Trends of clinical trials from 2017 to 2019 in Korea: an integrated analysis based on the Ministry of Food and Drug Safety (MFDS) and the Clinical Research Information Service (CRIS) registries

Prospective registration of clinical trials enables unbiased reporting of the results and ensures transparency [1]. Since 2004, the International Committee of Medical Journal Editors (ICMJE) has required the registration of clinical trials in a public registry as a condition for publication [2]. Since then, prospective registration of clinical trials has increased remarkably [3] and has been accelerated by the legal mandates of registration by regulatory bodies including the United States [4].

Although the number of clinical trials in public registries has increased, the quality of the data in the registries still needs improvement [5]. A worldwide survey of clinical trial registries pointed out that a considerable number of trials missed important information and often was outdated while retrospective registration accounted for a fourth of the total trials [3]. Similar results were reported in a recent study that only 37% of trials were registered in a prospective manner, among which, only 31% of the trials provided study results [6].

In particular, phase 1 clinical trials, which often involve healthy volunteers, have several complicated concerns. Submission of results from phase 1 clinical trials is not mandatory as the trial is not considered as an applicable clinical trial under the Food and Drug Administration Amendments Act [7]. This could make phase 1 clinical trials more vulnerable to biases. Another concern is the overlapping enrollment of healthy volunteers [8] which necessitates sophisticated tracking of phase 1 clinical trials [9].

In Korea, two local databases provide fundamental information on clinical trials. The public database of the Ministry of Food and Drug Safety (MFDS) frequently updates information on the approved clinical trials, and the other is a local clinical trial registry named Clinical Research Information Service (CRIS) [10]. MFDS has provided the approval status of clinical trials, and made registration for the database mandatory since 2019 [11]. CRIS was developed in February 2010 by Korea Centers for Disease Control and Prevention to support registration and report of the study results [10]. Both trial databases are not identical as the former provides all of the information on the clinical trials approved by the regulatory agency whereas the latter is one of the possible public trial registries accepted by ICMJE.

To provide a consistent overview of the approved clinical trials along with its counterpart study for years 2014–2016 [12], we analyzed the characteristics of clinical trials approved in terms of the study phase, therapeutic area, and geographic distribution. The previous study demonstrated an increasing trend of phase 1 clinical trials and a geographic imbalance in Korea. In the current analysis, we set up the study period as 2017–2019, which was prior to the first coronavirus disease 2019 (COVID-19) outbreak in Korea, to avoid the significant impact of COVID-19 on the clinical trial environment [13]. In addition, we newly included CRIS registry into analysis to promote the establishment of a harmonized clinical trial registry in Korea.

We found that the overall number of clinical trials in Korea from 2017 to 2019 increased by 7.9% compared to previous three-year interval. The increase of phase 1 trials and IIT was another remarkable change, in contrast to the previous predominance of phase 3 trials. In terms of therapeutic area, the number of clinical trials in the Antineoplastic and immunomodulating agent area was still the largest as it was in the previous three years [12]. We also found a relative increase of trials in the Alimentary tract and metabolism area from 157 (8.0%) to 295 (14.2%).

We noted that relative increase in the Alimentary tract and metabolism area was a distinctive trend compared to that in 2014–2016. Despite the increase in phase 1 clinical trials in 2014–2016, the proportion of the Alimentary tract and metabolism was constant. In contrast, the proportion of the area, where most trials were at phase 1 (Supplementary Fig. 1), almost doubled in 2017–2019 (Table 2). We expected that increased development of several popular drug classes in the area (e.g., proton-pump inhibitors [18]) might contribute to the increase.

The predominance of phase 1 clinical trials was aligned with early-stage-focused development led by the domestic biopharmaceutical companies. In a review conducted by Korea National Enterprise for Clinical Trials (KoNECT), industry-sponsored domestic trials were mostly at phase 1; the ratio of phase 1, 2, and 3 trials was 6.7: 1.0: 1.3, respectively [19]. As more than half of the phase 1 trials were associated with fixed combination drugs or new formulations of marketed drugs in Korea [19], the trends of clinical trials could be influenced by several popular drug classes. Similarly, bioequivalence trials highly focused in several therapeutic areas were reported in another study [20].

