587-Q2

O R I G I N A L P A P E R

Adolescent Sexual Risk: Factors Predicting Condom Use Across the Stages of Change

Cynthia Grossman Æ Wendy Hadley Æ Larry K. Brown Æ Christopher D. Houck Æ April Peters Æ Marina Tolou-Shams Æ Project SHIELD Study Group

Published online: 22 April 2008

� Springer Science+Business Media, LLC 2008

Abstract This study examined factors associated with

high-risk adolescents’ movement toward or away from

adopting consistent condom use behavior using the

Transtheoretical Model Stages of Change. Participants

drawn from the inactive comparison condition of a ran-

domized HIV prevention trial (Project SHIELD) responded

to items assessing pros and cons of condom use, peer

norms, condom communication, and perceived invulnera-

bility to HIV. Participants were categorized based on their

condom use behavior using the Transtheoretical Model.

Multiple logistic regressions found that progression to

consistent condom use was predicted by continuing to

This work was not written as part of Cynthia I. Grossman’s official

duties as a Government employee. The views expressed in this

manuscript do not necessarily represent the views of the NIMH, NIH,

HHS, or the United States Government.

Project SHIELD Study Group Principal Investigators Larry Brown, M.D.—Rhode Island Hospital, Providence, RI Ralph DiClemente, Ph.D.—Emory University, Atlanta, GA M. Isabel Fernandez, Ph.D.—University of Miami, Miami, FL Timothy Flanigan, M.D.—Miriam Hospital, Providence, RI Deborah Haller, Ph.D.—Virginia Commonwealth University, Richmond, VA Lori Leonard, Sc.D.—University of Texas, Houston Health Science Center, Houston, TX Lydia O’Donnell, Ed.D.—Education Development Center, Inc., Newton, MA William E. Schlenger, Ph.D.—Research Triangle Institute, Research Triangle Park, NC Barbara Silver, Ph.D.—Substance Abuse and Mental Health Services Administration, Rockville, MD Site Investigators Richard Crosby, Ph.D.—Emory University, Atlanta, GA Caryl Gay, Ph.D.—University of Miami, Miami, FL Janet Knisely, Ph.D.—Virginia Commonwealth University, Richmond, VA Celia Lescano, Ph.D.—Rhode Island Hospital, Providence, RI Kevin Lourie, Ph.D.—Rhode Island Hospital, Providence, RI Louise Masse, Ph.D.—University of Texas, Houston Health Science Center, Houston, TX Janet O’Connell, MPH—Miriam Hospital, Providence, RI David Pugatch, M.D.—Miriam Hospital, Providence, RI Eve Rose, Ph.D.—Emory University, Atlanta, GA

Ann Stueve, Ph.D.—Columbia School of Public Health, New York, NY Leah Varga, M.A.—University of Miami, Miami, FL Sue Vargo, Ph.D.—Education Development Center, Inc., Newton, MA Gina Wingood, Sc.D., MPH—Emory University, Atlanta, GA Coordinating Center Investigators—Research Triangle Institute, Research Triangle Park, NC Jamia L. Bacharach, J.D. Sylvia Cohn Courtney Johnson, Ph.D. Jacquelyn R. Murphy, BSc Allison Rose, Ph.D. Scott Royal, Ph.D., MPH Consumer Representatives Christian Aldridge Beri Hull Sean Scott

C. Grossman (&) National Institute of Mental Health, 6001 Executive Blvd.

Rm 6201, Bethesda, MD 20892, USA

e-mail: grossmanc@mail.nih.gov

W. Hadley � L. K. Brown � C. D. Houck � A. Peters � M. Tolou-Shams Rhode Island Hospital, Brown Medical School, Providence,

RI, USA

Project SHIELD Study Group

Providence, RI, USA

123

AIDS Behav (2008) 12:913–922

DOI 10.1007/s10461-008-9396-y

perceive more advantages to condom use, reporting greater

condom use communication with partners, and less per-

ceived invulnerability to HIV. Movement away from

adopting consistent condom use was predicted by a

decrease in perceived advantages to condom use, increased

perceived condom disadvantages, and fewer condom dis-

cussions. Future interventions may be tailored to enhance

these factors that were found to change over time.

Keywords Adolescent � Sexual risk behavior � Stages of change � Condom use

Introduction

Nearly half of the estimated 19 million new STIs each year

are in youth ages 15–21 (Weinstock et al. 2004) and people

under the age of 25 comprise nearly fifty percent of new

HIV infections (Office of National AIDS Policy 2000). In

order to develop effective interventions to reduce this large

problem, it is essential to examine the factors associated

with adolescents’ adoption of safer sex behavior. Two

theories, the Transtheoretical Model (TTM; Prochaska and

DiClemente 1983; Prochaska et al. 1992) and the Theory

of Planned Behavior (TPB; Fishbein and Ajzen 1975),

provide a useful framework for classifying individuals

according to their progression from unsafe to safe behavior.

The Transtheoretical Model, with its description of the

Stages of Change, attempts to explain why and how indi-

viduals change and how this change can be predicted and

facilitated (Prochaska and DiClemente 1983; Prochaska

et al. 1992). Stages of Change (Prochaska and DiClemente

1983) has been used to classify individuals with regard

to their adoption of safe behaviors across multiple health

domains including smoking, diet, exercise, and more

recently, condom use (Schumann et al. 2005; Armitage

et al. 2004; Courneya et al. 2001; Cabral et al. 2004;

Hacker et al. 2005). An individual can be classified in one

of five stages. In the first three stages (precontemplation,

contemplation, and preparation), the individual may or may

not be considering adoption of the behavior and is not

consistently engaging in the target behavior. The final two

stages (action and maintenance) are marked by consistent

engagement in and intention to continue the behavior.

