Comparing rates of adverse events detected in incident reporting and the Global Trigger Tool: a systematic review.

Many hospitals continue to use incident reporting systems (IRSs) as their primary patient safety data source. The information IRSs collect on the frequency of harm to patients [adverse events (AEs)] is generally of poor quality, and some incident types (e.g. diagnostic errors) are under-reported. Other methods of collecting patient safety information using medical record review, such as the Global Trigger Tool (GTT), have been developed. The aim of this study was to undertake a systematic review to empirically quantify the gap between the percentage of AEs detected using the GTT to those that are also detected via IRSs. The review was conducted in adherence to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. Studies published in English, which collected AE data using the GTT and IRSs, were included. In total, 14 studies met the inclusion criteria. All studies were undertaken in hospitals and were published between 2006 and 2022. The studies were conducted in six countries, mainly in the USA (nine studies). Studies reviewed 22 589 medical records using the GTT across 107 institutions finding 7166 AEs. The percentage of AEs detected using the GTT that were also detected in corresponding IRSs ranged from 0% to 37.4% with an average of 7.0% (SD 9.1; median 3.9 and IQR 5.2). Twelve of the fourteen studies found <10% of the AEs detected using the GTT were also found in corresponding IRSs. The >10-fold gap between the detection rates of the GTT and IRSs is strong evidence that the rate of AEs collected in IRSs in hospitals should not be used to measure or as a proxy for the level of safety of a hospital. IRSs should be recognized for their strengths which are to detect rare, serious, and new incident types and to enable analysis of contributing and contextual factors to develop preventive and corrective strategies. Health systems should use multiple patient safety data sources to prioritize interventions and promote a cycle of action and improvement based on data rather than merely just collecting and analysing information.

Nuclear medicine incident reporting in Australia: control charts and notification rates inform quality improvement.

BACKGROUND: Australia has a statutory incident reporting system for radiopharmaceutical maladministrations, but additional research into registry data is required for the purpose of quality improvement in nuclear medicine. AIMS: We (i) used control charts to identify factors contributing to special cause variation (indicating higher than expected rates) in maladministrations and (ii) evaluated the impact of heterogeneous notification criteria and extent of underreporting among jurisdictions and individual facilities, respectively. METHODS: Anonymised summaries of Australian Radiation Incident Register reports permitted calculation of national monthly maladministration notification rates for 2007-2012 and preparation of control charts. Multivariate logistic regression assessed the association of population, insurance and regulatory characteristics with maladministration notifications in each Australian State and Territory. Maladministration notification rates from two facilities with familiarity of notification processes and commitment to radiation protection were compared with those elsewhere. RESULTS: Special cause variation occurred in only 3 months, but contributed to 21% of all incidents (42 of 197 patients), mainly because of 'clusters' of maladministrations (n = 24) arising from errors in bulk radiopharmaceutical dispensing. Maladministration notification rates varied significantly between jurisdictions (0 to 12.2 maladministrations per 100 000 procedures (P < 0.05)) and individual facilities (31.7 vs 5.8 per 100 000; χ(2) = 40; 1 degree of freedom, P < 0.001). CONCLUSIONS: Unexpected increases in maladministration notifications predominantly relate to incident 'clusters' affecting multiple patients. The bulk preparation of radiopharmaceuticals is a vulnerable process and merits additional safeguards. Maladministration notification rates in Australia are heterogeneous. Adopting uniform maladministration notification criteria among States and Territories and methods to overcome underreporting are warranted.

The missing evidence: a systematic review of patients' experiences of adverse events in health care.

