Prescriptions of NOACs for patients with acute ischemic stroke and atrial fibrillation have been well documented in literature on clinical pharmacy and medicine [
17‐
19]. Similarly, managing the adoption of NOACs, a first-in-class oral anticoagulant, in various clinical settings is an interesting research area in public health and health financing. However, little is known about how health care institutions approach prescribing new drugs, and how many high-priced new drugs are substituted for old and low-cost drugs in South Korea. We aimed to address the speed of adoption and prescription volume of high-priced new drugs. NOACs, presenting a better risk-benefit balance compared to a traditional and low-priced comparator, Warfarin, are an ideal example to study. We applied a survival analysis to elucidate factors that affected the speed of adoption of NOACs, and interrupted time series analysis to estimate the effect of policy options to manage pharmaceutical expenditures. Based on the empirical analysis, we suggest policy implications for the adoption of new drugs and the management of pharmaceutical expenditures.
Summary of findings
There are interesting findings in the study. First, we found that health care institutions are conservative in adopting new drugs. Among 934 health care institutions in the study population, 334 institutions (36%) had prescribed NOACs at least once during the study period. However, the speed of adoption was related to the characteristics of the health care institution. For instance, the time to prescription for primary, secondary, and tertiary institutions was significantly delayed compared to that of quaternary institutions. However, location was not significantly related to the speed of adoption. Second, we noted that prescriptions of NOACs before the withdrawal of the need for a letter of opinion were marginal, and the prescription volume of NOACs was significantly increased after the withdrawal of the letter of opinion. Prescriptions of NOACs before the withdrawal of the letter were less than 100 prescriptions per month and increased to more than 400 prescriptions per month after the withdrawal of the letter. Third, we found different trends in the speed of adoption and the prescription volume between quaternary and tertiary institutions. During the study period, most prescriptions were prescribed from quaternary (43%) and tertiary institutions (45%). However, the speed of adoption was faster for tertiary institutions in the short run, while the speed of adoption was faster for quaternary institutions in the long run.
Being conservative in prescribing NOACs
In our analysis, none of the health care institutions prescribed NOACs immediately after they were reimbursed on October 1, 2010. The first prescription for NOACs was observed after the extension of reimbursement coverage with the need for the letter of opinion (January 1, 2013) in our dataset. Furthermore, we found that most health care institutions (64%) have not prescribed NOACs even after the coverage extended with the need of a letter of opinion, indicating that health care institutions were conservative in adopting new drugs. Being conservative in adopting NOACs could be explained by a variety of factors, such as an intentionally cautious approach to adopting new drugs to ensure patient safety or a lack of awareness of a drug’s introduction [
11]. More specifically, physicians have incomplete information on the safety and effectiveness of new drugs [
23,
24], and initial clinical trials are often too small to detect rare adverse reactions or demonstrate comparative effectiveness [
31]. Furthermore, the HIRA requested physicians to provide a letter of opinion when prescribing NOACs to manage the prescription volume of NOACs in South Korea.
We also confirmed that the speed of adoption varied among the types of health care institutions. In particular, the adoption of NOACs among institutions was in the order of quaternary, tertiary, secondary, and primary institutions: quaternary and tertiary institutions were often faster to adopt new drugs; and primary and secondary institutions were often slower to adopt new drugs. This observation means that high-risk patients might visit quaternary and tertiary institutions or cardiologists at these institutions. Thus, quaternary, and tertiary institutions will be more likely to adopt new medicines earlier compared to primary institutions or primary care physicians.
Our observations were consistent with the previous literature presenting that physicians were conservative in prescribing new drugs and that physician specialty was the most consistent predictors of adopting new drugs [
10,
11,
13,
23,
32,
33]. For instance, Anderson et al. (2018) examined the adoption and prescription of three cardiovascular drugs of differing novelty, including Dabigatran, Aliskiren, and Pitavastatin in Pennsylvania in the United States. Note that Dabigatran, which is one of the drugs belonging to NOACs, is a first-in-class drug approved by the FDA in 2010. In the literature, most physicians (73%) did not adopt Dabigatran in the first 15 months after marketplace introduction, and cardiologists were more likely to adopt Dabigatran compared to primary care physicians.
