Design and quality considerations for randomized controlled trials in systemic sclerosis (Journal article)

Kyriakidi, M./ Ioannidis, J. P.

OBJECTIVE: To appraise systematically randomized controlled trials (RCTs) on systemic sclerosis (SSc) in order to determine whether the parameter of study design and its quality may influence the reporting of efficacy for tested interventions. METHODS: Seventy RCTs were analyzed (1965-2000) in terms of design, patient characteristics, outcomes, and reported results. RESULTS: Median sample size was 28 patients. Fifty-nine trials were double blind, but only 16 mentioned the randomization mode and only 7 described allocation concealment. There was sufficient information on withdrawals in 37 trials. Larger trials with longer followup scored higher on quality characteristics, but had higher withdrawal rates. Only 8 trials had a followup of more than 1 year. Significant efficacy was less likely to be reported in double-blind studies (P = 0.029) and in studies with larger rates of withdrawal (P = 0.032). Specification of the following parameters improved over time: power calculations (P = 0.0003), outcomes (P = 0.001), and sample size per arm (P = 0.011). CONCLUSIONS: Several aspects of the quality of design and conduct of SSc RCTs can be improved. Adequately powered trials with longer followup and clear outcomes are needed.
Institution and School/Department of submitter: Πανεπιστήμιο Ιωαννίνων. Σχολή Επιστημών Υγείας. Τμήμα Ιατρικής
Keywords: Databases, Factual,Double-Blind Method,Female,Humans,Male,Middle Aged,Quality Control,Randomized Controlled Trials as Topic/*methods/standards,Research Design/standards/*trends,Sample Size,Scleroderma, Systemic/*therapy,Time Factors
ISSN: 0004-3591
Appears in Collections:Άρθρα σε επιστημονικά περιοδικά ( Ανοικτά)

Files in This Item:
There are no files associated with this item.

 Please use this identifier to cite or link to this item:
  This item is a favorite for 0 people.

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.