Interventions for the uptake of evidence-based recommendations in acute stroke settings

High-performance automated abstract screening with large language model ensembles

OBJECTIVE

screening is a labor-intensive component of systematic review involving repetitive application of inclusion and exclusion criteria on a large volume of studies. We aimed to validate large language models (LLMs) used to automate abstract screening.

MATERIALS AND METHODS

LLMs (GPT-3.5 Turbo, GPT-4 Turbo, GPT-4o, Llama 3 70B, Gemini 1.5 Pro, and Claude Sonnet 3.5) were trialed across 23 Cochrane Library systematic reviews to evaluate their accuracy in zero-shot binary classification for abstract screening. Initial evaluation on a balanced development dataset (n = 800) identified optimal prompting strategies, and the best performing LLM-prompt combinations were then validated on a comprehensive dataset of replicated search results (n = 119 695).

RESULTS

On the development dataset, LLMs exhibited superior performance to human researchers in terms of sensitivity (LLMmax = 1.000, humanmax = 0.775), precision (LLMmax = 0.927, humanmax = 0.911), and balanced accuracy (LLMmax = 0.904, humanmax = 0.865). When evaluated on the comprehensive dataset, the best performing LLM-prompt combinations exhibited consistent sensitivity (range 0.756-1.000) but diminished precision (range 0.004-0.096) due to class imbalance. In addition, 66 LLM-human and LLM-LLM ensembles exhibited perfect sensitivity with a maximal precision of 0.458 with the development dataset, decreasing to 0.1450 over the comprehensive dataset; but conferring workload reductions ranging between 37.55% and 99.11%.

DISCUSSION

Automated abstract screening can reduce the screening workload in systematic review while maintaining quality. Performance variation between reviews highlights the importance of domain-specific validation before autonomous deployment. LLM-human ensembles can achieve similar benefits while maintaining human oversight over all records.

CONCLUSION

LLMs may reduce the human labor cost of systematic review with maintained or improved accuracy, thereby increasing the efficiency and quality of evidence synthesis.

Global, regional, and national burden and trends of stroke among youths and young adults aged 15-39 years from 1990 to 2021: findings from the Global Burden of Disease study 2021

Background

Stroke is a leading cause of disability and mortality worldwide, with rising incidence rates among youths and young adults aged 15-39 years. However, comprehensive assessments of stroke burden in this age group at global, regional, and national levels are limited. This study examines trends in stroke incidence, mortality, and disability-adjusted life years (DALYs) from 1990 to 2021 using data from the Global Burden of Disease (GBD) study.

Methods

Data from the GBD study (1990-2021) were analyzed to assess the age-standardized incidence, mortality, and DALYs related to stroke in individuals aged 15-39 years. The relationship between stroke burden and the Socio-Demographic Index (SDI) was explored across 204 countries and 21 regions. Trends were analyzed using the estimated annual percentage change (EAPC) and average annual percentage change (AAPC).

Results

This study reveals global, regional, and national trends in stroke burden among youths and young adults (15-39 years) from 1990 to 2021. In 2021, the global age-standardized stroke incidence was 757,234.61 cases, with 8.72 million DALYs and 122,742 stroke-related deaths. Although global incidence increased by 19.09%, age-standardized rates (ASRs) declined by 0.67% annually. DALYs and mortality rates also decreased globally. Notably, stroke burden increased in low and low-middle SDI regions. South Asia had the highest number of cases, while Oceania reported the highest mortality rate. These findings underscore regional disparities in stroke trends. Globally, metabolic risks (46.2%) and high systolic blood pressure (37.87%) are major contributors to stroke-related mortality.

Acute Stroke Units Nested within Broader Neurology: Care Bundles for Nursing to Enhance Competence and Interdisciplinary Coordination

PURPOSE OF REVIEW

The benefits of acute stroke unit care, with nurses as central figures, are well-documented. As care bundles gain traction to enhance evidence-based nursing care, this review explores their development and adaptation in a setting where stroke care is integrated into general neurology. It also highlights key elements for reinforcing competence and interdisciplinary support.

