<scp>PMDA</scp> Perspective on <scp>RWD</scp>/<scp>RWE</scp> Utilization for Regulatory Purposes Including Assessment on the Impacts of Regulatory Actions and Safety Risk of a Drug at Postmarketing Stage

Real-world data (RWD) and real-world evidence (RWE) have been used for regulatory decision-making as one of useful information for not only the assessment of postmarketing drug safety but also the review of new drug approval in Japan, the United States, and European Union (EU) [1-3]. When focusing on postmarketing drug safety assessment, there various reasons for using RWD, mainly including (1) understanding real-world situation in clinical practice (e.g., drug utilization study), (2) safety risk assessment, and (3) assessing impacts of regulatory action (e.g., package insert [PI] revision). Different sources of RWD, such as electronic medical records, claims, and registries, have been used in different regions including the Sentinel in the United States [4] and multiple registries in EU [5]. In Japan, the Pharmaceuticals and Medical Devices Agency (PMDA) mainly utilizes two types of RWD sources, namely, MID-NET [6, 7] and the National Database of Health Insurance Claims and Specific Health Checkups of Japan (NDB) [8] (Figure 1). In particular, MID-NET has unique characteristics which contains many types of laboratory test results for analysis in addition to the normal data categories from hospital information such as electronic medical record, claims, and diagnostic procedure combination data from 10 collaborative organizations including 23 hospitals. NDB is the largest health-related database in Japan with comprehensive electronic health insurance claims and specific health check-up data on almost all Japanese nationals from medical facilities including hospitals, clinics, pharmacies, and dental clinics across Japan, but no laboratory test results are available for analysis. The database used for a study is usually selected based on a research question and results of feasibility assessment. Hence, appropriate target populations, outcomes, and covariates can be identified and made available for analysis [2, 8]. For example, MID-NET is generally used for a study whose target outcome is measurable on laboratory tests, but has a limitation in examining rare events due to relatively smaller sample size compared with NDB. In recent years, the use of anti-coronavirus disease 2019 (COVID-19) drugs has been approved even with limited data under special approval processes (e.g., emergency approval) as a regulatory decision for responding to public health emergencies. With the utilization of these drugs, continuous monitoring with rapid accumulation of safety information from an early stage after approval is essential for ensuring benefit and risk balance and promoting the proper use of a drug. RWD/RWE plays an important role in drug safety monitoring under these circumstances and in evaluating the impacts of such special regulatory decisions in real-world clinical practice. To understand the real-world situation in clinical practice, various types of laboratory test results available in MID-NET have been used for monitoring the postmarketing safety of anti-COVID-19 drugs from an early stage after approval via periodic data analysis [9]. Further, to elucidate important information for the proper use of anti-COVID-19 drugs in clinical practice, the PMDA investigated the trend in the real-world prescription of anti-COVID-19 drugs in hospitalized patients with COVID-19 based on data from MID-NET [10]. In the meantime, no regulatory actions have been made based on these study results. However, if unexpected use of a drug and/or marked increase of adverse events by the drug are identified in these studies, a regulatory action such as a PI revision to manage a risk of the drug can be taken in an early stage after approval. With the use of RWD for assessing the impacts of regulatory action, various studies with different objectives, such as checking the compliance status of precautions for use described in the PI and examining trend changes in prescription and/or a particular safety risk before and after the PI revision, can be performed. For example, MID-NET data were analyzed by the PMDA to characterize granulocytopenia in patients receiving thiamazole, whose PI require to undergo laboratory test specifically for neutrophil count during treatment with thiamazole to prevent adverse events [11]. Briefly, in 4371 patients with thiamazole, the occurrence of granulocytopenia in patients who had undergone blood tests at all previous periods was similar or higher than that among those who had not undergone blood test in all previous periods. The neutrophil count was relatively lower in the group of patients with granulocytopenia even before the occurrence of granulocytopenia. With a new user cohort and a nested-case control design, this study provided important information regarding the effects of routine blood tests and the potential risk factors for granulocytopenia. This information facilitates a proper use of thiamazole by understanding the characteristics of granulocytopenia during thiamazole treatment in real-world clinical practice. As another example, PMDA utilized MID-NET to examine the impact of the dear healthcare professional letter (DHCP) on the incidence of hypocalcemia associated with denosumab using an interrupted time series design [7]. This study indicated a decreased trend in the occurrence of hypocalcemia after issuing DHCP. The frequency of routine laboratory testing for serum calcium was high before the issuance of the DHCP letter and remained high afterward, with no trend change. Furthermore, the impacts of