Volume 19, Number 95 (11-2011)                   daneshvarmed 2011, 19(95): 9-16 | Back to browse issues page


XML Persian Abstract Print


Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:

Ghaheri A, Faghihzadeh S, Ghazanfari T, Zayeri F, Soroush M R. Fitting logistic model to some quantitative and qualitative variables to discriminate between mustard-exposed and non-exposed individuals. daneshvarmed. 2011; 19 (95) :9-16
URL: http://daneshvarmed.shahed.ac.ir/article-1-367-en.html

Professor of Biostatistics Department of Biostatistics, Tarbiat Modares University
Abstract:   (6711 Views)

Background and Objective: Sulfur mustard (SM) is a chemical warfare agent used by the Iraqi forces during the Iraq-Iran war. So far some statistical methods have only been applied to discriminate between exposed and non-exposed Sardasht inhabitants, using immunological variables. This paper focuses on allocating objects to exposed and non-exposed groups and choosing the best model, using a variety of qualitative and quantitative variables available.

 

Materials and Methods: The participants had been recruited from Sardasht (exposed) and Rabat (non-exposed) residents. In general, a sample of 372 and 128 in case and control groups were assessed in this study. To allocate the persons to the exposed group, a logistic regression model was used. Models with main effects of the independent variables were compared using sensitivity, specificity and the area under their ROC curves.

 

Results: Sensitivities of quantitative, qualitative and mixed models are as follows: 0.953, 0.948 and 0.952. Specificities of these models are 0.225, 0.524 and 0.542 respectively. Variables included in the mixed model are IL-18BP (p<0.001), sP-selectin (p=.003), itching (p<0.001), plaque (p=0.017), chronic cough (p=0.029), sputum (p=0.017), hyperpigmentation (p=0.026) and bulbar conjunctive (p=0.025).

 

Conclusion: According to the models' sensitivities, specificities, R2 and the area under their ROC curves on the one hand and the simplicity of qualitative variables usage on the other hand, the qualitative model could be preferred to as the mixed one.

Full-Text [PDF 196 kb]   (818 Downloads)    
Type of Study: Research |

Add your comments about this article : Your username or email:
Write the security code in the box

Send email to the article author


© 2015 All Rights Reserved | Daneshvar

Designed & Developed by : Yektaweb