PURPOSE: Preventable patient harm due to adverse events (AEs) is a significant health problem today facing contemporary health care. Knowledge of patients' experiences of AEs is critical to improving health care safety and quality. A systematic review of studies of patients' experiences of AEs was conducted to report their experiences, knowledge gaps and any challenges encountered when capturing patient experience data. DATA SOURCES: Key words, synonyms and subject headings were used to search eight electronic databases from January 2000 to February 2015, in addition to hand-searching of reference lists and relevant journals. STUDY SELECTION: Titles and abstracts of publications were screened by two reviewers and checked by a third. Full-text articles were screened against the eligibility criteria. DATA EXTRACTION: Data on design, methods and key findings were extracted and collated. RESULTS: Thirty-three publications demonstrated patients identifying a range of problems in their care; most commonly identified were medication errors, communication and coordination of care problems. Patients' income, education, health burden and marital status influence likelihood of reporting. Patients report distress after an AE, often exacerbated by receiving inadequate information about the cause. Investigating patients' experiences is hampered by the lack of large representative patient samples, data over sufficient time periods and varying definitions of an AE. CONCLUSION: Despite the emergence of policy initiatives to enhance patient engagement, few studies report patients' experiences of AEs. This information must be routinely captured and utilized to develop effective, patient-centred and system-wide policies to minimize and manage AEs.

Understanding the hospital sharps injury reporting pathway.

BACKGROUND: Patient-care workers are frequently exposed to sharps injuries, which can involve the risk of serious illness. Underreporting of these injuries can compromise prevention efforts. METHODS: We linked survey responses of 1,572 non-physician patient-care workers with the Occupational Health Services (OHS) database at two academic hospitals. We determined whether survey respondents who said they had sharps injuries indicated that they had reported them and whether reported injuries were recorded in the OHS database. RESULTS: Respondents said that they reported 62 of 78 sharps injuries occurring over a 12-month period. Only 28 appeared in the OHS data. Safety practices were positively associated with respondents' saying they reported sharps injuries but not with whether reported injuries appeared in the OHS data. CONCLUSIONS: Administrators should consider creating reporting mechanisms that are simpler and more direct. Administrators and researchers should attempt to understand how incidents might be lost before they are recorded.

Medical Errors and Barriers to Reporting in Ten Hospitals in Southern Iran.

BACKGROUND: International research shows that medical errors (MEs) are a major threat to patient safety. The present study aimed to describe MEs and barriers to reporting them in Shiraz public hospitals, Iran. METHODS: A cross-sectional, retrospective study was conducted in 10 Shiraz public hospitals in the south of Iran, 2013. Using the standardised checklist of Shiraz University of Medical Sciences (referred to the Clinical Governance Department and recorded documentations) and Uribe questionnaire, we gathered the data in the hospitals. RESULTS: A total of 4379 MEs were recorded in 10 hospitals. The highest frequency (27.1%) was related to systematic errors. Besides, most of the errors had occurred in the largest hospital (54.9%), internal wards (36.3%), and morning shifts (55.0%). The results revealed a significant association between the MEs and wards and hospitals (p < 0.001). Moreover, individual and organisational factors were the barriers to reporting ME in the studied hospitals. Also, a significant correlation was observed between the ME reporting barriers and the participants' job experiences (p < 0.001). CONCLUSION: The medical errors were highly frequent in the studied hospitals especially in the larger hospitals, morning shift and in the nursing practice. Moreover, individual and organisational factors were considered as the barriers to reporting MEs.

Impacts of Just Culture on Perioperative Nurses' Attitudes and Behaviors with regard to Patient Safety Incident Reporting: Cross-sectional Nationwide Survey.

PURPOSE: Just culture refers to a culture that encourages members of an organization to exchange important safety information and compensates them when they perform such information exchanges. The establishment of a just culture in hospital organizations might be an important means of enhancing patient safety incident reporting. This study aimed to investigate the impact of just culture on the attitudes and behaviors toward patient safety incident reporting in perioperative nurses. METHODS: A nationwide cross-sectional survey was performed using structured questionnaires. The participants were 208 perioperative nurses in tertiary general hospitals in South Korea. Data were collected by self-reported on-line questionnaires, from August to September 2020. Data were analyzed with descriptive statistics, independent t-test, chi-square test, Fisher's exact test, one-way ANOVA, Scheffé test, Pearson's correlation analysis, Spearman rank correlation analysis, hierarchical multiple regression, and hierarchical logistic regression using the SPSS WIN 23.0 program. RESULTS: Hierarchical multiple regression analysis revealed that just culture explained an additional 34.5%p of the attitudes on patient safety incident reporting. Hierarchical logistic regression analysis showed that just culture was a significant predictor of behaviors regarding patient safety incident reporting (odds ratio = 2.25, p = 0.017). The final regression model accounted for 16.0% of the behaviors regarding patient safety incident reporting. CONCLUSIONS: This study empirically shows that just culture impacted the attitudes and behaviors regarding patient safety incident reporting in perioperative nurses. This study provides an evidence about the importance of the just culture in every day nursing practice setting. Personnel and organizational efforts for improving or implementing just culture are required to ensure greater patient safety by enhancing the patient safety incident reporting of perioperative nurses in hospitals.

