Study-level Data Research Articles

Combining individual participant and aggregated data in a meta-analysis with correlational studies.

This paper presents methods for combining individual participant data (IPD) with aggregated study level data (AD) in a meta-analysis of correlational studies. Although medical researchers have employed IPD in a wide range of studies, only a single example exists in the social sciences. New policies at the National Science Foundation requiring grantees to submit data archiving plans may increase social scientists' access to individual level data that could be combined with traditional meta-analysis. The methods presented here extend prior work on IPD to meta-analyses using correlational studies. The examples presented illustrate the synthesis of publicly available national datasets in education with aggregated study data from a meta-analysis examining the correlation of socioeconomic status measures and academic achievement. The major benefit of the inclusion of the individual level is that both within-study and between-study interactions among moderators of effect size can be estimated. Given the potential growth in data archives in the social sciences, we should see a corresponding increase in the ability to synthesize IPD and AD in a single meta-analysis, leading to a more complete understanding of how within-study and between-study moderators relate to effect size. Copyright © 2012 John Wiley & Sons, Ltd.

Dose of Erythropoiesis-Stimulating Agents and Adverse Outcomes in CKD: A Metaregression Analysis

Targeting higher hemoglobin levels with erythropoiesis-stimulating agents (ESAs) to treat the anemia of chronic kidney disease (CKD) is associated with increased cardiovascular risk. Metaregression analysis examining the association of ESA dose with adverse outcomes independent of target or achieved hemoglobin level. Patients with anemia of CKD irrespective of dialysis status. We searched MEDLINE (inception to August 2010) and bibliographies of published meta-analyses and selected randomized controlled trials assessing the efficacy of ESAs for the treatment of anemia in adults with CKD, with a minimum 3-month duration. Two authors independently screened citations and extracted relevant data. Individual study arms were treated as cohorts and constituted the unit of analysis. ESA dose standardized to a weekly epoetin alfa equivalent, and hemoglobin levels. All-cause and cardiovascular mortality, cardiovascular events, kidney disease progression, or transfusion requirement. 31 trials (12,956 patients) met the criteria. All-cause mortality was associated with higher (per epoetin alfa-equivalent 10,000-U/wk increment) first-3-month mean ESA dose (incidence rate ratio [IRR], 1.42; 95% CI, 1.10-1.83) and higher total-study-period mean ESA dose (IRR, 1.09; 95% CI, 1.02-1.18). First-3-month ESA dose remained significant after adjusting for first-3-month mean hemoglobin level (IRR, 1.48; 95% CI, 1.02-2.14), as did total-study-period mean ESA dose adjusting for target hemoglobin level (IRR, 1.41; 95% CI, 1.08-1.82). Parameter estimates between ESA dose and cardiovascular mortality were similar in magnitude and direction, but not statistically significant. Higher total-study-period mean ESA dose also was associated with increased rate of hypertension, stroke, and thrombotic events, including dialysis vascular access-related thrombotic events. Use of study-level aggregated data; use of epoetin alfa-equivalent doses; lack of adjustment for confounders. In patients with CKD, higher ESA dose might be associated with all-cause mortality and cardiovascular complications independent of hemoglobin level.

Coronary heart disease in primary care: accuracy of medical history and physical findings in patients with chest pain – a study protocol for a systematic review with individual patient data

