Some biomedical research should be permitted in US prisons if the benefits to the prisoners themselves clearly outweigh the risks, recommends the US Institute of Medicine (IoM).

But drug trials in this setting should be confined to Phase III studies, with the ratio of prisoner to non-prisoner subjects not exceeding 50%, the IoM says.

The recommendations come in a report on Ethical Considerations for Research Involving Prisoners commissioned by the US Department of Health and Human Services’ (DHHS) Office for Human Research Protections. The IoM was asked to consider whether the present rules governing scientific research on prisoners needed updating.

Given the existing restrictions on biomedical research involving prisoners in the Code of Federal Regulations (Title 1, Article 46, Subpart C), the IoM’s review has prompted concerns that conditions for drug trials in prisons are being eased to satisfy the huge increase in demand for clinical study participants.

While the Institute’s recommendations would provide some relief from the strong federal restraints on biomedical research in prisons, they also stress that any such research should be “severely limited” because of the risks involved and the particular vulnerability of the prison population. Moreover, they call for uniform standards and a national system of oversight for all research programmes enrolling prisoners, as well as a detailed public database to track these studies.

Under the current patchwork approach to protecting prisoners involved in research, the IoM notes, most studies occur outside the scope of federal regulations and often without the scrutiny of institutional review boards.

According to the report, 15 US states already permit therapeutic biomedical research on prisoners, while three allow “biomedical studies of a non-therapeutic nature." Very few of these trials, however, appear in the published literature

The possibility of regulatory endorsement for more extensive use of prisons as trial sites has raised the spectre of past abuses that saw anything from acne treatments to syphilis vaccines and hallucinogenics tested on prisoners – often without informed consent or proper understanding of the risk/benefit equation.

According to testimony by the Alliance for Human Research Protection (AHRP) before the IoM committee reviewing research on prisoners in July 2005 – a “very low-profile” public meeting at which, the AHRP claimed, 12 out of 14 invited presenters favoured “rolling back” existing federal restrictions – in 1976 the president of the US pharmaceutical industry association estimated that 85% of Phase I trials were conducted on prisoners.

Revelations of research abuses in and outside US prisons led to the introduction of legislation to protect human participants in the late 1970s, including the Subpart C provisions mentioned above. These limit biomedical or behavioural research conducted or supported by the DHHS to areas such as the causes and effects of incarceration – providing there is “minimal risk” to the prisoners – and practices “which have the intent and reasonable probability of improving the health or well-being of the subject."

Also permitted is research on “conditions particularly affecting prisoners as a class”, such as hepatitis vaccine trials, subject to expert consultation and a published notice of the planned study in the Federal Register.

The IoM found these categories to be poorly defined, arguing instead for a system where all proposals for research enrolling prisoners were reviewed based on the risks and benefits for individual subjects.

It also pointed to the more than four-fold increase in the US prison population over the last 30 years, aggravating the vulnerability of prisoners whose privations – restrictions on liberty or autonomy, limited privacy or inadequate healthcare – were barriers to fundamental research ethics, such as voluntary informed consent.

In this context Phase I and II clinical studies, in which lack clarity on safety and effectiveness would tip the balance towards risk over benefit, were out of the question, the IoM felt. Even at Phase III the proposed ‘50% rule’ should apply, to “minimise the possibility of using prisoners as human subjects because they are a more convenient or accessible population."