Validation of a research classification tool for primary care residents according to the "Jarde" law

Aim. To validate a tool for classification of research protocols into two categories: research involving the human person (RIPH) or not (no-RIPH) according to the legislation in force since April 2018, for the theses of the general practice residents (IMG). To identify among the no-RIPH those that are under the supervision of an ethics committee. Method. Independent classification of thesis by two IMG using the decision-making tool based on legal texts. The classification was compared to the experts' classification. Divergent classifications were discussed among experts during deliberation meetings to reach consensus. Each time the decision-making tool was optimized, the classification of the thesis was fully reproduced by the two IMG. Meanwhile, no-RIPH that required the advice of an ethics committee were identified. Results. 254 theses were classified. The experts' classification found a RIPH prevalence of 9%. Doubts persisted for two theses; two others were unclassifiable. The deliberation meetings allowed two changes to the decision-making tool. The final classification of IMG found a prevalence of RIPH at 11%. The informational indices of the final tool found a sensitivity of 1, a specificity of 0.98, a VPP of 0.82, a VPN of 1. Half of the no-RIPH files, although not under the responsibility of a CPP, raised ethical issues regarding the international founding texts requiring the approval of an ethics committee for publication. Conclusion. There was no validated tool to classify research projects as belonging or not to the Jarde law. This decision-making tool has satisfactory intrinsic validity. Although 91% of the plans were classified as no-RIPH, half of them would be subject to the opinion of an ethics committee in accordance with international research ethics principles.