BackgroundChest pain is a common complaint in primary care, with coronary heart disease (CHD) being the most concerning of many potential causes. Systematic reviews on the sensitivity and specificity of symptoms and signs summarize the evidence about which of them are most useful in making a diagnosis. Previous meta-analyses are dominated by studies of patients referred to specialists. Moreover, as the analysis is typically based on study-level data, the statistical analyses in these reviews are limited while meta-analyses based on individual patient data can provide additional information. Our patient-level meta-analysis has three unique aims. First, we strive to determine the diagnostic accuracy of symptoms and signs for myocardial ischemia in primary care. Second, we investigate associations between study- or patient-level characteristics and measures of diagnostic accuracy. Third, we aim to validate existing clinical prediction rules for diagnosing myocardial ischemia in primary care. This article describes the methods of our study and six prospective studies of primary care patients with chest pain. Later articles will describe the main results.Methods/DesignWe will conduct a systematic review and IPD meta-analysis of studies evaluating the diagnostic accuracy of symptoms and signs for diagnosing coronary heart disease in primary care. We will perform bivariate analyses to determine the sensitivity, specificity and likelihood ratios of individual symptoms and signs and multivariate analyses to explore the diagnostic value of an optimal combination of all symptoms and signs based on all data of all studies. We will validate existing clinical prediction rules from each of the included studies by calculating measures of diagnostic accuracy separately by study.DiscussionOur study will face several methodological challenges. First, the number of studies will be limited. Second, the investigators of original studies defined some outcomes and predictors differently. Third, the studies did not collect the same standard clinical data set. Fourth, missing data, varying from partly missing to fully missing, will have to be dealt with.Despite these limitations, we aim to summarize the available evidence regarding the diagnostic accuracy of symptoms and signs for diagnosing CHD in patients presenting with chest pain in primary care.Review registrationCentre for Reviews and Dissemination (University of York): CRD42011001170

Safety profile of statins alone or combined with ezetimibe: a pooled analysis of 27 studies including over 22,000 patients treated for 6-24 weeks

Aims: The aim of this analysis was to assess the overall safety and tolerability profiles of various statins + ezetimibe vs. statin monotherapy and to explore tolerability in sub-populations grouped by age, race, and sex. Methods: Study-level data were combined from 27 double-blind, placebo-controlled or active-comparator trials that randomized adult hypercholesterolemic patients to statin or statin + ezetimibe for 6-24 weeks. In the full cohort, % patients with AEs within treatment groups (statin: N = 10,517; statin + ezetimibe: N = 11,714) was assessed by logistic regression with terms for first-/second-line therapy (first line = drug-naïve or rendered drug-naïve by washout at study entry; second line = ongoing statin at study entry or statin run-in), trial within first-/second-line therapy, and treatment. The same model was fitted for age (< 65, ≥ 65 years), sex, race (white, black, other) and first-/second-line subgroups with additional terms for subgroup and subgroup-by-treatment interaction. Results: In the full cohort, the only significant difference between treatments was consecutive AST or ALT elevations ≥ 3 × upper limit of normal (ULN) (statin: 0.35%, statin + ezetimibe: 0.56%; p = 0.017). Significantly more subjects reported ≥ 1 AE; drug-related, hepatitis-related and gastrointestinal-related AEs; and CK elevations ≥ 10 × ULN (all p ≤ 0.008) in first-line vs. second-line therapy studies with both treatments. AEs were generally similar between treatments in subgroups, and similar rates of AEs were reported within age and race subgroups; however, women reported generally higher AE rates. Conclusions: In conclusion, in second-line studies, ongoing statin treatment at study entry likely screened out participants for previous statin-related AEs and tolerability issues. These results describe the safety profiles of widely used lipid-lowering therapies and encourage their appropriate and judicious use in certain subpopulations.

Effects of ventilator circuit change frequency on ventilator-associated pneumonia: Meta-analysis

Objective To evaluate the effects of ventilator circuit change frequency on ventilatorassociated pneumonia (VAP).Methods Meta-analysis of effects of ventilator circuit change frequency on VAP was conducted with study-level data from 1995 to 2010 in Pubmed,Embase,Web of Science databases.Results Nine articles were included (sample size:20 326 mechanically ventilated patients).Analysis of six articles showed that the incidence of VAP in ventilator circuit change every 2 or 3 days was 4.05％,while 3.65％ in ventilator circuit change every 7 days.Compared with change ventilator circuit every 2 or 3 days,the risk ratio (RR) of VAP in weekly changes was 0.77 [0.54,1.09] ( P =0.14 ).Analysis three articles showed that compared to ventilator circuit change every 7 days with 15.89％ incidence of VAP,the incidence of VAP in circuit change more than 14 days was 14.9％,and RR was 0.98 [0.69,1.39] (P =0.91 ).Conclusions Regular ventilator circuit change frequency in various intervals can't difference in the incidence of VAP in mechanical ventilation patients. Key words: Mechanical ventilation; Ventilator circuit change; RCT; Ventilator-associated pneumonia; Incidence; Meta-analysis

