Construction of a Multi-Label Classifier for Extracting Multiple Incident Factors From Medication Incident Reports in Residential Care Facilities: Natural Language Processing Approach.

BACKGROUND: Medication safety in residential care facilities is a critical concern, particularly when nonmedical staff provide medication assistance. The complex nature of medication-related incidents in these settings, coupled with the psychological impact on health care providers, underscores the need for effective incident analysis and preventive strategies. A thorough understanding of the root causes, typically through incident-report analysis, is essential for mitigating medication-related incidents. OBJECTIVE: We aimed to develop and evaluate a multilabel classifier using natural language processing to identify factors contributing to medication-related incidents using incident report descriptions from residential care facilities, with a focus on incidents involving nonmedical staff. METHODS: We analyzed 2143 incident reports, comprising 7121 sentences, from residential care facilities in Japan between April 1, 2015, and March 31, 2016. The incident factors were annotated using sentences based on an established organizational factor model and previous research findings. The following 9 factors were defined: procedure adherence, medicine, resident, resident family, nonmedical staff, medical staff, team, environment, and organizational management. To assess the label criteria, 2 researchers with relevant medical knowledge annotated a subset of 50 reports; the interannotator agreement was measured using Cohen κ. The entire data set was subsequently annotated by 1 researcher. Multiple labels were assigned to each sentence. A multilabel classifier was developed using deep learning models, including 2 Bidirectional Encoder Representations From Transformers (BERT)-type models (Tohoku-BERT and a University of Tokyo Hospital BERT pretrained with Japanese clinical text: UTH-BERT) and an Efficiently Learning Encoder That Classifies Token Replacements Accurately (ELECTRA), pretrained on Japanese text. Both sentence- and report-level training were performed; the performance was evaluated by the F1-score and exact match accuracy through 5-fold cross-validation. RESULTS: Among all 7121 sentences, 1167, 694, 2455, 23, 1905, 46, 195, 1104, and 195 included "procedure adherence," "medicine," "resident," "resident family," "nonmedical staff," "medical staff," "team," "environment," and "organizational management," respectively. Owing to limited labels, "resident family" and "medical staff" were omitted from the model development process. The interannotator agreement values were higher than 0.6 for each label. A total of 10, 278, and 1855 reports contained no, 1, and multiple labels, respectively. The models trained using the report data outperformed those trained using sentences, with macro F1-scores of 0.744, 0.675, and 0.735 for Tohoku-BERT, UTH-BERT, and ELECTRA, respectively. The report-trained models also demonstrated better exact match accuracy, with 0.411, 0.389, and 0.399 for Tohoku-BERT, UTH-BERT, and ELECTRA, respectively. Notably, the accuracy was consistent even when the analysis was confined to reports containing multiple labels. CONCLUSIONS: The multilabel classifier developed in our study demonstrated potential for identifying various factors associated with medication-related incidents using incident reports from residential care facilities. Thus, this classifier can facilitate prompt analysis of incident factors, thereby contributing to risk management and the development of preventive strategies.

External Validation and Update of the Risk Prediction Model for Denosumab-Induced Hypocalcemia Developed From a Hospital-Based Administrative Database.

PURPOSE: Denosumab is used to treat patients with bone metastasis from solid tumors, but sometimes causes severe hypocalcemia, so careful clinical management is important. This study aims to externally validate our previously developed risk prediction model for denosumab-induced hypocalcemia by using data from two facilities with different characteristics in Japan and to develop an updated model with improved performance and generalizability. METHODS: In the external validation, retrospective data of Kameda General Hospital (KGH) and Miyagi Cancer Center (MCC) between June 2013 and June 2022 were used and receiver operating characteristic (ROC)-AUC was mainly evaluated. A scoring-based updated model was developed using the same data set from a hospital-based administrative database as previously employed. Selection of variables related to prediction of hypocalcemia was based on the results of external validation. RESULTS: For the external validation, data from 235 KGH patients and 224 MCC patients were collected. ROC-AUC values in the original model were 0.879 and 0.774, respectively. The updated model consisting of clinical laboratory tests (calcium, albumin, and alkaline phosphatase) afforded similar ROC-AUC values in the two facilities (KGH, 0.837; MCC, 0.856). CONCLUSION: We developed an updated risk prediction model for denosumab-induced hypocalcemia with small interfacility differences. Our results indicate the importance of using data from plural facilities with different characteristics in the external validation of generalized prediction models and may be generally relevant to the clinical application of risk prediction models. Our findings are expected to contribute to improved management of bone metastasis treatment.

