When Jesse Gelsinger, an 18-year-old Arizona man, died during a gene-therapy experiment last autumn, the scientific community first rallied to find out what went wrong. Then politicians and bureaucrats started looking for someone to blame.

Fault has come to rest on the University of Pennsylvania's Institute for Human Gene Therapy (IHGT), whose clinical-trials programme has now been terminated. The IHGT had failed to inform the US Food and Drug Administration (FDA) of adverse events experienced by other patients treated before Gelsinger, according to an FDA letter to the university. But other institutions, agencies and individuals may share some moral culpability: when oversight breaks down, everyone involved in it must examine their role.

The sense that this process is far from settled emerged at a recent Senate hearing. When officials from the Department of Health and Human Services, National Institutes of Health (NIH) and FDA were asked whether recommendations made in 1998 about clinical-trial oversight had been implemented, none responded definitively. Weeks earlier, the health department's Office of Inspector General pointed out that the 1998 recommendations had been largely ignored. Those recommendations spoke directly to many of the problem areas in the Gelsinger trial: informed consent, adverse-events reporting and clinical-trial oversight. The agencies' failure to address issues raised years earlier may have indirectly contributed to the trial's tragic outcome.

Adverse-events reporting may be the clearest example of this. Hundreds of previously unreported adverse events poured in to the NIH once it asked for them following Gelsinger's death. Many gene-therapy researchers privately object to filing adverse-events reports either to the FDA, where they remain confidential, or to the NIH, where they are made public. But, in retrospect, having more public records on adverse events associated with the vector used in Gelsinger's trial would have been useful. Perhaps if enough data on immune responses associated with the vector had been made public before his death, rather than after, the trial would never have been launched.

So whose fault is it that many gene-therapy investigators didn't report their adverse events earlier? The answer is hard to determine, but the NIH leadership may have sent an inadvertent message that reporting to the Recombinant DNA Advisory Committee (RAC) isn't compulsory when in 1996 they tried to reduce the committee's scope.

And although IHGT investigators didn't notify the FDA of all adverse events immediately, they did notify them of similar immune responses. In each case, the FDA let the trial resume. The IHGT and FDA changed the route of vector administration, from intravenously to directly into patients' livers, without notifying the RAC. While that adjustment was intended to limit the vector to the liver, Gelsinger's autopsy showed that it had the opposite effect.

Clearly, the IHGT's closure does not obviate the need for other institutions, agencies and individuals to assess their roles in the tragedy and to make their conclusions public.