Dataset Information

Methods to Account for Uncertainty in Latent Class Assignments When Using Latent Classes as Predictors in Regression Models, with Application to Acculturation Strategy Measures.

ABSTRACT: Latent class models have become a popular means of summarizing survey questionnaires and other large sets of categorical variables. Often these classes are of primary interest to better understand complex patterns in data. Increasingly, these latent classes are reified into predictors of other outcomes of interests, treating the most likely class as the true class to which an individual belongs even though there is uncertainty in class membership. This uncertainty can be viewed as a form of measurement error in predictors, leading to bias in the estimates of the regression parameters associated with the latent classes. Despite this fact, there is very limited literature treating latent class predictors as measurement error models. Most applications ignore this issue and fit a two-stage model that treats the modal class prediction as truth. Here, we develop two approaches-one likelihood-based, the other Bayesian-to implement a joint model for latent class analysis and outcome prediction. We apply these methods to an analysis of how acculturation behaviors predict depression in South Asian immigrants to the United States. A simulation study gives guidance for when a two-stage model can be safely implemented and when the joint model may be required.

SUBMITTER: Elliott MR

PROVIDER: S-EPMC7480960 | biostudies-literature | 2020 Mar

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Methods to Account for Uncertainty in Latent Class Assignments When Using Latent Classes as Predictors in Regression Models, with Application to Acculturation Strategy Measures.

Elliott Michael R MR Zhao Zhangchen Z Mukherjee Bhramar B Kanaya Alka A Needham Belinda L BL

Epidemiology (Cambridge, Mass.) 20200301 2

Latent class models have become a popular means of summarizing survey questionnaires and other large sets of categorical variables. Often these classes are of primary interest to better understand complex patterns in data. Increasingly, these latent classes are reified into predictors of other outcomes of interests, treating the most likely class as the true class to which an individual belongs even though there is uncertainty in class membership. This uncertainty can be viewed as a form of meas ...[more]

PMID: 31809338

Similar Datasets

Project description:Background: Misuse of antibiotics is prevalent worldwide and primary care is a major contributor. Although a clear diagnosis is fundamental for rational antibiotic use, primary care physicians often struggle with diagnostic uncertainty. However, we know little about how physicians cope with this situation and its association with antibiotic prescribing. Methods: A total of 583 primary care physicians were surveyed using the Dealing with Uncertainty Questionnaire. Their prescriptions (n = 949,181) over the year 2018 were retrieved retrospectively. Two categories of behavioral patterns of participants were identified based on latent class analyses (high vs. low openness and collaborativeness) in responding to diagnostic uncertainty. Multi-level logistic regression models were established to determine the associations between these behavioral patterns and antibiotic prescribing (overall and broad-spectrum antibiotics) for illness without an indication for antibiotics and those with a conditional indication for antibiotics, respectively, after adjustment for variations of patient (level one) and physician (level two) characteristics. Results: Most physicians reported open communications with their patients (80.96%), collected further information (85.08%), and referred patients to specialists (68.95%) in dealing with diagnostic uncertainly. More than half (56.95%) sought help from colleagues. Less than 20% acted on intuition or adopted a "wait and see" strategy. About 40% participants (n = 238) were classified into the group of low openness and collaborativeness in coping with diagnostic uncertainty. They were more likely to prescribe antibiotics for the recorded illness without an indication for antibiotics (AOR = 1.013 for all antibiotics, p = 0.024; AOR = 1.047 for broad-spectrum antibiotics, p < 0.001), as well as for the recorded illness with a conditional indication for antibiotics (AOR = 1.226 for all antibiotic, p < 0.001; AOR = 1.257 for broad-spectrum antibiotics, p < 0.001). Conclusion: Low tolerance with diagnostic uncertainty is evident in primary care. Inappropriate and over antibiotic prescribing is shaped by physicians' coping methods of diagnostic uncertainty. Physicians' openness and collaborativeness in responding to diagnostic uncertainty is associated with lower antibiotic prescribing in primary care. Interventions targeting on better management of diagnostic uncertainty may offer a promising approach in reducing antibiotic use in primary care.

Dataset Information

Methods to Account for Uncertainty in Latent Class Assignments When Using Latent Classes as Predictors in Regression Models, with Application to Acculturation Strategy Measures.

Publications

Methods to Account for Uncertainty in Latent Class Assignments When Using Latent Classes as Predictors in Regression Models, with Application to Acculturation Strategy Measures.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure