Japan's PHR landscape is fragmented across private vendors, hindering data portability. In the PHOENICS project, we addressed this by adopting FHIR Storage to exchange health data. We identified 30 document types; half lacked internal metadata for schema and versioning. We set up a FHIR CodeSystem naming method to enable document parsers to programmatically identify document types. Our findings highlight the need for stakeholders to keep a central registry for the metadata uniquely identifiable for automated document routing.

This study established a standard format for progress notes as a training dataset for LLMs. We reviewed existing guidelines and analyzed the structure of the EMR system at the Medical School of Ehime University. The Subjective, Objective, Assessment, and Plan (SOAP) sections in our system were successfully mapped to C-CDA sections. A DocumentReference extension was introduced to handle attachments, such as DICOM images. The results found substantial discrepancies between Japanese EMRs and C-CDA on FHIR, emphasizing the need for further study across multiple vendors to achieve progress in note standardization in Japan.

Observational Health Data Sciences and Informatics (OHDSI) is an international research community dedicated to large-scale observational studies using real-world healthcare data. Participation in OHDSI requires mapping local terminology systems to the OHDSI Standard Vocabulary (OSV) and transforming healthcare data into the Observational Medical Outcomes Partnership Common Data Model (OMOP CDM), a standardized database schema. Adherence to the OSV and CDM enables the integration of datasets from different countries and regions, facilitating international cross-sectional analyses and supporting the discovery of large-scale evidence and new medical knowledge.Despite the globally recognized healthcare technology and systems excellence in Japan, Japanese real-world data (RWD) remain underutilized in international research. This is primarily due to reliance on domestically managed controlled terminologies in Japan that are not aligned with international controlled terminologies such as SNOMED CT, making mapping Japanese RWD to the CDM challenging. In addition, the wide variety of pharmaceutical products in Japan has hindered the establishment of mappings to RxNorm, the standardized drug terminology used in OHDSI.Previously, we used a Large Language Model (LLM) to map Japanese pharmaceutical data to RxNorm. A sampling-based evaluation confirmed that the LLM could accurately identify mapping candidates. Pharmacists and a medical informatics researcher validated these mappings, resulting in an ingredient-based mapping of Japanese pharmaceutical terms to RxNorm.Researchers interested in pharmacoepidemiology, pharmacoeconomics, and drug-related clinical decision support systems integrated with Japanese RWD can benefit significantly from this dataset. It also contains information about the target drugs, their translated names, LLM-generated suggestions, and reference data, making it suitable for developing and validating natural language processing and machine learning techniques for terminology mapping.

This study developed and preliminarily evaluated a prototype tool that applies a large language model-based named entity recognition (NER) approach to pseudonymize Japanese clinical text. Japanese is an agglutinative language. The system was built using the Presidio framework with the GiNZA BERT model, which was trained for NER tasks. The approach achieved a precision of 0.672, a recall of 0.995, and an F-score of 0.802. The findings demonstrate the potential for further accuracy improvements through NER task training with a medical domain-specific LLM and rule-based processing that incorporates morphological analysis.

No study has shown the relationship between alanine-glyoxylate aminotransferase 2 ( AGXT2 ) single nucleotide polymorphisms (SNPs) and depressive symptoms. The present case–control study examined this relationship in Japanese adults. Cases and control participants were selected from those who participated in the baseline survey of the Aidai Cohort Study, which is an ongoing cohort study. Cases comprised 280 participants with depressive symptoms based on a Center for Epidemiologic Studies Depression Scale (CES-D) score ≥ 16. Control participants comprised 2034 participants without depressive symptoms based on the CES-D who had not been diagnosed by a physician as having depression or who had not been currently taking medication for depression. Adjustment was made for age, sex, smoking status, alcohol consumption, leisure time physical activity, education, body mass index, hypertension, dyslipidemia, and diabetes mellitus. Compared with the GG genotype of rs180749, both the GA and AA genotypes were significantly positively associated with the risk of depressive symptoms assessed by the CES-D: the adjusted odds ratios for the GA and AA genotypes were 2.83 (95% confidence interval [CI] 1.23–8.24) and 3.10 (95% CI 1.37–8.92), respectively. The TGC haplotype of rs37370, rs180749, and rs16899974 was significantly inversely related to depressive symptoms (crude OR 0.67; 95% CI 0.49–0.90), whereas the TAC haplotype was significantly positively associated with depressive symptoms (crude OR 1.24; 95% CI 1.01–1.52). This is the first study to show significant associations between AGXT2 SNP rs180749, the TGC haplotype, and the TAC haplotype and depressive symptoms.

