Do patients’ SEP influence therapeutic strategy in acute myeloid leukaemia: Results from a French longitudinal observational study

Abstract

Introduction Several studies highlight the influence of socioeconomic position (SEP) on cancer survival, especially through treatment choice. However, data on hemopathy are scarce. Specific studies on acute myeloid leukaemia (AML) found an association between educational level and access to transplantation, SEP and survival or neighborhood SEP and access to chemotherapy and transplantation. However, these studies were done in countries with heath systems that are not the same as in France. In this study, we investigate the influence of SEP on the treatment received among patients of at least 60 y with AML using data from France, i.e. in settings of tax-supported Health Care System and policy aiming at harmonizing the quality of care as the National ‘cancer plans’. Methods We used data from an observational cohort representative of AML patients of at least 60y treated in MP. This represented 697 patients with an AML diagnosed between 1st January 2009 to 31st December 2014, excluding AML 3 - promyelocytic - subtypes. We built a theoretical model of clinical decision process in which we distinguished 2 steps: access to chemotherapy (first step), and access to an alternative to best supportive care among patients who were not eligible for intensive chemotherapy (second step). Logistic models were used to test, independently for each step, the effect of PSE on treatment received, first without adjustment, and then, after successive adjustment for confounders identified with bivariate analyses (a = 0.2). We measured PSE by the French version of the European Deprivation Index (EDI), an ecological deprivation index already used in several studies addressing social inequalities in cancer management and outcome. Results Analyses on complete data (n = 593 - 85% of included patients) show that, in the first step, disadvantaged patients had a lower probability to be treated by intensive chemotherapy compared to most advantaged patients. However, this association did not persist after adjusting for ‘initial status’, i.e. whether it was an AML de novo or a post-affection AML. No statistically significant association was found between SEP and treatment choice in the second step among patients who was not eligible to intensive chemotherapy. Consistent results were found in sensitivity analyses using alternative coding of EDI and after dealing with missing data using multiple imputation. Conclusion No direct influence of PSE on treatment choice was found in our study among a sample representative of patients of at least 60 y treated for AML in the Midi-Pyrenees region. These results might be seen as an evidence of the efficacy of the regional organization in care network to provide homogeneous quality of care in the territory. Nevertheless, further studies are needed to explore the influence of SEP on the initial status, i.e. AML de novo or AML post-affection.