A survey of randomised controlled trials in HIV has found that almost half of their potential participants are excluded from enrolling due to strict entry criteria. This may lead to under-representation of certain groups and explain the differences between the results of clinical trials and the outcomes in the real world. The findings were published in the 4th November edition of AIDS.
The survey also found that many journal articles describing a trial's results fail to list all of the study protocol’s criteria, which are determined when the trial is designed. This may mislead the reader into thinking that the trial results are more generalisable than they really are.
The investigators call for a critical appraisal of the necessity and effects of selection criteria and more openness about the criteria in journal publications.
Randomised controlled trials can provide strong evidence on the efficacy and safety of medical interventions, including drugs, medical devices and surgical procedures. However, their use is limited by how similar the participants of the trial are to the population who may eventually use the intervention.
In HIV research, as well as in other areas of medicine, the effects of drugs are often more impressive in clinical trials than in the real world setting. To allow readers to assess how generalisable an intervention is, the publication of a study must include a full description of the criteria used for selection of the trial’s participants.
Investigators from the United States wished to determine the impact of selection criteria on the participants in HIV-related trials. They examined the entry criteria listed in the protocols for 32 trials and in their published results, including 20 trials from the Adults AIDS Clinical Trial Group (ACTG) and twelve from the Terry Beirn Community Programs for Clinical Research on AIDS (CPCRA).
The researchers applied the studies’ criteria to the Women’s Interagency HIV Study (WIHS), the largest cohort of HIV-infected women in the United States. This was chosen as its only exclusion criterion is age under 13 years.
“We determined that a range of 0 to 67.6% (median 42%) of participants in the WIHS observational cohort would have been excluded from participating in these influential HIV randomised controlled trials per the original protocols,” the researchers write. “Re-examination of these criteria may lead to increased participation of women in HIV randomised controlled trials.”
They also compared the descriptions of the criteria in the protocols, obtained from the trial co-ordinators, with those in the journal articles describing their findings. They found that, on average, only 60% of the women who were actually excluded from the trials were excluded based on the criteria listed in the published articles.
“Significant disparities in eligibility criteria reporting between protocols and publications were found,” they write. “A reader referring to methods in HIV trial publications would have underestimated the percentage of WIHS participants actually excluded from these trials … by 40%."
“The Consolidated Standards of Reporting Trials (CONSORT) statement in 1996 identifies key information about entry criteria (including inclusion and exclusion criteria), randomisation methods, patient follow-up, and outcome determination that should be reported for every randomised controlled trial publication. Our analysis demonstrates that current reporting of eligibility criteria, at least for these HIV clinical trials, may not be meeting these standards,” they explain.
The researchers suggest that eligibility criteria for randomised controlled trials be minimised to improve the applicability of trial results. In particular, they suggest that “broad and / or arbitrary eligibility criteria,” such as mild neuropathy or low white blood cell counts, should be reconsidered or narrowed in order to diversify trial participants.
They also suggest that subjective criteria relying on the investigators’ judgement should be minimised. These include judgements as to whether a participant would be able to adhere to the study’s protocol, and “may introduce social bias into HIV randomised controlled trials,” they write.
“Standardisation and openness of clinical trial reporting will allow clinicians to more fully assess the applicability of a trial finding to an individual patient and allow researchers and policymakers to expand scientific investigation for all representative populations with HIV,” they conclude.
The investigators point out that their study only examined the effects of inclusion criteria on a cohort of HIV-infected women. “Our analysis did not examine the impact of HIV randomised controlled trial eligibility criteria on the exclusion of HIV-infected men, so we cannot comment on whether enrolment criteria in these HIV clinical trials would have differentially excluded women,” they write.
Studies examining the effects on cohorts of men with HIV are underway.
Gandhi M et al. Eligibility criteria for HIV clinical trials and generalizability of results: the gap between published reports and study protocols. AIDS 19: 1885-1896, 2005.