[Evaluation of Patient Outcomes after Pharmacological Intervention in Home Health Care Utilizing Pharmaceutical Care Records].

The purpose of this study was to identify patient outcomes after pharmacist interventions in the home health care context using pharmaceutical care records accumulated during daily operations. We focused on 591 cases at Nakajima Pharmacy from April 2020 to December 2021, where dispensing fees were charged to prevent duplication of medication and unnecessary interactions of home patients (excluding those related to adjustment of ongoing medications). The study investigated the content and background of prescription changes, the follow-up rate, and patient outcomes. The most common circumstances that led to pharmacist intervention for homebound patients were symptom occurrence (uncontrolled symptom, new symptom, drug adverse event). Of the patients for whom pharmacist intervention was provided for symptoms, 72.8% received follow-up according to the pharmaceutical care records. Furthermore, 59.2% of patients with follow-up showed an improvement of their symptoms. In addition, many patients had their medications discontinued or the dosage reduced by the pharmacist despite stable symptoms. More than 90% of these patients showed no change in symptoms. Besides interventions associated with the occurrence of symptoms, many interventions related to medication adherence were found to result from the patient's physical condition, such as poor swallowing function. The results suggest that tracking pharmacy drug histories may help pharmacists to better understand the need for follow-up implementation and the changes in patient outcomes after interventions.

Concomitant use of lansoprazole and ceftriaxone is associated with an increased risk of ventricular arrhythmias and cardiac arrest in a large Japanese hospital database.

OBJECTIVES: To determine whether concomitant use of ceftriaxone and oral or intravenous lansoprazole increases the risk of ventricular arrhythmia and cardiac arrest in the real-world setting in Japan. METHODS: The data analyzed were obtained from the JMDC hospital-based administrative claims database for the period April 2014 to August 2022. Patients who received a proton pump inhibitor (PPI) while receiving ceftriaxone or sulbactam/ampicillin were identified. The frequency of ventricular arrhythmia and cardiac arrest was analyzed according to whether oral or intravenous PPI was concomitant with ceftriaxone or sulbactam/ampicillin. Estimates of the incidence of ventricular arrhythmia and cardiac arrest were then compared among the groups, using the Fine-Gray competing risk regression model. RESULTS: The results showed that the risk of ventricular arrhythmia and cardiac arrest was significantly higher with concomitant ceftriaxone and oral lansoprazole (hazard ratio 2.92, 95% confidence interval 1.99-4.29, P < 0.01) or intravenous lansoprazole (hazard ratio 4.57, 95% confidence interval 1.24-16.80, P = 0.02) than with concomitant sulbactam/ampicillin and oral or intravenous lansoprazole. CONCLUSIONS: Oral and intravenous lansoprazole may increase the risk of ventricular arrhythmia and cardiac arrest in patients who are receiving ceftriaxone.

A cross-sectional survey of hepatitis B virus screening in patients who received immunosuppressive therapy for rheumatoid arthritis in Japan.

