Quality of reporting randomised controlled trials in dental public health.

Methods: Electronic (PubMed and Medline) and hand searching of three dental public health journals, Community Dental Health, Community Dentistry and Oral Epidemiology and the Journal of Public Health Dentistry 1993 - 2008. Methods: Studies defined as controlled experiments using a random method to assign study units were included: papers reporting secondary analysis, pilot studies and those that provided brief methodological descriptions were excluded. Methods: Each article included in the study was assessed and scored independently by two observers, with any discrepancies being resolved by discussion Articles were assessed using the CONSORT criteria.

Results: 119 RCTs were initially identified (85 electronically and 33 from hand searching) and of these 48 were assessed. Of the 56 CONSORT criteria there were on average 27 (SD 6.9) present. The average number of criteria increased over the period.

Conclusions: There were inadequacies in the reporting of trials in dental public health journals. The quality of the reporting could be improved if the CONSORT statement were followed more closely.