Effects of Condom Wrapper Graphics
and Scent on Condom Use in the
Botswana Defence Force
Bonnie Tran, PhD
Department of Defense HIV/AIDS Prevention Program
20 July 2012
1
Background


Correct and consistent condom use is highly
effective in preventing HIV infection
Condom use among military personnel is
generally inconsistent

Military personnel at higher HIV risk
1.
2.
3.
4.
Davis K., Weller S .(1999)
Nwokoji U., Ajuwon A .(2004)
Bing E,. Ortiz D., Ovalle-Bahamon R,. et al. (2008)
Miller N, Yeager R (1995)
2
Rationale and Objective

Rationale




HIV prevalence in Botswana (17.6%) is high
Botswana Defence Force (BDF) provides free condoms
Government-issued condoms (Lorato/Carex) often not
used due to unpleasant scent and unattractive wrapper
Formative work



Objective

1.
Scented condoms highly preferred
Camouflaged wrapper was appealing
Determine if condom-wrapper graphics and scent
improved condom use in BDF
Personal communication with BDF (2010)
3
Study Design

Non-randomized intervention study conducted from Oct
2010 to Apr 2011

Pre-intervention condom: those typically used by participant
Intervention condom: scented/unscented, packaged in 2
wrappers

4 BDF sites selected

Base
Location
Intervention
Scent
1
South
Unscented
2
South
Scented
3
North
Unscented
4
North
Scented
4
Participants
Eligibility criteria
 Male BDF personnel
 Sexually active
 Aged 18 to 30
 4 selected sites
Recruitment
 Flyers
 BDF newsletters
 Standard military
communication channels
Target sample size = 260 (65 per site)
211 (81%) consented
120
vs. 91
5
Study Procedures
Baseline
survey
Intervention
condoms
distributed
Training
session
Diary 1
Day −14
Wrap-up
Diary 2
Day −7
Diary 3
Day 0
Pre-Intervention Period
Day 7
2 weeks
Day 14
Diary 4
Day 21
Day 28
Post-Intervention Period
6
Study Procedures
Baseline
survey
Intervention
condoms
distributed
Training
session
Diary 1
Day −14
Wrap-up
Diary 2
Day −7
Diary 3
Day 0
Pre-Intervention Period
Day 7
2 weeks
Day 14
Diary 4
Day 21
Day 28
Post-Intervention Period
7
Baseline Survey







Demographics
Sexual behavior history
HIV risk perception
HIV transmission knowledge
Alcohol use
Circumcision status
Condom use frequency,
attitudes, & behaviors
8
Sexual Behavior Diary

Measured sexual activity and condom use

Completed on daily basis




Up to 3 sexual occasions
Type of sexual partner (spouse, regular cohabitating,
regular non-cohabitating, casual)
Condom used (yes/no) and reasons why
Participants instructed to not modify their sexual
behaviors while participating in study
9
Study Procedures
Baseline
survey
Intervention
condoms
distributed
Training
session
Diary 1
Wrap up
Diary 2
Diary 3
Diary 4
Establish pre-intervention
condom use
Day −14
Day −7
Day 0
Pre-Intervention Period
Day 7
2 weeks
Day 14
Day 21
Day 28
Post-Intervention Period
10
Study Procedures
Baseline
survey
Training
session
Diary 1
Diary 2
Intervention
condoms
distributed
Wrap up
Diary 3
Diary 4
Establish pre-intervention
condom use
Day −14
Day −7
Day 0
Pre-Intervention Period
Day 7
2 weeks
Day 14
Day 21
Day 28
Post-Intervention Period
11
Study Procedures
Baseline
survey
Intervention
condoms
distributed
Training
session
Diary 1
Wrap up
Diary 2
Diary 3
Establish post-intervention
condom use
Establish pre-intervention
condom use
Day −14
Day −7
Diary 4
Day 0
Pre-Intervention Period
Day 7
2 weeks
Day 14
Day 21
Day 28
Post-Intervention Period
12
Study Procedures
Baseline
survey
Intervention
condoms
distributed
Training
session
Diary 1
Wrap up
Diary 2
Diary 3
Establish pre-intervention
condom use
Day −14
Day −7
Diary 4
Establish post-intervention
condom use
Day 0
Pre-Intervention Period
Day 7
2 weeks
Day 14
Day 21
Day 28
Post-Intervention Period
13
Data Analyses


Analyses conducted using data from 4 diaries
Based on 155 sexually active participants completed
both pre- and post-intervention diaries


Diaries 1 & 2 established pre-intervention condom use
Diaries 3 & 4 established post-intervention condom use
Condom use rate (CUR) = frequency of protected sex
total frequency of sex

