浅谈系统评价的文献偏倚问题

文献偏倚是指由于无法全面地获得相关的临床研究资料,从而影响系统评价结果的真实性.常见的文献偏倚主要包括(1)文献发表偏倚,指某些研究始终不能被发表或被延迟发表;(2)文献查寻偏倚,指系统评价过程中,没有全面地获得已发表的相关文献;(3)文献筛选偏倚,是由于文献筛选不当,错误地剔除了某些有用的文献.文献偏倚评估的常用方法是以各研究的样本量为纵坐标,作用大小为横坐标,绘制漏斗形图.     

偏倚对系统评价质量的影响

系统评价被认为是当前提供治疗性干预的最佳证据,但是由于无法全面获得相关的研究资料,仍然不可避免的存在偏倚。偏倚的类型主要包括文献发表性偏倚、文献查寻偏倚和文献筛选偏倚,其中最难克服的是文献发表性偏倚。偏倚评估的方法常用漏斗图,控制方法包括预先注册临床试验、发布研究信息、系统评价时全面收集所有发表和未发表的临床试验。     

脑胶质瘤干预类系统评价/Meta分析的方法学和报告质量评价

目的系统评价全世界脑胶质瘤干预类系统评价/Meta分析（SR/MA）的方法学和报告质量。方法计算机检索Pub Med、EMbase、h e Cochrane Library、CNKI、CBM等数据库,纳入脑胶质瘤干预类SR/MA,检索时限截至2013年7月。由2位研究者独立筛选文献,而后采用AMSTAR和PRISMA清单对纳入研究的方法学与报告质量进行评价与分析。结果共纳入51个SR/MA。结果显示：纳入研究中方法学质量存在的主要问题有无研究设计方案、检索策略不全面、纳入研究出版物形式局限、未评价文章发表偏倚及未说明相关利益冲突;纳入研究的报告质量存在的主要问题是检索策略的报告不规范、纳入研究质量和偏倚风险报道不全面和研究结果表述不全（部分缺乏森林图、综合结果的估计值和可信区间、异质性检验结果）。结论脑胶质瘤干预类SR/MA的方法学质量和报告质量还存在不同程度的问题,该领域研究者应提高SR/MA制作的科学性和规范性,并遵循PRISMA进行报告。     

Methodological approaches in analysing observational data: A practical example on how to address clustering and selection bias

BackgroundBecause not every scientific question on effectiveness can be answered with randomised controlled trials, research methods that minimise bias in observational studies are required. Two major concerns influence the internal validity of effect estimates: selection bias and clustering. Hence, to reduce the bias of the effect estimates, more sophisticated statistical methods are needed.AimTo introduce statistical approaches such as propensity score matching and mixed models into representative real-world analysis and to conduct the implementation in statistical software R to reproduce the results. Additionally, the implementation in R is presented to allow the results to be reproduced.MethodWe perform a two-level analytic strategy to address the problems of bias and clustering: (i) generalised models with different abilities to adjust for dependencies are used to analyse binary data and (ii) the genetic matching and covariate adjustment methods are used to adjust for selection bias. Hence, we analyse the data from two population samples, the sample produced by the matching method and the full sample.ResultsThe different analysis methods in this article present different results but still point in the same direction. In our example, the estimate of the probability of receiving a case conference is higher in the treatment group than in the control group. Both strategies, genetic matching and covariate adjustment, have their limitations but complement each other to provide the whole picture.ConclusionThe statistical approaches were feasible for reducing bias but were nevertheless limited by the sample used. For each study and obtained sample, the pros and cons of the different methods have to be weighted.     

Implications of systematic review for breast cancer prediction

Highly accurate and predictive models are essential components to promote early breast cancer screening in primary care or home care settings. This study was conducted to demonstrate how the relevant variable selection process influenced the predictive performance of the model to identify individuals at high risk for breast cancer. As such, as a strategy to increase the predictive performance of the models, a systematic review of previously published articles was conducted to select important risk factors for breast cancer. Through the systematic literature review and the application of variable selection methods, 13 final risk factors were identified. Logistic regression and naive Bayes predictive modeling techniques were used. Both models had higher predictive performances than previously developed models. It is believed that the systematic literature review process contributed to the identification of relevant variables and increased the predictive performance of the models. This study also implies that the naive Bayes was equivalent to and could be preferred over logistic regression.     