BACKGROUND: Patients with a history of hepatitis B virus (HBV) infection who are receiving immunosuppressive therapy are at risk of HBV reactivation and disease. Therefore, HBV screening is required prior to administering antirheumatic drugs with immunosuppressive effects. This study aimed to determine the status of hepatitis B surface antigen (HBsAg), hepatitis B core antibody (HBcAb), and hepatitis B surface antibody (HBsAb) screening prior to the initiation of drug therapy, including new antirheumatic drugs, in patients with rheumatoid arthritis. METHODS: This retrospective cross-sectional study used data from April 2014 to August 2022 from the Japanese hospital-based administrative claims database. The inclusion criteria were rheumatoid arthritis and first prescription date of antirheumatic drugs. RESULTS: A total of 82,282 patients with rheumatoid arthritis who were first prescribed antirheumatic drugs between April 2016 and August 2022 were included. Of the eligible patients, 9.7% (n=7,959) were screened for all HBV (HBsAg, HBsAb, and HbcAb) within 12 months prior to the date of initial prescription. The HBsAg test was performed in 30.0% (n=24,700), HBsAb test in 11.8% (n=9,717), and HBcAb test in 13.1% (n=10,824) of patients. The proportion of patients screened for HBV infection has been increasing since 2018; however, the proportion of patients screened for rheumatoid arthritis remains low. CONCLUSIONS: Our findings suggest that HBV screening may be insufficient in patients who received antirheumatic drugs. With the increasing use of new immunosuppressive antirheumatic drugs, including biological agents, healthcare providers should understand the risk of HBV reactivation and conduct appropriate screening.

Factor Analysis of Patients Who Find Tablets or Capsules Difficult to Swallow Due to Their Large Size: Using the Personal Health Record Infrastructure of Electronic Medication Notebooks.

BACKGROUND: Understanding patient preference regarding taking tablet or capsule formulations plays a pivotal role in treatment efficacy and adherence. Therefore, these preferences should be taken into account when designing formulations and prescriptions. OBJECTIVE: This study investigates the factors affecting patient preference in patients who have difficulties swallowing large tablets or capsules and aims to identify appropriate sizes for tablets and capsules. METHODS: A robust data set was developed based on a questionnaire survey conducted from December 1, 2022, to December 7, 2022, using the harmo smartphone app operated by harmo Co, Ltd. The data set included patient input regarding their tablet and capsule preferences, personal health records (including dispensing history), and drug formulation information (available from package inserts). Based on the medication formulation information, 6 indices were set for each of the tablets or capsules that were considered difficult to swallow owing to their large size and concomitant tablets or capsules (used as controls). Receiver operating characteristic (ROC) analysis was used to evaluate the performance of each index. The index demonstrating the highest area under the curve of the ROC was selected as the best index to determine the tablet or capsule size that leads to swallowing difficulties. From the generated ROCs, the point with the highest discriminative performance that maximized the Youden index was identified, and the optimal threshold for each index was calculated. Multivariate logistic regression analysis was performed to identify the risk factors contributing to difficulty in swallowing oversized tablets or capsules. Additionally, decision tree analysis was performed to estimate the combined risk from several factors, using risk factors that were significant in the multivariate logistic regression analysis. RESULTS: This study analyzed 147 large tablets or capsules and 624 control tablets or capsules. The "long diameter + short diameter + thickness" index (with a 21.5 mm threshold) was identified as the best indicator for causing swallowing difficulties in patients. The multivariate logistic regression analysis (including 132 patients with swallowing difficulties and 1283 patients without) results identified the following contributory risk factors: aged <50 years (odds ratio [OR] 1.59, 95% CI 1.03-2.44), female (OR 2.54, 95% CI 1.70-3.78), dysphagia (OR 3.54, 95% CI 2.22-5.65), and taking large tablets or capsules (OR 9.74, 95% CI 5.19-18.29). The decision tree analysis results suggested an elevated risk of swallowing difficulties for patients with taking large tablets or capsules. CONCLUSIONS: This study identified the most appropriate index and threshold for indicating that a given tablet or capsule size will cause swallowing difficulties, as well as the contributory risk factors. Although some sampling biases (eg, only including smartphone users) may exist, our results can guide the design of patient-friendly formulations and prescriptions, promoting better medication adherence.

