Ljubo Znaor

Background/Objectives: Discharge letters are essential for continuity of care but are often time-consuming to prepare and variable in quality. Large language models (LLMs) may help standardize and support this process, yet evidence in ophthalmology remains limited. This study compared the quality of resident-written and GPT-5.2-generated ophthalmology discharge letters derived from the same de-identified clinical data. Methods: This retrospective blinded study was conducted at a tertiary hospital in Croatia. For 146 consecutive inpatient discharges, original resident-written letters were paired with GPT-5.2-generated letters created using a standardized prompt; 142 complete pairs were available for the primary analysis. Three board-certified ophthalmologists evaluated anonymized letters using a structured assessment of accuracy, completeness, clarity/structure, tone/professional phrasing, conciseness, global quality, errors, omissions, and key content elements. Results: In the primary paired analysis, GPT-5.2-generated letters performed similarly to resident-written letters across accuracy, completeness, clarity/structure, errors, omissions, and overall quality. GPT-5.2-generated letters received higher ratings for tone/professional phrasing, whereas resident-written letters were rated as more concise, although inter-rater agreement was poor on these stylistic domains (at or below chance for conciseness) and these findings should therefore be interpreted as exploratory. Resident-written letters more often documented operations, while GPT-5.2-generated letters more consistently included findings. Reviewer-adjusted sensitivity analyses were less favorable to GPT-5.2 for several domains. Conclusions: GPT-5.2-generated ophthalmology discharge letters showed similar performance to resident-written letters in several evaluated domains in the primary paired analysis, but differences in specific content elements and less favorable sensitivity analyses indicate that clinician oversight remains necessary to ensure accuracy, procedural completeness, and clinical usability.

Background/Objectives: The ethical climate in a healthcare institution encompasses the shared perceptions of how ethical issues are managed in everyday practice. Our prior survey at the University Hospital of Split, Croatia, showed a simultaneous predominance of the “Rules” and “Laws and professional codes” ethical climates. Building on these findings, we explored how these climates manifest in everyday practice, how they align with staff values and guide their ethical decision-making, and how they are shaped by external factors. Methods: We conducted seven focus groups with 31 participants: nurses, residents, specialists, and members of the Hospital Ethics Committee (HEC). We identified patterns in the data using Graneheim and Lundman’s qualitative content analysis. Results: Three themes emerged from our analysis. We observed that the ethical climate was shaped predominantly by healthcare professionals themselves based on shared professional values and informal norms, rather than explicit institutional rules. Nurses, positioned as frontline workers, felt particularly exposed to ethical dilemmas, reporting perceived subordination to physicians, increased pressures from patients, and vulnerability in ethically ambiguous situations. The participants generally believed that institutional leadership insufficiently utilised existing tools, bodies, and mechanisms to support ethical behaviour and sanction misdemeanors, resulting in gaps in human resource management, a lack of practical protocols, and a weak HEC. Conclusions: To strengthen the ethical climate, institutional leadership should provide clear and practical guidelines, effectively utilise regulating bodies and support services, establish dedicated mechanisms to support nurses, and consistently enforce sanctions for unethical behaviour.

Introduction: The European Union’s Health Technology Assessment Regulation (HTAR) and its implementing acts foresee various forms of clinician involvement, such as joint clinical assessment or Joint Scientific Consultation. However, considering the varying preparedness levels for HTAR, as well as the understanding of the health technology assessment (HTA) principles and processes, this study aimed to evaluate the levels of HTA-related skills among medical and dental doctors in Croatia. Methods: A cross-sectional survey study was conducted among medical and dental medicine doctors in Croatia. The survey recorded respondents’ relevant experience with HTA processes along with skill levels across the entire HTA process, mainly for acting as individual clinical experts or on behalf of their professional organizations, as well as potential HTA doers. Skill levels were evaluated using a 5-point scale (1 – no knowledge to 5 – full expertise ). Results: Among the 376 respondents included, only 6.1 percent had previous involvement in HTA, and 2.2 percent were familiar with HTAR. Related to the HTA process, the highest scores were observed in the understanding of key concepts and results of searching for studies, critical appraisal, study synthesis preparation, and ethics. The lowest scores were recorded in health economics, evidence grading, qualitative synthesis, and public/patient involvement. Respondents with prior research experience and those who reported frequent research use had significantly higher HTA skill scores. Conclusions: A significant gap in HTA-related skills highlights the need for targeted professional development programs and long-term educational reforms to build the capacity for various modes of involvement in HTA processes and their implementation.