RECENT FINDINGS

Most evidenced based acute stroke unit recommendations focus on hyperacute medical management. In comparison, the literature on decision-making for selecting and evaluating key components of nursing surveillance to support specialized stroke care in geographically smaller settings is sparse although the benefits of nursing care bundles is emerging. Well-structured care bundles, grounded in robust evidence and supported by thorough documentation and effective implementation strategies, provide a clear framework for nursing care, facilitate continuous monitoring, and are useful for enhancing practices especially in smaller stroke units that lack the resources of more comprehensive state-of-the-art units. Tailoring stroke nursing care bundles to local contexts requires an adaptable approach.

Impact of academic center for evidence-based practice star model on door-to-needle times in patients with acute ischemic stroke

Objective

To explore the impact of the academic center for evidence-based practice (ACE) star model on the door-to-needle times (DNT) in patients with acute ischemic stroke (AIS).

Methods

Clinical data of 159 patients with AIS, treated in Sir Run Run Shaw Hospital Afffliated to Zhejiang University School of Medicine Alar Hospital from March 2022 to March 2024, were retrospectively analyzed. Seventy-eight patients received routine care (routine group), and 81 patients were treated using a combination of routine care with the ACE star model (ACE star group). Operating time, intervention effects, activities of daily living (ADL), neurologic outcomes, and incidence of adverse events of the two groups were compared.

Results

The duration of venous opening, computed tomography (CT) examination, and DNT in the ACE star group were shorter than those in the routine group (P<0.05). The DNT<45 minutes compliance rate, thrombolytic efficacy, and vascular recanalization in the ACE star group were higher than those in the Routine group (P<0.05). After the intervention, the ADL score of the ACE star group was significantly higher than that of the control group, while the NIHSS score was significantly lower than that of the control group (P<0.05). There was no significant difference in the incidence of adverse events between the two groups (P>0.05).

Conclusions

Adopting routine nursing care and intervention based on the ACE-star model for patients with AIS can shorten DNT, improve thrombolytic effect and vascular recanalization rate. ACE-star model is beneficial for restoring ADL ability and improving neurological function, without significant changes in the occurrence of adverse events.

Finding the right dose: a scoping review examining facilitation as an implementation strategy for evidence-based stroke care

BACKGROUND

Despite evidence supporting interventions that improve outcomes for patients with stroke, their implementation remains suboptimal. Facilitation can support implementation of research into clinical practice by helping people develop the strategies to implement change. However, variability in the amount (dose) and type of facilitation activities/facilitator roles that make up the facilitation strategies (content), may affect the effectiveness of facilitation. This review aimed to determine if, and how, facilitation dose is measured or reported and the type of facilitation strategies used to support adoption of stroke interventions in hospitals and subacute settings. We also assessed whether the included studies had reporting checklists or guidelines.

METHODS

The scoping review was based on Arksey and O'Malley's framework. Cochrane, CINAHL and MEDLINE databases were searched to identify randomised trials and quasi-experimental studies of stroke interventions published between January 2017 and July 2023. Accompanying publications (quantitative, qualitative, mixed methods or process evaluation papers) from eligible studies were also included. Narrative data synthesis was undertaken.

RESULTS

Ten studies (23 papers) from 649 full-text papers met the inclusion criteria. Only two studies reported the total facilitation dose, measured as the frequency and duration of facilitation encounters. Authors of the remaining eight studies reported only the frequency and/or duration of varying facilitation activities but not the total dose. The facilitation activities included remote external facilitator support via ongoing telecommunication (phone calls, emails, teleconferences), continuous engagement from on-site internal facilitators, face-to-face workshops and/or education sessions from external or internal facilitators. Facilitator roles were broad: site-specific briefing, action planning and/or goal setting; identifying enablers and barriers to change; coaching, training, education or feedback; and network support. Only two studies included reporting checklists/guidelines to support researchers to describe interventions and implementation studies in sufficient detail to enable replication.