According to the TTM, two salient factors for predicting

movement towards adoption of safer behavior are self-

efficacy and decisional balance. One’s self efficacy for

performing the behavior has been shown to have a linear

relationship with more movement to safer behavior

(Bandura 1977, 1982); as individuals progress toward

maintenance of positive health behavior, their confidence

in their ability to carry out the behavior increases (Grimley

et al. 1994). Decisional balance, or the perceived pros and

cons of behavior change, examines how cognitions about

the health behavior relate to stage of change (Janis and

Mann 1977; Velicer et al. 1985). Typically, people in the

precontemplation stage identify more cons associated with

the behavior, whereas those in the action or maintenance

stage perceive greater pros of engaging in the behavior

(Velicer et al. 1985). A crossover, in which pros begin to

outweigh cons, occurs during the preparation stage, per-

haps signaling readiness for change (Prochaska et al.

1994). This pattern has been demonstrated in condom and

contraceptive use among high-risk women and college

students (Bowen and Trotter 1995; Grimley et al. 1995;

Polacsek et al. 1999; Stark et al. 1998). In longitudinal

studies, these attitudes have predicted progression to later

stages and regression to earlier stages of change for both

exercise adoption and condom use behavior over a 3-month

period (Courneya et al. 2001; Malotte et al. 2000).

Another model of health behavior change, the Theory of

Planned Behavior (Fishbein and Ajzen 1975), has posited

that subjective norms, including perceived peer norms, are

important factors related to stages of change. For example,

the perception that peers engage in and value a low-fat diet

has been associated with remaining in the maintenance

phase for eating a low-fat diet (Armitage et al. 2004).

Brown and colleagues (2000) found that among adoles-

cents with hemophilia and infected with HIV, adolescent

perceptions of peer support for safer sex was associated

with improvement and maintenance of safer sexual

behavior. This is consistent with cross-sectional studies

that find associations of safer sexual behavior with healthy

peer norms (Santelli et al. 2004; Sieving et al. 2006).

Among an inner-city sample of women, greater perception

of peer support for condom use was associated with being

in a later stage of change (Cabral et al. 2004), but among

adults attending inner-city health clinics, peer norms were

not associated with transitions in stage of change for con-

dom use behavior (Malotte et al. 2000). Thus, the findings

for peer norms have been mixed and studies have been

limited to mainly cross-sectional reports.

With smoking, diet, and exercise, behavior change is

largely an individual choice. The adoption of condom use,

however, is a behavior that often demands communication

and agreement between partners. It may necessitate an

assessment of one’s own risk, and also the risk of one’s

partner(s). Therefore, communication with partners about

condom use and perceived personal vulnerability to STIs

and HIV must be taken into account. Polacsek and col-

leagues (1999) found that women who reported higher

perceived vulnerability to HIV infection were more likely

to be in a later stage of change for condom use behavior.

The influence of partner communication has yet to be

examined in relation to stages of change for condom use

behavior.

914 AIDS Behav (2008) 12:913–922

123

Few studies have examined the longitudinal process of

change among adolescents, despite the fact that behavior

change is likely to have the most enduring effect for this

population. Those studies examining stage of change

movement for condom use behavior in an adolescent

sample examined movement as part of an intervention

targeting factors associated with behavior change. How-

ever, these studies were not designed to assess naturally

occurring change in the absence of intervention (Hacker

et al. 2005). Understanding variables naturally associated

with change and relapse among adolescents not receiving

an intervention is crucial in order to identify salient factors

to be targeted by the next generation of interventions.

The literature on adolescent condom use behavior indi-

cates that different groups of adolescents have different

condom use patterns. In general, males use condoms

more often than females (Everett et al. 2000). Youth Risk

Behavior Surveillance (YRBS) data suggests that African

American teens report greater condom use than White

teens, with White and African American teens using con-

doms more often than Hispanic youth (Everett et al. 2000).

A study of male adolescents found that young African

American men were more likely to have used a condom at

last intercourse than their Hispanic or White peers (Marsi-

glio 1993). Younger teens are more likely to use condoms

than older ones, though sexual activity increases with age

(Ku et al. 1993). Relationship status also appears to play a

role, with teens in established relationships demonstrating

less condom use than teens in new relationships (Forten-

berry et al. 2002). These established patterns however, do

little to suggest how teens’ condom use behavior changes.

The current study restricts analyses to participants from

the inactive comparison condition to ensure the examina-

tion of naturally changing factors that predict progression

from early stages to later stages, as well as to identify the

factors that predict relapse from the action and maintenance

stages. Based on previous studies (Courneya et al. 2001;

Cabral et al., 2004; Janis and Mann 1977; Velicer et al.,

1985), we hypothesized significant baseline differences

among those in the early stages of change (precontempla-

tion, contemplation, and preparation) compared to those in

later stages of condom use behavior (action and mainte-

nance) across the four attitudinal variables (i.e. pros, cons,

peer norms, and perceived invulnerability) and condom

communication. Specifically, we expected those in the

early stages of change would report riskier attitudes and

less condom communication. It was expected that longitu-

dinally, the development of attitudes and communication

favorable to condom use would be associated with pro-

gression to later stages of condom use behavior, while an

increase in attitudes unfavorable to condom use was

hypothesized to predict relapse among those already in later

stages.

