BACKGROUND: Diagnostic errors are an underappreciated cause of preventable mortality in hospitals and pose a risk for severe patient harm and increase hospital length of stay.
OBJECTIVE: This study aims to explore the potential of machine learning and natural language processing techniques in improving diagnostic safety surveillance. We conducted a rigorous evaluation of the feasibility and potential to use electronic health records clinical notes and existing case review data.
METHODS: Safety Learning System case review data from 1 large health system composed of 10 hospitals in the mid-Atlantic region of the United States from February 2016 to September 2021 were analyzed. The case review outcome included opportunities for improvement including diagnostic opportunities for improvement. To supplement case review data, electronic health record clinical notes were extracted and analyzed. A simple logistic regression model along with 3 forms of logistic regression models (ie, Least Absolute Shrinkage and Selection Operator, Ridge, and Elastic Net) with regularization functions was trained on this data to compare classification performances in classifying patients who experienced diagnostic errors during hospitalization. Further, statistical tests were conducted to find significant differences between female and male patients who experienced diagnostic errors.
RESULTS: In total, 126 (7.4%) patients (of 1704) had been identified by case reviewers as having experienced at least 1 diagnostic error. Patients who had experienced diagnostic error were grouped by sex: 59 (7.1%) of the 830 women and 67 (7.7%) of the 874 men. Among the patients who experienced a diagnostic error, female patients were older (median 72, IQR 66-80 vs median 67, IQR 57-76; P=.02), had higher rates of being admitted through general or internal medicine (69.5% vs 47.8%; P=.01), lower rates of cardiovascular-related admitted diagnosis (11.9% vs 28.4%; P=.02), and lower rates of being admitted through neurology department (2.3% vs 13.4%; P=.04). The Ridge model achieved the highest area under the receiver operating characteristic curve (0.885), specificity (0.797), positive predictive value (PPV; 0.24), and F1-score (0.369) in classifying patients who were at higher risk of diagnostic errors among hospitalized patients.
CONCLUSIONS: Our findings demonstrate that natural language processing can be a potential solution to more effectively identifying and selecting potential diagnostic error cases for review and therefore reducing the case review burden.

The association between doctors\' long working hours and the seriousness of adverse events with high patient impact has not been fully confirmed. Most previous studies were based on work hour regulations using more than 80 hours per week as an indicator of long working hours. We aimed to assess the association using a shorter indicator as the cut-off for long working hours among hospital doctors including senior doctors. This cross-sectional study used 12,245 adverse event reports from the Japan Council for Quality Health Care. We defined long working hours as 55 hours or more in the week before the adverse event and assessed the association with the seriousness of adverse events with high patient impact. The results showed that doctors working 55 or more hours in the preceding week were more likely to be involved in serious adverse events than those working fewer hours (odds ratio (OR) 1.22, 95% confidence interval (CI): 1.12-1.32). This association remained significant after adjusting for all covariates (OR 1.18, 95% CI: 1.08-1.28). Senior doctors were more likely to be involved in serious adverse events. Long working hours among doctors were associated with the seriousness of adverse events.

BACKGROUND: At least 10% of hospital admissions in high-income countries, including Australia, are associated with patient safety incidents, which contribute to patient harm (\'adverse events\'). When a patient is seriously harmed, an investigation or review is undertaken to reduce the risk of further incidents occurring. Despite 20 years of investigations into adverse events in healthcare, few evaluations provide evidence of their quality and effectiveness in reducing preventable harm.This study aims to develop consistent, informed and robust best practice guidance, at state and national levels, that will improve the response, learning and health system improvements arising from adverse events.
METHODS: The setting will be healthcare organisations in Australian public health systems in the states of New South Wales, Queensland, Victoria and the Australian Capital Territory. We will apply a multistage mixed-methods research design with evaluation and in-situ feasibility testing. This will include literature reviews (stage 1), an assessment of the quality of 300 adverse event investigation reports from participating hospitals (stage 2), and a policy/procedure document review from participating hospitals (stage 3) as well as focus groups and interviews on perspectives and experiences of investigations with healthcare staff and consumers (stage 4). After triangulating results from stages 1-4, we will then codesign tools and guidance for the conduct of investigations with staff and consumers (stage 5) and conduct feasibility testing on the guidance (stage 6). Participants will include healthcare safety systems policymakers and staff (n=120-255) who commission, undertake or review investigations and consumers (n=20-32) who have been impacted by adverse events.
BACKGROUND: Ethics approval has been granted by the Northern Sydney Local Health District Human Research Ethics Committee (2023/ETH02007 and 2023/ETH02341).The research findings will be incorporated into best practice guidance, published in international and national journals and disseminated through conferences.

