The latest emergency response guidelines for clinical research

The latest version of the core framework for emergency response in clinical research can be directly summarized as "pre-risk classification + scenario-based response + full-link traceability". The core basis is the "Technical Guidance Principles for Emergency Management of Drug Clinical Trials (Trial)" issued by the National Medical Products Administration (NMPA) in December 2023, superimposed on the 2024 ICH GCP Compared with the 2020 version of the old guidelines, the biggest adjustment to the cross-regional disclosure requirements for emergencies added to R2 is to completely cancel the unified disposal process and instead match the corresponding disposal plans according to the trial risk level. The current disputes that have not yet reached a consensus in the industry are mainly concentrated in the two directions of "whether AI-assisted early warning should be included in the mandatory requirements" and "the priority definition of subject rights and trial progress under public health events".

To be honest, the previous old guidelines were like uniform-sized school uniforms. Whether you were conducting a high-risk CAR-T trial or a low-risk vitamin supplement trial, you had to follow the same process. Not only was it a waste of manpower, but it was easy to get stuck in the process when things went wrong. Now the new guidelines are more like customized suits. What risk levels are matched with what disposal requirements are much more flexible. Last month, I checked the reporting process with a teacher from a tertiary hospital. I happened to encounter an emergency situation they were dealing with in a Phase I trial in the gastroenterology department: 3 subjects developed grade 2 diarrhea after taking the drug. According to the requirements of the old guidelines, as long as it does not reach the SAE (serious adverse event) level, only routine records need to be reported. However, according to the latest guidelines, a group AE of the same type in the Phase I trial directly triggered a Level 2 warning, and the institution suspended it that day. For the administration of subjects in the same batch, the stool samples of three subjects were first submitted for inspection, and the medication records and drug batch information were synchronized to the sponsor's pharmaceutical department. The entire response process was completed in less than 2 hours. Finally, it was found that there was a 2-hour temperature deviation during the cold chain transportation of the drug, which happened to be stuck at the activity impact threshold stipulated in the guidelines. In the end, no more serious adverse reactions occurred, and the progress of the entire trial was not delayed.

Of course, not all institutions find this new requirement easy to use. I talked to the PI of a local county-level hospital before, and he had a headache: their department only accepts low-risk trials of external preparations for dermatology, and the department only has 2 CRCs. If the high-risk projects in the guidelines are required to be equipped with 24-hour emergency duty and AI real-time warning, it will not be able to spare manpower, but it is easy to make mistakes due to the handling process. The views of both factions in the industry are actually quite realistic: Head sponsors and top tertiary institutions are conducting high-risk trials such as oncology and cell therapy. They have sufficient manpower and material resources, and they would like to raise the warning line as high as possible. After all, the price of an SAE may be to shut down the entire project. ； However, the incidence of adverse reactions in low-risk trials conducted by grassroots institutions is not as high as that of the common cold. Excessive prevention and control has only increased unnecessary costs. Now NMPA does not provide a one-size-fits-all mandatory standard. Each agency adjusts it according to its own project risks. In fact, it is more in line with the actual situation.

Many new CRAs ask me, what should I do as soon as possible in an emergency? To put it bluntly, don’t think about flipping through the guide and filling out forms first, first check to see if the person is OK. Last year, I was at the site of a CAR-T trial and happened to encounter a subject who developed grade 3 cytokine release syndrome (CRS) after reinfusion. The first reaction of the research nurse at the time was not to find an SAE reporting template, but to push the IL-6 antagonist directly according to the previously rehearsed process, and at the same time contacted In the green channel of the ICU, the reporting process will only start after the subject's blood pressure stabilizes and the condition improves. This is also clearly mentioned in the latest guidelines: the rights of the subject are the absolute first priority, and all administrative processes must make way for first aid. The old requirement of reporting first and then handling has been completely abolished.

There is also a new requirement that is easily ignored by everyone: all emergency response records, including WeChat communication records, call recordings, drug testing reports, and even on-site surveillance videos, must be bound and archived with the subject data in the EDC (electronic data collection system), and cannot be stored separately in the institution's local computer. In the past, a sponsor was given a yellow card during on-site inspection because the emergency response recordings were not synchronized to the EDC. Many institutions have not corrected this, so we must be reminded.

I have been in this business for almost 7 years, and my deepest feeling is that the emergency guide is never used to deceive people, but is used to give everyone the bottom line. I encountered a project before, where the CRC accidentally distributed the experimental drug to the subjects in the placebo group. According to the old guidelines, it would definitely be considered a serious protocol violation, and it would have to be reported to NMPA, and the entire trial might be suspended. ； But according to the latest guidelines, as long as you have made a contingency plan for drug dispensing errors, conducted a full set of examinations on the subjects as soon as possible, and did not experience any adverse reactions, and proactively informed the subjects and the ethics committee, you only need to submit a rectification report in the end. It will not affect the progress of the trial at all, but is much more humane than the previous rigid requirements.

Oh, by the way, many colleagues have recently been discussing whether to include the subject’s emergency contact information as a routine verification item. After all, many subjects may conceal their basic medical history when enrolling. Contacting family members can save a lot of trouble when an accident occurs. However, there is no clear written regulation yet. Everyone can flexibly adjust according to their own project conditions.