Adverse Event Signal Detection Using Patients' Concerns in Pharmaceutical Care Records: Evaluation of Deep Learning Models.

BACKGROUND: Early detection of adverse events and their management are crucial to improving anticancer treatment outcomes, and listening to patients' subjective opinions (patients' voices) can make a major contribution to improving safety management. Recent progress in deep learning technologies has enabled various new approaches for the evaluation of safety-related events based on patient-generated text data, but few studies have focused on the improvement of real-time safety monitoring for individual patients. In addition, no study has yet been performed to validate deep learning models for screening patients' narratives for clinically important adverse event signals that require medical intervention. In our previous work, novel deep learning models have been developed to detect adverse event signals for hand-foot syndrome or adverse events limiting patients' daily lives from the authored narratives of patients with cancer, aiming ultimately to use them as safety monitoring support tools for individual patients. OBJECTIVE: This study was designed to evaluate whether our deep learning models can screen clinically important adverse event signals that require intervention by health care professionals. The applicability of our deep learning models to data on patients' concerns at pharmacies was also assessed. METHODS: Pharmaceutical care records at community pharmacies were used for the evaluation of our deep learning models. The records followed the SOAP format, consisting of subjective (S), objective (O), assessment (A), and plan (P) columns. Because of the unique combination of patients' concerns in the S column and the professional records of the pharmacists, this was considered a suitable data for the present purpose. Our deep learning models were applied to the S records of patients with cancer, and the extracted adverse event signals were assessed in relation to medical actions and prescribed drugs. RESULTS: From 30,784 S records of 2479 patients with at least 1 prescription of anticancer drugs, our deep learning models extracted true adverse event signals with more than 80% accuracy for both hand-foot syndrome (n=152, 91%) and adverse events limiting patients' daily lives (n=157, 80.1%). The deep learning models were also able to screen adverse event signals that require medical intervention by health care providers. The extracted adverse event signals could reflect the side effects of anticancer drugs used by the patients based on analysis of prescribed anticancer drugs. "Pain or numbness" (n=57, 36.3%), "fever" (n=46, 29.3%), and "nausea" (n=40, 25.5%) were common symptoms out of the true adverse event signals identified by the model for adverse events limiting patients' daily lives. CONCLUSIONS: Our deep learning models were able to screen clinically important adverse event signals that require intervention for symptoms. It was also confirmed that these deep learning models could be applied to patients' subjective information recorded in pharmaceutical care records accumulated during pharmacists' daily work.

Platelets Affect the Activity of Amino Acid Transporter SNAT4 in HuH-7 Human Hepatoma Cells.

Platelets have been reported to exert diverse actions besides hemostasis and thrombus formation in the body. However, whether platelets affect transporter activity remains to be determined. In this study, we examined the effects of platelets on the activity of amino acid transporter system A, which is known to be changed by various factors, and we clarified the mechanism by which platelets affect system A activity. Among system A subtypes, we found that sodium-coupled neutral amino acid transporter (SNAT) 4 played a central role in the transport activity of system A in HuH-7 human hepatoma cells. Interestingly, platelets showed a biphasic effect on system A activity: activated platelet supernatants (APS) including the granule contents released from platelets downregulated system A activity at lower concentrations and the downregulation was suppressed at higher concentrations. The downregulation was due to a decrease in the affinity of SNAT4 for its substrate and not a decrease in the SNAT4 abundance on the plasma membrane. In addition, APS did not decrease the expression level of SNAT4 mRNA. On the other hand, platelets did not affect system A activity when the platelet suspension was added to HuH-7 cells. These results indicate that platelets indirectly affect the transport activity of system A by releasing bioactive substances but do not directly affect it by binding to HuH-7 cells.

Immunoregulatory Effects of Elemental Diet and Its Ingredient, Tryptophan, via Activation of the Aryl Hydrocarbon Receptor in Mice.

