A retrospective cohort of patients living in the French region Languedoc-Roussillon was built. Inclusion criteria were mainly: first admission for HF in 2012; follow-up more than 1 year. The database we used (the SNIIRAM) include outpatient care claim data and hospitalization data. Patients were classified by a hierarchical ascending classification on principal components, using variables describing clinical status, use of specialized and non-specialized care, and main clinical outcomes (hospitalization, death).
2751 patients were included and followed during a median of 22 months. Mean age was 78, 484 (18%) died, 818 (30%) were readmitted at least once for heart failure. The cluster analysis revealed three different significant groups: 1/ group 1 (N=734) is characterized by a younger age, more cares with cardiologists and less main clinical outcomes. Groups 2 and 3 are not different as regards age and comorbidities. 2/group 2 (N=1060) differs from the group 3 (N=957): less cares out-of the hospital contrasting with more cares at hospital, especially at emergencies.
This cohort enables to evaluate consistency and adequation between cares and clinical status, following mail clinical outcomes. As this cohort is built in a heterogeneous region in terms of medical density or health organization, it should be a powerful tool for the study of spatial determinants.