The risk assessment tool for intensive care unit readmission: A systematic review and meta-analysis

ObjectiveTo review and evaluate existing risk assessment tools for intensive care unitreadmission.MethodsNine electronic databases (Medline, CINAHL, Web of Science, Cochrane Library, Embase, Sino Med, CNKI, VIP, and Wan fang) were systematically searched from their inception to September 2022. Two authors independently extracted data from the literature included. Meta-analysis was performed under the bivariate modeling and summary receiver operating characteristic curve method.ResultsA total of 29 studies were included in this review, among which 11 were quantitatively Meta-analyzed. The results showed Stability and Workload Index for Transfer: Sensitivity = 0.55, Specificity = 0.65, Area under curve = 0.63. And Early warning score: Sensitivity = 0.78, Specificity = 0.83, Area under curve = 0.88. The remaining tools included scores, nomograms, machine learning models, and deep learning models. These studies, with varying reports on thresholds, case selection, data preprocessing, and model performance, have a high risk of bias.ConclusionWe cannot identify a tool that can be used directly in intensive care unit readmission risk assessment. Scores based on early warning score are moderately accurate in predicting readmission, but there is heterogeneity and publication bias that requires model adjustment for local factors such as resources, demographics, and case mix. Machine learning models present a promising modeling technique but have a high methodological bias and require further validation.Implications for clinical practiceUsing reliable risk assessment tools is essential for the early identification of unplanned intensive care unit readmission risk in critically ill patients.  A reliable risk assessment tool must be developed, which is the focus of further research.     

The representation of healthcare end users' perspectives by surrogates in healthcare decisions: a literature review

Background: The representation of end users’ perspectives in healthcare decisions requires involvement of their surrogates when the end users, i.e. certain patients, elderly people, children and people with disabilities, are unable to present their views. Aims: To review critical issues, and the advantages and disadvantages of involving surrogates in representing end users’ perspectives in healthcare decisions. Methods: A systematic review of literature published in peer‐reviewed journals from 1990 to 2005. Results: Findings show that surrogates are used widely in health care and that they are necessary to represent end users’ perspectives in healthcare decisions when the latter are unable to do so themselves. Critical issues in using surrogates include key ethical, social, cultural, legal and medico‐technological factors; ascertaining the best interest of end users; potential conflict of interest; possible biased decisions and the burden on surrogates. The key advantage of surrogate involvement in healthcare decisions is their ability to represent end users’ needs, values and wishes. The main disadvantages include potential discrepancies between the decisions and conclusions of surrogates and end users; the failure of surrogates to predict end users’ preferences accurately and the lack of certainty that useful information will be obtained through the surrogacy process. Conclusion: This systematic review has revealed that the involvement of surrogates is an additional vital way to represent end users’ perspectives in healthcare decisions where for a range of reasons their opinions are unable to be effectively ascertained. However, because of the heterogeneity of surrogates and end users, the selection of appropriate surrogates and deploying surrogate decisions require particularly careful consideration of their value in individual cases; thus, subsequent decision‐making must be reviewed on a case‐to‐case basis to seek to ensure that the best interests, needs and wishes of the end user are fully and accurately represented.     

卒中后谵妄危险因素的Meta分析

目的 通过系统评价和Meta分析,明确卒中后谵妄（post stroke delirium, PSD）的危险因素,为有效预防脑卒中患者卒中后谵妄提供科学依据。方法 在PubMed、EMbase、CINAHL、 Cochrane Library、PsycINFO、中国知网、万方数据库、中国生物医学文献数据库（CBM）、维普中文生物医学期刊数据库（VIP）中进行文献检索,收集2018年2月之前公开发表的关于卒中后谵妄危险因素的研究,并进行文献质量评价,应用Stata 14.0进行Meta分析,采用固定效应或随机效应模型进行数据合并,计算OR值和95%可信区间,用Begg秩相关和漏斗图判定发表偏倚。结果 共纳入14篇文献,包括4 807例卒中患者,其中发生卒中后谵妄的患者784例（16.31%）,文献质量评价,均为中等或高质量研究。危险因素合并的OR值分别为：视觉障碍31.76（95%CI：11.61~86.86）、C反应蛋白升高7.64（95%CI：1.83~31.90）、认知障碍 7.48（95%CI：3.63~15.42）、前循环梗死5.24（95%CI：2.89~9.49）、出血性卒中3.70（95%CI：3.64~3.77）、感染3.67（95%CI：2.19~6.16）、卒中严重程度1.23（95%CI：1.06~1.42）、年龄1.07（95%CI：1.02~1.13）。Begg秩相关分析显示,前循环梗死存在一定的发表偏倚。结论 视觉障碍、C反应蛋白升高、认知障碍、出血性卒中、感染、卒中严重程度、年龄等是卒中后谵妄的独立预测因子,而前循环梗死是卒中后谵妄危险因素的证据尚不充分。应针对脑卒中人群中存在的可改变的危险因素进行早期干预。     