Methods

Procedure

Analyses for the current study included 446 participants in

the delayed intervention comparison condition of a ran-

domized control trial of an HIV prevention program

(Project SHIELD) conducted in three US cities: Atlanta

GA, Providence RI, and Miami FL (Crosby et al. 2005).

Adolescents were recruited using a variety of strategies,

including contact at medical clinics, community outreach,

posters and flyers, and self-referrals. Fifty dollars com-

pensation for time and effort was provided for participation

at each assessment. All study procedures were approved by

the institutional review boards at each site, and informed

consent was obtained from all participants or, in the case of

minors, their guardians.

Participants completed assessments via audio computer-

assisted self-interview (ACASI) using laptop computers

administered individually or in a group format (with moni-

tors ensuring privacy). To assist in recall of the past 90 days,

participants were provided a 90-day calendar and a standard

script was used, instructing participants to recall significant

events of the past 90 days to help with recollection of the

time period. Research staff remained in the area to answer

questions that arose while participants completed measures.

Participants completed a 30-min baseline assessment, and

were then randomized to either a 2-h HIV prevention pro-

gram focusing on HIV education and communication skills

or an inactive comparison condition. Data used in the current

analyses represent baseline and 6-month follow-up assess-

ments for those participants who were randomized to the

inactive comparison (no intervention) condition of the study.

Participants

Inclusion criteria for the study included being between the

ages of 15 and 21 and having had unprotected vaginal or

anal sex in the past 90 days. Adolescents were excluded if

they were currently pregnant or within 90 days of giving

birth, actively attempting to get pregnant, HIV positive by

self-report, or currently participating in another HIV pre-

vention study. Of 1,867 eligible adolescents, 1,386 (74%)

were enrolled and completed ACASI. Of these, 685 were

randomized to the comparison condition of the project,

which only required computer assessments and urine

screening for STIs at each study time point. A 77% follow-

up rate was achieved at 6 months. Participants were com-

pared based on completion of primary demographic data

and condom use behavior and no significant differences

were found. However, those with complete data were more

likely to be African American than those with incomplete

data. From these participants, 446 adolescents had

AIDS Behav (2008) 12:913–922 915

123

complete data at both baseline and 6-month follow-up,

which comprised the current study sample.

Measures

Measures used in Project SHIELD were largely derived

from Project LIGHT, a NIMH/NIH-funded multisite HIV

prevention trial for high-risk young adults, where they

demonstrated satisfactory internal reliability and sensitivity

to intervention impact (NIMH Multisite HIV Prevention

Trial Group 1998).

Demographic Variables

Participants reported age, gender, race/ethnicity, sexual

orientation, education, employment, and current living

situation (whether they were living with a sexual partner).

Stage of Change for Safer Sex

Stage of change items were derived from the Transtheo-

retical Model of Prochaska and DiClemente (1983) as

adapted by Schnell and colleagues (1996). The items

assessed condom use in the past 90 days (never, less than

half the time, about half the time, more than half the time,

always) as well as intentions to begin consistent condom

use (yes/no) in the next 90 days. At baseline, subjects in

precontemplation used condoms less than half the time and

were not planning to start using condoms all of the time in

the next 90 days. Those in contemplation used condoms

less than half the time and were planning to start consis-

tently using condoms in the next 90 days. Those in

preparation used condoms about half the time and were

planning to start using condoms in the next 90 days. Those

in the action stage used condoms more than half of the time

and those in the maintenance stage reported always using

condoms during sex.

Sexual Attitude Scales

Items assessing attitudes and perceptions related to condom

use were assessed at baseline. Pros of Condom Use

(a = .66, range = 5–25) were assessed using five items related to using condoms (‘‘I would feel guilty if my main

partner and I didn’t use a condom.’’) with a 5-point

(Strongly disagree to Strongly agree) Likert scale. Cons of

Condom Use (a = .85; range = 4–20) were measured with a scale assessing unpleasurable expectations regarding

condom use (e.g., ‘‘Sex doesn’t feel as good when you use

a condom.’’). The four items of this scale were assessed on

a 5-point (Strongly disagree to Strongly Agree) Likert

scale. Subjective norms for condom use were examined

using an eight-item measure of perceived Peer Norms

(a = .71; range = 8–40) for sexual activity and condom use (e.g., ‘‘How many of your friends think that it’s fine to

have vaginal or anal sex without a condom?’’). Responses

were on a five-point scale ranging from None to All.

Communication with partners about condoms in the past

90 days was measured via six yes/no items of the Condom

Use Communication and Negotiation Checklist (range 0–6;

e.g., ‘‘In the past 90 days, did you ever tell any partner you

wanted to use a condom?’’). Perceived Invulnerability to

HIV was assessed via two items rated on a 5-point

(Strongly disagree to Strongly agree) Likert scale (range 2–

10; e.g., ‘‘I don’t need to use a condom because people like

me don’t get HIV.’’).

Substance Use

Alcohol and Marijuana Use was assessed by asking

adolescents how many days of the last 30 they used each of

the two substances. Because the data on this variable was

skewed, responses were coded as 0 for no use, 1 for

1–4 days, and 2 for 5 or more days for each substance. In

order to minimize the competition for variance between

alcohol and marijuana use, and the positive correlation

found in adolescent samples, the coded responses were

summed to create a summary score reflecting any substance

use. Summary scores for this variable ranged from 0 to 4

with higher scores indicating more frequent substance use.

