Methods: We used retrospective cohort analysis to assess GC-CT testing in patients establishing care at an urban HIV clinic in San Francisco between February, 2013 – December, 2014 who had at least one subsequent primary care visit. The predictor variable of homelessness was defined as having stayed outdoors, in shelters, in vehicles, or in places not made for habitation during the last year and was evaluated during social work intake. The primary outcome was having GC-CT testing at a primary care visit. The probability of GC-CT testing at a primary care visit was calculated using logistic regression modeling with random effects to handle intra-subject correlations for repeated measurements. We adjusted for age, race/ethnicity, methamphetamine use, alcohol use, sexual orientation, prior positive GC-CT test, recent GC-CT test and visit frequency. Comparison of GC-CT test results by housing status was made using chi2.
Results: Of 323 patients, mean age was 41.4 years, 91% were male, 50% non-Latino White, and 45% were homeless. 204 patients (63%) had GC-CT testing done, and 138 (37%) had testing at primary care visits. A median visit frequency for housed patients was 1.42 visits/180 days v 1.61 for homeless patients (p-value 0.36). Average follow-up time was 472 days (sd 292 days). Homelessness was associated with lower odds of GC-CT testing during follow-up (aOR 0.60; 95% CI 0.40, 0.96 p-value 0.032). 28% of first GC-CT tests were positive in homeless v. 12% in housed patients (p-value 0.025).
Conclusion: Homeless patients had a higher rate of positive GC-CT on first test but 40% lower odds of having GC-CT testing at primary care visits despite controlling for visit frequency, recent testing and history of positive GC-CT. Further evaluation of disparities in GC-CT testing for homeless patients is warranted.
A. A. Clemenzi-Allen,
E. Geng, None
K. Christopolous, None