Variables
We defined three types of variables (i) At each consultation during the follow-up, oncologist pediatricians carried out a psychosocial evaluation. During this specific dedicated time, pediatricians, children and their parents discuss the potential psychosocial difficulties. At the end of the discussion, a synthesis was reported using as a binary form (yes/no) to indicate if learning or psychological difficulty was present or not. Whether the patient received assistance for each difficulty was also reported. (ii) The clinical variables were composed of the patients and their tumors characteristics. We merged the tumor types into five categories; i.e., blood disorders, CNS tumors, solid tumors, bone tumors, and other tumors (iii) Two variables evaluated the SES of the children and their household. First, the European Deprivation Index (EDI) score assessed the socioeconomic environment of each patient. This deprivation indicator is constructed by selecting fundamental needs associated both with objective and subjective poverty(22). This score is determined by an ecological measure using the IRIS-scale (Ilôts Regroupés pour l’Information Statistique ), which represents the smallest French geographical area for which there is a statistical evaluation to estimate social deprivation. Based on their addresses at diagnosis, each patient can be associated with an IRIS, and thus, their EDI score can be established(22,23). For interpretation, the highest EDI score is associated with greater social deprivation. Second, we used the travel time defined as the shortest time to travel by car from the patient’s address to the reference care center to assess geographical disparities. A geographical information system (ArcGIS 10.5®—Esri France) associated with a road map database (Navstreets®, provided by HERE and Esri France) was used for this estimation. For the RECAPGO database, each patient’s EDI score and travel time to reference care center were estimated from the address at diagnosis.