Calculated CUR over a two week period each for pre- and postintervention periods
14
Statistical Analyses
Goal
Statistical Plan
Overall intervention effect,
accounting for paired data
Wilcoxon signed-rank test
Mixed-effects logistic regressions
(MELR)
Effect of wrapper and scent
on CURs, accounting for
correlated data
MELR
• Wrapper effect: Wrapper*time
interaction
• Scent effect: Scent treated as time
varying covariate
15
Demographics
n
(%)
Age in years, Mean (SD)
25.1
(2.4)
Marital status
Single
Married/Cohabitating
173
38
(82.0)
(18.0)
Education
Junior (secondary)
Senior (secondary)
Tertiary
Vocational
1
157
44
9
(0.4)
(74.4)
(20.9)
(4.3)
Religion
Christian
Traditional
African Traditional
No religious affiliation
Other non-Christian
174
10
7
19
1
(82.5)
(4.7)
(3.3)
(9.0)
(0.5)
16
Baseline Differences Between
Wrapper Groups

Blue wrapper group more likely to be


Single, never married
Educated (completed tertiary school or higher)
17
Overall Intervention Effect
Condom Use Rate (%)
100
95%
95
90
∆9*
86%
85
Crude MELR Model
80
75
*p < 0.001
Post vs. pre
Pre-Intervention
Time
OR
(95% CI)
3.48*
(2.46-4.92)
Post-Intervention
18
Overall Intervention Effect

What could explain increase in condom use over
time in study population?


Used intervention condoms highly preferred by BDF
Handed out condoms to participants in person, which
could have promoted condom awareness and
encouraged use
19
Condom Use Rates Over Time by
Wrapper
Condom Use Rate (%)
100
Cami Wrapper
95%
95
90
94%
88%
85
83%
80
75
p = 0.27
Blue Wrapper
Pre-Intervention
Time
Post-Intervention
20
Differential Change in Condom
Use Rates Over Time by Wrapper
Condom Use Rate (%)
100
Blue Wrapper
Cami Wrapper
95
90
85
Wrapper*time interaction p = 0.002
80
75
Pre-Intervention
Time
Post-Intervention
21
Wrapper Effect – Crude MELR
Model
Condom Use Rate (%)
100
Blue Wrapper
95
Cami Wrapper
OR = 2.13**
90
OR = 6.65***
85
OR Interaction = OR Cami = 6.65 = 3.12**
OR Blue
2.13
80
75
Pre-Intervention
*p < 0.05, **p < 0.01, ***p < 0.001
Time
Post-Intervention
22
Wrapper Effect

Why did camouflage wrapper have larger effect on
condom use?

Cami pattern is linked to military environment and unique
to BDF military
23
Condom Scent Effect – Crude
MELR Model
Unadjusted Model
Reported scented condoms
OR
95% CI
2.40***
(1.57-3.66)
*p < 0.05, **p < 0.01, ***p < 0.001
24
Final MELR Model – Effect of
Scent and Wrapper
Adjusted Model†
OR
95% CI
Reported scented condoms
2.28**
(1.40-3.71)
Camouflage wrapper
Post vs. pre
6.41***
(3.62-11.34)
Blue wrapper
Post vs. pre
1.98**
(1.25-3.13)
Wrapper*time interaction
3.24**
(1.56-6.76)
†Results
of mixed effects logistic regression analysis, adjusting for baseline differences
(marital status, education, military unit, sexual partner types, baseline condom brand), and
other variables in model
*p < 0.05, **p < 0.01, ***p < 0.001
25
Limitations & Strengths
Limitations
 Intervention not
randomized
 Lack of control group
 2-week diaries not long
enough to capture average
behaviors
 Short follow-up time to
examine intervention effect
Strengths
 High participation rate
 Innovative and novel
intervention
26
Summary


Condom use rates higher among those who
received condoms packaged in camouflage
wrappers
Condom use rates also higher among those who
reported using scented condoms
27
Take Home Message

Other militaries and organizations providing
free condoms should consider conducting
social marketing research in target population
regarding which condoms to distribute
28
Acknowledgments

Committee members






Richard Shaffer, PhD (chair)
Anne Thomas, PhD
Florin Vaida, PhD
Christina Chambers, PhD
Richard Haubrich, MD
Botswana Defence Force







DHAPP







Judy Harbertson, PhD
Marni Jacobs, MPH
Stan Ito, MPH
Jenny McAnany, PhD
Audrey Djibo, M.S.
Mike Grillo, PhD
Building 304
Participants
Maj. Mooketsi Ditsela
Capt. Timothy Maje
Capt. Segolame Phirinyane
Lt. Moitshepi Sepopo
Lt. Zibanani Elias

Population Services International

US Embassy Botswana

David Kelapile
29
Disclaimer

The views expressed in this article are those of the authors and do not
necessarily reflect the official policy or position of the Department of the
Navy, Department of Defense, or the U.S. Government. Human subjects
participated in this study after giving their free and informed consent. This
research has been conducted in compliance with all applicable Federal
Regulations governing the Protection of Human Subjects in Research.
30