Procedures

P219C was a longitudinal observational study of children and adolescents with PHIV exposure designed to examine long-term effects of exposure to antiretroviral (ARV) medications and complications of HIV infection. It was conducted from September 2000 to May 2007 at over 80 participating sites in the US and Puerto Rico. Informed consent and assent were obtained according to local institutional review board guidelines. Participants' medical records were abstracted at study entry and follow-up visits to obtain medical and treatment histories, including markers of immune functioning, neurologic and psychiatric diagnoses, and ARV and concomitant medications, including stimulants. Follow-up visits included standardized assessment of verbal and nonverbal cognitive abilities, parent or primary caregiver (caregiver) report of child behavior, and reports from caregivers and older children regarding the children's QOL and current demographic and medical information. Respondents completed questionnaires independently or as interviews if they had difficulty reading. Medication start and stop dates were recorded as reported by caregivers and older participants or as documented in medical records. The baseline measure was defined as the measure completed at the visit preceding or on the same day as the visit at which the participant's first prescription was recorded. The first available outcome measure after baseline was selected as the follow-up measure.

Age-appropriate Wechsler scales^24-29 were employed to assess verbal and nonverbal intellectual functioning. These measures were administered according to standardized procedures at baseline and every three years following enrollment. Visits were scheduled around the children's birthdays at ages 3, 6, 9, 12, and 15 years. The Wechsler tests have excellent psychometric properties and were updated as revised test versions were published. The Verbal Scale Intelligence Quotient (VIQ) and Performance Scale Intelligence Quotient (PIQ) were used in the analyses; these are age-normed standardized scores with mean (M) = 100 and standard deviation (SD) = 15. The Full Scale IQ, a summary score, was not used in order to examine potential differential effects of stimulant use on verbal and nonverbal intellectual abilities. Data were reviewed and results were excluded if considered invalid due to a child's sensory or physical impairment or insufficient proficiency in English.

Conners' Parent Rating Scales (CPRS-48)³⁰ were administered to caregivers at baseline and every three years until their child was 15 years old. The CPRS-48 provides measures in five behavioral domains (labeled Conduct Problem, Learning Problem, Psychosomatic, Impulsive-Hyperactive, and Anxiety) and includes a Hyperactivity Index. Age-referenced T-scores (M = 50, SD = 10) were used in the analyses. Higher scores reflected greater frequency or intensity of problem behaviors.

Standardized measures of QOL, previously validated for use with children and adolescents with PHIV,³¹ were administered to caregivers (for ages 6 months to 20 years) and study participants (ages 12-20 years) at baseline and every 6-12 months following enrollment. The General Health Perception, Health Care Utilization, and Social/School Functioning domain scores were available for all participants; the Behavior Problem Index was available for those age 5-20 years. Domain composite scores (range = 0-100) were included in the analyses. Higher scores reflected better QOL.