Funnel plots are used to identify intensive care units (ICUs) with a higher than expected risk-adjusted mortality. ICUs with a standardised mortality ratio (SMR) within pre-defined control limits (often the 99.8% CL) are regarded as ‘in control’ and not labelled as a potential outlier for a particular calendar year. However, increased mortality rates not due to random fluctuations within and across the calendar years may be overlooked. We examined whether statistically significant and relevant differences in mortality over time between ICUs regarded as ‘in control’ are present.

A longitudinal register-based study.

88 ICUs in the Netherlands registering the admissions of all critically ill patients in the National Intensive Care Evaluation registry in the Netherlands from 2013 to 2023.

Hospital death analysed in a multivariable logistic regression analysis with a random intercept for ICU. The random intercept variance was translated to the median OR (MOR).

877 ICU-calendar year combinations were included, covering 759 498 unique admissions. The MOR increased from 1.12 (95% CI 1.10 to 1.15) for ICU-calendar year combinations with an SMR within the narrowest 95% CL (N=677) to 1.20 (1.17 to 1.24) for combinations with an SMR within the expanded 99.8% CL (including adjustment for overdispersion) (N=194) and to 1.21 (1.17 to 1.25) when including all ICU-calendar year combinations. Similar results were found for separate calendar years and separate diagnostic groups.

These results show differences in mortality between ICUs that were not labelled as outliers. Assessment of mortality performance should integrate cross-sectional funnel plots, the MOR and longitudinal trends in the SMR to better capture persistent patterns of excess risk.

Well-being of healthcare professionals (HCPs) is vital for care quality, staff retention and overall healthcare system effectiveness. This study aims to identify the organisational and workplace variables associated with sick leave and measures of engagement of HCPs on department level within a single Dutch academic hospital.

Cross-sectional study using routinely collected organisational data.

A tertiary-care academic hospital in the Netherlands.

25 clinical departments were included. Department level variables were derived from routinely collected hospital databases. Availability of data varied across variables. Analysis included information on patient population, human resources, care processes, quality of care and employee and patient experiences to assess differences, correlations and predictors for sick leave and engagement.

Primary outcome measures were (1) sick leave (%) and (2) engagement, assessed through two staff-survey items (vitality and connectedness; 0–10 Numeric Rating Scale). Both outcomes were analysed at department level.

Employee population data showed the most consistent patterns across analyses. Departments with higher staffing capacity had higher sick leave and lower engagement in group comparisons (p=0.009, p=0.030, respectively). In multivariable models, higher staffing capacity remained associated with increased sick leave (B=1.38, 95% CI 0.53 to 2.23, p=0.003). Engagement was positively associated with higher inflow (B=0.92, 95% CI 0.06 to 1.77, p=0.037) and negatively associated with outflow (B = –1.36, 95% CI –2.08 to –0.63, p=0.001). No consistent associations were found with patient population and patient experience measures.

Workforce-related factors, particularly staffing capacity and inflow and outflow, are strongly linked to sick leave and engagement. Routinely collected hospital data can be used to identify at-risk departments and inform targeted strategies for improving workforce sustainability. Future studies should explore more granular, team-level data to better support staff well-being and care quality.

The functional resonance analysis method (FRAM) is increasingly used to analyse healthcare processes. FRAM uses four steps to analyse a process and its potential variability. We systematically reviewed studies using FRAM in healthcare on how the four steps in FRAM are reported, defined and supported by data.

Systematic review following the preferred reporting items for systematic reviews and meta-analyses 2020 guidelines.

Web of Science, PubMed, Embase, Scopus, PsycINFO, Dimensions and Lens were searched up to December 2025.

All peer-reviewed studies using FRAM in a healthcare context that presented a FRAM visualisation were included. The papers had to be written in English.

Two independent reviewers screened titles and abstracts, and subsequently the full text of selected papers. Data was extracted reporting on the steps of FRAM, how functions were supported by data, and the functions and couplings of the visualisations.

Sixty-eight papers were included, of which 20 (29%) reported at least one aspect of all four steps in FRAM. While most studies (85%) described how functions were supported by data, the methods used varied widely. Terminology was interpreted differently concerning variability, the output of variability and the effect of combined variability.

Most FRAM studies in healthcare do not report all steps of FRAM, and interpretations of key terms differ. FRAM studies should more clearly describe which steps of the method are conducted, and how data is collected and analysed. Refinement of FRAM guidelines, particularly on data use and terminology, would enhance consistency and comparability across studies.

CRD42024592858.

To evaluate the effectiveness of the recommendation improvement matrix (RIM) methodology for improving the quality of recommendations resulting from sentinel event analysis, where we hypothesise that the use of the RIM methodology leads to better quality recommendations.

A before-and-after analysis of the quality of the formulated recommendations after sentinel event analysis.

The study was carried out during the 2023 Saudi Patient Safety Centre International Patient Safety Conference.

36 conference participants, including nurses, medical doctors, pharmacists, dentists, general practitioners and quality officers.

RIM methodology training during a structured 3-hour workshop.

The primary outcome was the proportion of recommendations that using the ² test, passed the RIM filter criteria before and after training. Secondary outcomes included changes in recommendation categorisation within the matrix and participant ratings of feasibility and usability on a five-point Likert scale using a t-test for comparison.

Prior to training, 49 recommendations were generated, of which 63.3% met the filter criteria. Post-training, the proportion of recommendations passing the filter increased significantly to 83% (p=0.00543). Adjustments to recommendations primarily improved alignment with the filter criteria, though limited improvements were observed in matrix categorisation. Participants rated the methodology’s feasibility and usability highly, with average scores of 4.39/5 and 4.43/5, respectively. However, 46% expressed a need for additional training, particularly on the matrix application.

The RIM methodology significantly improves the quality of recommendations following sentinel event analyses. To enhance its impact, further training focusing on matrix application is necessary. Incorporating the methodology into healthcare education and professional development could strengthen patient safety practices.