Do randomised controlled trials relevant to pharmacy meet best practice standards for quality conduct and reporting? A systematic review. (1st October 2019)

• Record Type: Journal Article
• Authors: Ritchie, Alison; Seubert, Liza; Clifford, Rhonda; Perry, Danae; Bond, Christine
• Is Part Of: International journal of pharmacy practice. Volume 28:Number 3(2020)
• Page Start: 220
• Page End: 232
• Publication Date: 2019-10-01
• Subjects: other -- pharmaceutical care -- RCT -- systematic review; Pharmacy -- Practice -- Periodicals; 615.1
• DOI: 10.1111/ijpp.12578
• Languages: English
• ISSNs: 0961-7671

Abstract

Abstract: Objectives: Evidence-based pharmacy practice requires a dependable evidence base. Randomised controlled trials (RCTs) are the gold standard of high-quality primary research, and tools exist to assist researchers in conducting and reporting high-quality RCTs. This review aimed to explore whether RCTs relevant to pharmacy are conducted and reported in line with Cochrane risk of bias and CONSORT standards, respectively. Methods: A MEDLINE search identified potential papers. After screening of titles, abstracts and full texts, the 50 most recent papers were reviewed and assessment of bias according to Cochrane domains and compliance with CONSORT checklist items was recorded. Each domain of the Cochrane tool and CONSORT checklist item and each article were given a percentage score, reported as median and interquartile range (IQR). Correlation between quality of conduct, quality of reporting, continent of origin, and journal impact factor was conducted using the R 2 statistic. The median domain score for risk of bias by paper according to the Cochrane risk of bias tool was 53.0% (IQR 38.5–68.5), while the median compliance score by paper for the CONSORT checklist was 64.0% (IQR 36.0–94.0%). key findings: The median Cochrane domain and median CONSORT item completion scores, respectively, were 50.0% (IQR 33.3–66.7%) and 59.5% (IQR 52.0–70.3%). The highest risk of bias was associated with allocation concealment and blinding, and the least well-reported items were randomisation details, sequence generation and allocation concealment. A positive relationship between conduct and reporting of RCTs was found ( R 2 = 0.75), while no correlation was found between quality of conduct or quality of reporting and journal impact factor, correlation coefficients ( R 2 = 0.06 and R 2 = 0.05, respectively). Summary: This review identified that issues related to randomisation and blinding are often inadequately conducted or not comprehensively reported by researchers conducting pharmacy relevant RCTs, providing useful information for education and future research.