Background and Objectives: To provide, for the first time, statistical data on risk factors for proliferative diabetic retinopathy (PDR) in young-onset type 1 diabetes (T1D) in Croatia, and to develop a predictive health economic model to evaluate the clinical and financial impact of implementing a national diabetic retinopathy (DR) screening. Materials and Methods: A cross-sectional study at University Hospital Split (June 2020–June 2022) analyzed 58 suitable T1D patients out of 562 screened. Patients were classified based on detailed fundus exams and photos into PDR and non-PDR groups. Clinical, demographic, and laboratory data were collected and analyzed using logistic regression. A health economic model was established to project the number of preventable PDR cases and the potential cost savings under various screening scenarios. Results: PDR was found in 47% of patients. Its presence was statistically significantly linked to longer diabetes duration, poor glycemic control, onset before 18 years of age, and irregular eye exams. Irregular ophthalmologic examinations increased the odds of PDR by nearly 30-fold. Health-economic modeling for 10,000 young-onset T1D patients showed that annual screening with 60% uptake could prevent about 1973 PDR cases and save €14.2 million annually. Screening remained cost-effective even with moderate uptake or less frequent intervals. Conclusions: Strict glycemic control and regular eye examinations are critical for preventing PDR in young T1D, and establishing a national screening program would be cost-effective, especially in resource-limited settings like Croatia, where providing appropriate, timely treatment may be challenging.

Accurate documentation is critical in ophthalmology, yet clinical notes often contain subtle errors that can affect decision-making. This study prospectively compared contemporary large language models (LLMs) for detecting clinically salient errors in emergency ophthalmology encounter notes and generating actionable corrections. 129 de-identified notes, each seeded with a predefined target error, were independently audited by four LLMs (o3 (OpenAI, closed-source), DeepSeek-v3-r1 (Deepseek, open-source), MedGemma-27B (Google, open-source), and GPT-4o (OpenAI, closed-source)) using a standardized prompt. Two masked ophthalmologists graded error localization, relevance of additional issues, and overall recommendation quality, with within-case analyses applying appropriate nonparametric tests. Performance varied significantly across models (Cochran’s Q = 71.13, p = 2.44 × 10⁻¹⁵). o3 achieved the highest error localization accuracy at 95.7% (95% CI, 89.5–98.8), followed by DeepSeek-v3-r1 (90.3%), MedGemma-27b (80.9%), and GPT-4o (53.2%). Ordinal outcomes similarly favored o3 and DeepSeek-v3-r1 (both p < 10⁻⁹ vs. GPT-4o), with mean recommendation quality scores of 3.35, 3.05, 2.54, and 2.11, respectively. These findings demonstrate that LLMs can serve as accurate “second-eyes” for ophthalmology documentation. A proprietary model led on all metrics, while a strong open-source alternative approached its performance, offering potential for privacy-preserving on-premise deployment. Clinical translation will require oversight, workflow integration, and careful attention to ethical considerations.

Background/Objectives: Ethical challenges in healthcare require awareness and adherence to professional codes of ethics, particularly in interdisciplinary settings such as tertiary hospitals. This study aimed to assess the attitudes and awareness of healthcare professionals regarding codes of ethics at the University Hospital of Split in Croatia, which did not have an institutional code at the time of the study. Methods: A cross-sectional survey using a structured, anonymous questionnaire was distributed both physically across hospital departments and online via email. Welch’s t-test, ANOVA, Kruskal–Wallis, and correlation tests were used to assess associations between favourability scores and participant characteristics. Linear and logistic regression analyses further examined predictors of favourable attitudes. Results: Of 442 returned questionnaires, 377 were complete and included in the analysis, mainly from nurses (56.5%) and physicians (42.7%). The median favourability score was 83.8% (88/105; IQR 78.1–88.6), with 87.0% scoring above the favourable threshold (≥75%). Female gender and higher education were significantly associated with more favourable attitudes. Participants strongly endorsed core principles such as patient confidentiality and autonomy, yet 57.6% considered ethics education during training inadequate, and only 36.3% viewed dual practices as a conflict of interest. Most respondents reported adherence to ethical standards (85.4%), while only over half were familiar with their professional ethics code (64.5%) and the hospital Ethics Committee (56.2%); a total of 66.3% supported introducing a hospital-specific code. Awareness and support for ethical structures were higher among women and those with more education. Conclusions: This study reveals a gap between personal ethical commitment and institutional ethical infrastructure. Strengthening ethics education and implementing a hospital-specific ethics code may enhance organisational ethical culture.