Data Analysis

For analyses, adolescents who were not using condoms

during the majority of their sexual activity at baseline (i.e.,

precontemplation, contemplation, or preparation stages)

were combined to form an Inconsistent Condom Users

subsample. These participants were then further catego-

rized based on movement across stages of change from

baseline to 6-month follow-up. Those who moved toward

more consistent condom use (e.g., precontemplation to

contemplation or preparation to action) were categorized

as Progressors and those who remained in the same stage

as baseline were categorized as Nonprogressors (i.e., those

who remained in precontemplation, contemplation, or

preparation).

Similarly, adolescents who were regularly using con-

doms at baseline (action and maintenance stages) were

grouped together to form the Consistent Condom Users

subsample. Examining their movement across stages of

change between baseline and 6 months, they were then

categorized as Maintainers, those who continued consistent

condom use and reported high intentions to use (e.g.,

stayed in action and/or maintenance), or Relapsers, those

who reverted to an earlier stage and/or reported less

916 AIDS Behav (2008) 12:913–922

123

intentions to use condoms (e.g. went from action to

contemplation).

Baseline differences between Consistent and Inconsis-

tent Condom Users were examined on demographic

variables and scale scores using Chi-square and t-tests.

Predictors of progression within the Inconsistent Condom

User group and maintenance within the Consistent Condom

User group were tested as follows. Within the Inconsistent

Condom Users subsample, bivariate tests (Chi-square and

t-tests) were first conducted to detect any baseline differ-

ences between the Progressor and Nonprogressor groups on

demographic variables (age, gender, race/ethnicity, sexual

orientation, education, employment, and current living

situation) and scale scores. Categorical variables of race/

ethnicity, sexual orientation, and education were collapsed

such that the referent categories were as follows: African-

American, heterosexual, and High School education or

greater. These analyses were followed by t-tests examining

group differences on residual change scores of the scales

(to account for baseline scores) from baseline to 6-month

follow-up assessment. Within the Consistent Condom

Users subsample, these same tests were performed to

examine differences between Maintainers and Relapsers.

Finally, to simultaneously examine the impact of scale

changes on progression or relapse, two separate multiple

logistic regressions predicting to membership in a transi-

tional group (e.g. Progression or Relapse) were conducted

using residual change scores of the examined variables.

Variables significantly associated with group membership

in the bivariate analyses were entered into the models and

residual scale scores were dichotomized using median

splits for ease of interpretation.

Results

Participants

The average age of the sample was 18.2 years (SD = 1.9)

and 62% were female. Fifty-three percent self-identified as

African-American, twenty-one percent as Hispanic, nine-

teen percent as White, and six percent as other racial/ethnic

category. The majority (92%) of adolescents identified

their sexual orientation as heterosexual. Fifty percent had

completed high school or greater at the time of participa-

tion. Forty-four percent of the sample was currently

employed, and sixteen percent were living with a sexual

partner. Participants categorized as Inconsistent Condom

Users (precontemplation, contemplation, and preparation

groups) made up 64% (n = 287) of the sample while

Consistent Condom Users (action and maintenance groups)

made up 36% (n = 159). Figure 1 shows the proportion of

participants in each stage at baseline and 6 month follow-

up.

Inconsistent Versus Consistent Condom Users

Analyses of baseline demographic variables revealed that

Inconsistent Condom Users, compared to Consistent Con-

dom Users, were more likely to be female (67% vs. 54%,

v2 = 6.53, P \ .05), less likely to report a racial identity of African-American (48% vs. 63%, v2 = 8.28, P \ .01), less likely to identify as heterosexual (86% vs. 96%, v2 = 4.97, P \ .05), and more likely to report that they were living with a partner (23% vs. 4%, v2 = 26.53, P \ .01). There were no significant baseline demographic differences

0

10

20

30

40

50

Precontemplation Contemplation Preparation Action Maintenance

Stage of Change

R ep

or ti

n g

(% )

Baseline

6-months

Fig. 1 Changes in stage of change from baseline to

6 months

AIDS Behav (2008) 12:913–922 917

123

between Inconsistent and Consistent Condom Users for

age, employment status, or level of education.

Significant baseline differences between the Inconsistent

Condom Users and the Consistent Condom Users were

found on all attitudinal variables (Table 1). Participants in

the Inconsistent Condom Use group reported riskier atti-

tudes, including fewer Pros of Condom Use, more Cons of

Condom Use, a greater number of friends who accepted

unsafe sex practices (Peer Norms), as well as greater per-

ceptions of themselves as invulnerable to HIV (Perceived

Invulnerability), and less discussion of condom use with

their partners (Condom Communication) compared to

those in the Consistent Condom Use group. Groups did not

differ on substance use at baseline.

Inconsistent Condom Users

Progressors (n = 122), those who moved towards a stage

reflecting more consistent condom use over a 6-month

period, were compared to Non-progressors (n = 165),

those who did not move toward more consistent condom

use. Progressors were significantly younger than Non-

progressors (t = 2.92, P \ .01; M = 17.9, SD = 1.85 vs. M = 18.5, SD = 1.77), but statistically there were no other

significant demographic differences.

Comparison of scale scores at baseline revealed that

Progressors reported higher scores on the Pros of Condom

Use Scale (t = -3.75, P \ .01) and lower scores on the Cons of Condom Use (t = 2.78, P \ .01) than Non-pro- gressors (see Table 2). No other significant baseline

differences were found between those who would become

Progressors and Non-progressors.

