Randomised trials in the South African Medical Journal, 1948-1997.

OBJECTIVE

To describe randomised controlled trials (RCTs) published in the South African Medical Journal (SAMJ) over a 50-year period from 1948 to 1997 with regard to number, topic and quality.

METHODS

We hand searched all issues of the SAMJ published during the study period to identify all published RCTs.

OUTCOME MEASURES

Number, topic and quality of RCTs published from 1948 to 1997.

RESULTS

Eight hundred and fifty-eight clinical trials were published during the period reviewed. Eighty-four per cent of RCTs were published as full articles. During the 1980s the number of RCTs published increased rapidly, with a peak of 35 in 1985, but then declined to only 5 in 1997. The majority (92%) of RCTs were conducted in a hospital setting. A varied range of subjects was covered, with gastroenterology taking the lead and no trials in public health. The sample size in more than 50% of RCTs was smaller than 50 patients. Fifty-one per cent (435 trials) used random allocation and 49% (423) quasi-random methods of allocation. Concealment of treatment allocation was judged to be adequate in 46% of studies (N = 200), blinding of observers assessing outcomes was adequate in 28% (123), and all the allocated test subjects were included in the primary analysis in 28% (123). The follow-up period was more than 1 year in 4% (17) and less than 6 days in 16% (71).

CONCLUSIONS

Compared with other international journals the SAMJ is highly regarded in terms of the number of trials published. There are, however, a number of deficiencies in the quality of the trials.