Journal of clinical epidemiology

Publications in this journal

• Article: Tailoring research to stakeholders.

  J André Knottnerus, Peter Tugwell
  Journal of clinical epidemiology 06/2013; 66(6):583-4.
• Article: Optimal type I and type II error pairs when the available sample size is fixed.

  John P A Ioannidis, Iztok Hozo, Benjamin Djulbegovic
  [show abstract] [hide abstract]
  ABSTRACT: OBJECTIVE: To model how to select the optimal pair of type I and type II errors that maximize study value when there are constrains on the available study sample size. STUDY DESIGN AND SETTING: Correct inferences [true positives (TPs) and true negatives (TNs)] increase and wrong inferences (false positives and false negatives) decrease the value of a study. We model the composite value of a study based on these four inferences, their relative importance, and relative frequency using multiplicative and additive models. Numerical examples are presented for randomized trials, epidemiologic studies, and agnostic omics investigations with massive testing and variable sample size constraints. RESULTS: The optimal choice of type I and type II errors varies a lot according to the available sample size and the plausible effect sizes in each field. We show how equations can be streamlined for special applications: when the value of all four inferences is considered equal, when the identification of TNs carries no value, and when a study carries no value unless at least one TP is discovered. CONCLUSION: The proposed optimization equations can be used to guide the selection of the optimal type I and type II errors of future studies in which sample size is constrained.
  Journal of clinical epidemiology 05/2013;
• Article: Not PEDro's bias: summary quality scores can be used in meta-analysis.

  Suhail A R Doi, Jan J Barendregt
  Journal of clinical epidemiology 05/2013;
• Article: Individual participant data meta-analyses should not ignore clustering.

  Ghada Abo-Zaid, Boliang Guo, Jonathan J Deeks, Thomas P A Debray, Ewout W Steyerberg, Karel G M Moons, Richard David Riley
  [show abstract] [hide abstract]
  ABSTRACT: OBJECTIVES: Individual participant data (IPD) meta-analyses often analyze their IPD as if coming from a single study. We compare this approach with analyses that rather account for clustering of patients within studies. STUDY DESIGN AND SETTING: Comparison of effect estimates from logistic regression models in real and simulated examples. RESULTS: The estimated prognostic effect of age in patients with traumatic brain injury is similar, regardless of whether clustering is accounted for. However, a family history of thrombophilia is found to be a diagnostic marker of deep vein thrombosis [odds ratio, 1.30; 95% confidence interval (CI): 1.00, 1.70; P = 0.05] when clustering is accounted for but not when it is ignored (odds ratio, 1.06; 95% CI: 0.83, 1.37; P = 0.64). Similarly, the treatment effect of nicotine gum on smoking cessation is severely attenuated when clustering is ignored (odds ratio, 1.40; 95% CI: 1.02, 1.92) rather than accounted for (odds ratio, 1.80; 95% CI: 1.29, 2.52). Simulations show models accounting for clustering perform consistently well, but downwardly biased effect estimates and low coverage can occur when ignoring clustering. CONCLUSION: Researchers must routinely account for clustering in IPD meta-analyses; otherwise, misleading effect estimates and conclusions may arise.
  Journal of clinical epidemiology 05/2013;
• Article: A framework for the analysis of heterogeneity of treatment effect in patient-centered outcomes research.

  Ravi Varadhan, Jodi B Segal, Cynthia M Boyd, Albert W Wu, Carlos O Weiss
  [show abstract] [hide abstract]
  ABSTRACT: Individuals vary in their response to a treatment. Understanding this heterogeneity of treatment effect is critical for evaluating how well a treatment can be expected to work for an individual or a subgroup of individuals. An overemphasis on hypothesis testing has resulted in a dichotomy of all heterogeneity of treatment effect analyses into confirmatory (hypothesis testing) and exploratory (hypothesis finding) analyses. This limited view of heterogeneity of treatment effect is inadequate for creating evidence that is useful for informing patient-centered decisions. An expanded framework for heterogeneity of treatment effect assessment is proposed. It recognizes four distinct goals of heterogeneity of treatment effect analyses: hypothesis testing, hypothesis finding, reporting subgroup effects for meta-analysis, and individual-level prediction. Accordingly, two new types of heterogeneity of treatment effect analyses are proposed: descriptive and predictive. Descriptive heterogeneity of treatment effect analyses report treatment effects for prespecified subgroups in accordance with prospectively specified analytic strategy. They need not be powered to detect heterogeneity of treatment effect. They emphasize estimation and reporting of subgroup effects rather than hypothesis testing. Sampling properties (e.g., standard error) of descriptive analysis can be characterized, thus facilitating meta-analysis of subgroup effects. Predictive heterogeneity of treatment effect analyses estimate probabilities of beneficial and adverse responses of individuals to treatments and facilitates optimal treatment decisions for different types of individuals. Procedures are also suggested to improve reliability of heterogeneity of treatment effect assessment from observational studies. Heterogeneity of treatment effect analysis should be identified as confirmatory, descriptive, exploratory, or predictive analysis. Evidence should be interpreted in a manner consistent with the analytic goal.
  Journal of clinical epidemiology 05/2013;
• Article: Interviewer effects when investigating abuse were not compatible with effect modification but instead with confounding.