New Data Rebut Claims About Antidepressants

Back to table of contents Previous article Next article Clinical & Research NewsFull AccessNew Data Rebut Claims About AntidepressantsMark MoranMark MoranSearch for more papers by this authorPublished Online:20 Apr 2012https://doi.org/10.1176/pn.47.8.psychnews_47_8_12-aAbstractA re-analysis of more than 40 studies refutes an earlier meta-analysis purporting to show that effectiveness of antidepressants was limited to severe depression.The antidepressants fluoxetine and venlafaxine were found to be efficacious for major depressive disorder in all age groups, although more so in youth and adults than in geriatric patients, and baseline severity was not significantly related to degree of treatment advantage over placebo.These results came from a re-analysis of data from more than 40 randomized, controlled trials of fluoxetine hydrochloride and venlafaxine hydrochloride in youth, adult, and geriatric subjects.The re-analysis was published online on March 5 in the Archives of General Psychiatry. The findings contradict a 2008 study in the journal PLoS (Public Library of Science) by Kirsch and colleagues that garnered much publicity for indicating that antidepressants appeared to be effective only for severe depression.The Archives study differed crucially from the 2008 PLoS article because it employed statistical models that allowed for patient-level data, incorporating baseline severity scores as both a discrete and a continuous variable and tracking change in symptoms over time to measure rates of relapse and remission; in contrast, the PLoS article simply looked at the average change of everyone in a study over the course of the study.“Unlike a typical meta-analysis, we got all the patient-level data as well as data at every point that patients were measured within a study,” lead author Robert Gibbons, Ph.D., told Psychiatric News. “Previous meta-analyses worked with study-level data, looking at average change from baseline to the end of the study and found that only those studies with more severely depressed patients showed a significant change.“We used a much more rigorous approach using intermediate data,” he said. “We were able to look at patient-level changes across time and compute response and remission rates. And we found that there was no significant difference in the amount of change depending on whether one was more sick or less sick at baseline.”He is a professor of biostatistics in the departments of medicine, health studies, and psychiatry at the University of Chicago.In the study, Children’s Depression Rating Scale–Revised scores, Hamilton Rating Scale for Depression scores for adult and geriatric populations, and estimated response and remission rates at six weeks were analyzed for 2,635 adults, 960 geriatric patients, and 708 youth receiving fluoxetine; for 2,421 adults receiving immediate-release venlafaxine; and for 2,461 adults receiving extended-release venlafaxine.Gibbons and colleagues found that patients in all age and drug groups had significantly greater improvement relative to control patients receiving placebo.Geriatric patients had the smallest drug-placebo differences, an 18.5 percent greater rate of improvement, 9.9 percent for response and 6.5 percent for remission. By contrast, child and adolescent patients had the largest overall drug placebo differences, a 29.0 percent greater rate of improvement, 24.1 percent for response and 30.1 percent for remission.Psychiatrist Michael Thase, M.D., who has conducted research in this area and was familiar with the methods and results of the Gibbons study, said it effectively refutes one of the more controversial conclusions of the PLoS paper, namely, that antidepressants are only effective in patients with severe depression.“Gibbons and colleagues were able to use statistical models that take into account each person’s level of pretreatment severity and how those people’s symptoms change across time, whereas Kirsch and colleagues worked with only study-level data,” Thase told Psychiatric News. “Average change from baseline gives you some information but much less than using individual patient data. Moreover, Gibbons and colleagues were able to compute response and remission rates, which are the more meaningful difference in outcome for people.”Thase, a professor of psychiatry in the Perelman School of Medicine at the University of Pennsylvania, spoke at APA’s 2010 annual meeting in New Orleans about efficacy of antidepressants. At that meeting, he said modern randomized, controlled studies of antidepressants demonstrate increasing effects of placebo because of patient selection and study design—a fact that has obscured the real value of antidepressants (Psychiatric News, July 2, 2010).Thase added that the PLoS analysis used last-observation-carried-forward—a statistical convention that “carries forward” the last scores of a patient who drops out of a study, as if it were an endpoint. “It’s a conservative convention to account for missing data that typically punishes the arm of active treatment in a study because drugs have side effects, so some greater percentage of people usually drop out of the active treatment before they have a chance to respond.” An abstract of “Benefits From Antidepressants: Synthesis of 6-Week Patient-Level Outcomes From Double-Blind Placebo-Controlled, Randomized Trials of Fluoxetine and Venlafaxine” is posted at http://archpsyc.ama-assn.org/cgi/content/abstract/archgenpsychiatry.2011.2044. ISSUES NewArchived

