Computerized counseling reduces HIV-1 viral load and sexual transmission risk: findings from a randomized controlled trial

Abstract

Objective: Evaluate a computerized intervention supporting antiretroviral therapy (ART) adherence and HIV transmission prevention.

Design: Longitudinal randomized controlled trial.

Settings: An academic HIV clinic and a community-based organization in Seattle.

Subjects: In a total of 240 HIV-positive adults on ART, 209 completed 9-month follow-up (87% retention).

Intervention: Randomization to computerized counseling or assessment only, 4 sessions over 9 months.

Main outcome measures: HIV-1 viral suppression, and self-reported ART adherence and transmission risks, compared using generalized estimating equations.

Results: Overall, intervention participants had reduced viral load: mean 0.17 log10 decline, versus 0.13 increase in controls, P = 0.053, and significant difference in ART adherence baseline to 9 months (P = 0.046). Their sexual transmission risk behaviors decreased (odds ratio = 0.55, P = 0.020), a reduction not seen among controls (odds ratio = 1.1, P = 0.664), and a significant difference in change (P = 0.040). Intervention effect was driven by those most in need; among those with detectable virus at baseline (>30 copies/mL, n = 89), intervention effect was mean 0.60 log10 viral load decline versus 0.15 increase in controls, P = 0.034. ART adherence at the final follow-up was 13 points higher among intervention participants versus controls, P = 0.038.

Conclusions: Computerized counseling is promising for integrated ART adherence and safer sex, especially for individuals with problems in these areas. This is the first intervention to report improved ART adherence, viral suppression, and reduced secondary sexual transmission risk behavior.

Figure 1

CARE+ Computerized Counseling Intervention Content By Session

Figure 2

CARE+ Computerized Counseling Intervention Trial, Four Sessions Over 9 Months HIV denotes human immunodeficiency virus-1.

Figure 3. Adjusted Means* by Time and Treatment Condition

* NOTE: For these independent 95% confidence intervals, non-overlap indicates p < .05, but overlap does not reliably indicate that p > .05 (Cumming, 2009). See Fig. 4 for contrasts. Detectable BL: had detectable HIV1 viral load at first baseline (BL) session Full sample: All participants regardless of viral load status at baseline

Figure 4

Comparison of Viral Load, ART Adherence, and Unprotected Sex for Each Follow-up Point and Changes within each Treatment Condition.