@article{fdi:010048385,
  title = {{L}'imputation multiple des donn{\'e}es manquantes al{\'e}atoirement : concepts g{\'e}n{\'e}raux et pr{\'e}sentation d'une m{\'e}thode {M}onte-{C}arlo = {M}ultiple imputation of missing at random data : general points and presentation of a {M}onte-{C}arlo method},
  author = {{C}ottrell, {G}illes and {C}ot, {M}ichel and {M}ary, {J}. {Y}.},
  editor = {},
  language = {{FRE}},
  abstract = {{B}ackground. - {S}tatistical analysis of a data set with missing data is a frequent problem to deal with in epidemiology. {M}ethods are available to manage incomplete observations, avoiding biased estimates and improving their precision, compared to more traditional methods, such as the analysis of the sub-sample of complete observations. {M}ethods. - {O}ne of these approaches is multiple imputation, which consists in imputing successively several values for each missing data item. {S}everal completed data sets having the same distribution characteristics as the observed data (variability and correlations) are thus generated. {S}tandard analyses are done separately on each completed dataset then combined to obtain a global result. {I}n this paper, we discuss the various assumptions made on the origin of missing data (at random or not), and we present in a pragmatic way the process of multiple imputation. {A} recent method, {M}ultiple {I}mputation by {C}hained {E}quations ({MICE}), based on a {M}onte-{C}arlo {M}arkov {C}hain algorithm under missing at random data ({MAR}) hypothesis, is described. {A}n illustrative example of the {MICE} method is detailed for the analysis of the relation between a dichotomous variable and two covariates presenting {MAR} data with no particular structure, through multivariate logistic regression. {R}esults. - {C}ompared with the original dataset without missing data, the results show a substantial improvement of the regression coefficient estimates with the {MICE} method, relatively to those obtained on the dataset with complete observations. {C}onclusion. - {T}his method does not require any direct assumption on joint distribution of the variables and it is presently implemented in standard statistical software ({S}plus, {S}tata). {I}t can be used for multiple imputation of missing data of several variables with no particular structure.},
  keywords = {{M}issing data ; {M}issing at random ; {M}ultiple imputation ; {MCMC}},
  booktitle = {},
  journal = {{R}evue d'{E}pid{\'e}miologie et de {S}ant{\'e} {P}ublique},
  volume = {57},
  numero = {5},
  pages = {361--372},
  ISSN = {0398-7620},
  year = {2009},
  DOI = {10.1016/j.respe.2009.04.011},
  URL = {https://www.documentation.ird.fr/hor/fdi:010048385},
}