Rituximab in combination with chemotherapy versus chemotherapy alone in HIV‐associated non‐Hodgkin lymphoma: A pooled analysis of 15 prospective studies

In HIV-positive patients with non-Hodgkin lymphoma (NHL), no benefit of adding rituximab to chemotherapy was seen in a randomized controlled trial (RCT). We performed a meta-analysis of prospective studies to ascertain outcomes in HIV-positive NHL patients treated with chemotherapy (chemo) versus rituximab and chemo (R-chemo). A literature search through September 2011 was performed using the key search "(HIV OR AIDS) AND lymphoma". The main outcomes were overall response rate (ORR), complete response rate (CRR) and 2-year overall survival (OS) and are reported as non-adjusted odds ratio (OR). We identified 15 prospective studies including 1,060 HIV-positive NHL patients, 675 treated with chemo and 385 with R-chemo. There was a higher proportion of HAART in R-chemo patients (82% vs. 68%; p < 0.01) but there were no differences in proportion of patients with advanced stage or high/high-intermediate age-adjusted International Prognostic Index (aaIPI) scores. Meta-analysis showed the OR for ORR, CRR and 2-year OS in patients treated with R-chemo was 1.39 (95% CI 0.79-2.47; p = 0.26), 1.66 (95% CI 0.98-2.82; p = 0.06) and 2.19 (95% CI 1.68-2.86; p < 0.001), respectively. HIV-positive lymphoma patients treated with R-chemo had higher odds for CR and 2-year OS when compared to chemo but also had a higher proportion of HAART usage.

Probiotics' effects on the incidence of nosocomial pneumonia in critically ill patients: a systematic review and meta-analysis

IntroductionTo evaluate the efficacy of probiotics in preventing nosocomial pneumonia in critically ill patients.MethodsWe searched PubMed, EMBASE, and the Web of Science for relevant studies. Two reviewers extracted data and reviewed the quality of the studies independently. The primary outcome was the incidence of nosocomial pneumonia. Study-level data were pooled using a random-effects model when I2 was > 50% or a fixed-effects model when I2 was < 50%.ResultsTwelve randomized controlled studies with a total of 1,546 patients were considered. Pooled analysis showed a statistically significant reduction in nosocomial pneumonia rates due to probiotics (odd ratio [OR]= 0.75, 95% CI 0.57 to 0.97, P = 0.03, I2 = 46%). However, no statistically significant difference was found between groups regarding in-hospital mortality (OR = 0.93, 95% CI 0.50 to 1.74, P = 0.82, I2 = 51%), intensive care unit mortality (OR = 0.84, 95% CI 0.55 to 1.29, P = 0.43, I2 = 0%), duration of stay in the hospital (mean difference [MD] in days = -0.13, 95% CI -0.93 to 0.67, P = 0.75, I2 = 46%), or duration of stay in the intensive care units (MD = -0.72, 95% CI -1.73 to 0.29, P = 0.16, I2 = 68%).ConclusionsThe use of probiotics was associated with a statistically significant reduction in the incidence of nosocomial pneumonia in critically ill patients. However, large, well-designed, randomized, multi-center trials are needed to confirm any effects of probiotics clinical endpoints such as mortality and length of ICU and hospital stay.

Characterization of the bronchodilatory dose response to indacaterol in patients with chronic obstructive pulmonary disease using model-based approaches.

