BMJ Evidence-Based Medicine最新文献

Objective: This study aims to comprehensively review the literature on the use of speech biomarkers in disease diagnosis and monitoring, focusing on recording protocols, speech tasks, speech features and processing algorithms.

Study design: Systematic review and meta-analysis.

Data sources: We conducted a search of six databases: PubMed, Embase, Scopus, Web of Science, PsycINFO and IEEE Xplore, covering studies published from database inception to May 2024.

Main outcome measures: The quality of the included studies was assessed using the Quality Assessment of Diagnostic Accuracy Studies tool (QUADAS-2) and the Quality Assessment of Prognostic Accuracy Studies (QUAPAS). Pooled sensitivity and specificity were calculated using a random-effects model. Subgroup analyses examined potential sources of heterogeneity, such as disease type, language, speech tasks, features and algorithms.

Results: A total of 96 studies were included, with 83 adopting a cross-sectional design and 50 having sample sizes of fewer than 100 participants. Assessment with QUADAS-2 and QUAPAS revealed that most included studies exhibited a high risk of bias in patient selection and index test domains, while concerns regarding applicability were generally low across studies. These studies covered 20 different diseases, with cognitive disorders, depression and Parkinson's disease being the most frequently studied. The pooled sensitivity and specificity for diagnostic models were 0.80 (95% CI 0.74 to 0.86) and 0.77 (95% CI 0.69 to 0.84) for psychiatric disorders (11 studies, n=2577); 0.85 (95% CI 0.83 to 0.88) and 0.83 (95% CI 0.79 to 0.86) for cognitive disorders (27 studies, n=2068); and 0.81 (95% CI 0.76 to 0.85) and 0.83 (95% CI 0.78 to 0.88) for movement disorders (20 studies, n=852). Further subgroup analyses identified recording device, language, speech task, speech features and algorithm selection as significant contributors to heterogeneity.

Conclusions: This review and meta-analysis of 96 studies highlights the influence of devices, environments, languages, tasks, features and algorithms on speech model performance across diseases. While speech biomarkers show promise for screening and monitoring-particularly via smartphones-the high risk of bias in many studies, especially in patient selection and index test interpretation, limits the strength of current evidence. Future large-scale, prospective studies are needed to validate generalisability and support clinical implementation.

Prospero registration number: CRD42024551962.

Background: Evaluation of the quality of evidence in systematic reviews (SRs) is essential for assertive decision-making. Although Grading of Recommendations Assessment, Development and Evaluation (GRADE) affords a consolidated approach for rating the level of evidence, its application is complex and time-consuming. Artificial intelligence (AI) can be used to overcome these barriers.

Design: Analytical experimental study.

Objective: The objective is to develop and appraise a proof-of-concept AI-powered tool for the semiautomation of an adaptation of the GRADE classification system to determine levels of evidence in SRs with meta-analyses compiled from randomised clinical trials.

Methods: The URSE-automated system was based on an algorithm created to enhance the objectivity of the GRADE classification. It was developed using the Python language and the React library to create user-friendly interfaces. Evaluation of the URSE-automated system was performed by analysing 115 SRs from the Cochrane Library and comparing the predicted levels of evidence with those generated by human evaluators.

Results: The open-source URSE code is available on GitHub (http://www.github.com/alisson-mfc/urse). The agreement between the URSE-automated GRADE system and human evaluators regarding the quality of evidence was 63.2% with a Cohen's kappa coefficient of 0.44. The metrics of the GRADE domains evaluated included accuracy and F1-scores, which were 0.97 and 0.94 for imprecision (number of participants), 0.73 and 0.7 for risk of bias, 0.9 and 0.9 for I² values (heterogeneity) and 0.98 and 0.99 for quality of methodology (A Measurement Tool to Assess Systematic Reviews), respectively.

Conclusion: The results demonstrate the potential use of AI in assessing the quality of evidence. However, in consideration of the emphasis of the GRADE approach on subjectivity and understanding the context of evidence production, full automation of the classification process is not opportune. Nevertheless, the combination of the URSE-automated system with human evaluation or the integration of this tool into other platforms represents interesting directions for the future.

Objectives: This study aims to address the status of children's and caregivers' participation in the development of paediatric core outcome sets (COS).

Methods: We included all paediatric COS from a previous systematic review and searched the Core Outcome Measures in Effectiveness Trials database to 26 February 2024 for recent paediatric COS. We used descriptive and thematic analysis methods to present the characteristics of the included COS and to describe children's and caregivers' participation in the development, including any facilitators and barriers. We assessed the degree of participation of children and caregivers in two steps: by rating whether their views were considered in forming the outcome list (yes/no) and then whether their views were integrated in determining the most important outcomes (fully integrated/partially integrated/not integrated).

Results: A total of 114 paediatric COS were included. 60 (53%) COS involved children and caregivers in the development process. 29 (48%) of the 60 COS considered children's and caregivers' views in forming the initial outcome list, which was most often conducted by interview (n=12 of 29, 41%). Regarding determining the most important outcomes, 35 (58%) of the 60 COS fully integrated children's and caregivers' views, and the most common method was the Delphi survey with consensus meeting (n=29 of 35, 83%); the youngest child participants were aged 7 years. The most frequently mentioned facilitator of children's and caregivers' participation was the engagement of patient groups or organisations.

Conclusion and relevance: We evaluated the degree of children's and caregivers' participation in the development of COS and found that strategies to promote children's and caregivers' participation should be constructed.

Objectives: To map the risk profile of drug classes at the population level by identifying associations between newly prescribed reimbursed drugs and sudden cardiac arrest (SCA), thereby providing a systematic basis for assessing individual drug effects.

Design: Population-based, matched case-control study.

