Latent Variable Model for Multivariate Data with Measure-specific Sample Weights and Its Application in Hospital Compare

10/18/2019 ∙ by Chengan Du, et al. ∙ 0

We developed a single factor model with measure-specific sample weights for multivariate data with multiple observed indicators clustered within a higher level subject. The factor is therefore a latent variable shared by multiple indicators within a same subject and the sample weights are different across different indicators and different subjects. Even after integrating out the latent variable, the likelihood of the data cannot be written as the sum of weighted likelihood of each subject because a subject has different sample weights respectively for its multiple indicators. In addition, the number of available indicators varies across subjects. We derive a pseudo likelihood for the latent variable model with measure-specific weights. We investigate various statistical properties of the latent variable model with measure-specific sample weights and its connection to the traditional factor analysis. We found that the latent variable model provides consistent estimates for its variances when the measure-specific sample weights are properly re-scaled. Two estimation procedures are developed - EM algorithm for the pseudo likelihood and marginalization of the pseudo likelihood by directly integrating out the latent variable to obtain the parameter estimates. This approach is illustrated by the analysis of publicly reported hospitals with indicators and sample weights. Numerical studies are conducted to investigate the influence of weights and their sample distribution.

READ FULL TEXT
POST COMMENT

Comments

There are no comments yet.

Authors

page 1

page 2

page 3

page 4

This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.