Journal of Medical Internet Research - International Scientific Journal for Medical Research, Information and Communication on the Internet #Preprint #PeerReviewMe: Warning: This is a unreviewed preprint. Readers are warned that the document has not been peer-reviewed by expert/patient reviewers or an academic editor, may contain misleading claims, and is likely to undergo changes before final publication, if accepted, or may have been rejected/withdrawn. Readers with interest and expertise are encouraged to sign up as peer-reviewer, if the paper is within an open peer-review period. Please cite this preprint only for review purposes or for grant applications and CVs (if you are the author). Background: Adverse drug reactions (ADRs) are common and they are the underlying cause of over a million serious injuries and deaths each year. The most familiar method to detect ADRs is relying on spontaneous reports. Unfortunately, the low reporting rate of spontaneous reports is a serious limitation of pharmacovigilance. Objective: Our objective was to identify a method to detect potential ADRs of drugs automatically using a deep neural network (DNN). Methods: We designed a DNN model that utilizes the chemical, biological, and biomedical information of drugs to detect ADRs. This model aimed to fulfil two main purposes, identifying the potential ADRs of drugs and predicting the possible ADRs of a new drug. For improving the detection performance, we distributed representations of the target drugs in a vector space to capture the drug relationships by using the word embedding approach to process amounts of biomedical literature. Moreover, we built a mapping function to address new drugs that do not appear in the dataset. Results: We predicted the ADRs of drugs recorded up to 2012, by using the drug information and the ADRs reported up to 2009. There were contained 746 drugs and 232 new drugs which only recorded in 2012 with 1,325 ADRs. The experimental results showed that the overall performance of our model with mean average precision (MAP) at top-10 is achieved 0.523 for ADR prediction on the dataset. Conclusions: Our model was effective in identifying the potential ADRs of a drug and the possible ADRs of a new drug. Most importantly, it can detect potential ADRs irrespective of whether they have been reported in the past.

Medication errors can have severe consequences and threaten patient safety. The patient safety-related benefits of automated dispensing cabinets (ADCs) have been reported by several previous studies, including a reduction in medication errors in intensive care units (ICUs) and emergency departments. However, the benefits of ADCs need to be assessed, given the different healthcare practice models. This study aimed to compare the rates of medication errors, including prescription, dispensing, and administrative, before and after using ADCs in intensive care units. The prescription, dispensing, and administrative error data before and after the adoption of ADCs were retrospectively collected from the medication error report system. The severity of medication errors was classified according to the National Coordinating Council for Medication Error Reporting and Prevention guidelines. The study outcome was the rate of medication errors. After the adoption of ADCs in the intensive care units, the rates of prescription and dispensing errors reduced from 3.03 to 1.75 per 100,000 prescriptions and 3.87 to 0 per 100,000 dispensations, respectively. The administrative error rate decreased from 0.046 to 0.026%. The ADCs decreased National Coordinating Council for Medication Error Reporting and Prevention category B and D errors by 75% and category C errors by 43%. To improve medication safety, multidisciplinary collaboration and strategies, such as the use of automated dispensing cabinets, education, and training programs from a systems perspective, are warranted.

Background: Geographic Information Systems (GIS) combined with spatial analytical methods could be helpful in examining patterns of drug use. Little attention has been paid to geographic variation of cardiovascular prescription use in Taiwan. The main objective was to use local spatial association statistics to test whether or not the cardiovascular medication-prescribing pattern is homogenous across 352 townships in Taiwan.Methods: The statistical methods used were the global measures of Moran'sIand Local Indicators of Spatial Association (LISA). While Moran'sIprovides information on the overall spatial distribution of the data, LISA provides information on types of spatial association at the local level. LISA statistics can also be used to identify influential locations in spatial association analysis. The major classes of prescription cardiovascular drugs were taken from Taiwan's National Health Insurance Research Database (NHIRD), which has a coverage rate of over 97%. The dosage of each prescription was converted into defined daily doses to measure the consumption of each class of drugs. Data were analyzed with ArcGIS and GeoDa at the township level.Results: The LISA statistics showed an unusual use of cardiovascular medications in the southern townships with high local variation. Patterns of drug use also showed more low-low spatial clusters (cold spots) than high-high spatial clusters (hot spots), and those low-low associations were clustered in the rural areas.Conclusions: The cardiovascular drug prescribing patterns were heterogeneous across Taiwan. In particular, a clear pattern of north-south disparity exists. Such spatial clustering helps prioritize the target areas that require better education concerning drug use.

There are limited population-based studies focusing on the chemopreventive effects of selective cyclooxygenase-2 (COX-2) inhibitors against colorectal cancer. The purpose of this study is to assess the trends and dose–response effects of various medication possession ratios (MPR) of selective COX-2 inhibitor used for chemoprevention of colorectal cancer.

Introduction Limited studies have evaluated the association between Clostridium difficile infection (CDI) and the duration of proton pump inhibitor (PPI) or histamine H2-receptor blocker (H2RA) use and provided a cutoff duration for PPI or H2RA use to mitigate a substantially increased risk of CDI. We aimed to evaluate these associations in hospitalized patients using a nationwide insurance claims database. Methods We conducted a nested case–control study to identify cases with a first ever record of CDI in a study cohort undergoing PPI or H2RA therapy from the National Health Insurance Database from 2012 to 2018. Each case was matched with one control by age, sex, and calendar year. We used conditional logistic regression to estimate the sensitivity, specificity, and area under the receiver operating characteristic curve (AUC ROC). Youden’s J statistic was used to identify the optimal cutoff duration in days for PPI or H2RA use. Results In the main analysis, the AUC ROC was 0.64 (95% CI 0.63–0.66) and optimal cutoff duration was 15 days for PPI users. The AUC ROC was 0.63 (95% CI 0.62–0.64) and optimal cutoff duration was 16 days for H2RA users. In the sensitivity analyses, the results were similar to those of the main analysis, and the optimal cutoff duration was in the range of 14–15 days. Conclusions The optimal cutoff duration for PPI and H2RA use was about 2 weeks. It is necessary to be cautious regarding the risk of CDI in patients taking PPIs or H2RAs for longer than 2 weeks.

Intravitreal bevacizumab therapy in preterm infants for retinopathy of prematurity (ROP) can be associated with hypotension. We report twin preterm infants who developed hypotension within 1 day after intravitreal bevacizumab therapy for ROP. Before receiving the medication, their clinical statuses were stable and similar. The dose, procedure, and premedication were the same; however, twin B presented with hypotension for 3 days. Although bevacizumab-related hypotension has been described in product information (incidence rate 7%–15%), this is the first case report of intravitreal bevacizumab for ROP inducing hypotension. Physicians should be aware of intravitreal bevacizumab therapy-related hypotension when treating ROP. We suggest conducting a postmarketing active surveillance on the systemic adverse effects of this regimen in preterm infants.