Safe delivery of care is a priority in dentistry, while basic epidemiological knowledge of patient safety incidents is still lacking. The objectives of this study were to (1) classify patient safety incidents related to primary dental care in Denmark in the period 2016-2020 and study the distribution of different types of dental treatment categories where harm occurred, (2) clarify treatment categories leading to \"nerve injury\" and \"tooth loss\" and (3) assess the financial cost of patient-harm claims. Data from the Danish Dental Compensation Act (DDCA) database was retrieved from all filed cases from 1st January 2016 until 31st December 2020 pertaining to: (1) The reason why the patient applied for treatment-related harm compensation, (2) the event that led to the alleged harm (treatment category), (3) the type of patient-harm, and (4) the financial cost of all harm compensations. A total of 9069 claims were retrieved, of which 5079 (56%) were found eligible for compensation. The three most frequent categories leading to compensation were \"Root canal treatment and post preparation\"(n = 2461, 48% of all approved claims), \"lack of timely diagnosis and initiation of treatment\" (n = 905, 18%) and \"surgery\" (n = 878, 17%). Damage to the root of the tooth accounted for more than half of all approved claims (54.36%), which was most frequently a result of either parietal perforation during endodontic treatment (18.54%) or instrument fracture (18.89%). Nerve injury accounted for 16.81% of the approved claims. Total cost of all compensation payments was €16,309,310, 41.1% of which was related to surgery (€6,707,430) and 20.4% (€3,322,927) to endodontic treatment. This comprehensive analysis documents that harm permeates all aspects of dentistry, especially in endodontics and surgery. Neglect or diagnostic delays contribute to 18% of claims, indicating that harm does not solely result from direct treatment. Treatment harm inflicts considerable societal costs.

BACKGROUND: Drug-drug interactions (DDIs) have potential to cause patient harm, including lowering therapeutic efficacy. This study aimed to (i) determine the prevalence of potential DDIs (pDDIs); clinically relevant DDIs (cDDIs), that is, DDIs that could lead to patient harm, taking into account a patient\'s individual clinical profile, drug effects and severity of potential harmful outcome; and subsequent actual harm among hospitalized patients and (ii) examine the impact of transitioning from paper-based medication charts to electronic medication management (eMM) on DDIs and patient harms.
METHODS: This was a secondary analysis of the control arm of a controlled pre-post study. Patients were randomly selected from three Australian hospitals. Retrospective chart review was conducted before and after the implementation of an eMM system, without accompanying clinical decision support alerts for DDIs. Harm was assessed by an expert panel.
RESULTS: Of 1186 patient admissions, 70.1% (n = 831) experienced a pDDI, 42.6% (n = 505) a cDDI and 0.9% (n = 11) an actual harm in hospital. Of 15,860 pDDIs identified, 27.0% (n = 4285) were classified as cDDIs. The median number of pDDIs and cDDIs per 10 drugs were 6 [interquartile range (IQR) 2-13] and 0 (IQR 0-2), respectively. In cases where a cDDI was identified, both drugs were 44% less likely to be co-administered following eMM (adjusted odds ratio 0.56, 95% confidence interval 0.46-0.73).
CONCLUSIONS: Although most patients experienced a pDDI during their hospital stay, less than one-third of pDDIs were clinically relevant. The low prevalence of harm identified raises questions about the value of incorporating DDI decision support into systems given the potential negative impacts of DDI alerts.

