|
Activity Number:
|
324
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
ENAR
|
| Abstract - #309611 |
|
Title:
|
Estimates of Intraclass Correlation from the Community Intervention Trial for Youth (CITY): An HIV Prevention Intervention for Young Men Who Have Sex with Men
|
|
Author(s):
|
Sherri L. Pals*+ and Lillian S. Lin and Carolyn Guenther-Grey and Johnson Wayne
|
|
Companies:
|
Centers for Disease Control and Prevention and Centers for Disease Control and Prevention and Centers for Disease Control and Prevention and Centers for Disease Control and Prevention
|
|
Address:
|
1600 Clifton Road MS E45, Atlanta, GA, 30333,
|
|
Keywords:
|
ICC ; HIV ; Sample Size ; Group-randomized trials
|
|
Abstract:
|
Sample size estimation for group-randomized trials (GRTs) must take the intraclass correlation coefficient (ICC) into account, but few ICC estimates from HIV/STD prevention GRTs have been published. The CITY project evaluated an HIV risk reduction intervention targeting young men who have sex with men. We used two waves of CITY data to estimate ICCs for sex behavior variables, unadjusted and adjusted for covariates. The unadjusted ICC for any unprotected sex (last 90 days) ranged from about 0.005 to 0.012. With 200 members per community, these ICCs would result in variance at least twice as large as expected under individual randomization . Ignoring an ICC of this magnitude in planning a GRT could result in a substantially underpowered study. Our analyses provide estimates of ICCs from an HIV/STD prevention trial that may be useful to investigators designing future trials.
|
