Behavioral Responses to Contextual Interventions: Evaluating In-Silico Public Health Interventions

• In Event: 40221 - Section on Mathematical Sociology Flash Talks
  In Refereed Roundtable (90 minute): Table 2: Behavioral Responses to Contextual Interventions: Evaluating In-Silico Public Health Interventions

Abstract

Computational modeling techniques and the capacity of computers to
simulate an intervention open up new avenues for assessing the
plausibility of proposed public health interventions (PHIs),
particularly in situations where individual behavior depends upon
complex contextual factors.

A PHI might be thought of as a set of actions intended to result in an
improvement in a health outcome of interest. Some examples of previous
PHIs include campaigns encouraging childhood vaccination, the addition
of fluoride to drinking water, and attempts to prevent adolescent from
taking up smoking. These interventions are frequently complicated to
implement, with context dependent effectiveness. They can also be slow
to gain approval for implementation, as well as expensive to perform.

While it is critical that PHIs continue to be empirically assessed for
effectiveness, an initial evaluation based on a simulated intervention
performed using well-validated models provides a means of determining
an intervention's plausible effectiveness. As these *in silico*
interventions are all but costless to perform, many more may be
considered than could be piloted and empirically tested. A failure of
a simulated intervention to produce a reasonable improvement in a
health outcome of interest would enable researchers to remove the
proposed intervention from consideration without needing to expend
resources by funding a pilot study. The saved resources could then be
used to pilot interventions that were found to be successful.

This paper examines and evaluates a proposed PHI (changing adolescent
smoking behavior by increasing the level of parental support received)
via a network based simulation. After briefly describing the
literature relating parental support to adolescent smoking behavior,
as well as the literature on assessing simulation validity, the paper
then describes the behavioral changes observed when a published
Stochastic Actor-Based Model is given an intervened on population and
discusses the results of the simulated intervention.

Authors

• Alexander Murray-Watters, University of California-Irvine

• Cheng Wang, Wayne State University

• John R. Hipp, University of California-Irvine

• Cynthia Lakon, University of California - Irvine

• Carter T. Butts, University of California-Irvine