Geospatial Frontiers in Health and Social Environments April 27-28, 2012

Modeling Contextual Neighborhood andModeling Contextual Neighborhood and Social Network Effects on Human

Behavior

Jeremy Mennis, Ph.D. Department of Geography and Urban Studies

Temple University

Context Matters for Understanding Addiction

“Given the complex interactions of biological, social, environmental, and developmental factors that underlie drug abuse and addiction, NIDA acknowledges the need to take a "whole systems" approach to treating thi di ” (NIDA St t i Pl 2010)this disease. .” (NIDA Strategic Plan, 2010)

But the prevailing paradigm is generally NOT geographic in nature: Drug ddi ti h i b i di ithi bi h i l d laddiction as a chronic brain disease within a biopsychosocial model.

…and Treatment

A New Geographic Emphasis in Behavioral Research?

Research on Substance Use: The Philadelphia Adolescent Lifestyle Study p y y

(NIDA, PI Mason, Co-PIs Mennis, Coatsworth, Valente, Lawrence, Zelenak)

Adolescents  and their social networks are constituted by the environment in which they operate:

which in turn influences their social network quality

RiskyRisky

which in turn influences their social network quality, sense of self, & social practices:

and ultimately their health behaviors

ProtectiveProtective

IdentityIdentity

RolesRoles

and ultimately their health behaviors.

Healthful BehaviorHealthful Behavior High Risk BehaviorHigh Risk Behavior

Research on Adolescent Substance Use

Sample:

Survey of 301 adolescent primary care patients at a Philadelphia public healthPhiladelphia public health care center.

Primarily African-American youth living in low to mid-income neighborhoods.

Features in the Environment

U.S. Census Tract-Level VariablesVariable Definition

Concentrated Disadvantage% High School % of the total population over age 25 with a high school diploma% Employed % of the total civilian population over age 16 who are employed% Public Assistance % of the total population receiving public assistance income% Female Headed % of households female headed with children under 16% Vacancy % of housing units vacantEthnic Diversityy% Foreign % of the total population who were born outside the U.S.% Linguistic Isolation % of the households with no one over five years old who speaks

EnglishRacial Diversity Index of racial diversityy yHispanic% Hispanic % of the total population that self-identifies as HispanicCrimeVi l t C i R t N b f i l t i it 2000 2002Violent Crime Rate Number of violent crimes per capita, 2000-2002Property Crime Rate Number of property crimes per capita, 2000-2002Residential Mobility% Renter % of the total households that are renter occupied% Same House % of the population over five years old who live in the same house

they occupied 5 years ago

Some Findings

1. Perceptions of safe and risky places differ between substance users and non-users, where non-users are more sensitive to their environment.,

2. The contextual characteristics of perceived risky places are particularly influential on substance use.

a. Perceived risk typically revolves around exposure to unstructured social interaction, as in malls and recreation centers, e.g. “There’s lots of dangerous people around and it’s a wide-open place.”

3. Proximity to bars, concentrated disadvantage, and other theorized negative influences at the perceived risky place enhances substance use.

4. Accessibility to certain pro-social activities that we theorized would y psuppress substance use (e.g. recreation centers) are actually associated with enhanced substance use.

5. Contextual effects are moderated by age and gender.y g g

Moderation of Contextual Effects by Age

Moderation of Concentrated Disadvantage by Age Moderation of Social Network Quality by Age

20

19

= age of adolescent

16

20

19

18

17

= age of adolescent

16

18

17

16

15

16

15

14

1315

14

13

Context: The Take-Home Messages

1. Geographic and social contexts are integral to understanding behavior and designing effective substance abuse prevention and treatment programsdesigning effective substance abuse prevention and treatment programs.

2. To capture geographic and social contexts, one must focus not only on the home location, but also on the activity space of individuals.

3. Geographic and social contexts are not independent but, rather, intertwined and interrelated forces on behavior.

4 The role of contextual mechanisms on behavior is manifested through the4. The role of contextual mechanisms on behavior is manifested through the perceptions and interactions of the individual with their social and geographic environments.

Challenges

1 H d ll t d t i t t d i l t k ti it d1. How do we collect data on integrated social networks, activity spaces, and perceptions and interactions of individuals with their geographic and social contexts?

2 H d d l hi ti t d t l d t ti ti l d l2. How do we develop more sophisticated conceptual and statistical models that will enable us to more fully test for cause and effect, processes of mediation and moderation, and mechanisms of selection versus influence?

3 Since m ch of these data ma be m ltidimensional and comple and ma3. Since much of these data may be multidimensional and complex, and may take the form of imagery, narrative text, and other non-quantitative formats, how can these data be integrated and analyzed effectively?

Social-spatial Risk and Protective Mechanisms in Urban Adolescent Substance Use (NIDA; PI Mason, Co-PIs Mennis, Light, Rusby, Westling, Crewe)

GPS

Space-time path (Kwan, 2008)

RSIENA for dynamic networks (Ripley, Snijders, Preciado, 2012)

Qualitative-Quantitative Data

EMA

Qualitative Quantitative Data Explorer

Do we need a new term?

Geobiopsychosocial?

Biogeopsychosocial?

Biopsychogeosocial?

Acknowledgments

Collaborators: M. Mason, P. Zelenak, D. Coatsworth, T. Valente, F. Lawrence, G. Stahler, D.Collaborators: M. Mason, P. Zelenak, D. Coatsworth, T. Valente, F. Lawrence, G. Stahler, D. Baron, S. Mazzella, R. Cotlar, Y. Cao, J. Light, J. Rusby, E. Westling, S. Crewe

Funding: NIH/NIDA, Temple University

S l t d P bli tiSelected Publications:

Mennis, J., Cao, Y., and Mason, M.J., in press. Qualitative GIS and the visualization of narrative activity space data. International Journal of Geographical Information Science.

2012 S fMennis, J., and Mason, M.J., 2012. Social and geographic contexts of adolescent substance use: the moderating effects of age and gender. Social Networks, 34(1): 150-157.

Mennis, J. and Mason, M.J., 2011. People, places, and adolescent substance use: Integrating activity space and social network data for analyzing health behavior. Annals of the Association of American Geographers, 101(2): 272-291.

Stahler, G., Mennis, J., Cotlar, R, and Baron, D., 2009. The influence of the neighborhood environment on treatment continuity and rehospitalization for dually diagnosed patients discharged from acute inpatient care. The American Journal of Psychiatry, 166(11): 1258-g p y y ( )1268.