OBJECTIVE:Although it has been suggested that a decline in oral function is one of the potential risk factors affecting mild cognitive impairment (MCI), evidence is insufficient to draw clear conclusions. This Japanese cross-sectional study examined the association between tongue pressure (TP) and MCI in middle-aged and older adults aged 36-84 years. METHODS:Study participants were 1019 (368 men and 651 women). TP was evaluated using a TP measurement device. The maximum value of three measurements was used for analysis. MCI was defined as being present if a participant had a Japanese version of the Montreal Cognitive Assessment score of <26. Adjustment was made for age, smoking status, alcohol consumption, leisure-time physical activity, body mass index, hypertension, dyslipidemia, diabetes mellitus, history of depression, number of teeth, employment, education, and household income. RESULTS:The prevalence of MCI was 45.3%. Among women, compared with the lowest tertile of TP, the second and highest tertiles were significantly associated with a lower prevalence of MCI with a clear dose-response relationship; the adjusted odds ratio (95% confidence intervals) in the second and highest tertiles of TP were 0.54 (0.36-0.83) and 0.55 (0.36-0.84), respectively (p for trend = 0.005). In contrast, no statistically significant association was observed between TP and the prevalence of MCI among men. CONCLUSIONS:Our findings suggest that higher TP might be inversely associated with the prevalence of MCI in middle-aged and older Japanese women.

Objectives: This study evaluated the efficacy of integrating a retrieval-augmented generation (RAG) model and a large language model (LLM) to improve the accuracy of drug name mapping across international vocabularies.Methods: Drug ingredient names were translated into English using the Japanese Accepted Names for Pharmaceuticals. Drug concepts were extracted from the standard vocabulary of OHDSI, and the accuracy of mappings between translated terms and RxNorm was assessed by vector similarity, using the BioBERT-generated embedded vectors as the baseline. Subsequently, we developed LLMs with RAG that distinguished the final candidates from the baseline. We assessed the efficacy of the LLM with RAG in candidate selection by comparing it with conventional methods based on vector similarity.Results: The evaluation metrics demonstrated the superior performance of the combined LLM + RAG over traditional vector similarity methods. Notably, the hit rates of the Mixtral 8x7b and GPT-3.5 models exceeded 90%, significantly outperforming the baseline rate of 64% across stratified groups of PO drugs, injections, and all interventions. Furthermore, the r-precision metric, which measures the alignment between model judgment and human evaluation, revealed a notable improvement in LLM performance, ranging from 41% to 50% compared to the baseline of 23%.Conclusions: Integrating an RAG and an LLM outperformed conventional string comparison and embedding vector similarity techniques, offering a more refined approach to global drug information mapping.

Abstract Disclosure: Y. Furukawa: None. K. Tanaka: None. O. Isozaki: None. A. Suzuki: None. T. Iburi: None. K. Tsuboi: None. M. Iguchi: None. N. Kanamoto: None. K. Minamitani: None. S. Wakino: None. T. Satoh: None. S. Teramukai: None. E. Kimura: None. Y. Miyake: None. T. Akamizu: None. Context: The mortality rate in thyroid storm (TS) has been reported to be higher than 10%. Even when patients survive, some have irreversible damages. In order to improve the prognosis of patients with TS, appropriate management as well as prompt and accurate diagnosis are needed. We created the diagnostic criteria in 2012 and established guidelines for the management of TS in 2016. Objective: We aimed to evaluate the effectiveness of the 2016 guidelines for the management of TS proposed by the Japan Thyroid Association and Japan Endocrine Society. Design: Prospective registry-based study through a secure web platform. Setting: Prospective multicenter registry. Patients and Measurements: Patients with new-onset TS were registered in the Research Electronic Data Capture (REDCap) from May 1, 2018 to April 30, 2022, over four years. On day 30 after admission, clinical information and prognosis of each patient were added to the platform. On day 180, the prognosis was described. The questionnaire consisted of 176 items totally. Results: This study included 110 patients with TS (93 definite and 17 suspected). The median of Acute Physiology and Chronic Health Evaluation (APACHE) II score was 13, higher than the score in the previous nationwide survey, 10 (p=0.001). Nonetheless, the mortality rate at day 30 was 5.5%, approximately half compared with 10.7% in the previous nationwide survey (p=0.13). Lower BMI, lower left ventricular ejection fraction (LVEF) and shock were positively associated with poor prognosis at day 30. The lack of fever ≥ 38°C was surprisingly associated to the outcome. Approximately 80% of doctors followed the guidelines. The mortality rate in patients with an APACHE II score ≥ 12 for whom the guidelines were not followed was significantly higher than the rate in patients for whom the guidelines were followed (50% vs. 4.7%) (p=0.01). Conclusions: Prognosis seemed better than in the previous nationwide survey, even though disease severity was higher. The mortality rate was lower when the guidelines were followed. Thus, the guidelines are useful for managing TS. Presentation: 6/2/2024

