Examining Additive Interaction of Risk Factors on Binary Dependent Variables to Address Racial, Social and Political Justice in Cross-Cultural Social Work Research

The events of the past year have highlighted that there continues to be a need to bend the arc of the moral universe toward racial, social and political justice. Now more than ever, there is momentum to take political action, and social workers are poised to advance health equity as leaders in social justice and health care. To do so, we must be able to demonstrate the evidence of the compounding effects of social factors and the environment that can lead to disadvantages in the lifecourse, and propose solutions that can have a community-level impact.

Most social work students and researchers are familiar with the use of interaction or moderation effects when examining risk factors to health outcomes in research. However, the approach that is used to examine interaction effects is primarily in multiplication interaction analysis and not additive interaction analysis. Multiplicative and additive are two types of interaction analyses, which apply fundamentally different approaches in estimation. For example, social work research can examine the multiplicative interaction of historical trauma through race as a modifier variable on cardiovascular health outcomes. Social work researchers can use additive interaction to identify the cumulative impact of an intervention on a specific outcome, such as mindfulness modalities when combined with chronic disease management and/or culturally-informed therapeutic approaches with attention to how clients can respond to political injustice.

This workshop aims to introduce the application of additive interaction analysis in social work research. In multiplicative interaction analysis, two variables are multiplied to estimate their interaction effect on a particular dependent variable. In an additive or biological interaction analysis, the effects of two or more variables are added or combined to estimate their interaction. In this workshop, we will look at the additive interaction effects between diabetes (no/ yes) and arthritis (no/ yes) on depression (no/ yes) for different race groups using the 2017 BRFSS data set. We will illustrate and explain key measures of additive interaction including the relative excess risk due to interaction (RERI), the proportion attributable to interaction (AP), and the synergy index (S). More specifically, we will demonstrate the computation process manually and then use Stata coding language to estimate these indices using the following equations:

Measure of “relative excess risk due to interaction (RERI):

RERI=(OR11)-(OR10)-(OR01)+1

Measure of the Proportion of Outcome Attributable to Interaction (AP). The AP index should differ than 0 when there is an additive interaction:

AP= ((OR11)-(OR10)-(OR01)+(OR00))/(OR11)

Measure of Interaction Synergy. This synergy index should differ than 1 when there is an additive interaction effect between to independent variables:

S=((OR11)-(OR00))/((OR10-(OR00)+(OR01-(OR00))

Participants will learn through examples provided through the workshop how to handle conditions where multiplicative and additive interactions can simultaneously be modeled, compared to when only one type of interaction if observed. The use of additive interaction can provide important insights for studying the impact of comorbidities in two or more risk factors on health and mental health outcomes for marginalized populations in the context of health equity and cross-cultural research.