However, O’Connor et al. (2018) reported rapid adoption of novel cancer therapeutics, specifically immune checkpoint inhibitors of programmed cell death 1 protein (anti-PD-1 agents), immediately after FDA approval in the United States [
12]. The authors explained that disease severity, preference for novelty, and clinical gains over existing treatments contributed to the rapid adoption of new drugs. Specifically, anti-PD-1 agents presented large gains in efficacy and effectiveness, including better survival in trials and durable responses to treatment, for selected patients [
34]. It is also noteworthy that anti-PD-1 agents are widely used to treat many types of solid tumors, including malignant melanoma and non-small cell lung cancer [
35].
Furthermore, we presented the difference in the speed of adoption among quaternary and tertiary institutions in this study. In the short term, the speed of adoption among quaternary institutions was like those of primary and secondary institutions, indicating that many quaternary institutions are reluctant to prescribe new medicines immediately after new drugs become available. Thus, the adoption curve for quaternary institutions in Fig.
2 was flat for the first 12 months after the first prescription occurred and significantly dropped after the first 12 months. However, the adoption curve for tertiary institutions smoothly dropped during the same period. Similarly, we analyzed the effect of the location of institutions in adopting new drugs. We found that location of institutions was not related to the adoption of new drugs, indicating the consistent results reported in the previous literature [
13,
23].
Managing pharmaceutical expenditure
We analyzed the prescription volume of NOACs to understand the adoption of new drugs and to manage pharmaceutical expenditure. During the study period, most prescriptions (88%) were prescribed from quaternary (43%) and tertiary (45%) institutions. Interestingly, we found different trends in prescribing NOACs among quaternary and tertiary institutions in our interrupted time series analysis. Before the withdrawal of a letter of opinion, the coefficient of the time trends for quaternary institutions (1.27, p-value < 0.0001) was higher than that of tertiary institutions (0.58, p-value 0.0054). However, the value of the coefficient of the time trends was reversed after the withdrawal of a letter of opinion; 9.01 (p-value < 0.0001) and 16.21 (p-value < 0.0001), respectively. Furthermore, the value of the immediate effect for tertiary institutions (141.15, p-value < 0.0001) was higher than that for quaternary institutions (128.04, p-value < 0.0001). Thus, we can conclude that the withdrawal of the letter of opinion had a stronger effect on tertiary institutions than on quaternary institutions.
This study also captured the effect of introducing a letter of opinion against extension of reimbursement of NOACs. As already demonstrated, a letter of opinion was quite effective in slow adoption of NOACs and managing prescription volume in South Korea. Particularly, all the slopes of the curve after the withdrawal of the letter of opinion increased in Fig.
2, and the prescription volume quite increased after the withdrawal of the letter in Fig.
3. These observations indicate that the introducing the letter of opinion were quite effective in the adoption of NOACs as well as managing prescription volume.
Study limitations
Our study has some limitations. First, we used a sample cohort and assumed that the cohort was reproducible. However, prescriptions for rare diseases were hard to be captured in the cohort, which was composed of 1 million individuals (2%) selected from South Koreans. Thus, prescriptions that had occurred before the extension of the reimbursement coverage on January 1, 2013 might have been omitted from our dataset. Second, we analyzed the adoption of new drugs in health care institutions. It means that we did not consider the characteristics of patients and their clinical information. Finally, we analyzed NOACs to understand the adoption of new drugs in South Korea, meaning that our findings might not be generalizable to other new drugs. For instance, the adoption of new drugs might vary by the characteristics of the drugs, including disease severity and clinical gains over existing treatments.