  Sílvia Fraga, Raquel Lucas, Diogo Costa, Henrique Barros
  [show abstract] [hide abstract]
  ABSTRACT: OBJECTIVES: To describe interviewer-related variability in abuse estimates and assess the nature of the interviewer effects on the associations between elder abuse and covariates. STUDY DESIGN AND SETTING: After intensive training, six interviewers administered structured questionnaires through face-to-face interviews to assess abuse in a population-based sample of 641 Portuguese individuals aged 60-84 years. RESULTS: The overall prevalence of abuse victimization during the previous year was 28.1%, but it differed significantly according to the interviewer, ranging from 16.9% to 36.8%. There was no statistical effect modification introduced by the interviewer on the association of abuse and its determinants. Additionally, interviewer-level variables (empathy and violence beliefs) showed no significant contribution to explain the variance attributable to potential interviewer effects. Adjusting for the interviewer had little or no effect on the odds ratio of abuse for gender, age, education, and quality of life. However, the interviewer introduced relevant confounding of the associations between abuse and other sensitive topics, such as somatic complaints. CONCLUSION: Although no relevant effect modification was observed, this study emphasizes the importance of the interviewer as a relevant confounder when estimating associations between sensitive variables, as it is the case of elder abuse.
  Journal of clinical epidemiology 05/2013;
• Article: The calibrated, unidimensional anxiety item bank for cardiovascular patients provided the basis for anxiety assessment in cardiovascular rehabilitation patients.

  Birgit Abberger, Anne Haschke, Claudia Krense, Markus Wirtz, Juergen Bengel, Harald Baumeister
  [show abstract] [hide abstract]
  ABSTRACT: OBJECTIVE: Computer adaptive tests (CATs) offer a flexible, test fair, and economic opportunity for accurate measurement of anxiety in patients with cardiovascular diseases (CVDs). The objective of this study was to develop and calibrate an item bank [anxiety item bank for cardiovascular patients (AIB-cardio)] as a prerequisite for an anxiety-CAT in CVD patients. STUDY DESIGN AND SETTING: After pretesting for relevance and comprehension, a pool of 155 anxiety items was answered on a five-point Likert scale. Sample consisted of 715 CVD patients, who were recruited in 14 German cardiac rehabilitation centers. A confirmatory factor analysis (CFA), Mokken analysis, and Rasch analysis were conducted. RESULTS: The results of CFA and Mokken analysis confirmed one factor structure and double monotonicity. In Rasch analysis, merging response categories and removing items with misfit, differential item functioning or local response dependency reduced the AIB-cardio to 37 items. The AIB-cardio fitted to the Rasch model with a nonsignificant item-trait interaction (chi-square, 133.89; degrees of freedom, 111; P = 0.07). Person separation reliability was 0.85, and unidimensionality could be verified. CONCLUSION: The calibrated, unidimensional AIB-cardio provides the basis for a CAT to assess anxiety in rehabilitation patients with CVD with good psychometric properties. Further testing in other cardiovascular patients is needed to increase generalizability.
  Journal of clinical epidemiology 05/2013;
• Article: Methods for setting priorities in systematic reviews.

  Peter Tugwell, J André Knottnerus, Leanne Idzerda
  Journal of clinical epidemiology 05/2013; 66(5):467-8.
• Article: Parents and school children reported symptoms and treatment of allergic disease differently.

