 |
Medication Safety Officers Society 4135 Members Strong A society of healthcare professionals dedicated to improving medication safety in healthcare organizations |
Scientific studies of artificial intelligence (AI) solutions in healthcare have been the subject of intense criticism—both in research publications and in the media.1–3 Early validations of predictive algorithms are criticised for not having meaningful clinical impact, and AI tools that make mistakes or fail to show immediate improvement in health outcomes are heralded as the first snowflakes in the next AI winter (a period of decreased interest in AI research and development). Scientific evidence is the language of trust in healthcare, and peer-reviewed studies evaluating AI solutions are key to fostering adoption. There are over two dozen reporting guidelines for AI in medicine,4 and many other consensus statements and standards that offer recommendations for the publication of research about medical AI.5 Despite such guidance, the average frontline clinician still struggles in interpreting the results of an AI study to...
‘We need bold, fundamental change that gets at the roots of the burnout crisis.’- US Surgeon General Vivek H. Murthy, MD, MBA.
Well-being was brought into clearer focus during the COVID-19 pandemic, during which the prevalence of healthcare worker (HCW) emotional exhaustion increased from 27%1 to 39%.2 Currently, there is not a coordinated effort to ensure HCW well-being interventions meet minimum standards of feasibility, accessibility and methodological rigour. In this issue of BMJ Quality and Safety, Melvin et al assessed perceptions of physician well-being programmes by interviewing physicians and people involved in these programmes.3 As is often the case with any real-world application of science, there are substantial gaps between the programmes as intended and the programmes in practice. The authors conclude that the ‘persistence of poor well-being outcomes suggests that current support initiatives are suboptimal’.
The key is understanding what is suboptimal....
In this issue of BMJ Quality and Safety, Black and colleagues present a qualitative study of healthcare teams working to uncover diagnoses in patients experiencing non-specific cancer symptoms.1 The study highlights the criticality of teams in helping support or derail diagnostic pathways. Overall, Black et al1 present unique insights that highlight the challenges clinical teams face when caring for patients with non-specific symptoms.
Unfortunately, we know that diagnostic processes such as those studied by Black et al are frequently unsafe. Diagnostic errors are ‘the single largest source of deaths across all (healthcare) settings,’ with estimates for cancer-related mistakes estimated at around 11.1%.2 A key challenge to making diagnoses in patients with non-specific symptoms is the presence of uncertainty throughout the diagnostic process.
As Black et al point out,1 uncertainty in the diagnostic process is felt by both patients and clinicians. It...
Surgical errors in ophthalmology can have devastating consequences. We developed an artificial intelligence (AI)-based surgical safety system to prevent errors in patient identification, surgical laterality and intraocular lens (IOL) selection. This study aimed to evaluate its effectiveness in real-world ophthalmic surgical settings.
MethodsIn this retrospective observational before-and-after implementation study, we analysed 37 529 ophthalmic surgeries (18 767 pre-implementation, 18 762 post implementation) performed at Tsukazaki Hospital, Japan, between 1 March 2019 and 31 March 2024. The AI system, integrated with the WHO surgical safety checklist, was implemented for patient identification, surgical laterality verification and IOL authentication.
ResultsPost implementation, five medical errors (0.027%) occurred, with four in non-authenticated cases (where the AI system was not fully implemented or properly used), compared with one (0.0053%) pre-implementation (p=0.125). Of the four non-authenticated errors, two were laterality errors during the initial implementation period and two were IOL implantation errors involving unlearned IOLs (7.3% of cases) due to delayed AI updates. The AI system identified 30 near misses (0.16%) post implementation, vs 9 (0.048%) pre-implementation (p=0.00067), surgical laterality errors/near misses occurred at 0.039% (7/18 762) and IOL recognition at 0.29% (28/9713). The system achieved>99% implementation after 3 months. Authentication performance metrics showed high efficiency: facial recognition (1.13 attempts, 11.8 s), surgical laterality (1.05 attempts, 3.10 s) and IOL recognition (1.15 attempts, 8.57 s). Cost–benefit analysis revealed potential benefits ranging from US$181 946.94 to US$2 769 129.12 in conservative and intermediate scenarios, respectively.
ConclusionsThe AI-based surgical safety system significantly increased near miss detection and showed potential economic benefits. However, errors in non-authenticated cases underscore the importance of consistent system use and integration with existing safety protocols. These findings emphasise that while AI can enhance surgical safety, its effectiveness depends on proper implementation and continuous refinement.
The vital role of medical workforce well-being for improving patient experience and population health while assuring safety and reducing costs is recognised internationally. Yet the persistence of poor well-being outcomes suggests that current support initiatives are suboptimal. The aim of this research study was to work with, and learn from, diverse hospital settings to understand how to optimise strategies to improve doctors’ well-being and reduce negative impacts on the workforce and patient care.
MethodsRealist evaluation consistent with the Realist And Meta-narrative Evidence Synthesis: Evolving Standards (RAMESES) II quality standards. Realist interviews (n=124) with doctors, well-being intervention implementers/practitioners and leaders in eight hospital settings (England) were analysed using realist logic.
