Latent Class Analysis of Sexual Partnerships and Substance Use Across Generations

Abstract

Sexually transmitted infections (STIs) remain an important public health concern for people of all age groups, with older age groups experiencing a notable increase in STI burden. Historically, most research into STI risk behaviors has focused on adolescents and young adults, leaving a paucity of research on the ways STI risk factors change over the life course. Additionally, age and cohort trends in STI risk factors can be challenging to investigate with standard statistical tools as they can be collinear and are subject to sociocultural and generational influences. To help address these issues, we used multi-group latent class analysis to identify and compare risk behavior profiles defined by responses to three sexual activity and three substance use variables, across and within four age groups. We identified six behavior profiles in the unstratified dataset and five behavior profiles in each of the four age stratified groups. The five behavior profiles identified in each of the age categories appear to reflect a similar set of five underlying profile “archetypes,” with the exact composition of each age category’s five profiles varying in the magnitude that specific behaviors are endorsed. Interestingly, despite the similarity of profiles across the four age groups, analyses indicate that the experience of belonging to any one of these five archetypes differs by age group. This variance is likely due group specific age, period, and cohort effects, and may indicate that, when estimating one’s STI risk, it is better to compare them to their peers than to the population as a whole.