Improving Resident Physician Participation in Reporting Patient Safety and Quality Concerns.

Background: Reporting medical errors, near misses, and adverse events is an important component of improving patient safety and resident learning. Studies have revealed that event reporting rates can be low for physicians, resident physicians, and fellows. The objective of this quality improvement project was to improve resident reporting of patient safety and quality events and engage residents in the analysis of events at a community-based teaching hospital in the United States. Methods: We developed a program to engage 122 residents from 6 Accreditation Council for Graduate Medical Education-accredited residency programs using a multifaceted approach that included instructing residents how to use the hospital's adverse event reporting system; requiring first-year residents to submit at least 1 report; reviewing all resident reports during a monthly multidisciplinary meeting; and ensuring that each resident who submitted a report received feedback on how the concern was being addressed. Results: The program resulted in a 41.8% (95% CI 31%-53%) absolute increase in the number of residents reporting a concern, and resident submissions led to several documented improvements in patient care. A survey was administered to the residents who submitted reports, and the majority (76.0% response rate) expressed satisfaction with both the reporting system and the feedback about how their submission was being addressed. The responding residents agreed that they were more likely to submit reports because of their experience with the program and that they felt the program would improve safety and the quality of care at the institution. Conclusion: This quality improvement project successfully increased resident event reporting and engaged residents in the review of submitted events. The program can serve as a model for other teaching hospitals.

In-patient falls: what can we learn from incident reports?

BACKGROUND: a previous analysis of 12 months data from the National Reporting and Learning System offered useful insights on contributory factors for patient falls but was limited due to the small data set of free-text analysis (n = 400). A subsequent pilot study of 4,571 reports found an apparent difference in the contributory factors for patients described as having cognitive and physical impairments. OBJECTIVE: to analyse 3 years national incident data (2005-08) to further explore the contributory factors of in-patient falls. METHODS: a total of 20,036 reports (15% sample) were analysed by coding the free-text data field. Contributory risk factors were compared with the whole sample and explored with the Chi-squared and Fisher's exact tests. RESULTS: data were reported about the degree of harm (100% of reports), (un)witnessed status of fall (78%), location (47%), patient activity (27%), physical impairment/frailty (9.5%) and cognitive impairment/confusion (9.2%). Less than 0.1% of reports provided data about dizziness, illness, vision/hearing, and medicines. Overall, patients were more likely to be harmed when away from the bed space, mobilising/walking and by falling from the bed when not intending to leave the bed. CONCLUSIONS: this analysis explored incident reports at a level of detail not previously achieved. It identifies significant contributory factors for fall locations and activities associated with physical and cognitive characteristics.

Patient safety incident-reporting items in Korean hospitals.