Context The mortality rate in thyroid storm (TS) has been reported to be higher than 10%.Objective We aimed to evaluate the effectiveness of the 2016 guidelines for the management of TS proposed by the Japan Thyroid Association and Japan Endocrine Society.Methods In this prospective multicenter registry-based study, patients with new-onset TS were registered in the Research Electronic Data Capture (REDCap), a secure web platform. On day 30 after admission, clinical information and prognosis of each patient were added to the platform. On day 180, the prognosis was described.Results This study included 110 patients with TS. The median of Acute Physiology and Chronic Health Evaluation (APACHE) II score was 13, higher than the score (10) in the previous nationwide epidemiological study (P = .001). Nonetheless, the mortality rate at day 30 was 5.5%, approximately half compared with 10.7% in the previous nationwide survey. Lower body mass index, shock, and lower left ventricular ejection fraction were positively associated with poor prognosis at day 30, while the lack of fever >= 38 degrees C was related to the outcome. The mortality rate in patients with an APACHE II score >= 12 for whom the guidelines were not followed was significantly higher than the rate in patients for whom the guidelines were followed (50% vs 4.7%) (P = .01).Conclusion Prognosis seemed better than in the previous nationwide survey, even though disease severity was higher. The mortality rate was lower when the guidelines were followed. Thus, the guidelines are useful for managing TS.

INTRODUCTION:Epidemiological evidence regarding the relationship between smoking and secondhand smoke (SHS) exposure and carotid intima-media thickness (CIMT) has been limited in Asian populations. Employing baseline data from the Aidai Cohort Study, Japan, we evaluated the evidence in this cross-sectional study.METHODS:Study subjects were 727 men aged 35-88 years and 1297 women aged 34-85 years. Information on smoking, SHS exposure, and confounders was obtained through a self-administered questionnaire. An automated carotid ultrasonography device was used to measure the right and left CIMT. The greatest CIMT measurement in the left or right common carotid artery was considered the maximum CIMT, and a maximum CIMT >1.0 mm was indicative of carotid wall thickening. Age, alcohol consumption, leisure time physical activity, hypertension, dyslipidemia, diabetes mellitus, body mass index, waist circumference, employment, and education level were adjusted at one time.RESULTS:The prevalence of carotid wall thickening was 13.0%. The prevalence of never smoking was 30.5% in men and 90.1% in women. Among those who had never smoked, the prevalence of never SHS exposure at home and work was 74.3% and 48.2% in men and 38.3% and 56.3% in women, respectively. Active smoking and pack-years of smoking were independently positively related to carotid wall thickening regardless of sex, although the association with current smoking in women was not significant. Independent positive relationships were shown between former smoking and pack-years of smoking and maximum CIMT in men but not in women. No significant relationships were found between SHS exposure at home and work and carotid wall thickening or maximum CIMT in either men or women.CONCLUSIONS:Active smoking, especially pack-years of smoking, was positively associated with carotid wall thickening in both sexes. Such positive associations with maximum CIMT were found only in men; however, interactions between smoking and sex were not significant.

Objectives Linking information on Japanese pharmaceutical products to global knowledge bases (KBs) would enhance international collaborative research and yield valuable insights. However, public access to mappings of Japanese pharmaceutical products that use international controlled vocabularies remains limited. This study mapped YJ codes to RxNorm ingredient classes, providing new insights by comparing Japanese and international drug-drug interaction (DDI) information using a case study methodology.Materials and Methods Tables linking YJ codes to RxNorm concepts were created using the application programming interfaces of the Kyoto Encyclopedia of Genes and Genomes and the National Library of Medicine. A comparative analysis of Japanese and international DDI information was thus performed by linking to an international DDI KB.Results There was limited agreement between the Japanese and international DDI severity classifications. Cross-tabulation of Japanese and international DDIs by severity showed that 213 combinations classified as serious DDIs by an international KB were missing from the Japanese DDI information.Discussion It is desirable that efforts be undertaken to standardize international criteria for DDIs to ensure consistency in the classification of their severity.Conclusion The classification of DDI severity remains highly variable. It is imperative to augment the repository of critical DDI information, which would revalidate the utility of fostering collaborations with global KBs.

