Owen Matthew Truscott Thomas

Background: The prevailing strategy for combating surgically related adverse events (AEs) is to anticipate and prevent them by learning from adverse events (Safety-I). However, not all AEs can be anticipated or prevented in complex systems. This requires more focus on how daily work succeeds, that is, resilient and adaptive behaviour (Safety-II). Learning from success and promoting a positive patient safety culture (PSC) are key global strategies for eliminating preventable AEs. As a patient safety initiative, Learning from Excellence (LfE) captures episodes of successful everyday work and is thus, a system for applying a Safety-II perspective. However, its effect on PSC has not been evaluated. This study aimed to explore whether the implementation of LfE in a postanaesthesia care unit over a 16-month period changed the nurses’ perceptions of the measurable PSC dimensions ‘work engagement’, ‘teamwork climate’, ‘working conditions’ and ‘safety climate’. Methods: This longitudinal two-group, controlled quasi-experimental study used an annual ‘Hospital Staff Survey’ to measure PSC factors before (2021–2022) and after (2023–2024) the implementation of LfE. The study included nurses from the surgical department of a Norwegian university hospital working in units specialised in postoperative care (intervention unit) and intensive care (comparison unit). The primary outcomes were binarised versions of individuals’ positive responses to Likert-type statements regarding ‘work engagement’, ‘teamwork climate’, ‘working conditions’ and ‘safety climate’. The data were analysed using logistic regression methods. Results: Relative to the changes in the comparison unit, the results within the intervention unit indicated non-significant small positive changes in ‘work engagement’, ‘teamwork climate’ and ‘working conditions’ and a non-significant small negative change in ‘safety climate’. Conclusions: LfE did not have a significant effect on PSC after 16 months implementation. Further longitudinal experimental studies are needed to measure the effect of LfE on PSC. Future research should focus on the role of psychological safety in LfE and assess whether reflexive spaces add value to LfE. Trial registration: ClinicalTrials.gov NCT05794490; registered date: 4 March 2023.

Background: Artificial intelligence (AI) is increasingly used in prostate cancer diagnostic workflows but remains insufficiently validated in high-prevalence cohorts often encountered in academic referral centers. Purpose: To compare the diagnostic performance of licensed AI software with routine radiologist readings of prostate MRI, using histopathology as the reference standard. Material and Methods: In this retrospective study, 1000 patients underwent prostate MRI for suspected prostate cancer (between May 2020 and December 2024), followed by transperineal biopsy; 391 subsequently underwent radical prostatectomy. Diagnostic performance was assessed using sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy across PI-RADS thresholds. Receiver operating characteristic (ROC) analysis, Cohen's kappa for inter-rater agreement, and paired McNemar's test were performed. Results: In total, 959 patients (mean age = 69.7 ± 8.4 years) were included; clinically significant prostate cancer (csPCa) was found in 830 (86.5%) patients. AI assigned more cases to PI-RADS 1–2 and fewer to PI-RADS 3 (κ = 0.388; P <.001). At the PI-RADS ≥3 threshold, radiologists showed sensitivity 96.0%, specificity 20.9%, accuracy 85.9%, PPV 88.7%, and NPV 45.0%, compared with 91.3%, 37.2%, 84.0%, 90.3%, and 40.0% for AI ( P <.001). ROC performance was comparable (AUC 0.763 vs. 0.739; P = .28). In the prostatectomy subgroup, AI demonstrated a higher false-negative rate (8.4% vs. 4.1%; odds ratio = 3.83, 95% confidence interval [CI] = 1.52–11.51). Conclusion: AI showed diagnostic performance comparable to that of radiologists for csPCa detection, with no significant difference in AUC. AI reduced indeterminate PI-RADS 3 scores but missed more significant cancers.