Clinical trial documents are complex and may have inconsistencies, leading to potential site implementation errors and may compromise participant safety. This study characterizes the frequency and type of administrative and potential patient safety interventions (PPSIs) made during the review of oncology trial documents for clinical trial implementation by centralized clinical content specialists.
A dedicated group of centralized clinical content specialists reviewed trial documents, including the protocol, laboratory manual, and pharmacy/cellular therapy manual, and collected intervention data over a 1-year period. Each trial was categorized by study phase and sponsor type, and multiple interventions could be identified per trial. Interventions were deemed administrative or PPSIs, with PPSIs further subcategorized as medication, laboratory, procedure related, or other.
Of 585 clinical trials reviewed, 269 (46%) required intervention(s). Among 1001 interventions, 171 (17.1%) were PPSIs. Most PPSIs were medication related (45.6%), with drug dosing interventions most frequently identified (53.8%). Phase 1 trials had the highest proportion of PPSIs (0.35:1) and administrative interventions (2:1) per trial compared with all other phases. Investigator-initiated trials saw the highest proportion of PPSIs per trial (0.44:1) of all sponsor types.
This study demonstrates a gap in patient safety when assessing trial documents for clinical trial implementation. One solution to address this gap is the utilization of a centralized team of clinical specialists to preemptively review trial documents, thereby enhancing patient safety during clinical trial conduct.

OBJECTIVE: Clinical trials represent a crucial step in the development and approval of medical devices. These trials involve evaluating the safety and efficacy of the device in a controlled setting with human subjects. However, traditional clinical trials can be expensive, time-consuming, and ethically challenging. Augmenting clinical trials with data from computer simulations, so called in silico clinical trials (ISCT), has the potential to address these challenges while satisfying regulatory requirements. However, determination of the patient harms in scope of an ISCT is necessary to ensure all harms are sufficiently addressed while maximizing the utility of the ISCT. This topic is currently lacking guidance. The objective of this work is to propose a general method to determine which patient harms should be included in an ISCT for a regulatory submission.
METHODS: The proposed method considers the risk associated with the harm, the impact of the device on the likelihood of occurrence of the harm and the technical feasibility of evaluating the harm via ISCT. Consideration of the risk associated with the harm provides maximum clinical impact of the ISCT, in terms of focusing on those failure modes which are most relevant to the patient population. Consideration of the impact of the device on a particular harm, and the technical feasibility of modeling a particular harm supports that the technical effort is devoted to a problem that (1) is relevant to the device in question, and (2) can be solved with contemporary modeling techniques.
CONCLUSIONS: As a case study, the proposed method is applied to a total shoulder replacement humeral system. With this framework, it is hoped that a consistent approach to scoping an ISCT can be adopted, supporting investment in ISCT by the industry, enabling consistent review of the ISCT approach across device disciplines by regulators, and providing maximum impact of modeling technologies in support of devices to improve patient outcomes.

UNASSIGNED: Patient safety culture is now at the forefront of the global health agenda and has been designated as a human right. Assessing safety culture is seen to be a prerequisite for improving safety culture in health-care organizations. However, no research has been conducted to examine the current study setup. Therefore, this study aims at assessing the status and factors influencing patient safety culture at Dilla University Teaching Hospital.
UNASSIGNED: This cross-sectional institutional-based study was conducted from February to March 2022 at Dilla University Hospital. The study used both qualitative and quantitative methods. A total of 272 health professionals were included in the survey. The qualitative data was collected using Key Informant Interviews and In-depth Interviews and 10 health professionals were selected purposively to meet the study objective.
UNASSIGNED: The overall composite positive patient safety culture response rate in the current study hospital was 37% (95% CI: 35.3, 38.8). Out of the 12 dimensions, teamwork within hospital units was the highest (75.3%), while frequency of event reporting was the lowest (20.7%) positive percentage response. Only two of the 12 dimensions scored above 50%. Factors affecting patient safety culture majorly at organizational and individual level were poor/low attitude of health professionals, poor documentation practice, and poor cooperation by clients, lack of training and continuous education, lack of standard operating procedure, Staff shortage and high work load.
UNASSIGNED: This study revealed that the overall composite positive patient safety culture response rate within the surveyed facility was alarmingly low compared to other hospitals in various countries. The results indicate that there is a need for improvement in areas such as event reporting, documentation, health-care workers\' attitude, and staff training. Hospitals must prioritize patient safety by cultivating a strong safety culture through effective leadership, adequate staffing, and education to enhance overall patient care.