CONCLUSIONS

There is a paucity of information on the measurement of facilitation dose and reporting on specific details of facilitation activities in stroke implementation studies. Detailed reporting of dose and content is needed to improve the scientific basis of facilitation as strategic support to enable improvements to stroke care. Development of a standardised measurement approach for facilitation dose would inform future research and translation of findings.

Do implementation interventions improve evidence-based care in acute stroke settings? A Cochrane Review summary with commentary

BACKGROUND

Evidence on acute stroke management is continuously growing. Stroke units are often associated with better access to high-level evidence-based practices, but even there, recommendations can be inconsistently delivered to patients with stroke. Implementation interventions are strategies designed to improve the application of evidence-based care.

OBJECTIVE

To provide a commentary on the Cochrane Review by Lynch et al. on the effects of implementation interventions on adherence to evidence-based recommendations by health professionals working in acute stroke units.

METHODS

A systematic search was performed in CENTRAL, MEDLINE, Embase, and other databases. The search was also performed in grey literature databases, trial registries, systematic reviews and primary studies, as well as in the reference list of identified studies.

RESULTS

The review included seven cluster-randomized trials (with 42,489 participants). Studies compared the implementation of strategies composed of different parts (multifaceted) to no intervention, or a multifaceted strategy vs another intervention. These strategies were aimed at changing and improving the delivery of care in the hospital. It included health professional participants, such as nurses, physicians and allied health professionals. The authors concluded that there was uncertainty whether implementation strategies compared with no intervention have any effect on patients receiving evidence-based care during their stroke unit admission. Implementation interventions compared to no intervention probably have little or no effect on the risk of patients dying or being disabled or dependent, and probably do not change patients' hospital length of stay.

CONCLUSION

Due to the very low certainty of evidence, there is uncertainty whether a multifaceted implementation intervention, compared to no intervention, can improve adherence to evidence-based recommendations in acute stroke settings.

A Process Evaluation of the National Implementation of a Bundle for Central Venous Catheter Care for Hemodialysis

Key Points

Health professionals resisted practice change in environments of low infection where the perception of a need to change is small. Standardizing care of central venous catheters for hemodialysis requires breaking down silos of practice to benefit all patients. Knowledge of and adherence to guidelines, formal change management, and ongoing facilitation are required to implement standardized care.

Background

Implementation of a care bundle standardizing insertion, management, and removal practices to reduce infection related to central venous catheters (CVCs) used for hemodialysis was evaluated in a stepped wedge, cluster randomized controlled trial conducted at 37 Australian hospitals providing kidney services, with no reduction in catheter-related blood stream infection detected. This process evaluation explored the barriers, enablers, and unintended consequences of the implementation to explain the trial outcomes.

Methods

Qualitative process evaluation using pre-post semistructured interviews with 38 (19 nursing and 19 medical) and 44 (25 nursing and 19 medical) Australian health professionals involved in hemodialysis CVC management. Analysis was guided by the process implementation domain of the Consolidated Framework for Implementation Research.

Results

Key influences on bundle uptake were that clinicians were open to change that was evidence-based and driven by guidelines and had a desire to improve practice and patient outcomes. However, resistance to change in environments of low infection, working in silos of practice, and a need for individualized delivery of patient education created barriers to uptake. Unintended effects of increased costs and lack of interoperability of systems for data collection were reported. Because the trial was in progress at the time of qualitative data collection, perceptions of the bundle may have been influenced by the fact that practices of participants were being observed as a part of the trial.

Conclusion

This national process evaluation revealed that health professionals who reported experiencing a benefit viewed the bundle positively. Those who already provided most of the recommended care or perceived that their patient population was not included in the research evidence that underpinned the interventions, resisted the implementation of the bundle. Potentially, formal change management processes using facilitation may improve implementation of evidence-based practice.

Clinical Trial registry name and registration number:

Australian New Zealand Clinical Trials Registry, ACTRN12616000830493 .