Social Stability as a Consistent Measure of Social Context in a Low-Income Social Network

Abstract

Background: Increasingly, studies of factors influencing health consider the importance of social contexts in which people are living. The selection of indicators to represent this social context in health research can seem arbitrary. This study examines the potential of social stability as a useful construct to represent social context in these studies. Purpose: This study applied a previous definition of social stability (SS) in a new population and examined its relationship to syndemic risk behaviors (sexual, substance use, and violence). Aim 1 examined whether SS (as measured by German, 2009) can identify distinct subgroups while describing SS prevalence and patterns. Aim 2 assessed whether measures of perceived or historical stability are related to SS status, and whether they influence latent SS classes. Aim 3 explored how SS level and subgroup are associated with risk behaviors, while examining the co-occurrence of these syndemic behaviors. Methods: A secondary analysis of data of heterosexuals at high-risk of HIV infection from the Baltimore site of National HIV Behavioral Surveillance Study was conducted. Descriptive and latent class analyses (LCA) were applied to characterize the prevalence and patterns of SS and risk behaviors and to identify SS subgroups. Logistic and latent class regression were applied to model the relationships of SS to risk behaviors and demographic covariates. Results: SS was more prevalent than expected, and co-occurrence of SS indicators was common in this population. LCA showed evidence for 3 sub-classes: high stability, residential instability, and income, employment instability. Perception or history of stability did not contribute to identifying latent classes. Education was an influential covariate in LCA. Co-occurrence and significant associations among risk behaviors are also common in the population. Ordinal and latent measurements of SS reliably predicted individual and combinations of sexual-substance use- and-violence risks. Relationships vary with the method of SS specification with stronger magnitude of odds of risk associated with the latent approach. Conclusion: Social stability well represented certain aspects of the social context in a new study population and demonstrated an influential relationship with syndemic risks. Consistent measures of SS should be considered for application in research including social conditions and health.