OBJECTIVE: To examine incident-reporting items in tertiary hospitals using the framework of the World Health Organization's International Classification for Patient Safety (ICPS). DESIGN: Cross-sectional survey. SETTING AND PARTICIPANTS: Forty acute-care tertiary hospitals in Korea (response rate = 90.9%). METHODS: Data were collected using a semistructured questionnaire during on-site interviews or via e-mail. Items were extracted from incident-reporting forms that required a reporter's input, and were analyzed using the ICPS framework. After removing redundant items, unique reporting items were mapped onto ICPS elements. The data are summarized using descriptive statistics. RESULTS: On average, hospitals used 2.4 incident-reporting forms (range = 1-9) and 136.7 reporting items (range = 31-310). All of the hospitals had incident-reporting items that described 'incident type' and 'incident characteristics'; however, only 7 hospitals (17.5%) had reporting items on incident 'detection', and 18 hospitals (45.0%) collected information on the 'organizational outcomes'. Of the 1145 unique reporting items, 297 (25.9%) were completely mapped onto ICPS elements at different levels of granularity, and 12.7% (n = 145) were mapped onto ICPS elements that had more granular subcategories. CONCLUSIONS: The ICPS framework is a useful reference model for the classification of incident-reporting items. However, further refinements to both the ICPS framework and incident-reporting items are needed in order to better represent data on patient safety. Furthermore, the use of a common reporting form at the national level is recommended for reducing variations in reporting items and facilitating the efficient collection and analysis of patient safety data.

Comparing the Safety Action Feedback and Engagement (SAFE) Loop with an established incident reporting system: Study protocol for a pragmatic cluster randomized controlled trial.

Background: Incident reporting is widely used in hospitals to improve patient safety, but current reporting systems do not function optimally. The utility of incident reports is limited because hospital staff may not know what to report, may fear retaliation, and may doubt whether administrators will review reports and respond effectively. Methods: This is a clustered randomized controlled trial of the Safety Action Feedback and Engagement (SAFE) Loop, an intervention designed to transform hospital incident reporting systems into effective tools for improving patient safety. The SAFE Loop has six key attributes: obtaining nurses' input about which safety problems to prioritize on their unit; focusing on learning about selected high-priority events; training nurses to write more informative event reports; prompting nurses to report high-priority events; integrating information about events from multiple sources; and providing feedback to nurses on findings and mitigation plans. The study will focus on medication errors and randomize 20 nursing units at a large academic/community hospital in Los Angeles. Outcomes include: (1) incident reporting practices (rates of high-priority reports, contributing factors described in reports), (2) nurses' attitudes toward incident reporting, and (3) rates of high-priority events. Quantitative analyses will compare changes in outcomes pre- and post-implementation between the intervention and control nursing units, and qualitative analyses will explore nurses' experiences with implementation. Conclusion: If effective, SAFE Loop will have several benefits: increasing nurses' engagement with reporting, producing more informative reports, enabling safety leaders to understand problems, designing system-based solutions more effectively, and lowering rates of high-priority patient safety events.

Reasons for medication administration errors, barriers to reporting them and the number of reported medication administration errors from the perspective of nurses: A cross-sectional survey.

The aim of the study was to identify the reasons for medication administration errors, describe the barriers in their reporting and estimate the number of reported medication administration errors. BACKGROUND: Providing quality and safe healthcare is a key priority for all health systems. Medication administration error belongs to the more common mistakes committed in nursing practice. Prevention of medication administration errors must therefore be an integral part of nursing education. DESIGN: A descriptive and cross-sectional design was used for this study. METHODS: Sociological representative research was carried out using the standardized Medication Administration Error Survey. The research study involved 1205 nurses working in hospitals in the Czech Republic. Field surveys were carried out in September and October 2021. Descriptive statistics, Pearson's and Chi-square automatic interaction detection were used to analyze the data. The STROBE guideline was used. RESULTS: Among the most frequent causes of medication administration errors belong name (4.1 ± 1.4) and packaging similarity between different drugs (3.7 ± 1.4), the substitution of brand drugs by cheaper generics (3.6 ± 1.5), frequent interruptions during the preparation and administration of drugs (3.6 ± 1.5) and illegible medical records (3.5 ± 1.5). Not all medication administration errors are reported by nurses. The reasons for non-reporting of such errors include fear of being blamed for a decline in patient health (3.5 ± 1.5), fear of negative feelings from patients or family towards the nurse or legal liability (3.5 ± 1.6) and repressive responses by hospital management (3.3 ± 1.5). Most nurses (two-thirds) stated that less than 20 % of medication administration errors were reported. Older nurses reported statistically significantly fewer medication administration errors concerning non-intravenous drugs than younger nurses (p < 0.001). At the same time, nurses with more clinical experience (≥ 21 years) give significantly lower estimates of medication administration errors than nurses with less clinical practice (p < 0.001). CONCLUSION: Patient safety training should take place at all levels of nursing education. The standardized Medication Administration Error survey is useful for clinical practice managers. It allows for the identification of medication administration error causes and offers preventive and corrective measures that can be implemented. Measures to reduce medication administration errors include developing a non-punitive adverse event reporting system, introducing electronic prescriptions of medicines, involving clinical pharmacists in the pharmacotherapy process and providing nurses with regular comprehensive training.

