Two-Part Pattern-Mixture Model for Longitudinal Incomplete Semi-Continuous Toenail Data
DOI:
https://doi.org/10.6000/1929-6029.2012.01.02.05Keywords:
Two-part model, semi-continuous data, gamma mixed model, dropout, Toenail dataAbstract
Longitudinal data with true zero values, known as longitudinal semi-continuous data, frequently occur in medical, environmental and biological studies. To model longitudinal semi-continuous data, two-part modelling approaches have been widely used in literature. In the first part of the two-part model, binary logistic regression is commonly used after converting the semi-continuous responses to binary responses. In the second part, the semi-continuous data are converted to positive continuous data after removing the true zero values from the responses. Although positive continuous or non-zero values tend to show a positively skewed distribution pattern, in the literature the normal distribution is commonly used to model them. Also, in longitudinal studies, data often suffer individual dropouts as they are collected overtime. In this paper, we propose a two-part pattern-mixture model to analyze longitudinal semi-continuous data with dropouts. In the proposed approach, we use pattern-mixture binary mixed models for the first part and positively continuous pattern-mixture gamma mixed models for the second part. Our approach can accommodate both subject- and time-specific correlation as well as dropout pattern. We also incorporate a computationally efficient estimation method for our models using a penalize quasi-likelihood approach. The proposed method is illustrated with an application to the longitudinal incomplete toenail data
References
Backer MDe, Vroery CDe, Lesaffre E, Scheys I, Keyser PDe. Twelve weeks of continuous oral therapy for toenail onychomycosis caused by dermatophytes: a double-blind comparative trial of Terbinafine 250 mg/day versus Itraconazole 200 mg/day. J Am Acad Dermatol 1998; 38: 57-63. http://dx.doi.org/10.1016/S0190-9622(98)70486-4 DOI: https://doi.org/10.1016/S0190-9622(98)70486-4
Komáreck A, Lesaffre E. Generalized linear mixed model with a penalized Gaussian mixture as a random effects distribution. Comput Statist Data Anal 2008; 52: 3441-58. http://dx.doi.org/10.1016/j.csda.2007.10.024 DOI: https://doi.org/10.1016/j.csda.2007.10.024
Gianni C. Update on antifungal therapy with Terbinafine. Giornale Italiano di Dermatologia e Venereologia 2010; 145: 415-24. http://ukpmc.ac.uk/abstract/MED/20461049
Newland JG, Abdel-Rahman SM. Update on Terbinafine with a focus on dermatophytoses. Clin Cosmetic Investig Dermatol 2009; 2: 49-63. http://www.ncbi.nlm.nih.gov/pmc/ articles/PMC3047923/ DOI: https://doi.org/10.2147/CCID.S3690
Zanini M. Surgical treatment of onychomycosis. Medicina Cutanea Ibero-Latino-Americana 2009; 37: 67-8. http://www.medcutan-ila.org/articulos/2009/1/pdf/mc371l.pdf
Lu SE, Lin Y, Shih WC. Analyzing Excessive No Changes in Clinical Trials with Clustered Data. Biometrics 2004; 60: 257-67. http://www.ncbi.nlm.nih.gov/pubmed/15032797 http://dx.doi.org/10.1111/j.0006-341X.2004.00155.x DOI: https://doi.org/10.1111/j.0006-341X.2004.00155.x
Olsen MK, Schafer JL. A Two-Part Random-Effects Model for Semicontinuous Longitudinal Data. J Am Stat Assoc 2011; 96: 730-45. http://www.tandfonline.com/doi/abs/10.1198/ 016214501753168389 http://dx.doi.org/10.1198/016214501753168389 DOI: https://doi.org/10.1198/016214501753168389
Tooze JA, Grunwald GK, Jones RH. Analysis of repeated measures data with clumping at zero. U.S. National Library of Medicine 2002. doi: 10.1191/0962280202sm291ra DOI: https://doi.org/10.1191/0962280202sm291ra
Anderson CJ, Verkuilen J, Johnson T. Applied generalized linear mixed models: continuous and discrete data. New York: Springer 2010.
Hasan MT, Sneddon G, Ma R. Pattern-mixture zero-inflated mixed models for longitudinal unbalanced count data with excessive zeros. Biometr J 2009; 51: 946-60. http://dx.doi.org/10.1002/bimj.200900093 DOI: https://doi.org/10.1002/bimj.200900093
Breslow NE, Clayton DG. Approximate inference in generalized linear mixed models. J of the Amer Stat Assoc 1993; 88: 9-25. http://dx.doi.org/10.2307/2290687 DOI: https://doi.org/10.1080/01621459.1993.10594284
Verbeke G, Molenberghs G. Linear mixed models for longitudinal data. New York: Springer-Verlag 2000. DOI: https://doi.org/10.1007/978-1-4419-0300-6
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2012 Rajibul Islam Mian, M. Tariqul Hasan
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
Policy for Journals/Articles with Open Access
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are permitted and encouraged to post links to their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work
Policy for Journals / Manuscript with Paid Access
Authors who publish with this journal agree to the following terms:
- Publisher retain copyright .
- Authors are permitted and encouraged to post links to their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work .