BackgroundIndacaterol is a once-daily long-acting inhaled β2-agonist indicated for maintenance treatment of moderate-to-severe chronic obstructive pulmonary disease (COPD). The large inter-patient and inter-study variability in forced expiratory volume in 1 second (FEV1) with bronchodilators makes determination of optimal doses difficult in conventional dose-ranging studies. We considered alternative methods of analysis.MethodsWe utilized a novel modelling approach to provide a robust analysis of the bronchodilatory dose response to indacaterol. This involved pooled analysis of study-level data to characterize the bronchodilatory dose response, and nonlinear mixed-effects analysis of patient-level data to characterize the impact of baseline covariates.ResultsThe study-level analysis pooled summary statistics for each steady-state visit in 11 placebo-controlled studies. These study-level summaries encompassed data from 7476 patients at indacaterol doses of 18.75-600 μg once daily, and showed that doses of 75 μg and above achieved clinically important improvements in predicted trough FEV1 response. Indacaterol 75 μg achieved 74% of the maximum effect on trough FEV1, and exceeded the midpoint of the 100-140 mL range that represents the minimal clinically important difference (MCID; ≥120 mL vs placebo), with a 90% probability that the mean improvement vs placebo exceeded the MCID. Indacaterol 150 μg achieved 85% of the model-predicted maximum effect on trough FEV1 and was numerically superior to all comparators (99.9% probability of exceeding MCID). Indacaterol 300 μg was the lowest dose that achieved the model-predicted maximum trough response.The patient-level analysis included data from 1835 patients from two dose-ranging studies of indacaterol 18.75-600 μg once daily. This analysis provided a characterization of dose response consistent with the study-level analysis, and demonstrated that disease severity, as captured by baseline FEV1, significantly affects the dose response, indicating that patients with more severe COPD require higher doses to achieve optimal bronchodilation.ConclusionsComprehensive assessment of the bronchodilatory dose response of indacaterol in COPD patients provided a robust confirmation that 75 μg is the minimum effective dose, and that 150 and 300 μg are expected to provide optimal bronchodilation, particularly in patients with severe disease.

Trends over time in the size and quality of randomised controlled trials of interventions for chronic low-back pain

Previous reviews of randomised controlled trials (RCTs) for low-back pain (LBP) have failed to identify any positive trend in study quality with more recent years of publication. This study aimed to identify and describe trends over time in the study design characteristics and risk of bias in chronic LBP trials performed over the past 30years. One fifty-seven randomised trials of interventions for chronic LBP were extracted from recently published systematic reviews. The reviews included RCTs on physical and rehabilitation interventions, injection therapy and denervation procedures, complementary and alternative therapies and pharmacological interventions for chronic LBP. Study level data were extracted and analysed for trends associated with year of publication. Overall, the mean sample size in the RCTs was 141 (median 70; range 17-3093). There was a slight increase in the median number of risk of bias criteria fulfilled from trials published prior to 1995 to those published after 1996. The analysis showed that in more recent years RCTs of medical interventions were more likely to be successfully blinded than RCTs of non-medical interventions. The continuing uncertainty regarding the efficacy of many interventions for chronic LBP again stresses the need for large RCTs with low risk of bias. Further research is needed into specific risks of bias within the RCTs for chronic LBP and the effect they have on the plausibility of the results.

Effects of antiresorptive treatment on nonvertebral fracture outcomes

Various definitions of nonvertebral fracture have been used in osteoporosis trials, precluding comparisons of efficacy. Using only subgroups of nonvertebral fractures for trial outcomes may underestimate the benefits and cost-effectiveness of treatments. The objectives of this study were to determine (1) the effect of antiresorptive treatment on various nonvertebral fracture outcomes, (2) whether risk reduction from antiresorptive treatment is greater for nonvertebral fractures that have stronger associations with low BMD, and (3) sample size estimates for clinical trials of osteoporosis treatments. Study-level data were combined from five randomized fracture-prevention trials of antiresorptive agents that reduce the risk of nonvertebral fracture in postmenopausal women: alendronate, clodronate, denosumab, lasofoxifene, and zoledronic acid. Pooled effect estimates were calculated with random-effects models. The five trials included 30,118 women; 2997 women had at least one nonvertebral fracture. There was no significant heterogeneity between treatments for any outcome (all p > 0.10). Antiresorptive treatment had similar effects on all fractures (summary hazard ratio [HR] = 0.76, 95% CI 0.70-0.81), high-trauma fractures (HR = 0.74, 95% CI 0.57-0.96), low-trauma fractures (HR = 0.77, (95% CI 0.71-0.83), nonvertebral six (ie, hip, pelvis, leg, wrist, humerus, and clavicle) fractures (HR = 0.73, 95% CI 0.66-0.80), other than nonvertebral six fractures (HR = 0.78, 95% CI 0.70-0.87), and all fractures other than finger, face, and toe (HR = 0.75, 95% CI 0.70-0.81). Risk reduction was not greater for fractures with stronger associations with low BMD (p = 0.77). A trial of all nonvertebral fractures would require fewer participants (n = 2641 per arm) than one of a subgroup of six fractures (n = 3289), for example. In summary, antiresorptive treatments reduced all nonvertebral fractures regardless of degree of trauma or special groupings, supporting the use of all nonvertebral fractures as a standard endpoint of osteoporosis trials and the basis for estimating the benefits and cost-effectiveness of treatments.

