Objective: To establish a prediction model of qi stagnation referring to two existing models.Design: Prospective observational study.Setting: We recruited patients who visited the Kampo Clinic at Keio University from February 2011 to March 2013.Methods: We constructed a random forest algorithm with 202 items as independent variables to predict qi stagnation patterns using full agreement data of the physicians' diagnosis and the result of two existing scores as a reference standard. To compare the new model with the two existing models, we calculated the discriminant ratio (prediction accuracy), precision, sensitivity (recall), specificity, and F-measure of these models.Results: The number of eligible participants was 1,194, and 29.1% of them were diagnosed with qi stagnation by Kampo physicians. The discriminant ratio, precision, sensitivity, specificity, and F-measure in our new model were 0.960, 0.672, 0.911, 0.964, and 0.774, respectively. Our new model had a significantly higher discriminant ratio than the two existing models.Conclusions: We constructed a better qi stagnation prediction model than the previously established ones. Our results can be utilized to reach an international agreement on qi stagnation pattern diagnosis in traditional East Asian medicine.

Background: Prostate cancer is one of the most common cancers among men worldwide and the fourth most common cause of death. The number of prostate cancer cases and deaths is increasing every year because of population aging. This study aimed to clarify the risk of developing prostate cancer due to fluctuations in Prostate Specific Antigen (PSA) levels in patients without a history of prostate cancer using large medical information data. Results: This retrospective cohort included 1707 male patients aged 60 years or older who had a PSA level measurement date (2-PSA) within 3 months or more and 2 years from the first PSA level measurement date (1-PSA) in the database between 2008 and 2019. We subtracted 1-PSA from 2-PSA and designated patients with a higher 2-PSA than 1-PSA to the “up” group (n = 967) and patients with a lower 2-PSA than 1-PSA to the “down” group (n = 740). By using Cox proportional hazards model, a significant increase in prostate cancer risk was observed in the up group compared with the down group (adjusted hazard ratio [HR] = 1.82, 95% confidence interval [CI] = 1.21–2.72; adjusted for patient background factors). Subgroup analysis showed that patients with PSA levels < 4 ng/mL had a significantly increased risk of developing prostate cancer if the next PSA level increases by approximately 20% (adjusted HR = 2.94, 95% CI = 1.14–7.58), and patients with PSA levels of 4 ng/mL or higher if the next PSA level is decreased by approximately 20% had a significantly reduced risk of developing prostate cancer (adjusted HR = 0.36, 95% CI = 0.18–0.74), compared to that with no change. Conclusions: This is the first study to clarify the association between PSA variability and risk of developing prostate cancer in patients without a history of prostate cancer. These results suggest that the suppression of elevated PSA levels may lead to the prevention of prostate cancer and that it would be better to perform a biopsy because the risk of developing prostate cancer may increase in the future if the PSA value increases above a certain level.

Kampo medicine is widely used in Japan; however, most physicians and pharmacists have insufficient knowledge and experience in it. Although a chatbot-style system using machine learning and natural language processing has been used in some clinical settings and proven useful, the system developed specifically for the Japanese language using this method has not been validated by research. The purpose of this study is to develop a novel drug information provision system for Kampo medicines using a natural language classifier® (NLC®) based on IBM Watson. The target Kampo formulas were 33 formulas listed in the 17th revision of the Japanese Pharmacopoeia. The information included in the system comes from the package inserts of Kampo medicines, Manuals for Management of Individual Serious Adverse Drug Reactions, and data on off-label usage. The system developed in this study classifies questions about the drug information of Kampo formulas input by natural language into preset questions and outputs preset answers for the questions. The system uses morphological analysis, synonym conversion by thesaurus, and NLC®. We fine-tuned the information registered into NLC® and increased the thesaurus. To validate the system, 900 validation questions were provided by six pharmacists who were classified into high or low levels of knowledge and experience of Kampo medicines and three pharmacy students. The precision, recall, and F-measure of the system performance were 0.986, 0.915, and 0.949, respectively. The results were stable even with differences in the amount of expertise of the question authors. We developed a system using natural language classification that can give appropriate answers to most of the validation questions.

Recent experimental studies have examined the efficacy of statins in preventing cancer, but the findings of clinical studies are inconsistent, and studies on Japanese patients are limited. This study aimed to clarify the association between statins and cancer risk among Japanese patients. We conducted a large population-based retrospective cohort study using the Japanese health insurance claims database, including patients newly diagnosed with dyslipidemia between 2005 and 2015. Patients who were on newly prescribed statins during the study period were designated as statin users. They were matched 1:1 with randomly selected drug nonusers who were not prescribed drugs for dyslipidemia according to age, sex, and year of first diagnosis of dyslipidemia. There were 23,746 patients in each group. The mean duration of follow-up for statin users and drug nonusers was approximately 2 years. Using a Cox proportional hazards model, significant reduction in cancer risk was observed in statin users compared with that in drug nonusers [adjusted HR = 0.84; 95% confidence interval (CI), 0.72–0.97; adjusted for patient background factors]. The results of subgroup analyses suggested that prescribed statins reduced the incidence of cancer of the digestive organs (adjusted HR = 0.79; 95% CI, 0.63–0.99) as well as reduced cancer risk in patients with nonsmokers (adjusted HR = 0.78, 95% CI = 0.65–0.92). Our results suggest that statin use may reduce cancer risk in patients with dyslipidemia. Prevention Relevance: This study clarified the relationship between statin use and cancer risk in patients with dyslipidemia. Our study will contribute to medicine selection in patients with hypercholesterolemia level. See related Spotlight, p. 1