Joe Vere1, Rajendra Joshi. 1. Department of Restorative Dentistry, School of Clinical Dentistry, University of Sheffield, Sheffield, UK. jwvere@yahoo.co.uk
Abstract
AIM: The aim of this systematic review was to assess the methodological and statistical quality of randomised controlled trials of treatment interventions of dental implant surgery and prosthodontics published between 2004 and 2008. METHOD: Randomised controlled trials were identified following detailed searches of Medline, EMBASE, CINHAL and the Cochrane Central Register of Controlled Trials. Identified studies were independently assessed by two non-blinded reviewers for 14 key methodological and statistical domains. Information was collected using a customised data collection sheet. RESULTS: The initial Medline search yielded 1484 studies. Following application of inclusion criteria 38 randomised controlled trials, not presenting the same patient material, were identified for this review. Inter-examiner agreement was assessed for all domains (median κ score 0.7). All disagreements were resolved by discussion. Randomisation was adequate in 16 trials (42%), allocation concealment was adequate in seven trials (18%) and outcome assessment was blind in 12 trials (31%). Seven trials (18%) tested the intra/inter-examiner reliability of at least one outcome of interest. Confidence intervals were reported in two trials (5%). Seventeen different statistical tests were used. However, inferential statistical analysis was considered appropriate in only 22 trials (57%). CONCLUSIONS: Randomised controlled trials of treatment interventions of dental implant surgery and prosthodontics published between 2004 and 2008 are poorly reported and, by themselves, provide little unbiased evidence to support the clinical decisions that we make.
AIM: The aim of this systematic review was to assess the methodological and statistical quality of randomised controlled trials of treatment interventions of dental implant surgery and prosthodontics published between 2004 and 2008. METHOD: Randomised controlled trials were identified following detailed searches of Medline, EMBASE, CINHAL and the Cochrane Central Register of Controlled Trials. Identified studies were independently assessed by two non-blinded reviewers for 14 key methodological and statistical domains. Information was collected using a customised data collection sheet. RESULTS: The initial Medline search yielded 1484 studies. Following application of inclusion criteria 38 randomised controlled trials, not presenting the same patient material, were identified for this review. Inter-examiner agreement was assessed for all domains (median κ score 0.7). All disagreements were resolved by discussion. Randomisation was adequate in 16 trials (42%), allocation concealment was adequate in seven trials (18%) and outcome assessment was blind in 12 trials (31%). Seven trials (18%) tested the intra/inter-examiner reliability of at least one outcome of interest. Confidence intervals were reported in two trials (5%). Seventeen different statistical tests were used. However, inferential statistical analysis was considered appropriate in only 22 trials (57%). CONCLUSIONS: Randomised controlled trials of treatment interventions of dental implant surgery and prosthodontics published between 2004 and 2008 are poorly reported and, by themselves, provide little unbiased evidence to support the clinical decisions that we make.