Information on Roundtables

Breakfast Roundtables
This year APS will offer Breakfast Roundtable discussions on Thursday, Friday and Saturday in two topic areas: Statistics and Best Practices.







Analysis of interactions in psychosomatic medicine: Do you really want the leftovers?
Julian F. Thayer
The analysis of interactions has remained a controversial issue in biobehavioral research. Interpretations based upon examination of cell means are problematic as they contain both main effect and interaction variance. The use of contrasts to examine patterns of cell means may represent an alternative approach. These issues are particularly important for psychosomatic research as the meaning of residual (interaction) variance can be nonsensical in typical research designs. In this discussion the many issues related to the analysis of interactions will be presented. The exact nature of interactions as residuals will be discussed. In addition, the use of contrasts to test patterns of cell means will be presented. Researchers will be guided on how to choose the correct analysis to answer their particular research question. Armed with these tools researchers should be able to construct more precise hypotheses and perform more accurate analyses of their data.

Confounding and Mediation
Michael Babyak
Psychosomatic researchers often rely on regression-type modeling to test hypotheses about relations among variables. In endeavoring to do so, the analyst is often confronted by significant challenges, not the least of which is how to select variables to include in these models. First and foremost, the choice of variables should be guided by the putative role that each plays in the causal system that the regression model attempts to represent. In this roundtable we will discuss the meaning of mediation and confounding in the context of causal processes, and how these meanings can be applied to variable selection. Specific tests for mediation also will be presented. Finally, we will consider how these concepts must be fitted into the context of model assumptions, sample size, power, and parameter stability in order to produce replicable models. The roundtable will be presented in lecture format with interaction and discussion encouraged.

What can latent variable analysis do for psychosomatic medicine research?
Maria Llabre
Latent variables are derived from the covariation among observed variables and are free from measurement error. When working with error-free variables researchers improve their ability to detect true effects or true associations increasing their power. In longitudinal data analysis, the measurement of change over time may be assessed with the use of latent variables that capture true change. Within a latent variable framework, parameters of change may be used as predictors or mediators of subsequent markers of disease. Latent variables are also useful in dealing with specific missing data problems. Thus, investing in latent variables offers advantages with respect to power, missing data problems, and complex models of change. During this breakfast roundtable, participants will be introduced to the latent variable framework through three specific examples from psychosomatic medicine research that illustrate its advantages.

Best Practices Roundtables
Treatment of Depression in Patients with Coronary Artery Disease and Congestive Heart Failure
Christoph Herrmann-Lingen, MD and Peter A. Shapiro, MD

Coronary artery disease and congestive heart failure are leading causes of disability and death in the developed world. Innumerable studies have demonstrated the high prevalence of depression in patients with these forms of heart disease, and the strong link between depression in heart disease patients and poor medical and quality of life outcomes. How should depression be treated in these patients, and with what expectations? We will review issues and current evidence regarding efficacy, side effects, and general health effects of psychotherapy and medication treatments for depression in cardiac patients and encourage discussion of clinical case material.

Best Practices for the Psychological Evaluation and Treatment of Adult Oncology Inpatients
Deidre B. Pereira, PhD and Karen Weihs, MD

Over the past 30 years, the sub-specialty of psycho-oncology has emerged as an integral component of comprehensive oncology research and treatment. The literature is replete with investigations examining relations among psychosocial factors, quality of life, intermediate endpoint biomarkers of health, and clinical outcomes in individuals with cancer. While the field of psycho-oncology has burgeoned in the research setting, it has perhaps more quietly established its role in the clinical setting. Increasingly, medical centers have staff designated to deal specifically with the psychosocial needs of individuals with cancer, their family members, and the healthcare providers that deliver medical services to them. Within the oncology inpatient setting, supportive services have traditionally been provided by social workers and nurses. While these services are vital, the discipline of clinical health psychology can be an important adjunct to these services, given its specialized training in (a) the role of biopsychosocial factors in the etiology, maintenance, exacerbation, and progression of disease, (b) the assessment and interpretation of empirically-validated psychological measures, and (c) the psychological treatment of diverse forms and degrees of psychopathology among the medically ill. In spite of this, little literature has specifically addressed the role of clinical health psychology in the context of the inpatient oncology setting. The goals of this Best Practices Roundtable are (a) to identify the unique psychosocial experiences of adults in the inpatient oncology setting, (b) to discuss the psychological evaluation and treatment of common mental health conditions among oncology inpatients, and (c) to explore the unique challenges associated with delivering psychological services to individuals in this setting.