Impact of allocation concealment on conclusions drawn from meta-analyses of randomized trials

IT University of Copenhagen, København, Capital Region, Denmark
International Journal of Epidemiology (Impact Factor: 9.2). 09/2007; 36(4). DOI: 10.1093/ije/dym087
Source: OAI

ABSTRACT Background Randomized trials without reported adequate allocation concealment have been shown to overestimate the benefit of experimental interventions. We investigated the robustness of conclusions drawn from meta-analyses to exclusion of such trials. Material Random sample of 38 reviews from The Cochrane Library 2003, issue 2 and 32 other reviews from PubMed accessed in 2002. Eligible reviews presented a binary effect estimate from a meta-analysis of randomized controlled trials as the first statistically significant result that supported a conclusion in favour of one of the interventions. Methods We assessed the methods sections of the trials in each included meta-analysis for adequacy of allocation concealment. We replicated each meta-analysis using the authors' methods but included only trials that had adequate allocation concealment. Conclusions were defined as not supported if our result was not statistically significant. Results Thirty-four of the 70 meta-analyses contained a mixture of trials with unclear or inadequate concealment as well as trials with adequate allocation concealment. Four meta-analyses only contained trials with adequate concealment, and 32, only trials with unclear or inadequate concealment. When only trials with adequate concealment were included, 48 of 70 conclusions (69%; 95% confidence interval: 56–79%) lost support. The loss of support mainly reflected loss of power (the total number of patients was reduced by 49%) but also a shift in the point estimate towards a less beneficial effect. Conclusion Two-thirds of conclusions in favour of one of the interventions were no longer supported if only trials with adequate allocation concealment were included.