Higher PEEP in Patients With Acute Lung Injury: A Systematic Review and Meta-Analysis

Studies of ventilation strategies that included higher PEEP in patients with acute lung injury (ALI) or acute respiratory distress syndrome (ARDS) have yielded conflicting results. To determine whether higher PEEP during volume-limited and pressure-limited ventilation is associated with 28-day mortality or barotrauma rates in patients with ALI/ARDS. We searched MEDLINE, CENTRAL, EMBASE, CINAHL, Web of Science, and the bibliographies of retrieved papers to identify randomized controlled trials that compared higher and lower PEEP in adult patients with ALI/ARDS who were already receiving volume-limited or pressure-limited ventilation. Two of us independently abstracted study-level data, including study design, patient characteristics, study methods, intervention, and main results. We pooled the study-level data with a random-effects model, unless heterogeneity was low (I(2) < 50%), in which case we used a fixed-effects model. The primary outcome was 28-day mortality. Four randomized trials (2,360 participants) were evaluated. Higher PEEP had a nonsignificant trend toward lower 28-day mortality (pooled relative risk 0.90, 95% CI 0.79-1.02). There was no difference in barotrauma between the 2 groups (pooled relative risk 1.17, 95% CI 0.90-1.52). Two studies reported an adjusted hospital death rate, and the pooled results of sensitivity analysis with those adjusted rates were identical to those of the unadjusted analysis. In 4 recent studies that used volume-limited or pressure-limited ventilation in ALI/ARDS patients, higher PEEP was not associated with significantly different short-term mortality or barotrauma. This study does not support the routine use of higher PEEP in patients with ALI/ARDS.

Confidence Distributions and a Unifying Framework for Meta-Analysis

This article develops a unifying framework, as well as robust meta-analysis approaches, for combining studies from independent sources. The device used in this combination is a confidence distribution (CD), which uses a distribution function, instead of a point (point estimator) or an interval (confidence interval), to estimate a parameter of interest. A CD function contains a wealth of information for inferences, and it is a useful device for combining studies from different sources. The proposed combining framework not only unifies most existing meta-analysis approaches, but also leads to development of new approaches. We illustrate in this article that this combining framework can include both the classical methods of combining p-values and modern model-based meta-analysis approaches. We also develop, under the unifying framework, two new robust meta-analysis approaches, with supporting asymptotic theory. In one approach each study size goes to infinity, and in the other approach the number of studies goes to infinity. Our theoretical development suggests that both these robust meta-analysis approaches have high breakdown points and are highly efficient for normal models. The new methodologies are applied to study-level data from publications on prophylactic use of lidocaine in heart attacks and a treatment of stomach ulcers. The robust methods performed well when data are contaminated and have realistic sample sizes and number of studies.

An Alternative to Pooling Kaplan-Meier Curves in Time-to-Event Meta-Analysis

A meta-analysis that uses individual-level data instead of study-level data is widely considered to be a gold standard approach, in part because it allows a time-to-event analysis. Unfortunately, with the common practice of presenting Kaplan-Meier survival curves after pooling subjects across randomized trials, using individual-level data can actually be a step backwards; a Simpson's paradox can occur in which pooling incorrectly reverses the direction of an association. We introduce a nonparametric procedure for synthesizing survival curves across studies that is designed to avoid this difficulty and preserve the integrity of randomization. The technique is based on a counterfactual formulation in which we ask what pooled survival curves would look like if all subjects in all studies had been assigned treatment, or if all subjects had been assigned to control arms. The method is related to a Kaplan-Meier adjustment proposed in 2005 by Xie and Liu to correct for confounding in nonrandomized studies, but is formulated for the meta-analysis setting. The procedure is discussed in the context of examining rosiglitazone and cardiovascular adverse events.