The concentration of clinical trials in urban areas (e.g., Seoul metropolitan area) was a definite trend in Korea. Seoul occupied more than half of the entire trials (52.6%), similar to 53.5% in the previous report. Another study of oncology clinical trial conducted in Korea between 2007 and 2013 also reported that six large volume hospitals each conducted more than 50 clinical trials while 45% of study centers conducted less than 10 trials [21]. A study on the approved oncology trials in 2019 also revealed that 92% trials were available in Seoul while only 33% in provincial areas [22].

Similar phenomena have been addressed in the United States [23, 24]. Volunteers in rural areas tended to participate in clinical trials significantly less than urban counterparts (odds ratio, 0.30–0.46) [24]. Furthermore, proximity to clinical sites were related to patient recruitment and retention. Thus, the geography of clinical trials needs to be taken into account when interpreting and generalizing the results of clinical trials [25].

We also found several discrepancies in the results by the clinical trial registries. For instance, the therapeutic area in Musculo-skeletal system ranked the 6^th in the MFDS public database whereas it was 2^nd (excluding ‘not specified’) in the CRIS registry. Discrepancies in other therapeutic areas and study phases were also noted, despite the differences in the coding systems (Tables 2 and 3). Especially, as shown in Table 1, most of the study phases (63.9%) were not specified in the CRIS registry.

The discrepancies and errors in the clinical trial registries have been continuously addressed. A cross-sectional study of registered trials in ClinicalTrials.gov and European Union Clinical Trials Register revealed that 16.2% of trials were discrepant on the completion status [26]. Another study regarding pediatric trials in a peer-reviewed journal found that 19 out 20 randomized-controlled trials had medium or high combined discrepancy scores [27]. Similar results were reported in other literature [28, 29].

The results of this study indicate that a harmonized clinical trial registry with a regulatory database is necessary for proper evaluation of the clinical trial landscape. Since 2019, all applicants are required to register clinical trial information to the MFDS database [11]. Thus, all clinical trials approved by MFDS could be identifiable. The applicants are recommended to register the information to other public registries including CRIS and ClinicalTrials.gov. However, as there have been no mandates and guidelines for which public registry to choose, each trial could be filed to any registry even in a duplicated manner.

Although the importance of clinical trial registries has been emphasized in a systemic review and meta-analysis [30], registry data have not been properly aligned with publications [31]. Furthermore, data are not often standardized. For example, a bioequivalence trial was previously classified as a separate entity in Korea; however, recently, bioequivalence trials are classified as phase 1 studies since October 2017 [32, 33]. Similarly, therapeutic area can be coded with various systems including the WHO-ATC and ICD-11. Duplicate registration of clinical trials (e.g., CRIS and ClinicalTrials.gov) should also be avoided [34]. Without such efforts, an appropriate evaluation for the trends of clinical studies would be seriously hampered.

Overall, the results of our study could help to facilitate standardization and harmonization of the clinical trial registries. In this study, we revealed that evaluation of multiple registries was necessary due to discrepancies for overview of the clinical trials in Korea. Solutions for the discrepancy would comprise standardized database coupled with streamlined regulations. Technical integration with other global registries (e.g., ClinicalTrials.gov) could also be a possible solution.

Our study has some limitations. As several IPs under the early stage of drug development were yet assigned ATC codes, the therapeutic area should be assigned by subjective judgment of the investigators. The missing information in the registries limited precise analysis of the trends. Further investigations on the study design (e.g., blinding and randomization) need to be performed in future research.

In conclusion, our study provided a comprehensive overview of clinical trials in Korea between 2017 and 2019. The discrepancy between clinical trial registries could be resolved by introducing standardized database and guidelines.

Supplementary Figure 1

The number of clinical trials according to therapeutic area and study phase (Ministry of Food and Drug Safety).