A randomized, controlled trial of a patient-centered disclosure counseling intervention for

Kenyan children living with HIV.Rachel C. Vreeman, MD, MS; Winstone M. Nyandiko, MBChB, MMED,

MPH; Irene Marete, MBChB, MMED; Ann Mwangi, PhD; Carole I. McAteer, MS; Alfred Keter, MS; Michael L. Scanlon, MA, MPH; Samuel O.

Ayaya, MBChB, MMED; Josephine Aluoch, MA; Joseph Hogan, ScD

Background

• For children living with HIV, learning about their HIV status (“disclosure”) is a critical process within the transition to adulthood

• Caregivers of perinatally HIV-infected children frequently worry about the impact of disclosure, particularly subsequent stigma

• Also report delayed disclosure can hurt medication adherence

• Social environment, perceived HIV-related stigma, and community beliefs shape caregivers’ disclosure decisions

• Prior work in western Kenya: only 55% disclosed by age 14 years in a sample of 748 children and youth across 4 large clinics

Study ObjectiveTo evaluate the impact of a patient-centered, culturally-appropriate and age-appropriate disclosure counseling intervention in a cluster-randomized trial among Kenyan children and their caregivers

1R01MH099747-01 (Vreeman)

Study Design• Prospective, cluster-randomized trial

• Enrolled child-caregiver dyads (children ages 10-14) attending eight clinics within the AMPATH HIV treatment program in Kenya

• Clinics were randomized to intervention or control

• Children and caregivers were followed for 24 months, with assessments every six months (0, 6, 12, 18, and 24 months)

• Curriculum for disclosure and adherence counseling: video-taped narratives; animated, tablet-based educational modules; print resources

• Counseling: Family and one-on-one

• Facilitated peer support groups

hiv-films.org

Primary Outcome: Disclosure Status

• Treated as a time-to-event outcome

Analyses• Disclosure treated as time-to-event outcome, measured on discrete time scale

• Discrete-time random-effects hazard models

• Results are summarized using two measures: • Time-specific hazard ratio: compares probability of new disclosure at each time point, where

denominator is those who have not yet disclosed• Time-specific prevalence of disclosure: compares cumulative proportion disclosed at each

time point post-baseline

• Clinical, mental, and behavioral outcomes: Independent sample t-test to compare normally distributed continuous variables, two-sample Wilcoxon rank-sum test to compare non-normally distributed variables. Comparison of proportions done with Wald test.

• Effect of intervention on depression (score on PHQ-9), and on emotional and behavioral symptoms (SDQ) assessed using a mixed effects ordinal logistic regression model.

• Included clinic-specific and subject-specific random effects with participants nested within clinics. The treatment arm and time variable, as well as interaction of the two, were included as the main effects.

Results: Participants• 285 children and their caregivers • Mean age 12.3 years, 52% female• Average time-on-treatment: 4.4 years

• 95% on first-line ART

• At baseline, 32% of children reported knowing their HIV status already (no difference between control and intervention groups)

• Children from control clinics were significantly more likely to have been orphaned (p=0.011), but otherwise demographic and clinical characteristics between groups not statistically different

• Majority of caregiver participants were biological mother of child (54%), but also many aunt/uncle caregivers (19%) and biological fathers (17%)

• During 24 months of follow-up, 25 patients withdrew from study and 7 patients died, with no significant differences by treatment group

Results: Impact on Disclosure

• Disclosures in both control and intervention arms increased over follow-up, but the intervention arm had significantly more disclosures

• Using child-reported disclosure, prevalence of disclosure increased significantly between the baseline and 24 months of follow-up from 29.2% to 58.5% in the control arm and from 33.2% to 74.0% in the intervention arm

• This was a significant difference in disclosure prevalence for the intervention group at 24 months (difference of 15.5%, 95% confidence interval: 3.7, 27.3)

• Both more disclosures and early disclosures for intervention group, with largest increase at 6 months

Results: Caregiver vs. Child Reports

• Caregiver and child reports of the child’s disclosure status were somewhat inconsistent.

• At baseline, 19% of child-caregiver dyads gave different answers within the dyad as to whether child knew his or her HIV status.

• In the majority of cases of disagreement (89%), caregiver reported that the child’s HIV status had been disclosed to child, while the child reported that they did not know their HIV status.

Results: Composite Measure of Disclosure

• Using a composite measure for positive disclosure (either caregiver or child reports disclosure), disclosure increased during study: • 47% to 85% in control arm

• 50% to 91% in the intervention arm

• Prevalence of disclosure was higher in intervention group at each time point, but these differences were only significant at the 6-month follow-up, when 70.0% of children in the intervention arm were disclosed after 6 months in the study, compared to 58.5% in the control arm (p=0.039).

Results: Mental and Behavioral Health

• Overall, there were not significant differences in mental and behavioral health outcomes at 24 months

• Trends suggested mental and behavioral distress increased at month 6 in the intervention group as disclosures increased, and then decreased compared to controls thereafter

Results: PHQ-9

• On the PHQ-9, children in intervention group had 2.1 times (95% CI 1.01, 4.25) odds of moving from a lower depression category to a higher (i.e., more severe) depression category compared to children in control group at month 6 (the same timepoint at which disclosures increased significantly)

• At months 12 and 18, children in intervention group had reduced odds of moving from a lower to higher depression category (though not statistically significant), and at month 24, the odds between intervention and controls were similar

Results: SDQ

• On the SDQ, children in the intervention group had 1.2 times (95% CI 0.55, 2.55) the odds of moving from a normal to borderline or borderline to abnormal category at 6 months

• At months 12, 18, and 24, children in the intervention group had trend towards reduced odds (ORs between 0.80 and 0.87), though these were also not statistically significant

Results: Viral Suppression

• Viral load measures drawn at 24-month final study visit

• 118/250 participants (47%) had detectable viral load (>40 copies/mL)

• In a comparison adjusted for CD4, individuals in the intervention group trended towards being less likely to have a detectable viral load (odds ratio = 0.80, 95% CI: 0.22-2.84) and more likely to achieve viral suppression (2.29, 95% CI: 0.89-5.39) but neither was statistically significant

Conclusions

• This study provides evidence for an effective, clinic-based intervention to increase disclosure of HIV status to children living with HIV

• Key components: Dedicated counselor available at clinic, supported by narrative-based counseling curriculum

• Making mental health support and peer support available throughout the disclosure process may be particularly important to navigate increased psychological distress immediately after disclosure and move towards resilience

Acknowledgements

This research was supported by a grant entitled “Patient-Centered Disclosure Intervention for HIV-Infected Children” (1R01MH099747-01) to Dr. Rachel Vreeman by the National Institute for Mental Health, Bethesda, Maryland, USA.