: As the amount of biomedical and healthcare data increases, data mining for medicine becomes more and more important for health improvement. At the same time, privacy concerns in data utilization have also been growing. The key concepts for privacy protection are k -anonymity and differential privacy, but k -anonymity alone cannot protect personal presence information, and differential privacy alone would leak the identity. To promote data sharing throughout the world, universal methods to release the entire data while satisfying both concepts are required, but such a method does not yet exist. Therefore, we propose a novel privacy-preserving method, ( ε , k ) -Randomized Anonymization. In this paper, we first present two methods that compose the Randomized Anonymization method. They perform k -anonymization and randomized response in sequence and have adequate randomness and high privacy guarantees, respectively. Then, we show the algorithm for ( ε , k ) -Randomized Anonymization, which can provide highly accurate outputs with both k -anonymity and differential privacy. In addition, we describe the analysis procedures for each method using an inverse matrix and expectation-maximization (EM) algorithm. In the experiments, we used real data to evaluate our methods’ anonymity, privacy level, and accuracy. Furthermore, we show several examples of analysis results to demonstrate high utility of the proposed methods.

ObjectivesWe evaluated the feasibility of using deep learning with a convolutional neural network for predicting bone mineral density (BMD) and bone microarchitecture from conventional computed tomography (CT) images acquired by multivendor scanners.MethodsWe enrolled 402 patients who underwent noncontrast CT examinations, including L1-L4 vertebrae, and dual-energy x-ray absorptiometry (DXA) examination. Among these, 280 patients (3360 sagittal vertebral images), 70 patients (280 sagittal vertebral images), and 52 patients (208 sagittal vertebral images) were assigned to the training data set for deep learning model development, the validation, and the test data set, respectively. Bone mineral density and the trabecular bone score (TBS), an index of bone microarchitecture, were assessed by DXA. BMDDL and TBSDL were predicted by deep learning with a convolutional neural network (ResNet50). Pearson correlation tests assessed the correlation between BMDDL and BMD, and TBSDL and TBS. The diagnostic performance of BMDDL for osteopenia/osteoporosis and that of TBSDL for bone microarchitecture impairment were evaluated using receiver operating characteristic curve analysis.ResultsBMD(DL) and BMD correlated strongly (r = 0.81, P < 0.01), whereas TBSDL and TBS correlated moderately (r = 0.54, P < 0.01). The sensitivity and specificity of BMDDL for identifying osteopenia or osteoporosis were 93% and 90%, and 100% and 94%, respectively. The sensitivity and specificity of TBSDL for identifying patients with bone microarchitecture impairment were 73% for all values.ConclusionsThe BMDDL and TBSDL derived from conventional CT images could identify patients who should undergo DXA, which could be a gatekeeper tool for detecting latent osteoporosis/osteopenia or bone microarchitecture impairment.

Aim: Epidemiological evidence regarding the relationship between fish and fatty acid intake and carotid intima–media thickness (CIMT) has been limited and inconsistent. The current cross-sectional study investigated this issue using baseline data from the Aidai Cohort Study. Methods: Study subjects were 2024 Japanese men and women aged 34–88 years. Right and left CIMT were measured at the common carotid artery using an automated carotid ultrasonography device. Maximum CIMT was defined as the largest CIMT value in either the left or right common carotid artery. Carotid wall thickening was defined as a maximum CIMT value ＞1.0 mm. Results: The prevalence of carotid wall thickening was 13.0%. In men, intake of n-3 polyunsaturated fatty acids (PUFA) was independently positively related to the prevalence of carotid wall thickening, while no associations were found between intake of fish and the other fatty acids and carotid wall thickening or maximum CIMT. In women, intake levels of fish, n-3 PUFA, eicosapentaenoic acid, docosahexaenoic acid, and arachidonic acid were independently inversely associated with carotid wall thickening and intake levels of fish, n-3 PUFA, α-linolenic acid, n-6 PUFA, and linoleic acid were independently inversely associated with the maximum CIMT. No significant relationships were found between intake of total fat, saturated fatty acids, or monounsaturated fatty acids and carotid wall thickening or maximum CIMT regardless of sex. Conclusions: In women, higher intake of fish and n-3 and n-6 PUFA may be associated with a lower prevalence of carotid wall thickening and a decrease in maximum CIMT.