Analyses of the residual change scores over time

(baseline to 6 months) reflected significant differences

between Progressors and Non-progressors for Pros of

Condom Use and Condom Communication. Over the 6-

month period, Progressors reported stable perceptions of

the Pros of Condom use, whereas Non-progressors dem-

onstrated a reduction in Pros of Condom Use, Progressors

showed greater increases in their communication with

partners about condoms (Condom Communication) relative

to Non-progressors. There was a trend in residual change

scores for Perceived Invulnerability whereby Progressors

reported a greater reduction in their perceptions of invul-

nerability to HIV than Non-progressors. No significant

differences between groups were found in residual change

scores for Cons of Condom Use, Alcohol/Marijuana Use,

or Peer Norms from baseline to 6 months (see Table 2).

A multiple logistic regression, controlling for age,

revealed that residual change scores for Condom

Communication, Pros of Condom Use, and Perceived

Invulnerability were predictive of membership in the Pro-

gressor group (v2 (4, 272) = 67.43, P \ .01; see Table 3). Participants reporting greater increases in Condom

Table 1 Baseline scale scores for inconsistent and consistent

condom users

** P \ .01

Variable Inconsistent condom use (n = 287) Consistent condom use (n = 159)

M SD M SD df t

Pros of condoms 13.88 3.69 17.83 3.80 443 -10.70**

Cons of condoms 11.75 3.97 9.07 3.68 439 7.00**

Peer norms 23.90 5.01 21.85 5.15 439 4.07**

Condom communication 2.66 1.01 3.95 1.17 438 -12.11**

Perceived invulnerability 3.04 1.52 2.68 1.20 391 2.71**

Alcohol/marijuana use 1.98 1.51 1.80 1.48 429 1.18

Table 2 Scale scores for non-progressors and progressors over 6 months

Variable Non-progressors (n = 165) Progressors (n = 122) Differences in residual change scores

Base Post Base Post

M SD M SD M SD M SD t

Pros of condoms 14.6 3.80 13.29 3.25 16.10 4.40 16.19 4.52 -6.37**

Cons of condoms 12.31 3.96 11.39 4.30 11.01 3.86 9.99 3.92 1.10

Peer norms 23.61 5.14 23.84 5.23 24.28 4.82 23.35 5.57 1.06

Condom communication 2.67 0.96 2.81 0.96 2.65 1.09 3.62 1.19 -4.33**

Perceived invulnerability 2.97 1.44 2.98 1.53 3.13 1.62 2.74 1.23 1.97*

Alcohol/marijuana use 2.01 1.52 1.88 1.47 1.94 1.50 1.80 1.54 0.36

* P \ .05, ** P \ .01

918 AIDS Behav (2008) 12:913–922

123

Communication were four and a half times more likely to

be among the Progressors and those reporting greater sta-

bility in the Pros of Condom Use were nearly three times

more likely to be in the Progressor group. Individuals who

reported greater reductions in their Perceptions of Invul-

nerability were two times more likely to be Progressors.

Age was not associated with increased odds of being

among the Progressors. Baseline stage was not included in

the model as it was not found to be a significant predictor

of being in the Progressor group.

Consistent Condom Users

No significant differences between the Relapse and Main-

tenance groups were found on baseline demographics. On

baseline scale scores (see Table 4), those in the Relapse

group reported fewer perceived Pros of Condom Use

(t = -4.27; P \ .01), more Cons of Condom Use (t = 2.39, P \ .05) and less Condom Communication (t = -3.48, P \ .01). No significant between group differences were found for Peer Norms, Perceived

Invulnerability, or Alcohol/Marijuana Use.

Analysis of the residual change scores found that

Relapsers reported greater changes in decreasing Pros of

Condom Use, increasing their perceptions of the Cons of

Condom Use, decreasing their Condom Communication,

and increasing their Perceived Invulnerability over the 6-

month period compared to those in the Maintenance

group (see Table 4). No significant differences in Alcohol/

Marijuana Use or Peer Norms emerged.

A multiple logistic regression of the residual change

scores predicting to the Relapse group (v2 (4,148) = 46.86, P \ .01; see Table 5) revealed that participants endorsing greater decreases over 6 months in Condom

Table 3 Multivariate logistic regression analyses predicting membership in progressor group

Variable Progressor group (N = 272)

b Wald df O.R. 95% C.I.

Age (B18) .19 .43 1 1.20 .69–2.10

Pros of condoms (greater positive change group) 1.06 13.29 1 2.87** 1.63–5.07

Condom communication (greater positive change group) 1.52 27.28 1 4.55** 2.58–8.04

Perceived invulnerability (greater negative change group) .82 8.02 1 2.28* 4.02–1.29

O.R. = Adjusted odds ratio; * P \ .05, ** P \ .01

Table 4 Scale scores for relapsers and maintainers over 6 months

Variable Relapsers (n = 66) Maintainers (n = 93) Differences in residual change scores

Base Post Base Post

M SD M SD M SD M SD t

Pros of condoms 16.38 3.71 15.11 2.97 18.86 3.53 18.70 4.19 -4.60**

Cons of condoms 9.88 3.58 10.67 3.55 8.48 3.67 8.16 3.71 3.60**

Peer norms 21.54 4.67 23.29 4.83 22.06 5.48 22.36 5.25 1.63

Condom communication 3.58 1.23 2.85 1.23 4.22 1.05 4.14 1.05 -6.01**

Perceived invulnerability 2.65 1.12 2.97 1.55 2.71 1.25 2.57 .92 2.06*

Alcohol/marijuana use 1.76 1.42 1.80 1.42 1.83 1.52 1.74 1.32 0.85

* P = .05, ** P = .01

Table 5 Multivariate logistic regression analyses predicting membership in relapse group

Variable Relapse group (N = 148)

b Wald df O.R. 95% C.I.