Inflammatory bowel disease (IBD) is characterized by chronic intestinal inflammation and its treatment varies widely; however, when inflammation is high, a complete nutrient containing pre-digested elemental diet (ED) is used to preserve the intestinal tract. In this study, we investigated the mechanisms underlying the effectiveness of EDs for IBD using mice. C57BL/6 mice were orally treated with the ED (5 mL/day) and its ingredient L-tryptophan (Trp) (1-100 mg/kg), respectively. Flow cytometry analysis revealed that treatment with the ED and Trp (10 and 100 mg/kg) significantly increased the percentage of splenic CD4+-/CD25+-/Foxp3+ regulatory T cells (Tregs). In the 2% DSS-induced colitis-mouse model, Trp administration (100 mg/kg) led to a significant decrease in TNF-α and increase in IL-10 in the serum as well as a significant decrease in the inflammation score. Furthermore, the aryl hydrocarbon receptor (AhR) agonistic activity, which is a key function of Treg induction, of Trp and 15 Trp metabolites was characterized using a highly sensitive DR-EcoScreen cell assay. Five Trp metabolites, including L-kynurenine, acted as AhR agonists, while Trp did not. Taken together, these results suggest that the ED treatment has a Trp-dependent immunoregulatory effect, and several Trp metabolites that activate the AhR might contribute to induction of remission in patients with IBD.

A Survey of the Status of Methadone Switching in Japan Using a Hospital-Based Administrative Claims Database.

Methadone is generally used for the management of cancer pain in patients who cannot obtain adequate analgesia from other strong opioids; however, it has a complicated and inconsistent conversion ratio from pre-switching opioid dosage to methadone. This issue may be pronounced in Japan because only oral tablets are commercially available. We aimed to elucidate the status of methadone switching in Japan, focusing on its dosage. Using a Japanese hospital-based administrative claims database, we included patients who switched to methadone between April 2008 and January 2021. The proportion of methadone switching completion that required more than the defined conversion ratio in the Japanese package insert (called "high-dose methadone switching") was evaluated as a primary endpoint. Other endpoints included "the duration from initiation to completion of methadone switching" and "factors affecting high-dose methadone switching by using multivariate logistic regression analysis". Of 1585 patients who received methadone, 370 were enrolled. Among those, 130 (35.1%) received high-dose methadone switching. The median duration of methadone switching completion (12 days) was longer in the high-dose methadone switching group than in other patients. Four variables were identified as factors affecting high-dose methadone switching. Younger age and outpatient status increased the risk of requiring high-dose methadone switching, whereas the concomitant use of nonsteroidal anti-inflammatory drugs and fentanyl as a pre-switching opioid decreased the risk. In conclusion, more than 30% of the patients underwent high-dose methadone switching and required long completion periods, suggesting that methadone switching remains challenging in Japan.

Comparison of different sustained-release opioids and acute respiratory conditions in patients with cancer and chronic kidney disease.

STUDY OBJECTIVE: Few data are available on the association between the use of oxycodone in patients with chronic kidney disease (CKD) and acute respiratory conditions. The aim of this study was to investigate whether oxycodone is associated with an increased risk of acute respiratory conditions in patients with cancer and CKD compared with other opioids. DESIGN AND SETTING: The data were obtained from a claims database in Japan. Patients with cancer and CKD who had received sustained-release opioids, including oral oxycodone, oral morphine, or transdermal fentanyl, between April 2014 and May 2021 were selected. The primary outcome was defined as an acute respiratory condition. Data for age and sex, morphine equivalent daily dose, concomitant use of specified medications, comorbidities defined based on the modified Charlson comorbidity index, substance use disorder, and lung cancer or metastatic lung cancer were investigated as covariates. Distribution of acute respiratory conditions was compared among the three sustained-release opioid groups using the log-rank test. Estimates of the incidence of acute respiratory conditions were compared among the groups using a Cox proportional hazards model with time-varying variables. MAIN RESULTS: A significant difference in the distribution of acute respiratory conditions was found among the three groups (p < 0.01). Cox regression analysis showed a significantly higher risk of acute respiratory conditions with morphine (hazard ratio [HR]: 3.04, 95% confidence interval [CI]: 1.07-8.65, p = 0.04) compared with oxycodone but no significant difference in risk with oxycodone (HR 0.67, 95% CI: 0.32-1.38, p = 0.27) compared with fentanyl. CONCLUSIONS: The findings suggest that the risk of acute respiratory conditions may be lower in patients with CKD who use oxycodone for cancer pain than in those who use morphine. Additionally, no difference in the risk of acute respiratory conditions was found between oxycodone and fentanyl use.