Publishing Clinical Trials and Dealing with External Comparisons

This brief note is designed to provide research teams that are involved with the publication of clinical trials in peer-reviewed medical journals with a reference for the rebuttal of referees’ criticism that your results are in some way inferior overall to those of others. In these reviewers’ eyes, a trial that shows less benefit than other trials has no public health benefit, and hence they recommend rejection. In fact, such policies lead to publication bias, the systematic overestimation of benefits of therapy in the published literature.     

系统评价控制偏倚的方法

系统评价与传统综述的最大区别在于其质量不同,即在于其减少偏倚的程度不同.一般而言,偏倚的来源主要包括以下方面:文献检索过程、研究选择过程和数据提取过程产生的偏倚以及所选择的各个研究内在的偏倚.由于系统评价采取了诸如制订检索策略、进行"倒漏斗"分析、预先确立研究选择的纳入与排除标准以及评估纳入研究的方法学质量等有效措施对可能存在的偏倚进行控制、识别以及描述,因而能够大大地减少偏倚.     

Does Implementation of Biomathematical Models Mitigate Fatigue and Fatigue-related Risks in Emergency Medical Services Operations? A Systematic Review

Abstract

Background: Work schedules like those of Emergency Medical Services (EMS) personnel have been associated with increased risk of fatigue-related impairment. Biomathematical modeling is a means of objectively estimating the potential impacts of fatigue on performance, which may be used in the mitigation of fatigue-related safety risks. In the context of EMS operations, our objective was to assess the evidence in the literature regarding the effectiveness of using biomathematical models to help mitigate fatigue and fatigue-related risks. Methods: A systematic review of the evidence evaluating the use of biomathematical models to manage fatigue in EMS personnel or similar shift workers was performed. Procedures proposed by the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) methodology were used to summarize and rate the certainty in the evidence. Potential bias attached to retained studies was documented using the Cochrane Collaboration's Risk of Bias tool for experimental studies. Results: The literature search strategy, which focused on both EMS personnel and non-EMS shift workers, yielded n = 2,777 unique records. One paper, which investigated non-EMS shift workers, met inclusion criteria. As part of a larger effort, managers and dispatchers of a trucking operation were provided with monthly biomathematical model analyses of predicted fatigue in the driver workforce, and educated on how they could reduce predicted fatigue by means of schedule adjustments. The intervention showed a significant reduction in the number and cost of vehicular accidents during the period in which biomathematical modeling was used. The overall GRADE assessment of evidence quality was very low due to risk of bias, indirectness, imprecision, and publication bias. Conclusions: This systematic review identified no studies that investigated the impact of biomathematical models in EMS operations. Findings from one study of non-EMS shift workers were favorable toward use of biomathematical models as a fatigue mitigation scheduling aid, albeit with very low quality of evidence pertaining to EMS operations. We propose three focus areas of research priorities that, if addressed, could help better elucidate the utility and impact of biomathematical models as a fatigue-mitigation tool in the EMS environment.     

Publication bias: a brief review for clinicians

Systematic reviews and meta-analyses provide the highest level of evidence to guide clinical decisions and inform practice guidelines. Publication bias results from the selective publication of studies based on the direction and magnitude of their results--studies without statistical significance (negative studies) are less likely to be published. Bias results from pooling the results from published studies alone leading to overestimation of the effectiveness of the intervention. In this review we define publication bias, how it affects the results of systematic reviews, how it can be detected and minimized, and how it can be prevented.     

