Objective: Currently, Risk Management Plans (RMPs), plan that provide new risk information about drugs, are published on the Pharmaceutical and Medical Devices Agency (PMDA) website.  The aim of this study was to compare enumerated risks in RMPs to the product labeling (PL) of the drug package insert.
Methods: The risks listed in RMPs on the PMDA website were assessed on February 10, 2014.  We investigated the documentation of these risks on the PL.
Results: Seven-hundred and eighty-five risks were enumerated in the RMPs of 77 drugs.  The enumerated risks were classified as “important identified risks” (66%), “important potential risks” (22%), and “important missing information” (12%).  Ninety-four percent of risks listed in RMPs were documented on the PL.  A portion of both the “important identified risks” and “important potential risks” groups were not documented on the PL.
Conclusion: This study was clearly the relation between risks listed in RMPs and documents on the PL.  Because a portion of the risks listed in RMPs was not documented on the PL, RMPs provide more safety information.  It is necessary to better understanding their characteristics, considering RMPs are a new source of drug information.

Objective: Similarity in drug appearance is one of the major environmental factors influencing dispensing errors, such as picking the wrong medication (picking error).  The purpose of this study is to verify if the index values of appearance similarity calculated objectively for multiple-specification drugs are the factors of picking error.
Methods: Four variables (number of total prescription, deviation of prescriptions between the specifications, sheet size, and color similarity of the sheet surface) were calculated.  The number of total prescription and deviation of prescriptions were extracted from the dispensing system.  Sheet size and color similarity were calculated, respectively, from the area ratio and by the Histogram Intersection method using the press through package (PTP) sheet image.  To evaluate the relationship between the picking error rate and these four variables, univariate and multivariate analyses were performed.
Results: The number of total prescription and the deviation of prescriptions were not significant factors.  In contrast, sheet size and color similarity significantly influenced the picking error rates.
Conclusion: Similarity in appearance between multiple-specification drugs is a risk factor of picking error.  When the multiple-specification pair has the same sheet size or high color similarity, one needs to be caution of picking error.  Further, in the pharmaceutical industry, to reduce the risk of dispensing errors, it is desirable to carry out the devise to enhance the identity of each specification.

Decreased adherence to medications among the elderly has become a problem in recent years. To overcome this problem, the dispensing of one-dose packages of medicines has been suggested. The purpose of this study was to survey elderly outpatients in a regional hospital at the Tokachi Development and Promotion Bureau and to analyze the factors that influence patients’ preference for a one-dose package medicine. The dispensation of one-dose package medication was observed in 20% of elderly outpatients. A correlation between the percentage of dispensed one-dose package medications and the number of agents was observed. Furthermore, dispensing rates of one-dose package medications increased with aging. Within the various departments analyzed, one-dose packaging rate was high in the Departments of Cardiology and Psychiatry. In this study, the correlation between residential municipalities of outpatients and one-dose package medications was examined. In the municipalities with a high rate of aged individuals and high average of household members, one-dose package medication rate was low. Multivariate logistic regression analysis confirmed these factors as significant. That one-dose packaging rates increased with the number of drugs and aging are consistent with the notion that one-dose package medications are designed to increase drug adherence among the elderly. Interestingly, one-dose package dispensing rate was low in areas with advanced aging. Because aging in Japan is expected to advance in the future, it is important to survey one-dose package dispensations.

Objective: Currently, the creation of a pharmaceutical risk management plan (RMP) for new drug information is obliged to pharmaceutical companies.  The created RMP is published on the Pharmaceuticals and Medical Devices Agency (PMDA) website.  RMP is a useful information source to ensure drug safety by healthcare professionals, including pharmacists.  “Risk minimization activities” of the RMP are especially important elements for healthcare professionals because they describe measures to minimize risk to patients.  We conducted a cross-sectional survey of the description of the contents of “risk minimization activities” in the RMP.
Methods: The RMP of 177 drugs that had been published in February 22, 2016 were investigated.
Results: Total risks enumerated for the study drugs were 1,678.  “Routine risk minimization activities” constituted 92.0% of total risks.  The most listed item on “routine risk minimization activities” was “attention on the product labeling of the drug package insert” (91.3%).  Differences in the expression level on “attention on the product labeling” were observed.  On the other hand, the most listed item of “additional risk minimization activities” was “the creation of documents for healthcare professionals” (38.3%) and “implementation of Early Post-marketing Phase Vigilance” (27.1%).
Conclusion: A clear understanding of RMP by healthcare professionals is important.  In the RMP, “risk minimization activities” (especially “additional risk minimization activities”) are the most important contents for healthcare professionals, because they include information of documents created by the pharmaceutical company for patient safety.  The level of description of the contents of RMP varies between drugs.  It is essential that these descriptions be uniform the expression level to be easily and accurately utilized by healthcare professionals.