Serious Bacterial Infections Associated with Eculizumab: A Pharmacovigilance Study.

Objective Molecular-targeted agents, including eculizumab and rituximab, are considered treatment options for refractory myasthenia gravis (MG), but bacterial infections can occur as serious adverse events when using these agents. The present study elucidated the relative risks of bacterial infections associated with eculizumab and rituximab using a pharmacovigilance database. Methods We analyzed eculizumab- and rituximab-associated adverse events reported between 2007 and 2021 in the US Food and Drug Administration Adverse Event Reporting System (FAERS) and herein report a refractory MG patient who developed streptococcal toxic shock syndrome during eculizumab treatment. Patients We evaluated a 74-year-old Japanese woman with refractory MG who developed severe bacteremia after receiving eculizumab. Results A total of 44,215 and 108,485 adverse events were reported with eculizumab and rituximab, respectively, from among 13,742,321 individual case safety reports in the FAERS database after data cleaning. We found a strong association between eculizumab and Neisseria infections. In contrast, we found only one case of meningococcal meningitis treated with rituximab. Both eculizumab and rituximab were weakly associated with streptococcal infections. Two cases of streptococcal toxic shock syndrome were associated with rituximab. Conclusion Careful monitoring of serious bacterial infections associated with eculizumab treatment is warranted.

Comparison of the incidence of nausea and vomiting between linezolid and vancomycin using claims database: a retrospective cohort study.

BACKGROUND: Nausea and vomiting during linezolid therapy have been reported as part of safety analyses in clinical trials. We have previously examined the incidence of vomiting during linezolid therapy (18.1%). A previous study conducted at a single hospital showed low external validity. It is necessary to verify whether these results can be reproduced using generalizable data sources. AIM: To evaluate the incidence of nausea and vomiting during linezolid therapy compared with vancomycin using a Japanese claims database. METHOD: Patients administered linezolid or vancomycin were selected from the database between January 2005 and June 2017. The primary endpoint was the comparison of nausea and vomiting between the linezolid and vancomycin groups. We conducted propensity score matching (PSM) to adjust for patient characteristics. To assess risk factors for nausea and vomiting, logistic regression was conducted as the secondary endpoint. We defined nausea and vomiting as the first prescription of antiemetics during linezolid or vancomycin therapy as a surrogate endpoint. RESULTS: In total, 1215 patients were enrolled. After PSM, the number of patients in the linezolid and vancomycin groups was 241. Nausea and vomiting were observed in 11.2% and 5.0% of patients in the linezolid and vancomycin groups, respectively (p < 0.05). Linezolid administration was extracted as a risk factor for nausea and vomiting (odds ratio, 2.09; 95% confidence interval, 1.02-4.30). CONCLUSION: This study clarified the relationship between linezolid and nausea and vomiting using a Japanese claims database. Further studies are required to elucidate the unknown mechanisms of linezolid-induced nausea and vomiting.

Effects of Body Composition on Renal Function Estimates in Older Patients.