Inconsistent results in meta-analyses for the prevention of falls are found between study-level data and patient-level data

Objective This study seeks to examine whether existing study-level data meta-analysis approaches can be used to produce unbiased and precise effect estimates relative to meta-analyses conducted using patient-level data, where a recurrent event is the outcome of interest. Study Design and Setting Data from two studies focusing on the prevention of falls in the hospital setting ( N = 1,838 total) was divided into the three hospital sites from which data were collected. Outcome data were considered as recurrent event survival data, single event survival data, count data, rate data, and binary data. A range of analysis approaches were considered. Results Andersen–Gill, negative binomial, bootstrap resampling, and modified relative risk analysis approaches produced congruous point estimates of effect, whereas modified relative risk analysis produced considerably smaller standard errors. Pooled effect point estimates derived from these approaches were not consistent when using study-level data as opposed to patient-level data, and 95% confidence intervals were excessively wide when between-study heterogeneity was present. Conclusion Conducting meta-analysis using patient-level data (if possible) or presenting results from individual trials without pooling of effect estimates may be preferable to presenting pooled effect estimates from meta-analysis of study-level data, where the outcome is a recurrent event.

Meta-analysis: Effects of Percutaneous Coronary Intervention Versus Medical Therapy on Angina Relief

Several meta-analyses have evaluated the efficacy of percutaneous coronary intervention (PCI) compared with medical therapy, but none has focused on angina relief. To summarize the evidence on the degree of angina relief from PCI compared with medical therapy in patients with stable coronary artery disease. The Cochrane Library (1993 to June 2009), EMBASE (1980 to June 2009), and MEDLINE (1950 to June 2009), with no language restrictions. Two independent reviewers screened citations to identify randomized, controlled trials of PCI versus medical therapy in patients with stable coronary artery disease. Two independent reviewers abstracted data on patient characteristics, study conduct, and outcomes. A random-effects model was used to combine data on freedom from angina and to perform stratified analyses based on duration of follow-up, inclusion of patients with recent myocardial infarction, coronary stent utilization, recruitment period, and utilization of evidence-based medications. A total of 14 trials, enrolling 7818 patients, met the inclusion criteria. Although PCI was associated with an overall benefit on angina relief (odds ratio, 1.69 [95% CI, 1.24 to 2.30]), important heterogeneity across trials was observed. The incremental benefit of PCI observed in older trials (odds ratio, 3.38 [CI, 1.89 to 6.04]) was substantially less and possibly absent in recent trials (odds ratio, 1.13 [CI, 0.76 to 1.68]). An inverse relationship between use of evidence-based therapies and the incremental benefit of PCI was observed. Information about the long-term use of medication was incomplete in most trials. Few trials used drug-eluting stents. Meta-regression analyses used aggregated study-level data from few trials. Percutaneous coronary intervention was associated with greater freedom from angina compared with medical therapy, but this benefit was largely attenuated in contemporary studies. This observation may be related to greater use of evidence-based medications in contemporary trials. Canadian Institutes of Health Research.

Understanding differences in results from literature-based and individual patient meta-analyses: An example from meta-analyses of observational data

