HAART and Sexual Behaviour: A Comparison of HAART-Initiated and HAART-Naïve Clients in The AIDS Support Organisation, Uganda

AIDS

HAART and Sexual Behaviour: A Comparison of HAART-Initiated and HAART-NaïveClients in The AIDS Support Organisation, Uganda

--Manuscript Draft--

Manuscript Number:

Full Title: HAART and Sexual Behaviour: A Comparison of HAART-Initiated and HAART-NaïveClients in The AIDS Support Organisation, Uganda

Article Type: Original paper (Epidemiology / Social)

Keywords: HIV; AIDS; Sexual Behaviour; ART; HAART; Prevention

Corresponding Author: Francis Wasagami, MPH (cand.), M.Sc (MUK), B.Stat (MUK)The AIDS Support Organisation (TASO) Uganda LimitedKampala, UGANDA

Corresponding Author SecondaryInformation:

Corresponding Author's Institution: The AIDS Support Organisation (TASO) Uganda Limited

Corresponding Author's SecondaryInstitution:

First Author: Francis Wasagami, MPH (cand.), M.Sc (MUK), B.Stat (MUK)

First Author Secondary Information:

Order of Authors: Francis Wasagami, MPH (cand.), M.Sc (MUK), B.Stat (MUK)

Andy Beke, MBChB (Ghana), MMed (CommHealth) (Medunsa), FCPHM

Order of Authors Secondary Information:

Abstract: Objectives: This study sought to establish whether there is a relationship betweenHAART initiation and change in risky sexual behaviour using HAART-naïve clients as acontrol group.Design: A cohort study was retrospectively carried out. This design was appropriateboth ethically and to improve the validity of the findings through use of the appropriatecounter-factual.Methods: Stratified sampling with simple random sampling within strata was used toselect the sample of size 340 (170 on HAART and 170 HAART-naïve; 68 males and272 females) with HAART status as the stratifying variable. Only clients aged 15through 40 at the start of the study were included studied. Data analysis wasperformed using binary logistic regression.Results: HAART was not a significant predictor of change in risky sexual behaviour(OR: 0.976; 95% CI: 0.63 to 1.511; p=0.912). Significant co-factors were: baseline risklevel (adjusted OR for a 0.0001 change in baseline risk: 1.603; 95% CI: 1.181 to 2.175;p=0.003); age (adjusted OR: 1.11; 95% CI: 1.048 to 1.176; p<0.0001); being in amonogamous marriage (adjusted OR: 0.487; 95% CI: 0.263 to 0.9; p=0.022); being ina polygamous marriage (adjusted OR: 0.325; 95% CI: 0.13 to 0.811; p=0.016); andhaving never been in marriage (adjusted OR: 0.159; 95% CI: 0.031 to 0.833; p=0.03).Conclusion: There was no statistically significant evidence for HAART as a predictor forchange in sexual behaviour among TASO clients. Prevention efforts should beconcerned with aspects like age, marital status and current sexual risk levels of clients.

Suggested Reviewers:

Opposed Reviewers:

Powered by Editorial Manager® and Preprint Manager® from Aries Systems Corporation

Abstract

Objectives: This study sought to establish whether there is a relationship

between HAART initiation and change in risky sexual behaviour using HAART-naïve

clients as a control group.

Design: A cohort study was retrospectively carried out. This design was

appropriate both ethically and to improve the validity of the findings through use of

the appropriate counterfactual.

Methods: Stratified sampling with simple random sampling within strata was

used to select the sample of size 340 (170 on HAART and 170 HAART-naïve; 68

males and 272 females) with HAART status as the stratifying variable. Only clients

aged 15 through 40 at the start of the study were included studied. Data analysis

was performed using binary logistic regression.

Results: HAART was not a significant predictor of change in risky sexual

behaviour (OR: 0.976; 95% CI: 0.63 to 1.511; p=0.912). Significant co-factors were:

baseline risk level (adjusted OR for a 0.0001 change in baseline risk: 1.603; 95% CI:

1.181 to 2.175; p=0.003); age (adjusted OR: 1.11; 95% CI: 1.048 to 1.176;

p<0.0001); being in a monogamous marriage (adjusted OR: 0.487; 95% CI: 0.263 to

0.9; p=0.022); being in a polygamous marriage (adjusted OR: 0.325; 95% CI: 0.13 to

0.811; p=0.016); and having never been in marriage (adjusted OR: 0.159; 95% CI:

0.031 to 0.833; p=0.03).