Objectives: The relationship between alanine-glyoxylate aminotransferase 2 (AGXT2) single-nucleotide polymorphisms (SNPs) and diabetes mellitus (DM) has not been investigated. Therefore, we performed a case-control study to examine this relationship. Methods: The study subjects included 2,390 Japanese men and women aged 34 to 88 years. In total, 190 cases were defined as having a fasting plasma glucose level >126 mg/dL, having a glycated hemoglobin >6.5% or currently using diabetic medication. The 2,200 remaining participants served as control subjects.Results: Compared with study subjects with the CC genotype of AGXT2 SNP rs37369, those with the TT, but not CT, genotype had a significantly increased risk of DM: the adjusted odds ratio (OR) for the TT genotype was 1.83 (95% confidence interval [CI], 1.04 to 3.47). AGXT2 SNPs rs37370 and rs180749 were not significantly associated with the risk of DM. The CTA haplotype of rs37370, rs37369 and rs180749 was significantly positively associated with the risk of DM (crude OR, 1.25; 95% CI, 1.01 to 1.56), whereas the CCA haplotype was significantly inversely related to DM (crude OR, 0.53; 95% CI, 0.27 to 0.95). The multiplicative interaction between AGXT2 SNP rs37369 and smoking status with regard to the risk of DM was not significant (p=0.32 for interaction).Conclusions: This is the first study to show significant associations between AGXT2 SNP rs37369, the CTA haplotype, and the CCA haplotype and DM. No interaction with regard to the risk of DM was observed between rs37369 and smoking.(c) 2022 The Author(s). Published on behalf of the Canadian Diabetes Association. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

This study used data from a large-scale multicenter medical information database in Japan to estimate the incidence rate of herpes zoster (HZ) and to examine the relationship between hypertension, dyslipidemia, and diabetes mellitus (DM), and the risk of HZ among patients with rheumatoid arthritis (RA). The research dataset consisted of 221,196 records of potential target patients with RA extracted between April 1, 2008 and August 31, 2017 from the Medical Data Vision database. To assess the association between hypertension, dyslipidemia, and DM and the risk of HZ, a case-control study was set up. Records of 101,498 study subjects met the inclusion criteria. During the observation period, 2566 patients developed HZ and the overall incidence rate was 5.2 (95% confidence interval: 5.0-5.4 per 1000 patient-years). Hypertension, dyslipidemia, and DM were significantly associated with an increased risk of HZ after adjustment for sex, age, hospital size, and use of anti-rheumatic drugs. When mutual adjustment was made for hypertension, dyslipidemia, and DM, the positive associations between hypertension and dyslipidemia and the risk of HZ remained significant; however, the positive association with DM completely disappeared. RA patients with hypertension or dyslipidemia may be at higher risk of HZ.

Changes in the 11th revision of the International Classification of Diseases (ICD-11) from ICD-10 may significantly impact coding quality. We conducted a field trial in 2017 to evaluate the line coding quality of 19 cases coded using the coding method of the World Health Organization. The cases with low agreement between the accuracy rates of ICD-10 and ICD-11 were cases that required the extension code. We should prepare effective educational content about how to use the extension code for proper coding in ICD-11.

In this big data era, a high-speed web access technique is expected for sharing a variety of large-scale data, especially over an international network. There are several high-speed web access techniques, such as data caching, data compression, data visualization, and intelligent data processing. In addition, hypertext transfer protocol (HTTP)/1.1 provides a persistent connection for better throughput than the HTTP/1.0. However, using these techniques developed so far, there are many users who are not satisfied with the current web access speed. In this paper, we propose a novel protocol handler for international web accesses based on a high-speed data transfer protocol, namely high-performance and flexible protocol (HpFP). We first design our protocol handler on the basis of HTTP, and then implement it for the Internet Explorer browser in Windows operating system (OS). Using a laboratory experiment, we evaluate the performance of our protocol handler in terms of throughput on a real-time web for weather satellite data. The results show that the proposed protocol handler achieves higher throughput than the HTTP for website requests on international networks. The improvement gets more significant as either the latency or the packet loss ratio becomes higher. This implies that the proposed protocol handler especially devotes to international web accesses for big data.