PT  - JOURNAL ARTICLE
AU  - Dobies, David
AU  - Barber, Kimberly
TI  - 47 The impact of applying both a critical statistical and a critical clinical appraisal to the randomised trials that form the cardiology revascularizatio nguidelines: implications for clinical practice
AID  - 10.1136/bmjebm-2018-111024.47
DP  - 2018 Jun 01
TA  - BMJ Evidence-Based Medicine
PG  - A23--A24
VI  - 23
IP  - Suppl 1
4099  - http://ebm.bmj.com/content/23/Suppl_1/A23.2.short
4100  - http://ebm.bmj.com/content/23/Suppl_1/A23.2.full
SO  - BMJ EBM2018 Jun 01; 23
AB  - Objectives Randomised trials and guidelines are used by physicians, administrators, patients, and policy makers to make medical decisions. The assumption is that the randomised trial conclusions and the guideline recommendations are accurate and have been properly vetted. Furthermore, the randomised trial conclusions and guideline recommendations are used for certification examinations. Therefore, a critical appraisal should be performed. Both a critical statistical and a critical clinical evaluation should be performed.Method The 5 Year Syntax and the Freedom Trials were evaluated. These trials compared percutaneous coronary artery intervention (PCI) utilising drug eluting stents to coronary artery bypass (CABG). In both trials patients with multi-vessel coronary disease were studied. These two trials form the basis on the cardiology guidelines. Both trials reported a highly significant result favouring CABG over PCI for patients with mult-vessel coronary artery disease.Both a critical statistical and a critical clinical evaluation of these two trials was performed. From a statistical standpoint, a sensitivity analysis was performed that included Forest plots. Also, the number needed to treat vs the number needed to harm was calculated. From a clinical standpoint, the definition of a myocardial infarction was evaluated along with the clinical impact. Both trials had significant limitations due to missing data.Results The actual statistical findings of both the 5 Year Syntax Trial and the Freedom Trial demonstrated no difference in major adverse cardiovascular and cerebrovascular events (MACCE). The critical statistical analysis of the Syntax Trial differed significantly from the reported trial results because the effect of missing data was accounted for. The reported results counted all missing data as non-events. The missing data was calculated as non-events, study event average, and as having events. The clinical evaluation focused on the definition of a myocardial infarction that was the biggest component of the MACCE rate. The definition was different for the CABG patients compared to the PCI patients.The Freedom Trial reported results were severely limited by the calculation of ratio at risk, a high differential loss, and a high overall attrition rate. When these limitations are corrected for the actual results show no difference between PCI and CABG.Conclusions The reported results of both the 5 Year Syntax Trial and The Freedom Trial differed significantly from the actual trial findings since either subjects were excluded or missing data was only calculated as non-events. Both trial conclusions were severely compromised. From a clinical standpoint, the definition of a myocardial infarction was study dependent, a newer generation of stents are used which are safer, and newer and more effective anti- platelet medications are used in current practice.Given the concern over the accuracy of reported conclusions of randomised trials, guideline committees should utilise independent statisticians for accurate analysis.