Conclusion: There was no statistically significant evidence for HAART as a

predictor for change in sexual behaviour among TASO clients. Prevention efforts

Abstract

should be concerned with aspects like age, marital status and current sexual risk

levels of clients.

1

HAART and Sexual Behaviour: A Comparison of HAART-Initiated

and HAART-Naïve Clients in The AIDS Support Organisation,

Uganda

Authors: Francis WASAGAMI 1,2

, Andy BEKE 2

Affiliation: 1. The AIDS Support Organisation (TASO) Uganda

2. School of Health Systems and Public Health, University of

Pretoria; Email: [email protected]

Correspondence to: Francis Wasagami

Email: [email protected]

This study was funded under a Master of Public Health Fellowship offered to Francis Wasagami

by the USAID through MEASURE Evaluation Project

Total number of words: Abstract: 249; Text: 3,481

Article

2

Introduction

Background

Highly Active Antiretroviral Therapy (HAART) is known to delay progression

to full-blown AIDS disease and prolong the lives of persons with HIV. This may

revive a desire for sexual relationships. This revived sexual desire in the case of

persons living with HIV (PLWH) may result in increased spread of the virus.

A number of studies with varying results have been carried out to

scientifically assess changes in sexual behaviour after commencement of HAART

[1-4]. In a prospective cohort study to assess change in sexual risky behavior

among largely heterosexual HIV-positive persons on ART for 12 months in

Mombasa, Kenya, Sarna et al. [3] found no statistically significant difference in

self-reported sexual activity at baseline (48%) and at 12 months (58%) (p=0.07).

Similarly there was no significant change in the number of sexual partners at

baseline (90% had one partner) and at 12 months (94% with one partner). The

study design did not, however, allow comparison with HIV-positive people not on

HAART to enable conclusions on whether the risk of transmission was higher or

lower for non-HAART clients. For this reason, it could not be deduced whether

HAART clients had a different level of sexual urge or engage in a higher number of

encounters than their counterparts who were not HAART.

In a study carried out in rural Uganda [2], a cohort of 926 PLWH were

started on HAART and prospectively followed for six months to assess changes in

3

their sexual behaviour. In this study, Bunnell et al. [2] found that risky sexual

behavior dropped by up to 70% (adjusted risk ratio, 0.3; 95% confidence interval

(CI), 0.2-0.7; P = 0.0017) in six months. A possible limitation of this study was that

the follow-up period of only six months was too short for clients to exhibit

objectively observable changes. Besides, no comparison group was used for

comparison of the observed change in risky sexual behaviour so as to draw more

valid conclusions. Similar to Sarna et al.’s [3] study, Bunnell et al. [2] did not

compare the study participants with other PLWH receiving all other services

except HAART, which meant that these researchers could not conclusively deduce

that the 70% reduction in sexual risky behavior was significantly due to HAART.

In a study [5] done in UThukela health district in KwaZulu-Natal, South

Africa, Peltzer et al [5] found a significant reduction in risky sexual behavior after

initiation of HAART of and following up 735 PLWH at 6 and 12 months. This result

supports that of Bunnell et al [2]. Similarly, the design of the Peltzer et al [5] study

did not allow for comparison with PLWH who were not on HAART during the study

period.

The above studies show that the effect of HAART on sexual risk behavior is

varied. While some studies have found no significant change in sexual risky

behavior after HAART initiation, others have either reported an increase or a

reduction in the risky behavior. Also notable is the fact that all these studies did not

have control groups to enable more valid conclusions on the effect of HAART on

sexual behaviour.

4

The objectives of the current study were to determine and compare levels

of risky sexual behaviour of clients on HAART with those for clients not on HAART

during the baseline period (2005 – 2007); and to determine and compare changes

in the level of risky sexual behaviour for HAART and HAART-naïve clients after 24

months of care in TASO clinics (2007 to 2009). The study hypothesized that there

is no difference between clients on HAART and those not on HAART in terms of

change in risky sexual behaviour for clients receiving HIV/AIDS services in a

health service delivery setting of TASO.

Sexual Risk Index

This study used a technique suggested by Bell and Trevino [6] to quantify

clients’ two-year sexual risk indices during and after the base period. The Bell and

Trevino [6] sexual risk index (SRI) is given by:

……………..…… (1)

where;

is the probability (risk) of person i being infected through sexual

contact with partner k,

is the HIV prevalence within the sub-population from which sexual partner k

is selected,

is the known (published) probability (risk) of HIV transmission through

sexual behaviour of type j and

is the number of sexual contacts/events with partner k and practicing

sexual behaviour of type j.