Pros of condoms (greater negative change group) 1.10 7.59 1 3.01* 1.38–6.61

Cons of condoms (greater positive change group) .91 4.92 1 2.48* 1.11–5.55

Condom communication (greater negative change group) 1.86 21.26 1 6.40** 2.91–14.10

Perceived invulnerability (greater positive change group) .40 .82 1 1.49 .63–3.52

O.R. = Adjusted odds ratio; * P = .05, ** P = .01

AIDS Behav (2008) 12:913–922 919

123

Communication were nearly six and a half times more

likely to relapse to an earlier stage of change. Participants

reporting a greater reduction in Pros of Condom use were

three times more likely to be in the Relapse group, whereas

participants reporting a greater increase in the Cons of

Condom Use were two and a half times more likely to be in

the Relapse group. Changes in Perceived Invulnerability

were not a significant predictor of membership in the

Relapse group.

Discussion

The primary aim of the current study was to determine

predictors of movement across the stages of change for

condom use behavior among an at-risk sample of adoles-

cents. We employed a design which allowed for the

examination of variables associated with progression

towards greater condom use and to identify factors asso-

ciated with relapse from later stages (e.g., action and

maintenance) to earlier stages of change (e.g., precontem-

plation, contemplation, and preparation). In general, our

hypothesis that risk attitudes and condom communication

would differentiate those in early versus later stages of

change was supported. Additionally, we found that change

in both attitudinal variables and condom communication

was associated with movement across the stages of change

(i.e., progression and relapse).

Consistent with the Transtheoretical Model (Prochaska

and DiClemente 1983) and the Theory of Planned Behavior

(Fishbein and Ajzen 1975), differences between Consistent

and Inconsistent Condom Users were found in their per-

ceptions of the pros and cons of condom use, as well as

their peer norms for condom use. Likewise, perceptions of

invulnerability to HIV were higher among Inconsistent

condom users, and this supports previous findings among a

sample of inner city women for whom perceived invul-

nerability to HIV infection was strongly associated with

being in an earlier stage of change for condom use behavior

(Polacsek et al. 1999).

Although we did not hypothesize that there would be

baseline differences between those that did or did not

transition over time (Progressors vs. Non-progressors and

Relapsers vs. Maintainers), our findings reveal that several

attitudinal variables served as predictors of movement

across the stages of change over 6 months. These attitudes

appear to function as indicators of readiness for change.

Specifically, the pros and cons of condom use predicted

transitions among stages in that Progressors reported more

pros and fewer cons of condom use at baseline compared

with Non-progressors. Comparisons among Relapsers and

Maintainers revealed that those who relapsed to early

stages of change over 6 months had reported fewer pros

and more cons associated with condom use at baseline and

also reported fewer condom discussions with partners.

Notably, baseline peer norms, perceptions of invulnera-

bility and alcohol/marijuana use were less relevant in

predicting transitions across the stages of change. Only

those variables directly related to condom use were asso-

ciated with movement across the stages of change for

condom use behavior.

As anticipated, changes in all of the variables, with the

exception of peer norms, were associated with movement

in the stages of change. For example, among Progressors

condom communication with partners significantly

increased over the 6-month assessment period relative to

Non-progressors. Among Relapsers, pros of condom use

decreased and cons significantly increased compared with

Maintainers over this same time frame. These results

among a high-risk sample of adolescents are consistent

with a previous study of adults recruited from inner-city

health clinics (Malotte et al. 2000) and support the

assumption that attitudes about condom use are an impor-

tant target for change in condom use behavior and can

change outside of a formal intervention.

Unexpectedly, peer norms were not associated with

transition to or relapse from later stages of change (e.g.

action and maintenance). This is in strong contrast to

findings among adolescents with hemophilia who were also

HIV-infected, wherein changes in peer norms were asso-

ciated with both improvement and maintenance of safer sex

behavior (Brown et al. 2000). However, our study assessed

condom use attitudes specifically, whereas the study of

adolescents with hemophilia assessed norms for abstinence

and non-intercourse in relation to all safer sex behaviors.

Likewise, Malotte and colleagues (2000) found that peer

norms were not associated with transition among stages of

change among a high-risk sample of adults for whom peer

groups may be stable. Among younger adolescent samples,

for whom peer groups are in greater flux, peer norms may

show greater change and be more strongly associated with

changes in condom use behavior. Among older samples,

other factors such as the perceived advantages and disad-

vantages (e.g. perceived pleasure, discomfort, and/or hassle

of condom use), may be stronger influences.

This study was the first to examine condom communi-

cation with partners in relation to stages of change among

adolescents. Most notably, condom communication showed

significant changes across time for both Relapsers and Pro-

gressors. For the Relapse group, condom communication

significantly decreased over 6-months and among Progres-

sors condom communication significantly increased. In

other studies attitudes/cognitive factors have been found to

be most important for change among those in the early

stages, whereas behavior affects the greatest change among

those in later stages (e.g. action and maintenance;

920 AIDS Behav (2008) 12:913–922

123

(Prochaska et al. 1992). Perhaps among adolescents, for

whom communication with sexual partners is a new and less

formed behavior, communication about sexual behavior can

be enhanced quickly and can then serve to promote and

maintain condom use.

