Safety outcome trials in four mainline medical journals 30 years apart: a narrative review and the need for the transparency of informed consents.

BACKGROUND

It was our impression that safety outcome trials were getting more frequent, raising ethical issues mainly related to patient autonomy. We and others had also proposed this autonomy would be best served if wording of the informed consents would be in the public domain.

METHODS

Initially two observers and an arbiter tabulated the main aims of randomized controlled trials (RCTs) published in 1990-1991 vs. 2019-2020 as efficacy, safety, or undecided in four mainline medical journals, from the websites. A pragmatic design as well as other salient features was also tabulated. After noting too many trials were categorized as undecided, two additional independent observers and the arbiter did a reassessment.

RESULTS

In our reassessment of 889 RCTs, 309 in earlier and 580 in the later time period, 828 (93%) were categorized as efficacy and 47 (5%) as a safety trial. We were undecided in 14 (2%) trials. The proportion of safety outcome trials between the two time periods were similar [14/309 (5%) vs. 33/580 (6%)] while RCTs of any category conducted in the critical care settings notably increased in time [12/309 (4%) vs. 52/580 (9%) OR 2.4; 1.3-4.6]. Death was a primary outcome measure in 0/14 among the earlier and 16/33 (49%) among the later safety outcome trials. Stroke in 9 and myocardial infarction in 8 safety outcome trials were additional primary outcome measures in the same group. There were 2 pragmatic trials in the earlier and 93 in the later period.

CONCLUSION

Although we did not observe a differential increase among the safety outcome trials of all categories, those in critical care settings had significantly increased in time. So did the safety outcome trials with primary outcome measures like death, myocardial infarction, and stroke as well as randomized controlled trials with a pragmatic design. These raise the issue of autonomy related to how clearly the sought-after safety through designing empirical studies primarily to quantitate harm had been worded to the trial participants in plain language. We maintain this issue cannot be adequately addressed unless the informed consent forms, especially for safety outcome trials, are in the public domain.