As you might expect, the CDC released its report on how its influenza lab managed to accidentally ship an influenza sample contaminated with avian influenza late in the day on an August Friday. Shocking, I know (and pdf format makes it that much easier to work with too! Or something).
Anyway, it looks like a combination of all too-human errors and pressures led to the contamination (pdf; boldface mine):
The Team 1 laboratory scientist involved in the incident indicated that the following procedures were used: inoculation of the MDCK cell culture with the H9N2 virus; decontamination of the biosafety cabinet (BSC) using a standard protocol; and then inoculation of the MDCK cell culture with the two H5N1 viruses. A minimum of 1.5 hours (30 min for inoculation of H9N2; 30 min for decontamination of the BSC; and 30 min for inoculation of the two H5N1 strains) would have been required to process the specimens as described by the Team 1 laboratory scientist. Additional time would have been needed for set up time at the BSC, movement of materials from the BSC to the incubator, showering out of the BSL3-E laboratory suite, and other actions required to complete the work.
The card key readers indicated that the Team 1 laboratory scientist entered the BSL3-E suite at 10:13 AM, and accessed the freezer with the H5N1 and H9N2 virus strains twice (at 10:50 and 10:54 AM). The Team 1 laboratory scientist manually signed out of the BSL3-E suite at 11:45 AM. A total of 51 minutes elapsed from when the Team 1 laboratory scientist accessed the agents in the freezer to when this staff member exited the BSL3-E suite. During the 51 minutes, the Team 1 laboratory scientist also would have been required to shower out of the suite and change into street clothes. Therefore, the time that this staff member performed the cell culture work was substantially less than the 1.5 hours that would have been required if the protocol had been followed.
Because there is no written documentation (i.e., a laboratory notebook or other notes), it is not possible to say conclusively what actions the Team 1 laboratory scientist did and how they were done. However, the 1.5 hour-protocol could not have been followed during the 51 minutes that the Team 1 laboratory scientist was in the BSL3-E suite. When interviewed, the Team 1 laboratory scientist acknowledged being rushed to attend a laboratory meeting at noon. The Team 1 laboratory scientist also indicated being unable to specifically remember the events since they had taken place almost 6 months previously. The Team 1 laboratory scientist further described following a “best practices” protocol for temporal separation of LPAI and HPAI virus propagation; however, this laboratory did not have a written, approved laboratory team-specific SOP for the work that the Team 1 laboratory scientist was doing.
The Team 1 lead, the Team 1 laboratory scientist who conducted the cell culture inoculations, the VSDB Branch Chief, the ID Director, and others noted the Division’s heavy work load at the time of the incident. The ID laboratories were under substantive pressure to generate data for the upcoming WHO Vaccine Consultation Meeting in February 2014. The ID typically provides 50%-75% of the data reviewed at these meetings.
While there are some good recommendations in the report, including the testing of samples, the real problem here was human nature: someone was rushed, and the team was worked very hard.
But I’m sure this has no relevance whatsoever to the debate around recreating the Spanish influenza strain. Couldn’t possibly happen again.