A Bayesian approach to estimating the regression coefficients of a multinomial logit model

A Bayesian approach to estimation of the regression coefficients of a multinominal logit model with ordinal scale response categories is presented. A Monte Carlo method is used to construct the posterior distribution of the link function. The link function is treated as an arbitrary scalar function. Then the Gauss-Markov theorem is used to determine a function of the link which produces a random vector of coefficients. The posterior distribution of the random vector of coefficients is used to estimate the regression coefficients. The method described is referred to as a Bayesian generalized least square (BGLS) analysis. Two cases involving multinominal logit models are described. Case I involves a cumulative logit model and Case II involves a proportional-odds model. All inferences about the coefficients for both cases are described in terms of the posterior distribution of the regression coefficients. The results from the BGLS method are compared to maximum likelihood estimates of the regression coefficients. The BGLS method avoids the nonlinear problems encountered when estimating the regression coefficients of a generalized linear model. The method is not complex or computationally intensive. The BGLS method offers several advantages over Bayesian approaches. ^

Evaluating alternative therapies. Support/imagery and immune function in breast cancer: A pilot study

Cancer patients increasingly request alternative therapies such as imagery techniques and support groups. Although research suggests evidence of enhanced psychosocial functioning with supportive group therapy and enhanced immune function with imagery techniques, studies are anecdotal or limited to case studies or descriptive reports. The efficacy of these alternative therapies should be validated by randomized, controlled trials and the mechanisms of action mediating immune function and outcome examined.^ In a 12-month pilot study, we evaluate the feasibility of conducting a controlled study with clinical trial methodology to test the effects of imagery/relaxation and support on quality of life, emotional well-being, and immune function for women after breast cancer. Using a randomized pre-post test design with three intervention waves, we assigned women (n = 47) to either standard care (n = 15), standard care plus 6-weekly support sessions (n = 16) or imagery/relaxation sessions (n = 16).^ The primary aim of this pilot study is to determine the feasibility of conducting a clinical trial of alternative therapies in a clinical care setting. Secondary aims are to determine parameter estimates for the effects of the two treatment groups on quality of life, coping, social support, and immune function and describe methodology issues related to trials of alternative therapies.^ The research provides direction for future studies of alternative therapies by describing the recruitment, clinical trial experience, and related methodology issues. The study extends previous work by differentiating the effects of support group from mental imagery among outpatient groups who are homogeneous regarding cancer type and treatment stage. The study provides data for future longitudinal studies of disease progression by differentiating the effectiveness of interventions designed to enhance quality of life, coping, social support, and immune function and subsequently, alter the clinical course of disease. ^