Syndemics of psychosocial problems and HIV risk: A systematic review of empirical tests of the disease interaction concept

Highlights

• Most empirical studies on syndemics use the “count variable” approach.

• Few studies appropriately tested the concept of disease interaction.

• Few studies had sufficient detail to interpret interaction on an additive scale.

Abstract

In the theory of syndemics, diseases co-occur in particular temporal or geographical contexts due to harmful social conditions (disease concentration) and interact at the level of populations and individuals, with mutually enhancing deleterious consequences for health (disease interaction). This theory has widespread adherents in the field, but the extent to which there is empirical support for the concept of disease interaction remains unclear. In January 2015 we systematically searched 7 bibliographic databases and tracked citations to highly cited publications associated with the theory of syndemics. Of the 783 records, we ultimately included 34 published journal articles, 5 dissertations, and 1 conference abstract. Most studies were based on a cross-sectional design (32 [80%]), were conducted in the U.S. (32 [80%]), and focused on men who have sex with men (21 [53%]). The most frequently studied psychosocial problems were related to mental health (33 [83%]), substance abuse (36 [90%]), and violence (27 [68%]); while the most frequently studied outcome variables were HIV transmission risk behaviors (29 [73%]) or HIV infection (9 [23%]). To test the disease interaction concept, 11 (28%) studies used some variation of a product term, with less than half of these (5/11 [45%]) providing sufficient information to interpret interaction both on an additive and on a multiplicative scale. The most frequently used specification (31 [78%]) to test the disease interaction concept was the sum score corresponding to the total count of psychosocial problems. Although the count variable approach does not test hypotheses about interactions between psychosocial problems, these studies were much more likely than others (14/31 [45%] vs. 0/9 [0%]; χ2 = 6.25, P = 0.01) to incorporate language about “synergy” or “interaction” that was inconsistent with the statistical models used. Therefore, more evidence is needed to assess the extent to which diseases interact, either at the level of populations or individuals, to amplify HIV risk.

Introduction

The poor and underserved often face a complex constellation of health and social problems that conspire to undermine their wellbeing. The term “syndemic” was proposed by Singer (1994) to call attention to these issues, with a special focus on the frequently co-occurring problems of substance abuse, violence, and HIV (Singer, 1996, Singer, 2006). As originally described and summarized in several publications (Singer, 1994, Singer, 1996, Singer, 2006), two characteristics are central to his conceptualization of a syndemic. First, diseases co-occur in particular temporal or geographical contexts due to harmful social conditions (disease concentration). Second, diseases interact at the level of populations and individuals, with mutually enhancing deleterious consequences for health (disease interaction). As summarized by Singer and Clair (2003) in one of the more recent restatements of the theory, “a syndemic is a set of intertwined and mutually enhancing epidemics involving disease interactions at the biological level that develop and are sustained in a community/population because of harmful social conditions and injurious social connections” (p.429). In one of the first empirical studies in this literature, Stall et al. (2003) attempted to extend conceptual thinking about syndemics to understanding HIV risk among men who have sex with men in the U.S. While their analysis did not provide evidence of interacting epidemics, they did show that psychosocial problems were frequently co-occurring and that men with a greater number of psychosocial problems were more vulnerable to HIV.

The theory of syndemics is consistent with previously published theories that have been deployed to explain the existence of health disparities, including the social origins of greater health risks among the marginalized and dispossessed (Farmer, 1999, Farmer, 2004, Link and Phelan, 1995, Scheper-Hughes and Bourgois, 2004) and the role of historically ingrained forces in exerting conjoint influences on HIV risk (Farmer, 1996, Farmer et al., 1993). The harmful effects of co-occurring psychosocial problems have also been highlighted in other literature that have evolved largely in parallel to the literature on syndemics, including the concepts of multimorbidity (van den Akker et al., 1996, van den Akker et al., 1998) and dual diagnosis (Drake et al., 1991, Lehman et al., 1989). In this regard, the concept of disease concentration is empirically well established. However, the extent to which there is empirical support for disease interaction in HIV risk remains unclear. While the theory of syndemics would remain useful for its conceptualization of disease concentration even if the concept of disease interaction were to fail empirical testing, this is an important gap in the literature because of the overall programmatic and policy importance of the syndemics orientation. A decade ago, Gerberding (2005) described U.S. Centers for Disease Control and Prevention (CDC) efforts to transform health protection research with the aim of reducing health disparities. Although complex systems theories and a syndemics orientation were explicitly incorporated into her vision for the CDC's future, she also noted that “application of complex systems theories or syndemic science to health protection challenges is in its infancy” (p.1405). While public health researchers motivated by an interest in improving health and wellbeing in vulnerable populations have made important progress in utilizing and understanding the theory of syndemics, to date there has been no systematic summary of how the concept of disease interaction in syndemic theory has been tested in empirical work.