The modified Cockcroft-Gault (CG) equation, previously developed for an aged-oriented cohort, was validated in a newly obtained dataset. Estimates of creatinine clearance (CCr) using this equation were found to be more accurate than those determined using the conventional CG equation, particularly for patients exceeding 65 years of age. We identified a subset of patients in this cohort whose estimates were inadequate. Using statistical analysis, we found that the deviation from estimates was attributed to a decreased albumin level. In addition, we determined a reduced albumin cutoff value for the modified CG equation to obtain a good estimate. Univariate linear regression analysis was applied to measure the CCr in this cohort and identify parameters related to body composition, and we found that extracellular water (ECW)/total body water (TBW) and body fat (%) were relevant. Using measured values of ECW/TBW and body fat (%), a multivariate linear regression (MLR) estimating equation was developed based on the modified CG equation. This equation was applied to a cohort over 65 years of age, and it was found that a good estimate was obtained for older patients with low albumin levels. Thus, we propose a flow diagram that illustrates conditions for selecting an appropriate estimating equation from among the CG, modified CG, and MLR equations.

Influence of Voiced and Semi-voiced Sounds on the Subjective Similarity of Different Drug Names: A Cognitive Psychological Experiment.

The similarity of drug names is one of the common causes of medication error. In Japan, similarity evaluation is performed prior to approval of new drugs in order to avoid potential confusion. However, existing indices do not take account of the difference between characters that contain voiced or semi-voiced and unvoiced sounds, so it is not clear whether such sounds influence the subjective similarity of drug names. Thus, we performed a cognitive psychological experiment to investigate this issue, using participants who had not received any education in medicine, nursing, or pharmacy. An analogue scale questionnaire was used to evaluate the subjective similarity of the names of drug pairs. Drug pairs for the main analysis were prepared by matching the first 0 to 3 characters, and then varying the difference in the number of voiced and semi-voiced characters from 0 to 3 in these matched characters. By means of this procedure, the drug pairs were classified into a total of 10 groups. Then, a total of 60 drug pairs were created by assigning 6 drugs to each group. The subjective similarity tended to increase with increasing number of common characters among the first three characters. When classified according to the number of these common characters, the subjective similarity was significantly decreased when voiced or semi-voiced sounds were present, as compared with when they were absent. These results indicate that a new drug name similarity index that takes account of voiced and semi-voiced sound differences should be developed to minimize medication errors.

Development of a risk prediction model for surgical site infection after lower third molar surgery.

BACKGROUND: There is little evidence regarding risk prediction for surgical site infection (SSI) after lower third molar (L3M) surgery. METHODS: We conducted a nested case-control study to develop a multivariable logistic model for predicting the risk of SSI after L3M surgery. Data were obtained from Hokkaido University Hospital from April 2013 to March 2020. Multiple imputation was applied for the missing values. We conducted decision tree (DT) analysis to evaluate the combinations of factors affecting SSI risk. RESULTS: We identified 648 patients. The final model retained the available distal space (Pell & Gregory II [p = 0.05], Pell & Gregory III [p < 0.01]), depth (Pell & Gregory B [p < 0.01], Pell & Gregory C [p < 0.01]), surgeon's experience (3-10 years [p = 0.25], <3 years [p < 0.01]), and simultaneous extraction of both L3M [p < 0.01]; the concordance-statistic was 0.72. The DT analysis demonstrated that patients with Pell and Gregory B or C and simultaneous extraction of both L3M had the highest risk of SSI. CONCLUSIONS: We developed a model for predicting SSI after L3M surgery with adequate predictive metrics in a single center. This model will make the SSI risk prediction more accessible.

Risk Factor Analysis of Vancomycin-Induced Nephrotoxicity in Paediatric Patients Aged 0-1 Year Using Japanese Medical Database.

