The variability of factors in the explanation of the links between family configuration and the care mix
Abstract
Using data from the CARE survey, we aim to describe and explain the different configurations of family care and professional care. Based on the hypothesis that these two types of configuration are correlated, it seemed relevant to observe the mobilization of those two types of care considering the existing family structure, in other words the family carers potentially available. We therefore grouped the individuals by type of family structure in order to observe the variability of the professional and family response to the needs expressed by the elderly. This initial work made it possible to identify, for each given family structure, the most likely care configurations. Then, we sought to explain the variation of these care configuration depending on the characteristics of the individual being cared for (level of dependency, number of children, age, gender, place of residence, income, etc). Based on an unparametred classification model, we identify the most relevant variables to divide the individuals according to their care configurations. With each partitioning, the algorithm seeks to divide the population into two groups to minimise the heterogeneity of the sub groups formed. In conclusion, it should be noted that the level of dependency - a variable included in the model - did not emerge as a discriminating variable to explain the patterns of care mix characteristic in the different groups. We will also show that depending the family configuration, the variable that distinguish the care mix response are not the same. For each family configuration we can observe different spring to explain the care arrangement.