Re: Intraosseous versus intravenous vascular access in upper extremity among adults with out-of-hospital cardiac arrest: cluster randomised clinical trial (VICTOR trial)
CCBYNCOpen access
Rapid response to:
Research
Intraosseous versus intravenous vascular access in upper extremity among adults with out-of-hospital cardiac arrest: cluster randomised clinical trial (VICTOR trial)
Re: Intraosseous versus intravenous vascular access in upper extremity among adults with out-of-hospital cardiac arrest: cluster randomised clinical trial (VICTOR trial)
Dear Prof. Lars W. Andersen,
Thank you for your comments.
During the first bi-weekly period, all four ALS teams were assigned to perform IO procedures. This measure was intended to ensure that all paramedics in Taipei were aware of the trial’s commencement, particularly after the trial had been repeatedly postponed due to the threat of the COVID-19 outbreak. In practice, when collaborating with fire departments, we aimed to reduce the cognitive load on firefighters without affecting their work caused by frequent visits and changes. Therefore, the research assistant initiated the randomization process to allocate two ambulance teams (two clusters) out of four every two weeks (as a block), with an additional four blocks of IV added to every seven blocks of the randomization scheme to achieve an approximate 1:2 ratio of IO to IV.
The EMTs are independent operators and were only informed of which vascular procedure to perform when the research assistant announced the intervention at the start of each specific bi-weekly period, without knowing the randomization scheme in advance. Thus, the allocation sequence remained concealed from them. We believe this randomization process effectively reduced confounding factors and yielded valid results.
Regarding the intended allocation ratio of 1:2, we found a final count of 3,915 in the IV blocks and 3,865 in the IO blocks, resulting in a nearly 1:1 ratio (1:1.01). We were also surprised by this unexpected result. The attached figure shows the exact number of patients in each cluster (https://drive.google.com/file/d/1VX5oMe-u41fcpHysAPBmrjObhzDWDXDw/view?u...) Before submitting, we rechecked these numbers multiple times. We also verified the total number of 7,780 patients from both the dispatch center records and the EMTs’ ambulance reports to ensure consistency. Additionally, we inquired with many on-scene paramedics about any procedural variations during IO/IV interventions, but no specific findings were identified. Therefore, we acknowledge that the two potential explanations for this phenomenon mentioned in the manuscript are not strongly convincing. At this point, we are open to any new perspectives that might help us understand the existing results.
Finally, we believe that transparent data sharing is crucial for clinical trials. However, processes such as translating the codebook from Chinese to English or preparing a statement of collaboration and intellectual property among data owners and receivers will take some time. We will inform you immediately once we are ready.
Thank you for raising these questions, as they provide us with the opportunity to clarify our findings.
Rapid Response:
Re: Intraosseous versus intravenous vascular access in upper extremity among adults with out-of-hospital cardiac arrest: cluster randomised clinical trial (VICTOR trial)
Dear Prof. Lars W. Andersen,
Thank you for your comments.
During the first bi-weekly period, all four ALS teams were assigned to perform IO procedures. This measure was intended to ensure that all paramedics in Taipei were aware of the trial’s commencement, particularly after the trial had been repeatedly postponed due to the threat of the COVID-19 outbreak. In practice, when collaborating with fire departments, we aimed to reduce the cognitive load on firefighters without affecting their work caused by frequent visits and changes. Therefore, the research assistant initiated the randomization process to allocate two ambulance teams (two clusters) out of four every two weeks (as a block), with an additional four blocks of IV added to every seven blocks of the randomization scheme to achieve an approximate 1:2 ratio of IO to IV.
The EMTs are independent operators and were only informed of which vascular procedure to perform when the research assistant announced the intervention at the start of each specific bi-weekly period, without knowing the randomization scheme in advance. Thus, the allocation sequence remained concealed from them. We believe this randomization process effectively reduced confounding factors and yielded valid results.
Regarding the intended allocation ratio of 1:2, we found a final count of 3,915 in the IV blocks and 3,865 in the IO blocks, resulting in a nearly 1:1 ratio (1:1.01). We were also surprised by this unexpected result. The attached figure shows the exact number of patients in each cluster (https://drive.google.com/file/d/1VX5oMe-u41fcpHysAPBmrjObhzDWDXDw/view?u...) Before submitting, we rechecked these numbers multiple times. We also verified the total number of 7,780 patients from both the dispatch center records and the EMTs’ ambulance reports to ensure consistency. Additionally, we inquired with many on-scene paramedics about any procedural variations during IO/IV interventions, but no specific findings were identified. Therefore, we acknowledge that the two potential explanations for this phenomenon mentioned in the manuscript are not strongly convincing. At this point, we are open to any new perspectives that might help us understand the existing results.
Finally, we believe that transparent data sharing is crucial for clinical trials. However, processes such as translating the codebook from Chinese to English or preparing a statement of collaboration and intellectual property among data owners and receivers will take some time. We will inform you immediately once we are ready.
Thank you for raising these questions, as they provide us with the opportunity to clarify our findings.
Sincerely,
Dr. Ying-Chih Ko
Prof. Wen-Chu Chiang
Competing interests: No competing interests