Objective: Similarity of drug names is one factor of dispensing incidents.  The aim of this study was to survey the relation between sensual similarities of drug names and the occurrence of taking errors for pharmacists who actually prepare medicine.
Methods: A pair of drugs (15 incident pairs and 104 control pairs) was displayed on a computer screen at random.  The subject’s task was to determine the sensual similarity of them.  Thirteen pharmacists who prepared these pairs and caused their incidents participated in the experiment.
Results: The result showed that the sensual similarity of drug names of incident pairs was found to be highly significant in comparison to one of the control pairs [p=0.026].  However, the similarity in incident pairs is not necessarily high.  It was suggested that the similarity of drug name was not the only factor of taking error.  Multiple linear regression analyses of the sensual similarity in control pairs were performed, in which 10 variables were reported as quantitative indicators of similarity of drug name and were able to be measured on the internet.  The correlation was good [R²=0.828].  However, this regression model was not useful when adjusting to incident pairs.  In incident pairs, the similarity value calculated by the regression model was lower than the measured sensual similarity.
Conclusion: The result suggested that measured sensual similarity includes other risk factors of taking error, such as appearance similarity and/or efficacy similarity and/or short distance arrangement.  It seemed that the pharmacist’s ability complicated the factor of taking error.

Objective: Taking the wrong medicine or medication error is a serious concern to patient safety.  The aim of this study was to statistically survey the relation between the placement of drugs on medicine shelf and the occurrence of error in taking a medicine.
Methods: The study comprised 2 groups.  The incident group contained 43 cases that were erroneously taken in the Obihiro Kosei General Hospital.  The control group contained 43 drug pairs matched by the similarity index of the drug names from among the drugs used in the hospital at random.  The similarity index of drug names was based on 10 quantitative indicators.  The distance of medicine shelf arrangement was represented by three variables: the horizontal distance, the vertical distance and the distance of shelf block.  Conditional logistic regression analyses of the occurrence of medication errors were performed by evaluating the three variables of the distance factor and their interaction for error in taking a similar-sounding named drugs.
Results: Conditional logistic regression analysis revealed that the vertical distance (OR: 0.64, 95%CI: 0.42-0.99) and the distance of the shelf block (OR: 0.74, 95%CI: 0.57-0.97) were significant risk-reduction factors of medication errors.  Four variables were extracted as the most suitable logistic regression model in terms of the interaction between them.  As the interaction between 3 variables (the horizontal distance, the vertical distance and the distance of shelf block) was significant (OR: 0.93, 95%CI: 0.86-0.99), they may be considered as synergistic risk-reduction factors.  Moreover, the horizontal distance was found to be a risk-enhancement factor (OR: 1.52, 95%CI: 0.93-2.48).
Discussion: In order to reduce the risk of medication errors due to similar-sounding drug names, placement of drug on the medicine shelf should take into consideration the three coordinates of the distance factor.

Purpose: The objective of this study was to assess the performance of ChatGPT (GPT-4) on all items, including those with diagrams, in the Japanese National License Examination for Pharmacists (JNLEP) and compare it with the previous GPT-3.5 model’s performance. Methods: The 107th JNLEP, conducted in 2022, with 344 items input into the GPT-4 model, was targeted for this study. Separately, 284 items, excluding those with diagrams, were entered into the GPT-3.5 model. The answers were categorized and analyzed to determine accuracy rates based on categories, subjects, and presence or absence of diagrams. The accuracy rates were compared to the main passing criteria (overall accuracy rate ≥62.9%). Results: The overall accuracy rate for all items in the 107th JNLEP in GPT-4 was 72.5%, successfully meeting all the passing criteria. For the set of items without diagrams, the accuracy rate was 80.0%, which was significantly higher than that of the GPT-3.5 model (43.5%). The GPT-4 model demonstrated an accuracy rate of 36.1% for items that included diagrams. Conclusion Advancements that allow GPT-4 to process images have made it possible for LLMs to answer all items in medical-related license examinations. This study’s findings confirm that ChatGPT (GPT-4) possesses sufficient knowledge to meet the passing criteria.

