Presentation 20april 000

Condom use at first and most recent sex among unmarried people aged 15 to 24 in Uganda: A case of Kabale and Mukono districts: 

Condom use at first and most recent sex among unmarried people aged 15 to 24 in Uganda: A case of Kabale and Mukono districts Tumwesigye, NM Ingham, R Holmes, D A presentation in a seminar organized by the Centre for AIDS Research (CAR) University of Southampton on 20th April 2005

Outline of presentation: 

Outline of presentation Introduction Background Objectives Previous work Methods Results Discussion and conclusion

Introduction: 

Introduction Condoms offer best protection against HIV/STD and unplanned pregnancies for those that can’t abstain Consistent and correct use of condoms can prevent 90-95% of HIV infections There are claims and counter claims for their role HIV reduction in Uganda - Singh et al. 2003 that all Abstinence, being faithful, and condom (ABC) were equally important Vs Green et al. 2002 for AB and Wawer 2005 for C

Introduction cont’d: 

Introduction cont’d But few consistently use condoms: In field trial, Rakai- Uganda 4.4% (Ahmed, 2001), Bar and hotel workers, Tanzania 14% (Ao et al. 2003). Current condom use among young unmarried women is 41% and men 58% (UDHS 2000/1) higher than Kenya (men: 43%) and Ghana (Men: 38%) Little is known about consistency and changes in condom use from first to most recent sex among young people

Why young people?: 

Why young people? 56% of population is less than 18 years 20% of the population is aged 15-24 years Of those aged 15-24, 64% are unmarried Median age at first marriage is 18.1 for females and 22.8 for males [UDHS2000/1] Two thirds of women get pregnant by age 18 HIV prevalence is 4.9% among ANC clients aged 15-24 [MOH]

Why first and most recent event? Why Kabale and Mukono?: 

Why first and most recent event? Why Kabale and Mukono? Most easily remembered events Strong link between first use and life time use: 20 fold increase in life time and 10 fold increase in current use if condom was used at first sex (Miller 1998) Kabale and Mukono represent differences in urbanisation, economic level, health facilities. Mukono has better indicators

Map of Uganda showing Mukono and Kabale districts: 

Map of Uganda showing Mukono and Kabale districts Kabale Mukono Kampala

Objectives: 

Objectives Establish level of condom use at first and most recent sex Identify factors associated with condom use at each of the two events Compare influence of the factors at the two events Determine predictors of consistent condom use for combined events

Literature: Condom use at First sex: 

Literature: Condom use at First sex Level of condom use: A third in Lesotho (FHI et al. 2002) More likely to be reported by women (FHI et al. 2002) Increases with age, lower with alcohol (Grunseit,1999) Higher with more education (Abma andamp;Sonenstein, 2001) Higher with casual than regular partners (Manlove, 2001) More likely in urban areas and among those who attend religious services and more communication with parents (Abma andamp;sonenstein 2001, Karim et al.2003)

Literature: Condom use at most recent sex: 

Literature: Condom use at most recent sex Level of use: 41% women and 58% men (UDHS,01) Men more likely to report use andamp; increases with age (Fadiora et al. 2003) increases with education level (Karim et al.2003) lower with alcohol consumption (Hall et al. 1990) More use with casual than regular partners (Sheeran andamp; Abraham 1994) higher with parental support, in urban areas, attending religious services (Meekersandamp;Klein 2003 Hingson, 1990; Najjumba at al. 2003, Mensch,, 1999)

Literature: Consistent condom use: 

Literature: Consistent condom use 15 % among STD patients in a hospital, Uganda (Nuwaha et al. 1999) 2% among traditional healers in Uganda (Ssali et al. 2002) 14% among urban male workers in Zimbabwe (Meekers 2003) Consistent use is higher with age at first sex (Tavares, 2004) alcohol consumption (Ray, 1998), casual partnership (Catania et al. 1998)

Gap in Literature: 