[Professional and psychological impact of second victims among hospital healthcare professionals]. / Impacto laboral y psicológico de las segundas víctimas entre los profesionales sanitarios hospitalarios.

OBJECTIVE: To analyze the occupational and psychological consequences suffered by healthcare workers who are considered second victims (SV). MATERIAL AND METHODS: Observational, descriptive and cross-sectional study among the healthcare workers of a university hospital. The answers collected in a specifically designed questionnaire about psychological consequences at work and the result of a post-traumatic stress scale, "Impact of Event Scale-Revised (IES-R, spanish version)" were evaluated. The variables between the groups were compared using the Chi square test (or Fisher's exact test) when both were qualitative and with the Student's T (or the Mann-Whitney U test for independent data), when one of them was quantitative. The level of statistical significance was P<.05. RESULTS: 75.5% (148/207) of the participants in the study suffered some adverse event (AE) and, of these, 88.5% (131/148) were considered SV. Physicians had a 2.2 times higher risk of feeling SV than nurses (95% CI: 1.88-2.52). The impact on the patient related to the AE explained why the professionals involved in it felt SV (P=.037). 80.6% (N=104) of the SVs presented post-traumatic stress. Women were 2.4 times more likely to suffer from it (OR: 2.4; 95% CI: 1.5-4.0). Intrusive thoughts in the SV were almost three times more frequent when the damage suffered by the patient was permanent or death (OR: 2.5; 95% CI: 0.2-3.6). CONCLUSIONS: Many healthcare workers, especially physicians, considered themselves to be SV, and many of them suffered from post-traumatic stress. The impact on the patient related to the AE was a risk factor for being SV and for suffering psychological consequences.

Annotation Guidelines for Medication Errors in Incident Reports: Validation Through a Mixed Methods Approach.

At present no adequate annotation guidelines exists for incident report learning. This study aims at utilizing multiple quantitative and qualitative evidence to validate annotation guidelines for incident reporting of medication errors. Through multiple approaches via annotator training, annotation performance evaluation, exit surveys, and user and expert interviews, a mixed methods explanatory sequential design was utilized to collect 2-stage evidence for validation. We recruited two patient safety experts to participate in piloting, three annotators to receive annotation training and provide user feedback, and two incident report system designers to offer expert comments. Regarding the annotation performance evaluation, the overall accuracy reached 97% and 90% for named entity identification and attribute identification respectively. Participants provided invaluable comments and opinions towards improving the annotation methods. The mixed methods approach created a significant evidential basis for the use of annotation guidelines for incident report of medication errors. Further expansion of the guidelines and external validity present options for future research.

Threats to Narcotic Safety-A Narrative Review of Narcotic Incidents, Discrepancies and Near-Misses Within a Large Canadian Health System.

BACKGROUND: Canada is currently experiencing an opioid crisis. PURPOSE: Nurses are the largest number of frontline healthcare professionals in Canada who administer narcotic pharmacotherapy, hence, they are ideally placed to improve narcotic stewardship in hospitals. Our study aims to understand the characteristics of narcotic incidents and hence recommend interventions for narcotic stewardship. METHODS: Our study was conducted within a 442-bed academic health sciences center in Ontario. We extracted anonymized narcotic incident reports which occurred over a 3-year period from the SAFER System. Descriptive statistics were utilized to analyze narcotic incidents and their contributory factors. RESULTS: 272 narcotic incident reports were submitted to SAFER within the study period. Most incidents (51%) involved hydromorphone and morphine and were primarily categorized as Level I (n = 154) and Level II (n = 60). Incorrect narcotic dosing (44%), and narcotic count discrepancies (27%) were most commonly reported with active failures being the most commonly reported contributory factors such as failure to review medication orders prior to narcotic administration. CONCLUSIONS: Nurses have an important role in narcotic safety as an intermediary between narcotic administration and incident reporting. Further research is needed to understand the enablers, barriers and opportunities for nurses and other healthcare professionals to improve narcotic stewardship.