Two important findings emerged from the current study,

both of which are helpful in informing intervention

development. The pros of condom use and condom com-

munication were consistent in their power to predict

progression to and maintenance of consistent condom use.

Notably, the pros of condom use were relatively stable for

those with greater condom use across the 6-month window

(i.e., Progressors and Maintainers), whereas among those

demonstrating less condom use behavior over time (i.e.,

Non-progressors and Relapsers), the pros of condom use

decreased. Among the Non-progressors and Relapsers there

appears to be a shift towards the perception of condoms

being less pleasurable and more of a hassle. Additionally,

Condom Communication followed a similar trend whereby

Progressors reported increased communication and

Relapsers reported decreased communication.

Based on these findings, interventions could categorize

individuals according to their baseline scores into consis-

tent and inconsistent condom users and then tailor

interventions accordingly. For those youth using condoms

inconsistently, interventions should first target increasing

the advantages of using condoms (e.g. lengthen duration of

sexual intercourse, decrease messiness, increase female

pleasure) and then promote and model skills for commu-

nication with partners about condoms. Previous studies

have reported that for females such communication is

difficult, but our findings were not specific to males, opti-

mistically suggesting that interventions can reinforce these

naturally evolving skills for both genders. Maintaining

consistent condom use is a challenge and interventions will

need to focus on novel ways to sustain interest in the

effectiveness and positive aspects of condoms (Crosby

et al. 2003), as well as helping adolescents communicate

with all partners even when ‘‘in love’’ (Lescano et al. 2006)

to continue to prevent the onset of risk. For those indi-

viduals using condoms consistently, interventions might

adhere to a relapse prevention model whereby the goals are

to preserve positive attitudes toward condoms, maintain

consistent communication with partners about condoms,

and reinforce perceptions of vulnerability to HIV.

Despite the strengths of the study, there are limitations.

Although the current study included a sample of high-risk

adolescents and examined a number of important variables

associated with stage of condom use behavior, it was not

fully inclusive of all variables that might be associated with

the Stage of Change Model, such as self-efficacy for con-

dom use. Another limitation of the data includes the

necessity to use 3 month, rather than the standard 6 month,

interval to categorize participants in the Maintenance stage.

In terms of generalizability, the mean age of the sample was

eighteen and therefore these findings may not be general-

izable to younger adolescents who are just beginning to

engage in sexual behavior. In fact, within the univariate

analyses Progressors tended to be younger than Non-Pro-

gressors, suggesting that younger adolescents may be more

likely to spontaneously adopt condom use, at least tempo-

rarily. Finally, the voluntary nature of the recruitment

strategy for this study may further limit generalizability.

The study adds to the body of literature examining stage

of change in relation to condom use behavior. In particular,

it examines a sample of high-risk youth and provides

support for the importance of improving positive attitudes

and communication about condom use to influence safer

sexual behaviors. These factors were found to change

naturally and predicted more consistent condom use over

6 months. The findings suggest that interventions that tar-

get positive attitudes about condom use and that build skills

for condom communication are likely to promote consis-

tent condom use among adolescents.

Acknowledgment Research supported by SAMHSA grant U10 SMS2073 to the cooperating sites: Rhode Island Hospital, Miriam

Hospital, Emory University, and University of Miami.

References

Armitage, C. J., Sheeran, P., Conner, M., & Arden, M. A. (2004).

Stages of change or changes of stage? Predicting transitions in

transtheoretical model stages in relation to healthy food choice.

Journal of Consulting & Clinical Psychology, 72, 491–499. Bandura, A. (1977). Self-efficacy: Toward a unifying theory of

behavioral change. Psychological Review, 84, 191–215. Bandura, A. (1982). Self efficacy mechanism in human agency.

American Psychologist, 37, 122–147. Bowen, A. M., & Trotter, R. (1995). HIV risk in intravenous drug

users and crack cocaine smokers: Predicting stage of change for

condom use. Journal of Consulting & Clinical Psychology, 63, 238–248.

Brown, L. K., Schultz, J. R., Parsons, J. T., Butler, R. B., Forsberg, A.

D., Kocik, S. M., et al. (2000). Sexual behavior change among

HIV infected adolescents with hemophilia. Pediatrics, 106, E22. Cabral, R. J., Galavotti, C., Stark, M. J., Gargiullo, P. M., Semaan, S.,

Adams, J., et al. (2004). Psychosocial factors associated with

stage of change for contraceptive use among women at increased

risk for HIV and STDs. Journal of Applied Social Psychology, 34, 959–983.

Courneya, K. S., Plotnikoff, R. C., Hotz, S. B., & Birkett, N. J. (2001).

Predicting exercise stage transitions over two consecutive 6-

month periods: A test of the theory of planned behaviour in a

population-based sample. British Journal of Health Psychology, 6, 135–150.

Crosby, R. A., DiClemente, R. J., Wingood, G. M., Salazar, L. F.,

Harrington, K., Davies, S. L., et al. (2003). Identification of

strategies for promoting condom use: A prospective analysis of

high-risk African American female teens. Prevention Science, 4, 263–270.

AIDS Behav (2008) 12:913–922 921

123

Crosby, R., DiClemente, R. J., Wingood, G. M., Salazar, L. F., Rose,

E., Levine, D., et al. (2005). Condom failure among adolescents:

Implications for STD prevention. Journal of Adolescent Health, 36, 534–536.