Background/Objectives: The World Health Organization’s Surgical Safety Checklist (WHO SSC) is known to reduce surgical complications and mortality, yet its implementation remains inconsistent across institutions. This study evaluated compliance with a modified SSC and assessed the impact of structured interventions to improve adherence at the University Hospital of Split, Croatia. Methods: A before-and-after study analyzed a sample of 1437 completed SSCs over four time points between April 2024 and May 2025: the baseline and after three successive interventions (an official letter from the hospital director emphasizing mandatory SSC use, individual meetings with department heads and head nurses reinforcing its importance, and a quality audit review of SSC completeness with leadership). Checklist completeness was assessed across five SSC sections: General Information, Sign In, Time Out, Sign Out, and Staff Info. Results: Overall checklist completeness increased from 78.3 ± 8.5% at baseline to 86.3 ± 2.5%, 92.0 ± 3.8%, and 94.7 ± 4.8% after the first, second, and third interventions, respectively (p < 0.001). All checklist sections improved significantly: General Info rose from 91.1 ± 6.0% to 98.6 ± 2.6% (p < 0.001); Sign In from 85.0 ± 11.4% to 97.0 ± 3.5% (p = 0.002); Time Out from 79.0 ± 14.6% to 96.4 ± 6.5% (p < 0.001); Sign Out from 70.2 ± 11.2% to 87.7 ± 11.0% (p = 0.003); and Staff Info from 70.7 ± 12.9% to 100.0 ± 0.0% (p < 0.001). Post hoc testing confirmed significant gains versus baseline for all three interventions (Dunnett’s test), with a further significant improvement between the first and third interventions (Tukey’s HSD, p = 0.011). Conclusions: Structured, low-cost interventions emphasizing administrative support, education, and communication significantly improved SSC adherence across a large tertiary hospital. This initiative represents a pioneering effort in Croatia and offers a scalable model for similar public healthcare systems aiming to enhance patient safety.

Purpose This study aimed to evaluate the knowledge, attitudes, and experiences of glaucoma patients about the side effects of intraocular pressure (IOP)-lowering medications in glaucoma, focusing on how these factors affect treatment adherence. Methods We conducted a cross-sectional survey during routine visits to glaucoma clinic, with participants aged 18 years or older on IOP-lowering medications for at least 6 months. The survey, consisting of 22 questions about various aspects of IOP-lowering medications and side effects. Results Most out of 398 analysed patients were not informed about side effects when IOP-lowering medication was prescribed, and 68.1% reported experiencing side effects. A significant number of participants reported changing IOP-lowering medications due to side effects. Higher self-reported knowledge about side effects was associated with lower concern (coef. = − 0.1636, P = 0.048) about side effects. Conclusion The study identified significant gaps in patient education and support about IOP-lowering medications side effects. Better communication and education by healthcare providers could improve treatment adherence, consequentially enhancing clinical outcomes and quality of life for glaucoma patients.

Objective To explore the barriers and facilitators to implementing the WHO’s Surgical Safety Checklist (SSC) at a tertiary-level hospital in Croatia, a country with a publicly funded universal healthcare system. Design Focus group-based study which used Graneheim and Lundman’s qualitative content analysis approach. Setting and participants We used purposive sampling to gather participants who were representative of the usual surgical team: nurses, residents, anaesthesiologists and surgeons. Aside from this requirement, we set no specific inclusion or exclusion criteria. We conducted 6 focus groups (2 with nurses, 2 with residents, 1 with anaesthesiologists and 1 with surgeons) with 29 participants. The nurses were all female, the residents and surgeons were all male, and anaesthesiologists were mostly male (60%). Most participants had at least a bachelor’s degree (93%). The focus groups were held in person at the hospital’s Department of Surgery in Croatia and were conducted in the local language, based on a predeveloped topic guide. Results Our analysis revealed two distinct themes that were fundamentally independent of each other. The first concerned within-team hierarchies and relationships that prevented the SSC from being used as intended, while the second was related to a diverse set of resource and logistical constraints that had existed before the implementation of the SSC. Conclusions Understanding the barriers faced by healthcare providers is crucial in designing training programmes that could help integrate the SSC into routine surgical practice. Our findings highlight the importance of actively including members of the surgical team in the implementation of the SSC and implementing interventions for improving within-team communication.