One of the target points in the PMDA 2nd midterm plan (FY2009-2013) is reinforcement and enhancement of the system for safety evaluation for pharmaceuticals using expanded data sources beyond spontaneous reports of adverse drug reactions (ADRs). To achieve this goal, PMDA started investigation in FY2009 to develop methodology to utilize electronic medical information for secondary purpose of safety evaluation of pharmaceuticals. (MIHARI project- Medical Information for Risk Assessment Initiative)
Data sources targeted in MIHARI project are claims data, diagnosis procedure combination (DPC) data, hospital information system (HIS) data, etc.
Secondary use of electronic medical information for safety evaluation is expected to enable safety evaluation based on quantitative analysis, which has been difficult so far. It will also provide faster and easier way of evaluation compared to collecting primary data from study planned and conducted just for the purpose.
PMDA intends to establish the system to utilize electronic medical information (eg. claim data, DPC data, HIS data) practically for safety evaluation by the end of FY2013. Aiming for this, PMDA is conducting various pilot studies using currently available data in the MIHARI project. Here we report on recent developments of this project.

PMDA started MIHARI project in FY2009 to enhance drug safety assessment by developing ways to utilize electronic medical information as additional data sources to spontaneous adverse drug reaction reports. The project has been established according to PMDA's second midterm plan. In this article, we will introduce latest two studies (pilot studies No. 4 and 5) using data of standardized electronic medical record(EMR) called SS-MIX (standardized structured medical record information exchange) data out of our several pilot studies. SS-MIX is a standard specification published by the Ministry of Health, Labour and Welfare. In these studies, anonymized SS-MIX data were provided by six collaborative hospitals respectively. In pilot study No. 4, we explored approaches for evaluating the impact of regulatory action which instructed relevant manufacturers to revise package inserts of sitagliptin phosphate hydrate (sitagliptin) to call physician's attention. The revision was about reducing dose of sulfonylurea (SU) to avoid serious hypoglycemia when it is prescribed concomitantly with sitagliptin. As indicators of the impact, we evaluated changes in proportion of concomitant use and average SU dose before and after the action and estimated the risk of hypoglycemia in concomitant users compared to SU alone users before and after the action. In conclusion, evaluating impact of the regulatory action using SS-MIX data was technically feasible; however, it was difficult to analyze with adequate accuracy due to limited size of the data. In pilot study No. 5, we examined validity of outcome definitions for hyperthyroidism which were applied to combinations of some data elements of SS-MIX data to identify the patients. Three types of outcome definitions were prepared; 1) definitive diagnosis of hyperthyroidism, 2) prescription of medication for hyperthyroidism, 3) prescription of medication for hyperthyroidism in or after the month in which definitive diagnosis of hyperthyroidism was given. Criteria for case ascertainment were determined according to relevant clinical guidelines. After the cases were ascertained, positive predictives values were calculated. The results suggested that using information on prescription of medication improves validity of definition of outcome. The findings from pilot studies in MIHARI project have been utilized in another project which PMDA is carrying forward now (EMR network project). The findings would be also helpful when we use data from this network. (Jpn J Pharmacoepidemiol 2013；18(1)：23-29)

Purpose: Fine-needle aspiration cytology (FNAC) of axillary lymph nodes (AxLNs) is performed to diagnose nodal metastasis in patients with breast cancer. Although the sensitivity of ultrasound-guided FNAC for identifying AxLN metastasis is in the range of 36%–99%, whether sentinel lymph node biopsy (SLNB) should be performed for neoadjuvant chemotherapy (NAC) patients with negative FNAC results is uncertain. This study aimed to determine the role of FNAC before NAC in the evaluation and management of AxLN in early breast cancer patients. Methods: We retrospectively analyzed 3,810 clinically node-negative (a lymph node with no clinical metastasis without FNAC or radiological suspicion of metastasis with negative FNAC results) patients with breast cancer who underwent SLNB between 2008 and 2019. We compared the positivity rate of sentinel lymph nodes (SLNs) between patients who received and those who did not receive NAC with negative FNAC results or without FNAC and axillary recurrence rate in the neoadjuvant group with negative SLNB results. Results: In the non-neoadjuvant (primary surgery) group, the positivity rate of SLNs in patients with negative FNAC results was higher than that in patients without FNAC (33.2% vs. 12.9%; p < 0.001). However, the SLN positivity rate of patients with negative FNAC results (false-negative rate for FNAC) in the neoadjuvant group was lower than that in the primary surgery group (3.0% vs. 33.2%; p < 0.001). After a median follow-up of 3 years, one axillary nodal recurrence was observed, which was a case from the neoadjuvant non-FNAC group. None of the patients in the neoadjuvant group with negative FNAC results had axillary recurrence. Conclusion The false-negative rate for FNAC in the primary surgery group was high;however, SLNB was the proper axillary staging procedure for NAC patients who have clinically suspicious AxLN metastases on radiologic examination but negative FNAC results.

Purpose: Safely postponing the use of chemotherapy is important for quality of life maintenance in patients with hormone receptor-positive advanced breast cancer. In previous studies, a combination of cyclin-dependent kinase 4 and 6 inhibitors (CDK4/6i) and fulvestrant prolonged the time to chemotherapy (TTC). In this study, we used real-world data to evaluate TTC in the context of CDK4/6i therapy. Methods: We performed a retrospective chart review of women with estrogen receptorpositive, human epidermal growth factor receptor 2-negative advanced breast cancer treated at the Aichi Cancer Center Hospital. The patients were categorized into having received CDK4/6i therapy first (n = 41), second (n = 33), and none at all (n = 67). The change in TTC among the groups was examined. Results: The median follow-up time was 13.8, 27.5, and 30.3 months in the CDK4/6i (first), CDK4/6i (second), and non-CDK4/6i groups, respectively. The median progression-free survival (PFS) with first-line therapy for metastasis was 30.0, 11.9, and 13.0 months, respectively (CDK4/6i [first] vs. non-CDK4/6i; p = 0.018, CDK4/6i [second] vs. non-CDK4/6i;p = 0.383). The median TTC was not reached in the CDK4/6i (first) group, was 39.1 months in the CDK4/6i (second) group, and was 44.2 months in the non-CDK4/6i group (CDK4/6i [first] vs. non-CDK4/6i; p = 0.880; CDK4/6i [second] vs. non-CDK4/6i; p = 0.407). The nonCDK4/6i group with TTC ≥ 60 months included more cases of secondary endocrine therapy resistance (p = 0.017), no perioperative chemotherapy (p = 0.021), and a longer disease-free interval (p = 0.093). Conclusion Although PFS was significantly longer in the CDK4/6i (first) group than in the non-CDK4/6i group, TTC did not significantly differ among the three groups in real-world data. The non-CDK4/6i group showed a long TTC in patients with late recurrence and low risk at the primary lesion site, who benefited greatly from hormone monotherapy.