- F J García López: Department of Epidemiology, School of Hygiene and Public Health, Johns Hopkins University, Baltimore, Maryland.
BACKGROUND: In Spain since 1982 laws require that clinical trials involving drugs be approved before implementation. We studied the impact of regulations on the quality of published trials.
METHODS: Four sets of trials were chosen: trials implemented after 1982 and registered as approved; trials implemented after 1982 but not registered; trials implemented before 1982; and trials conducted outside Spain. Trials were identified via MEDLINE, EMBASE, and Indice Médico Español. Sets were compared with regard to indicators of quality, as obtained from the information in the published reports. The comparison was based on a total of 273 Spanish trials published between 1988 and 1990, 85 approved and 188 unregistered; 97 old trials, published between 1980 and 1982; and 152 non-Spanish trials published between 1988 and 1990.
RESULTS: Approved trials, compared to their unregistered and old counterparts, were more often informed randomised, more of their published reports included lists of reasons for exclusions and information on consent and achieved higher scores of a quality index. Approved trials, compared to non-Spanish trials, had lower proportion of sample size justification, greater discrepancies between randomised and analyzed cases and a trend to lower quality scores. Multiple logistic regression analysis of quality scores showed that approved trials had higher scores than unregistered trials when single-centre trials (odds ratio for reaching scores in the upper quartile: 2.90; 95% confidence interval: 1.27 - 6.64) and similar when multicentre trials.
CONCLUSIONS: Approved trials achieved better indicators of quality than unregistered trials but did not achieve the standards of quality prevailing in the international community.