Double-blind phase I-II clinical trials: preliminary risk analysis applied to the pharmaceutical circuit of injectable anticancer drugs in a preparation unit

Background
Our preparation unit has experienced an increase in clinical trial (CT) activity following approvals for phase I-II CTs. In view of this increase and in order to comply with the 2022 Good Preparation Practices, we conducted a Process Hazard Analysis (PHA) of the pharmaceutical circuit for injectable anticancer drugs used in phase I-II-III CTs in onco-hematology. Double-blind phase I-II CT implementation (excluding placebo-controlled CTs, which follow a different process) is complex, as a high risk of exclusion is unblinding. To our knowledge, no risk analysis of this process has been described. Our aim is to share the corrective measures drawn from our PHA, specific to these phase I-II CTs.

Methods
The PHA method was carried out by a working group of 3 pharmacists specialized in CT, pharmacotechnics and risk analysis. Eighteen meetings took place between April 2023 and May 2024, including 3 on double-blind, placebo-free phase I-II CTs. The aim was to identify, analyze, prioritize and manage the risks associated with this activity. Scenarios were developed and rated in terms of initial criticality (iC) based on the hazard situations identified. The impact of the corrective measures was assessed by estimating the residual criticality (rC).

Results
Three phases and 4 main risks were identified in the phase I-II-III PHA, leading to 77 priority 1 hazard situations, including 10 for double-blind, placebo-free phase I-II CTs. Of these, 12 scenarios were developed (9 of iC=2 ; 3 of iC =1). The phase “CT feasibility and computerized protocol drafting” registered the highest number of hazard situations, and the main risk was the lack of training. Ten corrective measures were taken (5 of which were easy to implement), grouped into 4 different themes (software configuration (adaptation to Chimio® limitations), secure computer selection, training and organization). For example: specific detailed operating procedure for double-blind, placebo-free CT computerized protocol drafting with a fictitious unit inaccessible to prescribers; standard wording allowing doses and INNs on the preparation sheet but not on the labels ; harmonization of computerized protocol wordings ; training sheet on the principle of double-blind CTs and English-French lexicon terms related to CTs were drawn up. After implementing our corrective measures, no scenario had a rC ≥ 2.

Discussion/conclusion
The implementation of our corrective measures has allowed us to make the risks acceptable and to secure the phase I-II double-blind, placebo-free CT circuit in our preparation unit, especially in the critical preparation stage, by computerizing the preparation sheet in Chimio^® without removing the double-blind. Our PHA showed that corrective measures established for training are essential and need to be regularly updated.