metformin PI revision were evaluated in a new user cohort design based on MID-NET [ 12]. This study examined prescription trends and the occurrence of lactic acidosis in patients prescribed with metformin before and after PI revision in Japan on 18 June 2019, which allowed metformin use for patients with moderately decreased kidney function (30 <estimated glomerular filtration rate <60 mL/min/1.73 m2). In this study, 5874 patients were identified as new metformin users from 12 May 2016 to 31 March 2020, including 1145 patients with moderately decreased kidney function. The percentage of patients with lower daily metformin dose (≤500 mg) at the first prescription increased after PI revision, although no marked changes in the number of metformin prescription were observed before and after PI revision. No lactic acidosis was observed in patients with moderately decreased kidney function before and after PI revision. The results showed no further safety concerns regarding metformin use in patients with moderately decreased kidney function under the situation with careful use and safety monitoring of metformin based on the revised PI. As described above, RWD/RWE significantly contributes to postmarketing drug safety assessment. Figure 2 summarizes the purposes of RWD utilization in the field of drug safety. Understanding the impacts of regulatory action in clinical practice by assessing changes before and after safety-related regulatory action as described above is a key area in utilizing RWD for regulatory purposes. This type of study can provide useful evidence that can facilitate an accurate understanding of the impacts of regulatory action in clinical practice and can promote more efficient and timely safety-related regulatory actions by implementing appropriate measures reflecting real-world situations. If the study results indicate an insufficient regulatory action to manage the risks of a drug in clinical practice, additional safety-related regulatory actions may be implemented. Furthermore, understanding real-world situation in clinical practice by checking the compliance status of required tests (e.g., routine tests for serum calcium level as required in the PI) or examining prescription trends (including patient backgrounds), and investigating safety risks in real-world populations via safety risk comparison and/or safety signal detection are also major areas when utilizing RWD in drug safety. On the other hand, in our experiences, it may be usually difficult to use these databases for a study focusing on safety events which are generally managed by symptomatic treatments without specific medical interventions (e.g., acute liver failure managed by supportive care including hydration without specific pharmacotherapy and surgery), because clinically relevant cases cannot be accurately detected by using data elements available in these databases. For regulatory decision-making, all available scientific evidence and information including RWE, spontaneous adverse drug reaction reports, safety measures taken by foreign regulatory agencies, and literature, are used to evaluate current real-world situations and take necessary safety measures. In the postmarketing safety assessment, RWE based on the foreign databases are generally used as one of available evidences unless clear ethnic differences are identified. As a result, the possible conclusions include strengthening and degrading safety warnings as well as maintaining current safety measures. Continuous regulatory considerations on drug safety based on scientific evidence and information in the postmarketing stage can optimize pharmacovigilance activities. Several challenges in utilizing RWD for assessing regulatory actions have been reported, including the need to cover a wider range of medicines, selecting clinically relevant outcome measures, developing better methods to address potential confounders, assessing long-term outcomes, and using appropriate comparators [13]. Similar challenges have been observed in our experiences in utilizing RWD in Japan. Selecting better data source (NDB or MID-NET) fitted to the purpose of study may partly address these challenges but more comprehensive database in terms of data size and data varieties will be necessary. In particular, one of major challenges in Japan for further promoting the utilization of RWD/RWE for regulatory purposes is to build a linkage between different databases for enabling longer follow-up period of a patient and analyzing more enriched data, such as event of death and detailed clinical findings, together with the general hospital information data. As one of recent initiative by MHLW, the linkage of the NDB with other database such as the nursing care database has been discussed [14]. It is expected that such database linked with multiple data sources would be useful for assessments with wider scopes of regulatory needs in the future. In addition, the international cooperation and harmonization such as the discussions related to RWD/RWE in the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) [15, 16] will also be indispensable for increasing trust in RWD utilization and facilitating the regulatory acceptance of RWE on a global basis. The PMDA will continue exerting efforts in utilizing RWD/RWE for a better regulatory decision-making on drug safety. All authors equally contributed to this manuscript in the interpretation and writing of the manuscript. The views expressed herein are the result of independent work and do not necessarily represent the views and findings of the Pharmaceuticals & Medical Devices Agency. The authors have nothing to report. The authors declare no conflicts of interest.