Everett, S. A, Warren, C. W., Santelli, J. S., Kann, L., Collins, J. L., &

Kolbe, L. J. (2000). Use of birth control pills, condoms, and

Withdrawal among US high school students. Journal of Adoles- cent Health, 27, 112–118.

Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention and behavior: An introduction to theory and research. London: Addison Wesley.

Fortenberry, J. D., Tu, W., Harezlak, J., Katz, B. P., & Orr, D. P.

(2002). Condom use as a function of time in new and established

adolescent sexual relationships. American Journal of Public Health, 92, 211–213.

Grimley, D. M., Prochaska, J. O., Velicer, W. F., Blais, L., &

DiClemente, C. C. (1994). The transtheoretical model of

change. In T. Brinthaupt & R. Lipka (Eds.), Changing the self: Philosophies, techniques and experiences (pp. 201–227). Albany, NY: State University of New York Press.

Grimley, D. M., Prochaska, J. O., Velicer, W. F., & Prochaska, G. E.

(1995). Contraceptive and condom use adoption and mainte-

nance: A stage paradigm approach. Health Education Quarterly, 22, 20–35.

Hacker, K., Brown, E., Cabral, H., & Dodds, D. (2005). Applying a

transtheoretical behavioral change model to HIV/STD and

pregnancy prevention in adolescent clinics. Journal of Adoles- cent Health, 37, S80–S93.

Janis, I. L., & Mann, L. (1977). Decision making: A psychological analysis of conflict, choice and commitment. New York: Free Press.

Ku, L., Sonenstein, F. L., & Pleck, J. H. (1993). Factors influencing first

intercourse for teenage men. Public Health Report, 108, 680–694. Lescano, C., Vazquez, E. A., Brown, L. K., Litvin, E., Pugatch, D., &

Project SHIELD Study Group. (2006). Condom use with

‘‘Casual’’ and ‘‘Main’’ Partners: What’s in a Name? Journal of Adolescent Health, 39, 443.

Malotte, C. K., Jarvis, B., Fishbein, M., Kamb, M., Iatesta, M.,

Hoxworth, T., et al. (2000). Stage of change versus an integrated

psychosocial theory as a basis for developing effective behaviour

change interventions. AIDS Care, 12, 357–364. Marsiglio, W. (1993). Adolescent males’ orientation toward paternity

and contraception. Family Planning Perspectives, 25, 22–31. NIMH Multisite HIV Prevention Trial Group. (1998). The NIMH

multisite HIV prevention trial: Reducing HIV sexual risk

behavior. Science, 280, 1889–1894.

Office of National AIDS Policy. (2000). Youth and HIV/AIDS 2000: A new American agenda. Washington, DC: The White House.

Polacsek, M., Celentano, D., O’Campo, P., & Santelli, J. (1999).

Correlates of condom use stage of change: Implications for

interventions. AIDS Education and Prevention, 11, 38–52. Prochaska, J. O., & DiClemente, C. C. (1983). Stages and processes

of self-change of smoking: toward an integrative model of

change. Journal of Consulting & Clinical Psychology, 51, 390– 395.

Prochaska, J. O., DiClemente, C. C., & Norcross, J. C. (1992). In

search of how people change: Applications to addictive behav-

iors. American Psychologist, 47, 1102–1114. Prochaska, J. O., Redding, C. A., Harlow, L. L., Rossi, J. S., &

Velicer, W. F. (1994). The transtheoretical model of change

and HIV prevention: A review. Health Education Quarterly, 21, 471–486.

Santelli, J. S., Kaiser, J., Hirsch, L., Radosh, A., Simkin, L., &

Middlesteadt, S. (2004). Initiation of sexual intercourse among

middle school adolescents: The influence of psychosocial

factors. Journal of Adolescent Health, 34, 200–208. Schnell, D., Galavotti, C., Fishbein, M., & Chan, D. (1996).

Measuring the adoption of consistent use of condoms using the

stages of change model. Public Health Reports, 111, 59–68. Schumann, A., Meyer, C., Rumpf, H., Hannover, W., Hapke, U., &

John, U. (2005). Stage of change transitions and processes of

change, decisional balance, and self-efficacy in smokers: A

transtheoretical model of validation using longitudinal data.

Psychology of Addictive Behaviors, 19, 3–9. Sieving, R., Eisenberg, M., Pettingell, S., & Skay, C. (2006). Friends’

influence on adolescents’ first sexual intercourse. Perspectives on Sexual and Reproductive Health, 38, 13–19.

Stark, M. J., Tesselaar, H. M., O’Connell, A. A., Person, B.,

Galavotti, C., Cohen, A., et al. (1998). Psychosocial factors

associated with the stages of change for condom use among

women at risk for HIV and STDs: Implications for intervention

development. Journal of Consulting & Clinical Psychology, 66, 967–978.

Velicer, W. F., DiClemente, C. C., Prochaska, J. O., & Brandenberg,

N. (1985). Decisional balance measurement for assessing and

predicting smoking status. Journal of Personality and Social Psychology, 48, 1279–1289.

Weinstock, H., Berman, S., & Cates, W. (2004). Sexually transmitted

diseases among American youth: Incidence and prevalence

estimates 2000. Perspectives on Sexual and Reproductive Health, 36, 6–10.

922 AIDS Behav (2008) 12:913–922

123