Purpose: The objective of this study was to assess the performance of ChatGPT (GPT-4) on all items, including those with diagrams, in the Japanese National License Examination for Pharmacists (JNLEP) and compare it with the previous GPT-3.5 model’s performance. Methods: The 107th JNLEP, conducted in 2022, with 344 items input into the GPT-4 model, was targeted for this study. Separately, 284 items, excluding those with diagrams, were entered into the GPT-3.5 model. The answers were categorized and analyzed to determine accuracy rates based on categories, subjects, and presence or absence of diagrams. The accuracy rates were compared to the main passing criteria (overall accuracy rate ≥62.9%). Results: The overall accuracy rate for all items in the 107th JNLEP in GPT-4 was 72.5%, successfully meeting all the passing criteria. For the set of items without diagrams, the accuracy rate was 80.0%, which was significantly higher than that of the GPT-3.5 model (43.5%). The GPT-4 model demonstrated an accuracy rate of 36.1% for items that included diagrams. Conclusion Advancements that allow GPT-4 to process images have made it possible for LLMs to answer all items in medical-related license examinations. This study’s findings confirm that ChatGPT (GPT-4) possesses sufficient knowledge to meet the passing criteria.

Purpose: The objective of this study was to assess the performance of ChatGPT (GPT-4) on all items, including those with diagrams, in the Japanese National License Examination for Pharmacists (JNLEP) and compare it with the previous GPT-3.5 model’s performance. Methods: The 107th JNLEP, conducted in 2022, with 344 items input into the GPT-4 model, was targeted for this study. Separately, 284 items, excluding those with diagrams, were entered into the GPT-3.5 model. The answers were categorized and analyzed to determine accuracy rates based on categories, subjects, and presence or absence of diagrams. The accuracy rates were compared to the main passing criteria (overall accuracy rate ≥62.9%). Results: The overall accuracy rate for all items in the 107th JNLEP in GPT-4 was 72.5%, successfully meeting all the passing criteria. For the set of items without diagrams, the accuracy rate was 80.0%, which was significantly higher than that of the GPT-3.5 model (43.5%). The GPT-4 model demonstrated an accuracy rate of 36.1% for items that included diagrams. Conclusion Advancements that allow GPT-4 to process images have made it possible for LLMs to answer all items in medical-related license examinations. This study’s findings confirm that ChatGPT (GPT-4) possesses sufficient knowledge to meet the passing criteria.

Purpose: The objective of this study was to assess the performance of ChatGPT (GPT-4) on all items, including those with diagrams, in the Japanese National License Examination for Pharmacists (JNLEP) and compare it with the previous GPT-3.5 model’s performance. Methods: The 107th JNLEP, conducted in 2022, with 344 items input into the GPT-4 model, was targeted for this study. Separately, 284 items, excluding those with diagrams, were entered into the GPT-3.5 model. The answers were categorized and analyzed to determine accuracy rates based on categories, subjects, and presence or absence of diagrams. The accuracy rates were compared to the main passing criteria (overall accuracy rate ≥62.9%). Results: The overall accuracy rate for all items in the 107th JNLEP in GPT-4 was 72.5%, successfully meeting all the passing criteria. For the set of items without diagrams, the accuracy rate was 80.0%, which was significantly higher than that of the GPT-3.5 model (43.5%). The GPT-4 model demonstrated an accuracy rate of 36.1% for items that included diagrams. Conclusion Advancements that allow GPT-4 to process images have made it possible for LLMs to answer all items in medical-related license examinations. This study’s findings confirm that ChatGPT (GPT-4) possesses sufficient knowledge to meet the passing criteria.