Heat Maps: A Technique for Classifying and Analyzing Drinking Behavior

Abstract

Heat maps are presented here as an innovative technique for evaluating longitudinal drinking outcomes. The Life Transitions Study followed alcohol dependent individuals for 2.5 years during 2004–2009 in a Midwestern city (N = 364). The TimeLine Follow-Back obtained drinking information. Heat map results were compared with those obtained using growth mixture modeling. Heat map classes differed significantly on baseline clinical and demographic indicators. Data were gathered with support from NIAAA R01AA014442.

RÉSUMÉ

Cartes Thermiques: Une Technique de Classification et d’Analyse des Comportements face à l’Alcool

Des cartes thermiques sont présentées ici en tant que technique innovante pour l’évaluation longitudinale des résultats dans le domaine de la consommation de boissons alcoolisées. L’étude “Transition de vie” a suivi pendant 2.5 ans des individus dépendants à l’alcool au cours de la période 2004–2009 dans une ville du Midwest (N = 364). Des entrevues de suivi linéaire (TimeLine Follow-Back) ont ensuite permis l’obtention d’informations sur leur consommation d’alcool. Les résultats, présentés sous forme de cartes thermiques, ont été comparés à ceux obtenus en utilisant des modèles mixtes de croissance. Les classes des cartes thermiques diffèrent significativement en termes de niveaux de référence cliniques et démographiques. Les données ont été recueillies avec le soutien de NIAAA R01AA014442.

RESUMEN

Mapas de calor: Una técnica para clasificar y analizar longitudinalmente el consumo de alcohol

Mapas de calor son presentados aquí como una técnica innovadora para la evaluación de longitudinal del consumo de alcohol. El Estudio sobre Transiciones en la Vida siguió personas con dependencia de alcohol por 2.5 anos durante 2004–2009 en una cuidad del Medio Oeste del EEUU (N = 364). La información sobre hábitos de consumo se obtuvo usando el método de registro en calendario de eventos. Los resultados de los mapas de calor fueron comparados con los resultados obtenidos usando modelos mixtos de análisis de crecimiento. Las clases de los mapas de calor diferían significativamente en los primeros indicadores clínicos del pre test y en características demográficas de los participantes. Los datos fueron conseguidas con el financiamiento de NIAAA R01AA014442.

Keywords :

• heat maps
• longitudinal drinking data
• drinking outcomes
• alcoholism
• longitudinal research
• data visualization
• drinking patterns

THE AUTHORS

Dr. Amy R. Krentzman joined the University of Michigan Addiction Research Center as a Postdoctoral Research Fellow in 2009 after completing her Ph.D. at the Mandel School of Applied Social Sciences at Case Western Reserve University. Her research focuses on addiction and pathways to recovery as well as patterns of longitudinal data in alcoholism research.

Dr. Elizabeth A. R. Robinson is a Research Assistant Professor, funded by the NIAAA, Metanexus Institute and Fetzer Institute. Her recent work has focused on recovery from alcoholism and the role of spiritual and religious change. She is also interested in the potential usefulness of mindfulness-based strategies as an adjunct to treatment and/or AA involvement. Dr. Robinson's Ph.D. is in psychology and social work from the University of Michigan, as is her MSW and MPH.

Dr. Jennifer M. Jester joined the University of Michigan Addiction Research Center in 2000. Her Ph.D. is in biomedical engineering from the Johns Hopkins University. Her research focuses on longitudinal studies of children and adults in alcoholic families. She uses multiple statistical techniques such as growth mixture modeling to look at the effects of behavioral traits and environmental conditions on children's outcomes. She is also interested in interventions in these high-risk families to improve parenting and ameliorate ill effects for the children.

Dr. Brian E. Perron joined the faculty at the University of Michigan, School of Social Work after completing his Ph.D. at Washington University in 2007. Previously, he worked as a clinical social worker in community mental health, providing services to persons with serious mental illnesses and substance use disorders. His research focuses on issues related to the quality of care for persons with mental illnesses and substance use disorders. He is involved in a variety of research activities, including analysis of nationally representative data and clinic-based surveys and collaborating on field-based interventions. Dr. Perron is also interested in innovative research methodologies and provides statistical consultation for a number of projects.

GLOSSARY

• Heat Map: A heat map is a method of visualizing two-dimensional data where the hue or intensity of color varies according to a given value criterion.

• Clustered Heat Map: In a clustered heat map, similar color patterns are grouped together to reveal larger structural properties of the data.

• Growth Mixture Modeling (GMM): Growth mixture modeling is a procedure that examines cases in a given data set and identifies subpopulations that are referred to as “classes.” Classes are made up of cases that are similar to one another in either longitudinal change or differences in patterns of change.

• Latent Class Growth Analysis (LCGA): A form of growth mixture modeling in which the variance of parameters within each class is set to zero. This approach is used when the relationship between the independent and dependent variables differs in both strength and direction.