A new three-step method for using inverse propensity weighting with latent class analysis

Bias-adjusted three-step latent class analysis (LCA) is widely popular to relate covariates to class membership. However, if the causal effect of a treatment on class membership is of interest and only observational data is available, causal inference techniques such as inverse propensity weighting (IPW) need to be used. In this article, we extend the bias-adjusted three-step LCA to incorporate IPW. This approach separates the estimation of the measurement model from the estimation of the treatment effect using IPW only for the later step. Compared to previous methods, this solves several conceptual issues and more easily facilitates model selection and the use of multiple imputation. This new approach, implemented in the software Latent GOLD, is evaluated in a simulation study and its use is illustrated using data of prostate cancer patients.

[1]  J. Coebergh,et al.  Bowel, urinary, and sexual problems among long-term prostate cancer survivors: a population-based study. , 2009, International journal of radiation oncology, biology, physics.

[2]  A. Kasuya.  EuroQol--a new facility for the measurement of health-related quality of life. , 1990, Health policy.

[3]  J. Pater,et al.  A validation study of the domains of the core EORTC Quality of Life Questionnaire , 1993, Quality of Life Research.

[4]  A. D'Amico,et al.  Prostate Cancer, Version 2.2019, NCCN Clinical Practice Guidelines in Oncology. , 2019, Journal of the National Comprehensive Cancer Network : JNCCN.

[5]  J. Robins,et al.  Estimating causal effects from epidemiological data , 2006, Journal of Epidemiology and Community Health.

[6]  Jeroen K. Vermunt,et al.  Latent Class Modeling with Covariates: Two Improved Three-Step Approaches , 2010, Political Analysis.

[7]  George B. Macready,et al.  Concomitant-Variable Latent-Class Models , 1988 .

[8]  N. Aaronson,et al.  Long‐term quality of life among Dutch prostate cancer survivors , 2006, Cancer.

[9]  David C. Miller,et al.  Appropriateness Criteria for Active Surveillance of Prostate Cancer , 2017, The Journal of urology.

[10]  P. Rosenbaum.  Choice as an Alternative to Control in Observational Studies , 1999 .

[11]  J. Robins,et al.  Analysis of semiparametric regression models for repeated outcomes in the presence of missing data , 1995 .

[12]  Jeroen K. Vermunt,et al.  Power analysis for the bootstrap likelihood ratio test for the number of classes in latent class models , 2016, Advances in Data Analysis and Classification.

[13]  G. Shaw,et al.  Maternal pesticide exposure from multiple sources and selected congenital anomalies. , 1999 .

[14]  I. Korfage,et al.  Prostate cancer survivors who would be eligible for active surveillance but were either treated with radiotherapy or managed expectantly: comparisons on long‐term quality of life and symptom burden , 2010, BJU international.

[15]  G. Imbens.  The Role of the Propensity Score in Estimating Dose-Response Functions , 1999 .

[16]  Francesco Bartolucci,et al.  Causal Latent Markov Model for the Comparison of Multiple Treatments in Observational Longitudinal Studies , 2016 .

[17]  F. Bartolucci,et al.  Evaluating time-varying treatment effects in latent Markov models: An application to the effect of remittances on poverty dynamics , 2019 .

[18]  J. Robins,et al.  Doubly Robust Estimation in Missing Data and Causal Inference Models , 2005, Biometrics.

[19]  Helen Skerman,et al.  Latent Class Analysis Reveals Distinct Subgroups of Patients Based on Symptom Occurrence and Demographic and Clinical Characteristics. , 2015, Journal of pain and symptom management.

[20]  R J Glynn,et al.  Propensity scores for confounder adjustment when assessing the effects of medical interventions using nonexperimental study designs , 2014, Journal of internal medicine.

[21]  Jay Magidson,et al.  How to Perform Three-Step Latent Class Analysis in the Presence of Measurement Non-Invariance or Differential Item Functioning , 2020 .

[22]  J. Robins,et al.  Marginal Structural Models and Causal Inference in Epidemiology , 2000, Epidemiology.

[23]  P. Kelly,et al.  Quality of life of individuals seeking treatment at specialist non-government alcohol and other drug treatment services: A latent class analysis. , 2018, Journal of substance abuse treatment.

[24]  J. Vermunt,et al.  Testing for Measurement Invariance With Latent Class Analysis , 2010 .

[25]  J. Vermunt,et al.  Heterogeneity in Quality of Life of Long‐Term Colon Cancer Survivors: A Latent Class Analysis of the Population‐Based PROFILES Registry , 2020, The oncologist.

[26]  H. Akaike.  A new look at the statistical model identification , 1974 .

[27]  Karina Friis,et al.  A Latent Class Analysis of Multimorbidity and the Relationship to Socio-Demographic Factors and Health-Related Quality of Life. A National Population-Based Study of 162,283 Danish Adults , 2017, PloS one.

[28]  Geoffrey J. McLachlan,et al.  Finite Mixture Models , 2019, Annual Review of Statistics and Its Application.

[29]  D. Rubin,et al.  Reducing Bias in Observational Studies Using Subclassification on the Propensity Score , 1984 .

[30]  Stephanie T. Lanza,et al.  Causal Inference in Latent Class Analysis , 2013, Structural equation modeling : a multidisciplinary journal.

[31]  J. Coebergh,et al.  The Patient Reported Outcomes Following Initial treatment and Long term Evaluation of Survivorship registry: scope, rationale and design of an infrastructure for the study of physical and psychosocial outcomes in cancer survivorship cohorts. , 2011, European journal of cancer.

[32]  Marcel Croon,et al.  Estimating Latent Structure Models with Categorical Variables: One-Step Versus Three-Step Estimators , 2004, Political Analysis.

[33]  M. Cooperberg,et al.  Active surveillance for prostate cancer: a systematic review of the literature. , 2012, European urology.

[34]  Katherine E. Masyn,et al.  Covariates and Mixture Modeling: Results of a Simulation Study Exploring the Impact of Misspecified Effects on Class Enumeration , 2016 .

[35]  Jee-Seon Kim,et al.  Hybridizing Machine Learning Methods and Finite Mixture Models for Estimating Heterogeneous Treatment Effects in Latent Classes , 2020 .

[36]  A. Jemal,et al.  Cancer statistics, 2020 , 2020, CA: a cancer journal for clinicians.

[37]  Scott L. Zeger,et al.  Latent Variable Regression for Multiple Discrete Outcomes , 1997 .

[38]  P. Austin.  An Introduction to Propensity Score Methods for Reducing the Effects of Confounding in Observational Studies , 2011, Multivariate behavioral research.

[39]  José G. Dias,et al.  A bootstrap-based aggregate classifier for model-based clustering , 2008, Comput. Stat..

[40]  Paul F. Lazarsfeld,et al.  Latent Structure Analysis. , 1969 .

[41]  L. A. Goodman.  Exploratory latent structure analysis using both identifiable and unidentifiable models , 1974 .

[42]  A. Williams.  EuroQol : a new facility for the measurement of health-related quality of life , 1990 .

[43]  N. Aaronson,et al.  The European Organization for Research and treatment of cancer approach to quality of life assessment: guidelines for developing questionnaire modules , 1993, Quality of Life Research.

[44]  G. Schwarz.  Estimating the Dimension of a Model , 1978 .

[45]  Douglas L. Miller,et al.  A Practitioner’s Guide to Cluster-Robust Inference , 2015, The Journal of Human Resources.

[46]  M. Heymans,et al.  Different ways to estimate treatment effects in randomised controlled trials. , 2018 .