Critical incidents rates and types in Italian Intensive Care Units: A five-year analysis.

OBJECTIVE: To describe rates and types of critical incidents in Intensive Care Units. RESEARCH METHODOLOGY: A retrospective study in four intensive care units of an Academic Hospital located in the North-East of Italy. All critical incidents recorded in an incident reporting system database from 2013 to 2017 were collected. RESULTS: 160 critical incidents emerged. The rate was 1.7/100 intensive care-patient admissions, and 2.86/1000 in intensive care-patient days. Nurses reported most of the critical incidents (n = 113, 70.6%). In 2013 there were 19 (11.9%) critical incidents which significantly increased by 2017 (n = 38, 23.7%; p = 0.034). The most frequent critical incidents were medication/intravenous fluids issues (n = 35, 21.9%) and resources and organisational management (n = 35, 21.9%). Less frequently occurring incidents concerned medical devices/equipment (n = 29, 18.1%), clinical processes/procedures (n = 18, 11.3%), documentation (n = 14, 8.8%) and patient accidents (n = 13, 8.1%). Rare incidents included behaviour, clinical administration, nutrition, blood products and healthcare associated infection. CONCLUSION: Over a five-year period, documented incidents were steadily increasing in four Italian intensive care units. A voluntary incident reporting system might provide precious information on safety issues occurring in units. at both policy and professional levels.

Factors causing medication errors in an electronic reporting system.

AIM: To analyse medication error data from a hospital's electronic reporting system and identify the factors affecting error types and harmfulness. DESIGN: A retrospective study. METHODS: The 805 near misses and adverse events reported to the hospital's electronic reporting system between January 2014 and December 2018 were analysed using descriptive statistics, chi-square tests and logistic regression analyses. RESULTS: A total of 632 near misses and 173 adverse events were reported. Near misses and adverse events were the most common error type during the dispensing stage and medication administration, respectively. The odds of medication errors reported by nurses with 1-9 years of clinical experience were relatively low. After adjusting for confounders, the odds of medication errors directly observed by nurses were 65% lower than the odds of medication errors not directly detected. In clinical practice, nurses must be educated about errors in reporting depending on their degree of clinical experience.

Multicenter cross-sectional study on the reporting status and influencing factors of needlestick injuries caused by insulin injection devices among nurses in Peking, China.

BACKGROUND: Needlestick injuries (NSIs) most commonly occur in conjunction with insulin injections, and nurses had the highest rate of NSIs among different occupational groups. However, only few nurses have the awareness to report these incidents. We aimed to investigate the current reporting status and influencing factors of NSIs caused by insulin injection pen devices among nurses and nursing interns in Peking, China. METHODS: The study consisted of 4,609 nurses and nursing interns in different hospital departments, namely endocrinology wards, medical (except endocrinology) and surgical departments, emergency, outpatient departments, operation theaters, intensive care units, and other departments from 16 districts of Peking, China. A researcher-devised questionnaire was designed to assess the rate of reporting of accidental NSIs and its influencing factors. Descriptive statistics, the χ2 tests, and univariate and multivariate regression were used to analyze the independent influencing factors of reports on NSIs caused by insulin injection pen devices. RESULTS: Among all participants, most of them were women (97.61%). More than half of the participants had been working as nurses over 5 years (59.58%). Approximately 63.66% of participants had been trained to prevent and deal with accidental NSIs within 1 year. Approximately 19.33% of the participants experienced NSIs before, but only 30.30% of them reported the injury to the management department or director. The main reasons for not reporting were complex/cumbersome reporting procedure (317 of 621; 51.05%), being too busy at work at the time of injury (301 of 621; 48.47%), and low-risk for personal health (197 of 621; 31.72%). Multivariate analysis showed that age (P = .014; odds ratios [OR; 1.063, 1.736]), prior needlestick training education (P = .018; OR [0.406, 0.917]), and written system for preventing (P < .001; OR [0.289, 0.622]) were independent factors associated with reporting of NSIs caused by insulin injection pen devices. CONCLUSIONS: NSIs caused by insulin injection pen devices are common among nurses and are often not reported. It is necessary to provide nurses with regular training to prevent and deal with NSIs. Hospital management departments should adopt a simpler reporting procedure to understand the actual occurrence of NSIs to develop better prevention and improvement measures.

Medication-rights detection using incident reports: A natural language processing and deep neural network approach.

Medication errors often occurred due to the breach of medication rights that are the right patient, the right drug, the right time, the right dose and the right route. The aim of this study was to develop a medication-rights detection system using natural language processing and deep neural networks to automate medication-incident identification using free-text incident reports. We assessed the performance of deep neural network models in classifying the Advanced Incident Reporting System reports and compared the models' performance with that of other common classification methods (including logistic regression, support vector machines and the decision-tree method). We also evaluated the effects on prediction outcomes of several deep neural network model settings, including number of layers, number of neurons and activation regularisation functions. The accuracy of the models was measured at 0.9 or above across model settings and algorithms. The average values obtained for accuracy and area under the curve were 0.940 (standard deviation: 0.011) and 0.911 (standard deviation: 0.019), respectively. It is shown that deep neural network models were more accurate than the other classifiers across all of the tested class labels (including wrong patient, wrong drug, wrong time, wrong dose and wrong route). The deep neural network method outperformed other binary classifiers and our default base case model, and parameter arguments setting generally performed well for the five medication-rights datasets. The medication-rights detection system developed in this study successfully uses a natural language processing and deep-learning approach to classify patient-safety incidents using the Advanced Incident Reporting System reports, which may be transferable to other mandatory and voluntary incident reporting systems worldwide.

Near-miss analysis in a large hospital in southern Brazil: A 5-year retrospective study.

BACKGROUND: Near-miss analysis is an effective method for preventing serious adverse events, including never events such as wrong-site surgery. OBJECTIVE: To analyze all near-miss incidents reported in a large general hospital in southern Brazil between January 2013 and August 2017. METHOD: We performed a descriptive retrospective study of near-miss incidents recorded in the hospital's electronic reporting system in a large non-profit hospital (497 beds). The results are expressed as absolute (n) and relative frequencies (%). Pearson's chi-square test, Fisher's exact test (Monte Carlo simulation) and linear regression were used. RESULTS: A total of 12,939 near-miss incidents were recorded during the study period, with linear growth in the number of reports. Near-miss incidents were most frequent for medication, followed by processes unspecified in the International Classification for Patient Safety framework, followed by information control (patient chart and fluid balance data), followed by venous/vascular puncture. The highest prevalence of reports was observed in inpatient wards, in adult, pediatric, and neonatal intensive care units, and in the surgical center/post-anesthesia care unit. Pharmacists and nursing personnel recorded most of the reports during the day shift. CONCLUSION: The most frequent categories of near-miss incidents were medication processes, other institutional protocols, information control issues, and venous/vascular puncture. The significant number of reported near-miss incidents reflects good adherence to the reporting system.

An analysis of electronic health record-related patient safety incidents.

The aim of this study was to analyse electronic health record-related patient safety incidents in the patient safety incident reporting database in fully digital hospitals in Finland. We compare Finnish data to similar international data and discuss their content with regard to the literature. We analysed the types of electronic health record-related patient safety incidents that occurred at 23 hospitals during a 2-year period. A procedure of taxonomy mapping served to allow comparisons. This study represents a rare examination of patient safety risks in a fully digital environment. The proportion of electronic health record-related incidents was markedly higher in our study than in previous studies with similar data. Human-computer interaction problems were the most frequently reported. The results show the possibility of error arising from the complex interaction between clinicians and computers.