ResultsThere were four key findings, underpinned by 21 context-mechanism-outcome configurations: (1) solutions needed to align with problems, to support doctor well-being and avoid harm to doctors; (2) doctors needed to be involved in creating solutions to their well-being problems; (3) doctors often did not know what support was available to help them with well-being problems and (4) there were physical and psychological barriers to accessing well-being support.
Discussion and conclusionDoctors are mandated to ‘first, do no harm’ to their patients, and the same consideration should be extended to doctors themselves. Since doctors can be harmed by poorly designed or implemented well-being interventions, new approaches need careful planning and evaluation. Our research identified many ineffective or harmful interventions that could be stopped. The findings are likely transferable to other settings and countries, given the realist approach leading to principles and causal explanations.
Over the past two decades, the UK has actively developed policies to enhance early cancer diagnosis, particularly for individuals with non-specific cancer symptoms. Non-specific symptom (NSS) pathways were piloted and then implemented in 2015 to address delays in referral and diagnosis. The aim of this study was to outline the functions that enable NSS teams to investigate cancer and other diagnoses for patients with NSSs.
MethodsThe analysis was derived from a multisite ethnographic study conducted between 2020 and 2023 across four major National Health Service (NHS) trusts. Data collection encompassed observations, patient shadowing, interviews with clinicians and patients (n=54) and gathered documents. We used principles of the functional resonance analysis method to identify the functions of the NSS pathway and analyse their relevance to patient safety.
ResultsOur analysis produced 29 distinct functions within NSS pathways, organised into two clusters: pretesting assessment and information gathering, and post-testing interpretation and management. Safety-critical functions encompassed assessing the reason for referral, deciding on a plan of investigation and estimating the remaining cancer risk. We also identified ways that teams build and maintain safety across all functions, for example, by cultivating generalist-specialist expertise within the team and creating continuity through patient navigation. Variation in practice across sites revealed targets for an NSS pathway blueprint that would foster local development and quality improvement.
ConclusionsOur findings suggest that national and local improvement plans could differentiate specific policies to reduce unwarranted variation and support adaptive variation that facilitates the delivery of safe care within the local context. Enhancing multidisciplinary teams with additional consultants and deploying patient navigators with clinical backgrounds could improve safety within NSS pathways. Future research should investigate different models of generalist-specialist team composition.
To estimate and quantify the cost implications and health impacts of improving the performance of English endoscopy services to the optimum quality as defined by postcolonoscopy colorectal cancer (PCCRC) rates.
DesignA semi-Markov state-transition model was constructed, following the logical treatment pathway of individuals who could potentially undergo a diagnostic colonoscopy. The model consisted of three identical arms, each representing a high, middle or low-performing trust’s endoscopy service, defined by PCCRC rates. A cohort of 40-year-old individuals was simulated in each arm of the model. The model’s time horizon was when the cohort reached 90 years of age and the total costs and quality-adjusted life-years (QALYs) were calculated for all trusts. Scenario and sensitivity analyses were also conducted.
ResultsA 40-year-old individual gains 0.0006 QALYs and savings of £6.75 over the model lifetime by attending a high-performing trust compared with attending a middle-performing trust and gains 0.0012 QALYs and savings of £14.64 compared with attending a low-performing trust. For the population of England aged between 40 and 86, if all low and middle-performing trusts were improved to the level of a high-performing trust, QALY gains of 14 044 and cost savings of £249 311 295 are possible. Higher quality trusts dominated lower quality trusts; any improvement in the PCCRC rate was cost-effective.
ConclusionImproving the quality of endoscopy services would lead to QALY gains among the population, in addition to cost savings to the healthcare provider. If all middle and low-performing trusts were improved to the level of a high-performing trust, our results estimate that the English National Health Service would save approximately £5 million per year.
For patients in the intensive care unit (ICU), prolonged mechanical ventilation is associated with poor outcomes. A quality improvement (QI) initiative with the aim of reducing median time on the ventilator for tracheostomy patients was undertaken at a tertiary care ICU in Toronto, Canada. A QI team was formed, and using QI methodology, a deep understanding of our local process was achieved. Based on this information and on the latest evidence on weaning, a standard tracheostomy weaning protocol was designed. The protocol was refined through three developmental and two testing plan–do–study–act cycles. This study was a prospective time series showing the effect of the implementation of our intervention on tracheotomy patients’ time on the ventilator. The baseline median number of days on the ventilator after tracheostomy insertion was 17. Within 12 months of the introduction of the intervention, a shift in the data showing a reduction in the median time on the ventilator to 10.6 days had developed. Length of stay in the ICU was reduced by 4.3 days. Adherence and compliance to the protocol also improved over time. A standard tracheostomy weaning protocol was successfully developed, tested and implemented in a tertiary care ICU. Using strategies such as frequent communication with key stakeholders and incorporating a tracheostomy weaning progress sheet to document and track tracheostomy patients and their outcomes, this QI intervention has become engrained in the local culture at our centre. This weaning protocol has successfully reduced the median time on the ventilator for tracheostomy patients by over 6 days.