  Caroline S Danell, Anna Bergström, Carl-Fredrik Wahlgren, Eva Hallner, Maria Böhme, Inger Kull
  [show abstract] [hide abstract]
  ABSTRACT: OBJECTIVE: To examine the difference between children and their parents in reporting symptoms and treatment of allergic diseases within a longitudinal birth cohort. STUDY DESIGN AND SETTING: Information on symptoms and treatment of asthma, rhinitis, and eczema was obtained by questionnaire from 2,744 children (mean age: 12 years) and their parents. Differences between the responses were computed, and agreement assessed both absolutely and with kappa coefficient. RESULTS: On 12 of the 15 questions, children's and parents' reports differed significantly. Asthma-related issues appeared significantly more prevalent in the children's reports, although kappa values were fair to very good. For symptoms of allergic rhinitis, the prevalence pattern varied, and kappa values were moderate to good. Parents reported a higher prevalence of eczema-related issues, but the children reported a significantly higher prevalence of eczema itself. Kappa values ranged from moderate to good. CONCLUSION: Although reports of allergic symptoms and treatment by 12-year-old children and their parents were in moderate-to-good agreement, children reported more symptoms than their parents. Symptoms of allergic disease should be reported by children themselves, from the age of 11 years, whereas questions of prescribed pharmacological treatment could be answered either by the children or their parents.
  Journal of clinical epidemiology 04/2013;
• Article: Minimum clinically important differences identified for commonly used depression rating scales.

  Sarah C Masson, Aaron M Tejani
  Journal of clinical epidemiology 04/2013;
• Article: Clinical experience may affect clinician compliance with assigned treatment in randomized trials.

  Stephen D Walter, Afisi S Ismaila, Deborah J Cook, Mohit Bhandari, Kari A O Tikkinen, Gordon H Guyatt
  [show abstract] [hide abstract]
  ABSTRACT: OBJECTIVE: To examine the relationship between clinical experience and clinician compliance with the study protocol in randomized clinical trials. STUDY DESIGN AND SETTING: A recent randomized trial of surgical techniques for tibial fracture fixation. We consider rates of treatment crossovers and other noncompliance as a function of the relevant experience of the surgeon. We also examined the effects of noncompliance on patient outcomes. RESULTS: Crossovers from assigned treatment to the alternative occurred much more frequently in one arm than the other. The impact of surgical experience on crossovers was less clear, although there was some evidence that noncompliance with more difficult surgery was more frequent for less experienced surgeons. This raises the possibility that experience may be an important factor in other scenarios, affecting both compliance and patient outcomes. CONCLUSION: In randomized clinical trials, noncompliance by clinicians with the randomly assigned treatment can be highly detrimental to the power of the study. Further research is needed in this area to identify, quantify, and understand the factors associated with noncompliance, including clinical experience.
  Journal of clinical epidemiology 04/2013;
• Article: GRADE guidelines 15: Going from evidence to recommendation-determinants of a recommendation's direction and strength.

  Jeffrey C Andrews, Holger J Schünemann, Andrew D Oxman, Kevin Pottie, Joerg J Meerpohl, Pablo Alonso Coello, David Rind, Victor Montori, Juan Pablo Brito Campana, Susan Norris, Mahmoud Elbarbary, Piet Post, Mona Nasser, Vijay Shukla, Roman Jaeschke, Jan Brozek, Ben Djulbegovic, Gordon Guyatt
  [show abstract] [hide abstract]
  ABSTRACT: In the GRADE approach, the strength of a recommendation reflects the extent to which we can be confident that the composite desirable effects of a management strategy outweigh the composite undesirable effects. This article addresses GRADE's approach to determining the direction and strength of a recommendation. The GRADE describes the balance of desirable and undesirable outcomes of interest among alternative management strategies depending on four domains, namely estimates of effect for desirable and undesirable outcomes of interest, confidence in the estimates of effect, estimates of values and preferences, and resource use. Ultimately, guideline panels must use judgment in integrating these factors to make a strong or weak recommendation for or against an intervention.
  Journal of clinical epidemiology 04/2013;
• Article: Shrinkage methods enhanced the accuracy of parameter estimation using Cox models with small number of events.