Ordinal data are widespread in psychiatric research, especially in the form of Likert scales, but can be challenging to adequately model due to their ordered and discrete nature. In this work, we investigate the role of model specification for ordinal data in large-scale psychometric applications, focussing on the methodology of model-based clustering. Methods deliberately designed for ordinal data (ordinal likelihood) demonstrate the undesirable behaviours of instability and computational expense in large-scale data sets, while in comparison other more generic methods (e.g., Gaussian or multinomial likelihoods) offer preferable scalability and usability in the applications considered. In two “big data”-scale UK biobank data sets, using the PHQ-9 and a CIDI-SF-derived symptoms of depression, we examine the robustness of model-based clustering to likelihood choice (Gaussian vs Multinomial vs Ordinal). We further pursue an extensive, systematic simulation approach to understand the data properties driving the computational behaviour. Our results simultaneously uncover clinically relevant depression subtypes, and illuminate the pragmatic decision-making landscape spanning theoretical well-specification and practical convenience in two real-world large-scale applications and extensive simulations. The findings further highlight the demand for more robust and scalable tools designed for unsupervised learning of large-scale ordinal data.

Objectives: To determine the risk of pre‐eclampsia in ongoing pregnancies according to gestational age. We also studied whether the gestational‐age‐specific risks of developing pre‐eclampsia have changed over recent decades. Methods: This was a population‐based study using data from the Medical Birth Registry of Norway. The study population included all singleton pregnancies in Norway delivered between 21 and 43 weeks' gestation during the period 1984–2019. We calculated the proportion of pregnancies with pre‐eclampsia at each week of gestation. The numerator was the number of pre‐eclamptic pregnancies that were delivered during the gestational week being studied, and the denominator included all pregnancies still ongoing at the same gestational week. We analyzed the study period as a whole and subdivided it into four time periods (1984–1989, 1990–1999, 2000–2009 and 2010–2019). Results: A total of 1 955 082 pregnancies were included, of which 3.2% developed pre‐eclampsia. The risk of pre‐eclampsia increased with advancing gestational age. The risk per 1000 ongoing pregnancies was 0.2 at 28 weeks, 5.2 at 38 weeks and 13.6 at 42 weeks. The gestational‐age‐specific risk of developing pre‐eclampsia was lower at all gestational ages during the period 2010–2019 compared with 2000–2009, and the absolute decrease was greatest in term and post‐term pregnancies. Conclusions: The risk of developing pre‐eclampsia increased almost exponentially with advancing gestational age. During the last decade of the study period, the risk of pre‐eclampsia decreased, particularly in term pregnancies. This decrease may be explained by the use of obstetric interventions to shorten the duration of pregnancy among women at high risk of developing pre‐eclampsia. © 2026 The Author(s). Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of International Society of Ultrasound in Obstetrics and Gynecology.

Antiplatelet and anticoagulant therapy are associated with a significant risk of upper gastrointestinal bleeding (UGIB), as is the use of nonsteroidal anti-inflammatory drugs (NSAIDs). Concomitant treatment with proton pump inhibitors (PPIs) has been shown to reduce the risk of this complication. To compare the use of risk medications and PPIs in a cohort of adult patients with bleeding peptic ulcers or erosions in the stomach or duodenum with the background population. We performed a prospective observational study including 543 patients with endoscopically verified bleeding ulcers and/or erosions at two large hospitals in Norway. Information on risk medications taken prior to study enrolment was collected prospectively through structured interviews with patients and with the review of their medical records, whereas for the background population the information was obtained from The Norwegian Prescription Database (NorPD). Overall, 434 (80%) of the patients used risk medications, compared to 34% in the background population (p < 0.001). Only 39 (8.9%) of the patients received PPIs as co-medication, and 150 (34.6%) tested positive for Helicobacter pylori (H. pylori) infection. Among patients with bleeding ulcers and erosions, we found a high prevalence of risk medication use and an underutilization of prophylactic PPIs. H. pylori infection appears to play a minor role. Bleeding Ulcer and Erosions Study "BLUE Study", ClinicalTrials.gov Identifier: NCT03367897.