Systematic reviews in laboratory medicine: principles, processes and practical considerations

BACKGROUND: Systematic reviews and meta-analyses are generally accepted to represent the highest level of evidence, and are a cornerstone in practising evidence-based medicine. So far, these efforts have been largely confined to the evaluation of the efficacy and effectiveness of therapeutic and preventive interventions. Systematic reviews in laboratory medicine are scarce and many of them do not meet essential quality criteria [Clin. Chem. Lab. Med. 38 (2000) 577]. Most of these problems are related to the poor design and heterogeneity of primary research, and that there are no agreed methods or quality standards for making systematic reviews in laboratory medicine. AIMS AND OBJECTIVES: For better evidence in laboratory medicine, not only higher quality primary studies but also standardized methodologies for designing, conducting and reporting systematic reviews in diagnostics are needed. The aim of this review is to present the general principles and provide a step-by-step process of systematic reviewing in laboratory medicine. METHODS: This narrative review is based on the overview of the medical literature on the methodology of systematic reviewing and that of the "state of the art" of evidence-based diagnosis. RESULTS: Systematic reviews of diagnostic interventions differ from that of therapeutic interventions in the methods of question formulation, the choice of study design, the assessment of study quality and the statistical methods used to combine results. Therefore, the general principles of systematic reviewing are adapted to the specialist field of laboratory medicine. The process of systematic reviewing consists of six key steps: (1) preparation for the review, (2) systematic search of the primary literature, (3) selection of papers for review, (4) critical appraisal of the selected literature, (5) analysis and synthesis of data, and (6) interpretation of data. The most important technical and methodological aspects of each step and the essential elements of a good systematic review in laboratory medicine are presented. CONCLUSIONS: Systematic reviews of diagnostic interventions support clinical and policy decisions, the development of practice guidelines, clinical audit, technology assessment, economic evaluations, education and training, and identify gaps in our knowledge for future research. Systematic reviewing of laboratory data is expected to result in better, bigger and more reliable primary studies, which hopefully will support the diffusion of new diagnostic technologies with scientifically proven efficacy and effectiveness in the future.     

Conducting a systematic review.

In response to the growing volume of health care literature and the variable quality of reported studies, systematic reviews have increasingly been used to guide health care decisions because of their rigorous summary of the research. Systematic reviews utilise planned methods of identifying, appraising, then summarising the results from individual studies. The steps in performing a systematic review include: preparing a detailed research protocol; selecting criteria for inclusion of articles in the review; systematically searching the published and unpublished literature; determining which articles meet the predefined inclusion criteria; critically appraising the quality of the research; extracting outcome data from the research report and statistically combining data, where appropriate, in order to summarise the best available evidence on the topic of interest. These processes are documented in the systematic review report, and can be subject to peer review and critique like other research.     

Risk factors for candidemia in neonates: Systematic review and meta-analysis

Premature and low birth weight newborns are at greater risk of developing nosocomial infections, including Candida spp infections. This is a systematic literature review to identify risk factors for developing candidemia in neonates. The search was performed in the VHL, Cochrane, Embase, Pubmed, Scielo, Scopus and Web of Science databases and the terms were: Candidemia, Candidiasis, Candida spp. Invasive Candidiasis, Risk Factors, Infant, Newborn and Neonatology. Meta-analysis was performed for risk factor using Comprehensive Meta Analysis. A total of 2218 articles were identified and 42 papers were included. The main risk factors were: low weight, previous surgery, intubation, mechanical ventilation, central venous catheter, antibiotics and parenteral nutrition. Low birth weight and neonatal units care assistance factors were the main predictors of Candidemia in neonates. Prospective studies are necessary with systematic follow-up during hospitalization to define measures to prevent infection, since retrospective studies have inherent selection and information bias.     

