs. 3 and 4).
In this study, we examined the impact of financial incentives to prescribe biosimilars to outpatients who required self-injection in Japan by using the DeSC database. An interrupted time-series analysis revealed that the financial incentives introduced in April 2020 were not associated with an increase in biosimilar prescriptions.
The proportion of biosimilar prescriptions to those of overall insulin glargine products was approximately 65% from 2019 to 2020 according to a report by the MHLW6), whereas it was found to be approximately 56% in our study. The reason for this difference may be that the original products listed in the report did not include Lantus XR14). According to one report, the proportion of biosimilar prescriptions relative to all etanercept prescriptions was approximately 20% in 2019 and approximately 40% in 20206). In contrast, in this study, it was 6%, and etanercept was prescribed only for 5 months during the study period. Additionally, the number of etanercept prescriptions in 2020 was half that in 2019. There are three possible reasons why the proportion of biosimilar prescriptions and the number of prescriptions in 2020 were small. First, biosimilar etanercept was not fully supplied to medical institutions because one of the suppliers restricted its sales from July 2018 to July 2019. Second, our study did not include new biosimilars launched in 2019. Third, doctors are reluctant to switch from original products to biosimilars for patients with chronic inflammatory diseases, such as rheumatoid arthritis, who have already been effectively treated with the original products15).
Our interrupted time-series analyses revealed that the financial incentives introduced in April 2020 were not associated with an increase in biosimilar prescriptions. The proportion of medical institutions willing to adopt biosimilars was 73.5%, as reported by the MHLW in 202317). This was similar to the proportion of medical institutions prescribing biosimilars in 2020, which was found to be 69% in our study. The proportion of incentives reimbursed relative to the number of initial biosimilar prescriptions was approximately 6%, and the proportion of medical institutions that reimbursed incentives relative to those that prescribed biosimilars was approximately 9%. These findings suggest that, while biosimilars were prescribed, the associated incentives were not utilized. According to a report by the MHLW17), approximately 17% of doctors who prescribed biosimilars in clinics and approximately 8% of those in hospitals increased their prescription of biosimilars because the fee was initiated in 2021. According to another report by the MHLW18), approximately 22% of doctors recognized that the fee had started, and only 29% of them increased their prescription of biosimilars because of the fee in 2023. Therefore, the financial incentive scheme may not be sufficiently understood, which could explain its lack of impact, or the scheme itself may simply be ineffective.
According to a report by the MHLW18), the main reason for not switching to biosimilars for insulin and etanercept was concern about the stable supply of biosimilars. The most compelling reason for promoting the prescription of biosimilars in clinics is to reduce the economic burden on patients. Japan has eliminated this burden in terms of biosimilar pricing by setting prices 30% lower than those of the original products and strictly regulating biosimilar prices, as a supply-side measure, compared with the UK, France, and Korea. However, Japan lacks demand-side measures19). Therefore, Japan needs to explore other demand-side measures to reduce the economic burden on patients in addition to pricing biosimilars and providing financial incentives for doctors to promote biosimilar prescriptions.
This study has some limitations that should be considered when interpreting the results. First, new biopharmaceuticals launched during the target period were excluded. Second, although the financial incentive scheme targets four types of biopharmaceuticals, teriparatide was not included in our study. Third, the volume of biopharmaceuticals supplied may have influenced the proportion of biosimilar prescriptions. Forth, drug prices for all biopharmaceuticals included in this study were revised in April 2020 as part of a regular cost-reduction policy. However, these changes were not specific to biosimilars and thus unlikely to have influenced our findings20). Finally, the burden of medical expenditures differs depending on age and income. Patients with lower out-of-pocket expenses may be less sensitive to the price of biosimilars, which could have influenced the results. However, the database does not include this information, and we could not evaluate its impact.
Our study suggests that the financial incentives introduced in April 2020 in Japan were not associated with an increase in biosimilar prescriptions.
The first author is affiliated with the Ministry of Health, Labor, and Welfare in Japan at the time of submission.
The second author receives the research funding from DeSC Healthcare Inc.
This work was supported by a grant from the Ministry of Health, Labour and Welfare, Japan (23AA2003).
Takahito Morita: Concept and design, Acquisition of data, Analysis and interpretation of data, Drafting of the manuscript, Critical revision of the paper for important intellectual content, Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Yusuke Sasabuchi: Acquisition of data, Critical revision of the paper for important intellectual content, Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Hideo Yasunaga: Final approval of the version to be published, Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Yusuke Sasabuchi and Hideo Yasunaga are the Editorial Board members of Annals of Clinical Epidemiology (ACE). They were not involved in the peer-review or decision-making process for this paper.
