The most important resource needed to conduct research on humans, it is said, is not brainpower or money: it is trust. In the United States, as elsewhere, hundreds of institutions and thousands of investigators work to protect that trust by carefully evaluating proposals for clinical trials and other research that uses human subjects.

Each US institution hosting such a study typically conducts its own ethical review of the proposal. The review process serves many functions: it is an expression of the responsibility that these investigators feel towards protecting their local community, an opportunity to tweak protocols to adapt to the community’s specific needs, and a protection against potential lawsuits resulting from a flawed research protocol.

Sadly, evidence suggests that much of this effort is misplaced. A 2010 survey of 45 institutions reviewing the same protocol found that local scrutiny resulted in no substantial changes (B. Ravina et al. Ann. Neurol. 67, 258–260; 2010). Instead, most alterations simply inserted standardized institutional language — unrelated to the proposed study — to the informed-consent document signed by research participants before they enter a trial. The total cost of all that review: more than US$100,000.

On 3 December, the US National Institutes of Health (NIH) announced a draft policy intended to reduce that redundancy. Open for comment until 29 January, the proposal would require NIH-funded trials that are conducted at more than one site to be approved by a single institutional review board (IRB), which must be willing to shoulder responsibility for all of the sites. The intention is to speed up the approval process for trials that are conducted at multiple facilities. At present, each site may take a crack at reviewing a protocol, often delaying the start of a trial and introducing potential inconsistencies in study protocols and consent forms at different sites.

There is no evidence that multiple ethics reviews enhance protections for human subjects.

The NIH’s move is the latest in a string of efforts by US regulators to change this institutional practice. In 2006, the US Food and Drug Administration released guidance for clinical trials conducted at multiple sites. In it, the agency stated that this ethical review need not take place at every institution. Instead, each trial could designate an institution to conduct a central review for all participating sites. Four years later, the US Office of Human Research Protections wrote a letter stating its support for that guidance. Despite these assurances, however, it has been difficult to change entrenched institutional practices that have been solidified for more than 40 years.

The NIH’s proposal does not prohibit any participating site from conducting its own review, but clearly frowns on the practice — and explicitly pushes the cost of a duplicate review onto the institution.

Inertia is difficult to overcome, particularly at large institutions and with such a valuable resource at stake. Much of this stubbornness is due to an understandable desire by investigators to protect their patients and community. Some local IRB members feel that abdicating their review of research protocols is a violation of their responsibility to that community, and worry that standards will slip if they do not personally review the study.

As the NIH has said, there is no evidence that multiple ethics reviews enhance protections for human subjects. Centralized review may seem to save time and money, but there is no clear evidence that it protects study subjects any better. Still, the NIH’s move to encourage central review is the right one, given the available evidence.

Regulations that favoured local IRB reviews were developed in an era when studies were typically done at a single site. This is no longer the case. As therapies become more tailored to individual genetics, and diseases are subdivided into rarer subtypes, more sites are needed to enrol enough patients to evaluate an intervention.

Around the world, DNA sequencing labs are generating reams of genetic data that could hold the clues to the next medical revolution. Finding those clues quickly and ethically will require studies that combine data from across the globe. Investigators are clamouring for unified informed-consent documents that will allow them to compile genetic information into databases without creating a legal thicket of differing privacy protections. The NIH’s move is an important step in that direction, but there is much farther to go.