UNASSIGNED: The highly complex and technological environment of critical care manages the most critically unwell patients in the hospital system, as such there is a need for a highly trained nursing workforce. Intensive care is considered a high-risk area for errors and adverse events (AE) due to the severity of illness and number of procedures performed.
UNASSIGNED: To investigate if the percentage of Critical Care Registered Nurses (CCRN) within an Intensive Care Unit (ICU) is associated with an increased risk of patients experiencing an AE.
UNASSIGNED: We conducted a retrospective cohort study of patients admitted between January 2016 and December 2020 to a tertiary ICU in Australia. Descriptive statistics and multivariable logistic regression were used to investigate the relationship between the proportion of CCRNs each month and the occurrence of an AE defined as any one of a medication error, fall, pressure injury or unplanned removal of a central venous catheter or endotracheal tube per patient.
UNASSIGNED: A total of 13,560 patients were included in the study, with 854 (6.3%) experiencing one AE. Patients with an AE were associated with higher illness severity and frailty scores. They were more commonly admitted after medical emergency team response calls and were less commonly elective ICU admissions. Those with an AE had longer ICU and in-hospital length of stay, and higher ICU and in-hospital mortality, on average. After adjusting for ICU LOS and acute severity of illness, being admitted during a month of higher critical care nursing skill-mix was associated with a statistically significant lower odds of having a subsequent AE (OR 0.966 [95% CI: 0.944-0.988], p 0.003).
UNASSIGNED: An increasing percentage of CCRNs is independently associated with a lower risk-adjusted likelihood of an AE. Increasing the skill-mix of the ICU nursing staff may reduce the occurrence of AEs and lead to improved patient outcomes.

The aims of this study were (1) to identify older patients\' risk factors for drug-related readmissions and (2) to assess the preventability of older patients\' drug-related revisits.
Post hoc analysis of a randomized clinical trial with patients aged ≥65 years at eight wards within four hospitals in Sweden. (1) The primary outcome was risk factors for drug-related readmission within 12 months post-discharge. A Cox proportional hazards model was made with sociodemographic and clinical baseline characteristics. (2) Four hundred trial participants were randomly selected and their revisits (admissions and emergency department visits) were assessed to identify potentially preventable drug-related revisits, related diseases and causes.
(1) Among 2637 patients (median age 81 years), 582 (22%) experienced a drug-related readmission within 12 months. Sixteen risk factors (hazard ratio >1, P < 0.05) related to age, previous hospital visits, medication use, multimorbidity and cardiovascular, liver, lung and peptic ulcer disease were identified. (2) The 400 patients experienced a total of 522 hospital revisits, of which 85 (16%) were potentially preventable drug-related revisits. The two most prevalent related diseases were heart failure (n = 24, 28%) and chronic obstructive pulmonary disease (n = 13, 15%). The two most prevalent causes were inadequate treatment (n = 23, 27%) and insufficient or no follow-up (n = 22, 26%).
(1) Risk factors for drug-related readmissions in older hospitalized patients were age, previous hospital visits, medication use and multiple diseases. (2) Potentially preventable drug-related hospital revisits are common and might be prevented through adequate pharmacotherapy and continuity of care in older patients with cardiovascular or lung disease.