Background Meta-analyses are increasingly used to summarise observational data however a literature meta-analysis (LMA) may give different results to the corresponding individual patient meta-analysis (IPMA). This study compares the published results of equivalent LMAs and IPMAs, highlighting factors that can affect the results and therefore impact on clinical interpretation of meta-analyses. Method Univariate results from published meta-analyses of prospective observational outcome data were compared, as were the number of studies, patients and length of follow-up. The absolute difference in survival was calculated. The association between severe diastolic dysfunction (RFP) and death post acute myocardial infarction (AMI) and in chronic heart failure (HF) were used as clinical examples. Results The IPMA hazard ratio was lower that the LMA odds ratio: AMI hazard ratio 2.67 (95% confidence interval 2.23 to 3.20), odds ratio 4.10 (3.38 to 4.99); HF hazard ratio 2.42 (2.06 to 2.83), odds ratio 4.36 (3.60 to 5.04). The IPMAs contained most of the studies from the LMAs as well as additional unpublished data, and a longer length of follow-up was available in the IPMAs (AMI 3.7 vs 2.6 yr, HF 4.0 vs 1.5 yr). Restricting analysis to the same studies in both the LMA and IPMA resulted in a similar difference in effect sizes between methods to those found in the published analyses. Conclusions The result of a meta-analysis is affected by whether study level or individual patient data have been used, and the variant of analysis that is required. Awareness and consideration of these factors is important for clinical interpretation of meta-analyses.

The Search for Apnea Genes

The Search for Apnea Genes

Effects of sources of variability on sample sizes required for RCTs, applied to trials of lipid-altering therapies on carotid artery intima-media thickness

Studies measuring progression of carotid artery intima-media thickness (cIMT) have been used to estimate the effect of lipid-modifying therapies cardiovascular event risk. The likelihood that future cIMT clinical trials will detect a true treatment effect is estimated by leveraging results from prior studies. The present analyses assess the impact of between- and within-study variability based on currently published data from prior clinical studies on the likelihood that ongoing or future cIMT trials will detect the true treatment effect of lipid-modifying therapies. Published data from six contemporary cIMT studies (ASAP, ARBITER 2, RADIANCE 1, RADIANCE 2, ENHANCE, and METEOR) including data from a total of 3563 patients were examined. Bayesian and frequentist methods were used to assess the impact of between study variability on the likelihood of detecting true treatment effects on 1-year cIMT progression/regression and to provide a sample size estimate that would specifically compensate for the effect of between-study variability. In addition to the well-described within-study variability, there is considerable between-study variability associated with the measurement of annualized change in cIMT. Accounting for the additional between-study variability decreases the power for existing study designs. In order to account for the added between-study variability, it is likely that future cIMT studies would require a large increase in sample size in order to provide substantial probability (> or =90%) to have 90% power of detecting a true treatment effect.Limitation Analyses are based on study level data. Future meta-analyses incorporating patient-level data would be useful for confirmation. Due to substantial within- and between-study variability in the measure of 1-year change of cIMT, as well as uncertainty about progression rates in contemporary populations, future study designs evaluating the effect of new lipid-modifying therapies on atherosclerotic disease progression are likely to be challenged by large sample sizes in order to demonstrate a true treatment effect.

A meta-analysis of the placebo response in antimuscarinic drug trials for overactive bladder

BackgroundThe purpose of this analysis was to characterize the placebo response in antimuscarinic drug trials for OAB, based on changes in commonly-used efficacy endpoints.MethodsPlacebo arm data for incontinence episodes, micturitions, voided volume and study characteristics were extracted from randomized placebo controlled antimuscarinic drug trials in OAB, from studies identified in a prior meta-analysis, and from a systematic review of more recently published studies. Relationships between variables were examined using linear regression, and changes in endpoints were analyzed by a meta-analysis approach. The effect of placebo arm size and magnitude of placebo response on probability of successful study outcome was analyzed using an ANOVA model.ResultsChanges in the placebo arms for all 3 endpoints were substantial and statistically significant, and highly heterogeneous. There were significant associations between baseline and change scores for some but not all of the endpoints. More recent studies tended to have more subjects than earlier studies, and there were positive associations between probability of achieving statistically significant results and size of the placebo arm. The magnitude of changes in placebo arms did not appear to influence the likelihood of the study to be statistically significant.ConclusionThis analysis confirms earlier observation that the placebo response in OAB trials is substantial and highly heterogeneous. There are multiple potential reasons for this; however, these could not be explored in this analysis of study-level data. Two approaches may be used in clinical trials to manage high placebo effect: recruitment of 1) greater numbers of patients and/or 2) more severely affected patients; however, only the former approach is associated with increased probability of successful study outcome.