Background and Hypothesis: Schizophrenia's heterogeneity hampers research, prevention, and treatment. We previously identified a bimodal distribution of red blood cell polyunsaturated fatty acids (PUFA) during acute schizophrenia, defining Low PUFA (LP) and High PUFA (HP) groups. We hypothesized that LP patients exhibit glutathione system dysfunction representing a distinct biotype characterized by redox dysregulation.Study Design: In this cross-sectional follow-up study, we examined 55 schizophrenia spectrum patients five years after an acute episode alongside 51 healthy controls. We measured amino acids, glutathione and related enzymes, and mRNA expression of glutathione-related proteins. Regularized regression analysis identified key biochemical predictors of group (LP=0 or HP=1) and symptom severity.Study Results: The strongest biochemical predictor of the LP subgroup was elevated plasma 2-aminobutyrate (OR=0.0013). LP patients showed robustly lower gene expression of the cystine/glutamate antiporter (xCT) (OR=10.0), more homogeneous xCT distribution, moderately lower blood glutathione, and elevated plasma valine compared to HP patients. HP patients exhibited significantly lower plasma cystine compared to controls. PUFA group robustly modified the relationship between gene expression of glutathione-related proteins and PANSS symptom scores, with different correlation patterns observed between subgroups.Conclusions: Our findings suggest that schizophrenia comprises two biologically distinct subtypes: a Low PUFA biotype characterized by xCT-deficit and altered 2-aminobutyrate metabolism, and a High PUFA biotype with different glutathione regulatory mechanisms. The Low PUFA/xCT-deficit biotype represents a potential disease entity with distinct pathophysiology and prevention and treatment implications. Genotyping of xCT and glutathione-related enzymes in future studies may further validate these biotypes and guide personalized interventions.

Likelihood-free inference for simulator-based statistical models has developed rapidly from its infancy to a useful tool for practitioners. However, models with more than a handful of parameters still generally remain a challenge for the Approximate Bayesian Computation (ABC) based inference. To advance the possibilities for performing likelihood-free inference in higher dimensional parameter spaces, we introduce an extension of the popular Bayesian optimisation based approach to approximate discrepancy functions in a probabilistic manner which lends itself to an efficient exploration of the parameter space. Our approach achieves computational scalability for higher dimensional parameter spaces by using separate acquisition functions, discrepancies, and associated summary statistics for distinct subsets of the parameters. The efficient additive acquisition structure is combined with exponentiated loss-likelihood to provide a misspecification-robust characterisation of posterior distributions for subsets of model parameters. The method successfully performs computationally efficient inference in a moderately sized parameter space and compares favourably to existing modularised ABC methods. We further illustrate the potential of this approach by fitting a bacterial transmission dynamics model to a real data set, which provides biologically coherent results on strain competition in a 30-dimensional parameter space.

This is a semidouble-blind randomized controlled trial comparing the clinical and radiographic outcomes of fixation of displaced 3- and 4-part proximal humerus fractures with an intramedullary nail versus a locking plate. Seventy-nine patients aged 45 to 81 years (mean 66.5 years) were randomized to open reduction and osteosynthesis with either locking nails or plates. The primary outcome was the Disabilities of the Arm, Shoulder, and Hand (DASH) score at 2 years after surgery. The secondary outcomes were the Constant score (CS), Oxford shoulder score (OSS), visual analog scale for pain at rest and during activity, and complication and reoperation rates. The patients were assessed at 6 weeks, 12 weeks, 6 months, 1 year, and 2 years after inclusion. Physiotherapists who were blinded to the choice of implant performed assessments at 6 months and 2 years. Patients were also blinded to the choice of implant. Three patients were lost to follow-up, leaving data from 38 patients in each group available for analysis. There were no statistically significant differences in the DASH score, CS, OSS, or pain score at any time. The DASH score at 2 years was 12.7 (95% confidence interval [CI] 8.8-17.0) for the nail group and 10.0 (95% CI 5.6-16.1) for the plate group (p = 0.48). Twelve (32%) patients underwent reoperation in the nail group, whereas 2 (5%) patients underwent reoperation in the plate group (p = 0.006). Fourteen (37%) patients in the nail group and 4 (11%) patients in the plate group experienced complications (p = 0.05). There were no statistically significant differences in patient-reported outcomes or function 2 years after surgery for fixation of displaced 3- and 4-part part proximal humerus fractures with nails or plates. However, there were more complications and reoperations in the nail group. Therapeutic Level I. See Instructions for Authors for a complete description of levels of evidence.