Gap in Literature Most studies reviewed were from USA, Europe, Australia. Few in Africa like one in Lesotho, Ghana, South Africa, Zimbabwe. The studies are mostly on selected risk groups like patients/FP clinic clients, students Few are specific to condom use and young people Nothing seen yet on condom use at first sex in Uganda

Methods: 

Methods Primary data-Survey in Mukono and Kabale districts: Questions from UDHS, WHO andamp; modified with Pilot –May 2004 Sample size was 786 but only 445 ever had sex Sampling: cluster by village and systematic for household in 94 villages Analysis: binomial andamp; multinomial Logistic Regression Models. Consistency measured by condom use at first and most recent sex Model building: Backward andamp; forward

Description of the sample: 

Description of the sample

Level of condom use at first and most recent sex: 

Level of condom use at first and most recent sex

Condom use at first and most recent sex: 

Condom use at first and most recent sex

Logistic Reg. Models for condom use at first and latest sex: 

Logistic Reg. Models for condom use at first and latest sex

Consistency of condom use: 

Consistency of condom use

Multinomial model predicted probabilities of condom use : 

Multinomial model predicted probabilities of condom use

Discussion: 

Discussion Condom use at first sex is higher than in other countries in Africa like Ghana (18% men , 27% women) and Lesotho (34% men) Condom use at most recent sex higher than the country average in Mukono but lower in Kabale Women more likely to use condom at first sex than men. Similar results in Karim et al 2003, FHI et al. 2002

Discussion cont’d: 

Discussion cont’d Condom use at first sex is higher in Mukono district, among women (Mukono), in steady relationship, among those in school, among those attending public schools and with higher age at first sex Age at first sex, being in school, women cited in other studies (Karim et al.2003 Grunseit, 1999) Condom use in steady relationship differs from Manlove et al. 2001 Attending public schools- not surprising Women are more likely to use condom because their partners are older (16 Vs 20 years) are in steady relationship

Discussion cont’d: 

Discussion cont’d Condom use at most recent event is higher in Mukono district, among males, with lower age at first sex, among those in school, among those who listen to radio, among those who use condoms at first sex, in steady relationship Higher condom use at lower age at first sex different from several studies including Abma et al 2001 and Benefo 2004. Probably, getting to trust their partners Similar effect of radio, condom use at first sex cited in many studies including Yoder at al 1997 in Zambia, Abma et al. 2001,

Discussion cont’d: 

Discussion cont’d Reduced influence by most recent sex Age at first sex School foundation Type of partner Increased influence by most recent sex Gender. Radio

Discussion cont’d: 

Discussion cont’d Consistent use is higher in Mukono, probably due to accessibility among those in school (women only), probably due to influence of condom use at first sex where among men alcohol is less influencial with higher age at first sex same in Tavares, 2004 among non-alcohol consumers same as Ray, 1998

Conclusions and implications: 

Conclusions and implications Less condom use in Kabale and probably similar upcountry districts calls for more sensitization First relationships for young men tend to be casual or one time partnership. Important in HIV/STD sensitization Less condom use in mission and private schools. Design of ways to reach young people in the schools Keeping young people in school is protective against risky sex, hence support for Universal Primary Education (UPE) and proposed universal Secondary Education (USE)

Conclusions and implications: 

Conclusions and implications Increase in condom use with age at first sex renders support for campaigns for delayed first sex (primary abstinence) Less condom use in mission and private schools calls for ways of reaching young people in the schools Communication through radios continue to be effective Evidence of effect of alcohol on risky sex is important for health education messages

Constraints: 

Constraints Cross-sectional design: Attributes at latest event may not be the same with first event. E.g taking alcohol and listening to radio. Possible memory lapse for events at first sex especially if not consensual

Work in progress: 

Work in progress Multi-level effects –village andamp; parish Qualitative data to answer why Religious and private school less likely to use Being in school is protective Alcohol consumption Differences by district You may contact the author at nmbonat@soton.ac.uk