  I-Feng Lin, Wushou Peter Chang, Yi-Nan Liao
  [show abstract] [hide abstract]
  ABSTRACT: OBJECTIVE: When the number of events is small during Cox regression analysis, it is unclear what alternative analytical strategies can be used and when this type of alternative approach is needed. This study explores several analytical strategies in this situation. STUDY DESIGN AND SETTING: Simulations and sensitivity analyses were performed on data with numbers of events per predictive variable (EPVs) below 10 using a Cox model with a partial likelihood (PL), Firth's penalized likelihood, or the Bayesian approach. RESULTS: For scenarios involving binary predictors with an EPV of six or less, the simulations showed that the Firth and Bayesian approaches were more accurate than was PL. The performances of various approaches were similar when the EPV was greater than six in the binary predictor. Furthermore, the performances involving continuous predictors were similar, regardless of the EPV. The bias and precision of the parameter estimates using Bayesian analysis depended on the selection of priors. CONCLUSIONS: When the EPV is six or less, the results for categorical predictors tend to be too conservative. Firth's estimator may be a good alternative in this situation. Appropriate choices of priors when using Bayesian analysis should increase the accuracy of the parameter estimates, although this requires expertise.
  Journal of clinical epidemiology 04/2013;
• Article: Item reduction based on rigorous methodological guidelines is necessary to maintain validity when shortening composite measurement scales.

  Christophe Goetz, Joël Coste, Fabienne Lemetayer, Anne-Christine Rat, Sébastien Montel, Sophie Recchia, Marc Debouverie, Jacques Pouchot, Elisabeth Spitz, Francis Guillemin
  [show abstract] [hide abstract]
  ABSTRACT: OBJECTIVE: To review current practice and update guidelines for the methodology of shortening composite measurement scales (CMSs). STUDY DESIGN AND SETTING: A literature review gathered data on 91 shortening processes from 1995 to 2009. The validity of the initial CMS, the shortening methods, and the validity of the derived short-form scales were examined. The results were compared with those from a previous literature review (articles from 1985 to 1995) to develop updated guidelines for CMS shortening. RESULTS: The literature review revealed a persisting lack of use of rigorous methodology for CMS shortening. Of the 91 cases of CMS shortening, 36 combined a content approach and a statistical approach; 45 used only a statistical approach and 10 (11%) only a content approach. The updated guidelines deal with the validity and conceptual model of the initial CMS, the preservation of content and psychometric properties during shortening, the selection of items, and the validation of the short form. CONCLUSION: Item reduction based on a rigorous methodology is necessary if the short-form instrument aims to maintain the validity and other measurement properties of the parent instrument, which in turn supports application in research and clinical practice.
  Journal of clinical epidemiology 04/2013;
• Article: Effective writing and publishing scientific papers-part I: how to get started.

  Daniel Kotz, Jochen W L Cals
  Journal of clinical epidemiology 04/2013; 66(4):397.
• Article: Introducing a new series on effective writing and publishing of scientific papers.

  Daniel Kotz, Jochen W L Cals, Peter Tugwell, J André Knottnerus
  Journal of clinical epidemiology 04/2013; 66(4):359-60.
• Article: How to write a research paper.

  J André Knottnerus, Peter Tugwell
  Journal of clinical epidemiology 04/2013; 66(4):353-4.
• Article: The Short Form 36 English and Chinese versions were equivalent in a multiethnic Asian population.

  Maudrene L S Tan, Hwee-Lin Wee, Jeannette Lee, Stefan Ma, Derrick Heng, E-Shyong Tai, Julian Thumboo
  [show abstract] [hide abstract]
  ABSTRACT: OBJECTIVE: The primary aim of this article was to evaluate measurement equivalence of the English and Chinese versions of the Short Form 36 version 2 (SF-36v2) and Short Form 6D (SF-6D). STUDY DESIGN AND SETTING: In this cross-sectional study, health-related quality of life (HRQoL) was measured from 4,973 ethnic Chinese subjects using the SF-36v2 questionnaire. Measurement equivalence of domain and utility scores for the English- and Chinese-language SF-36v2 and SF-6D were assessed by examining the score differences between the two languages using linear regression models, with and without adjustment for known determinants of HRQoL. Equivalence was achieved if the 90% confidence interval (CI) of the differences in scores, due to language, fell within a predefined equivalence margin. RESULTS: Compared with English-speaking Chinese, Chinese-speaking Chinese were significantly older (47.6 vs. 55.5 years). All SF-36v2 domains were equivalent after adjusting for known HRQoL. SF-6D utility/items had the 90% CI either fully or partially overlap their predefined equivalence margin. CONCLUSION: The English- and Chinese-language versions of the SF-36v2 and SF-6D demonstrated equivalence.
  Journal of clinical epidemiology 03/2013;