  • [Show abstract] [Hide abstract]
    ABSTRACT: It remains unknown whether the combination of antiresorptive agents and exercise would generate additive effects on bone mineral density (BMD) in postmenopausal women, though their separate roles in preventing bone loss have been well established. This meta-analysis aimed to evaluate the combined impact of antiresorptive treatment and exercise on the lumbar spine and femoral neck BMD in postmenopausal women compared with an exercise-only intervention. A systematic literature search of PubMed, EMBASE, SportDiscus and ProQuest up to Jun 2014 was conducted to identify the influence of antiresorptive agents and exercise on BMD in postmenopausal women. The study quality of the included trials was evaluated. The effect sizes were estimated by calculating the standardized mean difference (SMD). Subgroup analyses were conducted by pharmacological regimens and exercise categories. Nine studies with a total of 1,248 postmenopausal women met the inclusion criteria. The heterogeneity between the studies was evident at the spine (I2 = 78.7%) and hip (I2 = 41.7%) measurements; random-effects models were used in the data analysis. The pooled effect sizes associated with the combined interventions of antiresorptive agents and exercise were significant at the lumbar spine BMD (SMD = 0.511, 95% CI = 0.118-0.904, p = 0.011). Combining hormone replacement therapy (HRT) and exercise training generated greater beneficial effects on lumbar spine (SMD = 0.729, 95% CI = 0.186-1.273, p = 0.009) and femoral neck BMD (SMD = 0.220, 95% CI = 0.0110-429, p = 0.039) than the exercise-only intervention. Impact exercise was sensitive to antiresorptive agents in preventing postmenopausal bone loss both at the spine (SMD = 1.252, 95%CI = 0.465-2.039, p = 0.002) and hips (SMD = 0.414, 95%CI = 0.106-0.723, p = 0.008). Our findings indicate that antiresorptive agents significantly increase the impact of exercise on the prevention of bone loss in postmenopausal women, which implies that the combination of antiresorptive agents and exercise may generate additive effects.
    PLoS ONE 01/2015; 10(1):e0116729. DOI:10.1371/journal.pone.0116729 · 3.53 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Abstract Background A critical part of future service delivery will involve improving the degree to which people become engaged in ‘self-management’. Providing better support for self-management has the potential to make a significant contribution to NHS efficiency, as well as providing benefits in patient health and quality of care. Objective To determine which models of self-management support are associated with significant reductions in health services utilisation (including hospital use) without compromising outcomes, among patients with long-term conditions. Data sources Cochrane Central Register of Controlled Trials, Cumulative Index to Nursing and Allied Health, EconLit (the American Economic Association’s electronic bibliography), EMBASE, Health Economics Evaluations Database, MEDLINE (the US National Library of Medicine’s database), MEDLINE In-Process & Other Non-Indexed Citations, NHS Economic Evaluation Database (NHS EED) and PsycINFO (the behavioural science and mental health database), as well as the reference lists of published reviews of self-management support. Methods We included patients with long-term conditions in all health-care settings and self-management support interventions with varying levels of additional professional support and input from multidisciplinary teams. Main outcome measures were quantitative measures of service utilisation (including hospital use) and quality of life (QoL). We presented the results for each condition group using a permutation plot, plotting the effect of interventions on utilisation and outcomes simultaneously and placing them in quadrants of the cost-effectiveness plane depending on the pattern of outcomes. We also conducted conventional meta-analyses of outcomes. Results We found 184 studies that met the inclusion criteria and provided data for analysis. The most common categories of long-term conditions included in the studies were cardiovascular (29%), respiratory (24%) and mental health (16%). Of the interventions, 5% were categorised as ‘pure self-management’ (without additional professional support), 20% as ‘supported self-management’ (< 2 hours’ support), 47% as ‘intensive self-management’ (> 2 hours’ support) and 28% as ‘case management’ (> 2 hours’ support including input from a multidisciplinary team). We analysed data across categories of long-term conditions and also analysed comparing self-management support (pure, supported, intense) with case management. Only a minority of self-management support studies reported reductions in health-care utilisation in association with decrements in health. Self-management support was associated with small but significant improvements in QoL. Evidence for significant reductions in utilisation following self-management support interventions were strongest for interventions in respiratory and cardiovascular disorders. Caution should be exercised in the interpretation of the results, as we found evidence that studies at higher risk of bias were more likely to report benefits on some outcomes. Data on hospital use outcomes were also consistent with the possibility of small-study bias. Limitations Self-management support is a complex area in which to undertake literature searches. Our analyses were limited by poor reporting of outcomes in the included studies, especially concerning health-care utilisation and costs. Conclusions Very few self-management support interventions achieve reductions in utilisation while compromising patient outcomes. Evidence for significant reductions in utilisation were strongest for respiratory disorders and cardiac disorders. Research priorities relate to better reporting of the content of self-management support, exploration of the impact of multimorbidity and assessment of factors influencing the wider implementation of self-management support. Study registration This study is registered as PROSPERO CRD42012002694. Funding The National Institute for Health Research Health Services and Delivery Research programme.
  • [Show abstract] [Hide abstract]
    ABSTRACT: Insomnia is an important adverse event of mechanical thromboprophylaxis. This sleep disorder has been reported as one of the commonest adverse events of the new oral anti-Xa anticoagulant darexaban, with similar rates to mechanical thromboprophylaxis in a randomized controlled trial (RCT). However, the perceived effect could have been biased because it was an open-label RCT. Therefore, we aimed to review the incidence of insomnia with non-vitamin K antagonist oral anticoagulants (NOACs). We performed a systematic review and meta-analysis of Phase III RCTs. Electronic databases MEDLINE and CENTRAL (inception to September 2013) were searched as well as review articles and references of included studies. We included phase III RCTs which compared NOACs with any other control group. Data were analyzed and pooled to estimate risk ratio (RR) with 95% confidence intervals (95%CI) for insomnia using inverse variance method. Statistical heterogeneity was evaluated with I (2) test. We included seven studies (two apixaban RCTs, two dabigatran RCTs, one darexaban RCTs, and two rivaroxaban RCTs), enrolling a total of 23,023 patients. Overall, NOACs were not associated to an increased risk of insomnia: RR 0.94 (95%CI 0.83-1.08; I (2) = 0%). In blinded studies (six studies), NOACs also did not show increased risk of insomnia (RR 0.94, 95%CI 0.83-1.08; I (2) = 0%). Results were similar irrespective of the comparators. NOACs (apixaban, dabigatran, darexaban, rivaroxaban) did not show increased risk of insomnia. Results according to study design (blinded vs. open-label trials) overlap the main analysis.

Full-text (2 Sources)

Available from
Jul 3, 2014