[Research reports in Chinese medical journals]

To assess research designs employed in Chinese medical journals over the last decade.

The types of study design used in all original articles published in five leading journals (sponsored by the Chinese Medical Association) in 1985 (n = 640) and 1995 (n = 954) were evaluated.

Of 1,594 articles, over 50% were case reports and cross-sectional studies, and one third were longitudinal studies. Compared with 1985, the proportion of clinical trials increased from 5.6% to 11.3%, especially that of RCT increased from 1.3% to 4.2%. The proportion of prospective studies also increased from 3.9% to 6.0% and that of retrospective studies decreased from 25.0% to 13.9%, particularly, that of case series decreased from 17.3% to 5.9%. The inappropriate use of research design was frequent, the proportion of this problem was 38.0% and 36.8% for 1985 and 1995 respectively. The longitudinal studies were most liable to have this problem. The five commonly problems were the lack or inappropriateness of control, non-representativeness of sample, unbelievable conclusion, no randomization, and no consideration of confounders. They accounted for 66.9%.

The quality of research designs used in Chinese medical research is gradually improving in the last decade, but the lack or inappropriate use remains a serious problem.