Setting: Greater Vienna area, Austria, using linked data from the Austrian Healthcare Reimbursement Database and the Vienna Ambulance Service between 2013 and 2020.

Participants: A total of 31 330 insured individuals were included from an estimated 14 448 612 person-years at risk. Cases (n=6266) were patients with SCA, each matched 1:4 to controls (n=25 064) without SCA by age, sex and event date. The estimated absolute case risk equals 43.4 per 100 000 person-years.

Interventions: Newly prescribed reimbursed drugs within 4 weeks before the SCA event, classified according to four-digit Anatomical Therapeutic Chemical (ATC) codes.

Main analyses: Associations between newly prescribed drug classes and SCA, estimated using conditional logistic regression and expressed as ORs with 95% CIs, adjusted for comorbidities and concomitant medications prescribed 120-29 days before the event.

Results: Among 322 relevant ATC drug classes, 245 were newly prescribed within 28 days prior to the SCA event. Of these, 57 (23%) were significantly associated with SCA. Eight drug classes demonstrated a markedly elevated risk (OR≥10), including oripavine derivatives (OR 64, 95% CI 8 to 486), other cardiac preparations (OR 22, 95% CI 2 to 204), plain antiandrogens (OR 18, 95% CI 1.2 to 275) and phenylpiperidine derivatives (OR 16, 95% CI 7 to 38). The most frequently prescribed associated drug classes were penicillins with beta-lactamase inhibitors (179 cases; OR 2.17, 95% CI 1.8 to 2.7), pyrazolones (167 cases; OR 2.14, 95% CI 1.7 to 2.7) and adrenergics combined with anticholinergics (130 cases; OR 2.94, 95% CI 2.2 to 3.9).

Conclusions: Numerous outpatient-prescribed drug classes were associated with SCA. This exploratory, population-based analysis provides a systematic map of potential safety signals to inform more detailed pharmaco-epidemiological investigations, in which confounding and causality can be examined more rigorously.

Objectives: To systematically review the utilisation of the Grading of Recommendations Assessment, Development and Evaluations (GRADE) framework within the context of WHO public health guidelines (PHGs), identifying key features, areas of concentration and potential deficiencies.

Design and setting: From 2007 to February 2024, a comprehensive search of the WHO website was conducted to identify PHGs that have incorporated the GRADE methodology.

Participants: The study focused on the PHGs identified through the above search.

Interventions: Data extraction and analysis were independently conducted by researchers using Microsoft Excel 2019.

Main outcome measures: For each PHG, key recorded characteristics included publication details, thematic areas, the strength of recommendations, the certainty of evidence, and characteristics of the GRADE downgrading/upgrading domains.

Results: Out of 228 PHGs examined, 9234 (90.55%) outcome indicators used the GRADE rating system, predominantly in the area of sexual and reproductive health (50%). Only 31.62% of the outcomes reported moderate and high certainty in evidence. The main clustering results were dominated by adverse events. Among the 4013 recommendations, 2067 (51.51%) were strong, while 1477 (36.81%) were weak/conditional. It is noteworthy that 46.83% of the strong recommendations were based on low or very low confidence in evidence. Among these strong recommendations, 119 met the criteria of five paradigmatic situations where it was necessary to issue a strong recommendation despite low or very low confidence in the effect estimates. Among the 13 230 instances of downgrading, 41.09% were due to the risk of bias and 35.90% were due to imprecision. Only 0.25% outcomes were upgraded by the magnitude of effect size and 0.03% by dose-response gradient.

Conclusion: The GRADE approach is widely used in the development of PHGs. More than half of the recommendations in PHGs are based on low or very low-quality evidence, primarily due to risks of bias and imprecision. Additionally, strong recommendations based on low-confidence or very low-confidence estimates are frequently made. Therefore, it is necessary to enhance guideline developers' understanding of the GRADE methodology and to further investigate and clarify discordant recommendations in PHGs.

Background: The adaptation of clinical practice guidelines (CPGs) is increasingly used to efficiently develop contextualised recommendations for local, regional and national settings. However, adaptation approaches lack standardisation, and their application remains unclear.

Objective: To identify existing literature on CPG adaptation approaches, synthesise and analyse the definitions, rationales, key steps and their applications.

Methods: We conducted a scoping review and identified studies through (1) database searches in MEDLINE and EMBASE (October 2024); (2) forward citation searches (January 2024); (3) manual searches; and (4) expert recommendations. We included documents describing adaptation approaches that allow reproducibility or adapted guidelines reflecting their application, without language restriction. We analysed data using descriptive and content analysis, with visualisations using Tableau.

Results: We identified 49 adaptation approaches (2000-2024) and 151 applications (2007-2024). 28 approaches originated from high-income countries (57.1%), primarily aimed at improving guideline development efficiency. We categorised these approaches into methodological frameworks (n=16, 32.7%), medical association frameworks (n=7, 14.3%), organisational handbooks (n=12, 24.5%), national or regional approaches (n=13, 26.5%) and one topic-specific methodology. From the key adaptation steps of the identified approaches, we found that 'judging quality of evidence' and 'establishing guideline group processes' were the least addressed. Approximately 20% of applications modified the original approaches during the adaptation process. ADAPTE (n=50, 33.1%) and the Grading of Recommendations, Assessment, Development and Evaluations (GRADE)-ADOLOPMENT (n=41, 27.2%) are the most commonly used adaptation approaches. Also, over half of the identified applications were adapted at the evidence level (n=88, 58.3%) and completed the process within 12 months (range: 2-36 months).

Conclusions: 49 adaptation approaches were identified, with a recent surge in interest and utilisation. Although adaptation can save time and resources, substantial variations among existing approaches and their applications remain. These findings highlight the need for comprehensive guidance to support standardisation and optimise the quality of the adaptation process.