Background: This is a semidouble-blind randomized controlled trial comparing the clinical and radiographic outcomes of fixation of displaced 3- and 4-part proximal humerus fractures with an intramedullary nail versus a locking plate. Methods: Seventy-nine patients aged 45 to 81 years (mean 66.5 years) were randomized to open reduction and osteosynthesis with either locking nails or plates. The primary outcome was the Disabilities of the Arm, Shoulder, and Hand (DASH) score at 2 years after surgery. The secondary outcomes were the Constant score (CS), Oxford shoulder score (OSS), visual analog scale for pain at rest and during activity, and complication and reoperation rates. The patients were assessed at 6 weeks, 12 weeks, 6 months, 1 year, and 2 years after inclusion. Physiotherapists who were blinded to the choice of implant performed assessments at 6 months and 2 years. Patients were also blinded to the choice of implant. Results: Three patients were lost to follow-up, leaving data from 38 patients in each group available for analysis. There were no statistically significant differences in the DASH score, CS, OSS, or pain score at any time. The DASH score at 2 years was 12.7 (95% confidence interval [CI] 8.8-17.0) for the nail group and 10.0 (95% CI 5.6-16.1) for the plate group (p = 0.48). Twelve (32%) patients underwent reoperation in the nail group, whereas 2 (5%) patients underwent reoperation in the plate group (p = 0.006). Fourteen (37%) patients in the nail group and 4 (11%) patients in the plate group experienced complications (p = 0.05). Conclusions: There were no statistically significant differences in patient-reported outcomes or function 2 years after surgery for fixation of displaced 3- and 4-part part proximal humerus fractures with nails or plates. However, there were more complications and reoperations in the nail group. Level of Evidence: Therapeutic Level I. See Instructions for Authors for a complete description of levels of evidence.

Background Hospitals should adopt multiple methods to monitor incidents for a comprehensive review of the types of incidents that occur. Contrary to traditional incident reporting systems, the Green Cross (GC) method is a simple visual method to recognise incidents based on teamwork and safety briefings. Its longitudinal effect on patient safety culture has not been previously assessed. This study aimed to explore whether the implementation of the GC method in a postanaesthesia care unit changed nurses’ perceptions of different factors associated with patient safety culture over 4 years. Methods A longitudinal quasi-experimental pre–post intervention design with a comparison group was used. The intervention unit and the comparison group, which consisted of nurses, were recruited from the surgical department of a Norwegian university hospital. The intervention unit implemented the GC method in February 2019. Both groups responded to the staff survey before and then annually between 2019 and 2022 on the factors ‘work engagement’, ‘teamwork climate’ and ‘safety climate’. The data were analysed using logistic regression models. Results Within the intervention unit, relative to the changes in the comparison group, the results indicated significant large positive changes in all factor scores in 2019, no changes in 2020, significant large positive changes in ‘work engagement’ and ‘safety climate’ scores in 2021 and a significant medium positive change in ‘work engagement’ in 2022. At baseline, the comparison group had a significantly lower score in ‘safety climate’ than the intervention unit, but no significant baseline differences were found between the groups regarding ‘work engagement’ and ‘teamwork climate’. Conclusion The results suggest that the GC method had a positive effect on the nurses’ perception of factors associated with patient safety culture over a period of 4 years. The positive effect was completely sustained in ‘work engagement’ but was somewhat less persistent in ‘teamwork climate’ and ‘safety climate’.