Studying outcomes of intensive care unit survivors: measuring exposures and outcomes

Background Measurement of long-term outcomes and the patient and intensive care unit (ICU) factors predicting them present investigators with unique challenges. There is little systematic guidance for measuring these outcomes and exposures within the ICU setting. As a result measurement methods are often variable and noncomparable across studies.Methods We use examples from the critical care literature to describe measurement as it relates to three key elements of clinical studies: subjects, outcomes and exposures, and time. Using this framework we review the principles and challenges of measurement and make recommendations for long-term outcomes research in the field of critical care medicine.Discussion Relevant challenges discussed include: (a) selection bias and heterogeneity of ICU research subjects, (b) appropriate selection and measurement of outcome and exposure variables, and (c) accounting for the effect of time in the exposure-outcome relationship, including measurement of baseline data and time-varying variables.Conclusions Addressing these methodological challenges will advance research aimed at improving the long-term outcomes of ICU survivors.Electronic Supplementary Material Electronic supplementary material to this paper can be obtained by using the Springer Link server located at .This research was supported by National Institutes of Health (ALI SCCOR Grant # P050 HL 73994-01). D.M.N. is supported by Clinician-Scientist Awards from the Canadian Institutes of Health Research and the University of Toronto, and a Detweiler Fellowship from the Royal College of Physicians and Surgeons of Canada.     

Review of the Efficacy and Effectiveness of Adaptive Seating for Children with Cerebral Palsy

The purpose of this article is to review the published research evidence for the efficacy and effectiveness of adaptive seating in the management of children with cerebral palsy. This review is undertaken to determine whether adaptive seating outcomes proposed in the clinical literature have been substantiated. The literature search was confined to English-language publications between 1982 and 1994. Thirty-seven research studies were identified through: 1) on-line search of Index Medicus, Cumulative Index to Nursing and Allied Health Literature, and Psychological Abstracts; 2) scanning conference proceedings and article reference lists; and, 3) tracking of author citations through the Science Citations Index. Eight studies met the selection criteria of using subjects who were children with cerebral palsy, including adaptive seating as an independent variable, and including an unsupported sitting comparison condition. Study quality was evaluated using an adaptation of Sackett's validity criteria. Sackett's classification system was utilized to identify levels of evidence and to grade recommendations. Grade A, B, and C recommendations were generated from Level I to Level V evidence for some of the clinically proposed seating outcomes. Additional seating research is required to strengthen the recommendations and investigate unexplored outcomes. Limitations of the review are potential reviewer and publication bias.     

Review of the efficacy and effectiveness of adaptive seating for children with cerebral palsy

The purpose of this article is to review the published research evidence for the efficacy and effectiveness of adaptive seating in the management of children with cerebral palsy. This review is undertaken to determine whether adaptive seating outcomes proposed in the clinical literature have been substantiated. The literature search was confined to English-language publications between 1982 and 1994. Thirty-seven research studies were identified through: 1) on-line search of Index Medicus, Cumulative Index to Nursing and Allied Health Literature, and Psychological Abstracts; 2) scanning conference proceedings and article reference lists; and, 3) tracking of author citations through the Science Citations Index. Eight studies met the selection criteria of using subjects who were children with cerebral palsy, including adaptive seating as an independent variable, and including an unsupported sitting comparison condition. Study quality was evaluated using an adaptation of Sackett's validity criteria. Sackett's classification system was utilized to identify levels of evidence and to grade recommendations. Grade A, B, and C recommendations were generated from Level I to Level V evidence for some of the clinically proposed seating outcomes. Additional seating research is required to strengthen the recommendations and investigate unexplored outcomes. Limitations of the review are potential reviewer and publication bias.     

康复期脑卒中患者跌倒风险预测模型的系统评价

目的 系统评价康复期脑卒中患者跌倒风险预测模型,以期为临床实践提供参考依据。 方法 系统检索中国知网、万方数据知识服务平台、PubMed、Cochrane和Embase数据库截至2021年6月发表的康复期脑卒中患者跌倒风险预测模型相关文献,语种限定为中文和英文。由2名研究者独立筛选文献和提取数据,对纳入研究的人群特征、研究类型、预测因素、模型构建方法及模型预测结果等进行分析和比较。 结果 共纳入12项研究,包含15个模型,其中10个模型报告了模型的曲线下面积,有4个模型同时报告了曲线下面积的可信区间,4个模型表现出良好的区分度。仅有6个模型进行了拟合优度检验,其中4个模型报告了P值,结果显示一致性良好。12项研究的适用性较好,但存在较高的偏倚风险,主要是因为分析领域的应变量事件数不足、忽略缺失数据、基于单变量分析筛选预测因子、缺乏模型性能评估以及模型过度拟合。 结论 康复期脑卒中患者跌倒风险预测模型尚存在一些不足,未来研究应完善研究设计和研究报告,并进行内部验证和外部验证。