Latent class analysis of a syndemic of risk factors on HIV testing among black men

ABSTRACT

Syndemic methodology has been employed in several studies of HIV-related outcomes affecting Black men who have sex with men (BMSM) and rarely in Black heterosexual men. In contrast to the most common method for assessing syndemics, the use of a syndemic component index, latent class analysis can identify unique combinations of risk factors that may form a syndemic. Analyzing a primarily heterosexual sample of 1,786 Black men from the 2015 Behavioral Risk Factor Surveillance System (BRFSS), we used a 4 latent class model based on depression diagnosis, poverty, and healthcare access to predict ever having been HIV tested. Class 1 was characterized by low proportions of all the risk factors. Class 2 had relatively high healthcare barriers, being the most likely to not have a personal doctor (.8175) and the most likely to have no routine checkup in the past year (.6327) but had relatively low depression diagnosis and poverty. Class 3 had relatively high poverty (.8853), but generally low barriers to healthcare access. Class 4 was characterized by high proportions of all the risk factors. Using log-binomial regression models, there was a significantly lower prevalence of ever having been HIV tested among class 3 (PR = 0.69, 95% CI 0.49, 0.98) and class 4 (PR = 0.49, 95% CI 0.28, 0.84) compared to class 1. When adjusting for education, age, and marital status, the associations were attenuated but still significant for class 3 (aPR = 0.71, 95% CI 0.52, 0.96) and class 4 (aPR = 0.60, 95% CI 0.46, 0.78). Latent class analysis may better serve syndemic research aims in understanding HIV-related outcomes among high-risk populations. Future research using this method to evaluate HIV testing outcomes among BMSM is recommended.

KEYWORDS:

• LCA
• syndemic
• HIV
• testing
• black

Acknowledgements

The authors would like to acknowledge the support from the University of Maryland, College Park. The 2015 Behavioral Risk Factor Surveillance System (BRFSS) dataset was used for this research. This dataset can be found at the following link: https://www.cdc.gov/brfss/annual_data/annual_2015.html

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Rodman E. Turpin http://orcid.org/0000-0003-0457-5584

Natalie Slopen http://orcid.org/0000-0002-5644-6007

Shuo Chen http://orcid.org/0000-0002-7990-4947

Bradley Boekeloo http://orcid.org/0000-0002-4182-6613

Cher Dallal http://orcid.org/0000-0002-7946-3969

Typhanye Dyer http://orcid.org/0000-0002-8493-5772

Additional information

Funding

This study was supported by the School of Public Health’s Dean’s Fellowship at the University of Maryland, College Park.