Nonadherence to medical appointments is associated with increased plasma HIV RNA and decreased CD4 cell counts in a community-based HIV primary care clinic

Abstract

This study examined the association of appointment nonadherence to markers of disease severity using one year of demographic and health information on 995 individuals with HIV in primary care at an urban community health centre. At the latest visit, 106 of 946 valid cases (11.2%) had a CD4 less than or equal to 200, and 454 of 936 valid cases (48.5%) had detectable plasma HIV RNA (greater than 50 copies/ml). Using logistic regression, appointment nonadherence (number of missed appointments) was a significant predictor (p<.03) of having an AIDS-defining CD4 count over and above the effects of number of kept appointments (p<.0001), and whether or not the patient was taking HAART (p<.002). Appointment nonadherence was also a significant predictor (p<.05) of having a detectable viral load over and above the effects of number of kept appointments (p<.003), HAART (p<.0001) and age (p<.004). Looking only at individuals with a detectable viral load at the earliest visit, the only significant unique predictor of improvement to an undetectable viral load at the latest visit was being on HAART (p<.05). Looking at those only with an undetectable viral load at the earliest visit, the only predictor of declining to a detectable viral load was number of kept appointments (p < 003), and being on HAART (p<.05).

Acknowledgments

Funding for this project was partially supported by grant MH 66660 from the National Institute of Mental Health and grant 5P30AI027763 from The National Institute of Allergy and Infectious Diseases.

Notes

1. We reran these analyses substituting number of months with HIV for age. Number of months with HIV was available for 757 individuals in the sample. The pattern of results for predictors of biomedical variables was the same except that number of months was significant for having an AIDS-defining CD4 count at the latest visit, but was not significant for having a detectable viral load. When entering both age and number of months and age, in general the two predictors cancelled each other out due to colinearity.