Vancomycin (VCM)-induced nephrotoxicity (VIN) is a major side effect in paediatric patients. However, most studies are limited to patients aged 0-18 years. We evaluated the risk factors of VIN in patients aged 0-1 year using Japanese electronic medical record database. We used RWD database which was contained electronic medical records and claims data of approximately 20 million people from 160 medical institutions. We targeted hospitalized patients who were administered VCM between June 2000 and December 2020. VIN was defined by two criteria: Criterion 1 was an increase in serum creatinine (Scr) ≥ 0.5 mg/dL or 50% during VCM treatment period compared to the Scr baseline; and criterion 2 was an increase in Scr ≥50% within seven days or Scr ≥0.3 mg/dL within two days during VCM treatment. The risk factors of VIN were evaluated using multivariate logistic regression analysis. We analysed 446 patients; patients with VIN in Criteria 1 and 2 were 33 and 58, respectively. In Criterion 1, multivariate logistic regression analysis identified four independent factors with p-value <0.05 (VCM concentration ≥20 mg/L, amphotericin B (AMPH-B), piperacillin-tazobactam (TAZ/PIPC), and vasopressor drugs). In Criterion 2, multivariate logistic regression analysis identified concomitant use of vasopressor drugs with p-value <0.05. Therefore, concomitant use of vasopressor drugs was suggested to affect the risk of VIN in patients aged 0-1 year. The findings may help in developing estimation models to assess the risk of VIN in paediatric patients.

[A Novel Approach to Analyze the Factors Affecting Adverse Drug Reactions by Combination of Electronic Medical Record Database and Machine Learning Method].

Decision tree analysis, a flowchart-like tree framework, is a typical machine learning method that is widely used in various fields. The most significant feature of this method is that independent variables (e.g., with or without concomitant use of vasopressor drugs) are extracted in order of the strength of their relationship with the dependent variable to be predicted (e.g., with or without adverse drug reactions), forming a tree-like model. Specifically, users can easily and quantitatively estimate the proportion of event occurrences considering "interrelationships among multiple combinations of factors" by answering the questions in the constructed flowchart. Previously, we applied the decision tree model to vancomycin-associated nephrotoxicity and demonstrated that this method can be used to analyze the factors affecting adverse drug reactions. However, the number of cases that can be analyzed decreases significantly as the number of branches increases. Thus, many cases are necessary to generate highly accurate findings. In attempt to solve this problem, we combined big data and decision tree analyses. In this review, we present the results of our research combining big data (electronic medical record database) and a machine learning method. Furthermore, we discuss the limitations of these methods and factors to consider when applying the results of big data and machine learning analyses to clinical practice.

Exploring Factors Affecting the Occurrence of Hypersensitivity Reactions Induced by Nonionic Iodine Contrast Media.

Hypersensitivity reactions induced by nonionic iodine contrast media sometimes occur and can be life threatening. However, independent factors affecting their occurrence remain to be fully established. Therefore, the purpose of this study was to clarify independent factors affecting the occurrence of hypersensitivity reactions induced by nonionic iodine contrast media. Patients who received nonionic iodine contrast media at Keiyu Hospital from April 2014 to December 2019 were included. The adjusted odds ratio (OR) and 95% confidence interval (CI) for factors affecting hypersensitivity reactions induced by contrast media were calculated by logistic regression analysis. The multiple imputation method was used to impute missing data. Hypersensitivity reactions occurred in 0.72% (163 cases) of 22,695 cases enrolled in this study. In univariate analysis, 10 variables met the criteria of P < .05 and proportion of missing data <50%. In multivariate analysis, age (OR, 0.98; 95% CI, 0.97-0.99), outpatient status (OR, 2.08; 95% CI, 1.20-3.60), contrast medium iodine content (OR, 1.02; 95% CI, 1.01-1.04), history of drug allergy (OR, 2.41; 95% CI, 1.50-3.88), and asthma (OR, 17.4; 95% CI, 7.53-40.1) were identified as independent factors affecting contrast media-induced hypersensitivity reactions. Among these factors, history of drug allergy and asthma appear to be clinically relevant and reliable due to their high OR and plausible biological mechanisms, but the other three factors require further validation.