5

If client i has sexual events with n partners in a given period of time, the overall

sexual risk of is given as,

………………………. (2)

Because TASO is a routine service delivery setting whereby data are not

gathered on all variables of interest, the following assumptions were made in order

to compute the sexual risk index:

1. The number of sexual events in each recall period (three months) for

every self-reported sexual partner is 1;

2. The number of sexual partners, n, in any period of consideration is equal to

the cumulative number of sexual partners reported for that entire period

irrespective of whether the reported partner in any one 3-month period was

reported in an another 3-month period;

3. The sub-population prevalence is equal to the national prevalence and

4. We made the assumption that all reported sexual events were heterosexual

and vaginal [7] i.e. , the probability of HIV transmission through

vaginal sex, which depends on whether sex was insertive or receptive, and

whether a condom was used.

With the above-mentioned assumptions, equation (1) becomes,

………………………….. (3)

6

Because we utilised routine service data, the number of quarterly records

on sexual activity per client per year were less than the expected four for some

clients since clients are under no obligation to make contact with clinic staff at

least once every quarter. Clients with missing data for certain quarters may have

lower risk indices, which may makes them appear to be at lower risk than those

with more records. The missing data would render comparison of the clients

invalid.

There was therefore a need to prepare the study data in such a way as to

ensure validity of comparisons made. To achieve this, a mean quarterly sexual risk

index (SRI) was computed for each client for each year using the available

quarterly SRIs. The resulting mean quarterly SRI was used as the SRI for the

quarters where data were missing for the particular year. In the cases where there

was completely no data on sexual risk in a given year, the quarterly risk indices for

that year were estimated using the quarterly average of the nearest year to the left

of the affected year where quarterly risk data were available.

Since sexual activity in TASO is reported using 3-monthly recall method,

only one record per quarter of a given year was considered in the analysis in case

the client had more than one record for a given quarter. In this case the record

with the smallest response code to the question "Have you had sex in the last

7

three months? 1.Yes 2. No 3. Not applicable" was considered in the estimation of

that quarter's sexual risk index.

To adjust for variation of risk according to type of sexual partner, the

weights were attached to types of sexual partners. A weight of 1 was used for

casual partners, 0.5 for steadies, 0.1 for spouses, and 0 if the client did not have

sex with anyone; in this case n is taken as equal to 1 for computation purposes.

The estimated prevalence for the sub-population from which partner k is selected

is therefore given by . Applying this prevalence in equation (3) gives a finer

estimate for the risk posed by partner k, as shown in equation (4).

……………………………………………. (4)

Equation (4) provides an estimate of the sexual risk of infection/re-infection

by partner k in a routine health care environment and was used with equation (2)

to compute the estimated sexual risk index for each participant in the study. That

is,

8

Methods

Study Design

A cohort study was retrospectively carried out. The change in risky sexual

behaviour for clients on HAART was compared with the change in risky sexual

behaviour among clients not initiated on HAART during the study period.

Change in risky sexual behaviour was determined by comparing the sexual

risk index two years before HAART initiation and two years after HAART initiation

for clients on HAART; and two years before the reference quarter and two years

after the reference quarter for HAART-naïve clients. The reference quarter was the

quarter of HAART initiation 2007 for clients on HAART while for all HAART naïve

clients, it was taken to be 15th June 2007 since this is mid-way between January

2005 and December 2009.

Clients in both groups were those who had been registered for care at

TASO in or before 2005 and had been retained in care up to the end of 2009.

Retention in care for the purpose of this study meant receiving at least one

counselling consultation per year from 2005 to 2009 and not being reported dead.

The treatment group was the cohort of clients initiated on HAART in 2007. The

control/comparison group comprised of clients who did not start HAART until after

2009.

9

Study Population

Clients started on HAART between January and December 2007 who meet

the inclusion criteria formed the sub-population for the treatment group. For the

control group, HAART-naïve clients who meet the inclusion criteria were eligible

for the study.

Inclusion and exclusion criteria

Records for clients started on HAART between January and December

2007, who had been receiving services in TASO for at least two years at the time

of HAART initiation; and remained on HAART up to 2009, were included in the

study as the treatment group. Clients for the control group were those who had

been in care by 2005, were retained in care in TASO till 2009 and did not initiate

HAART until after 2009.

Both categories included only clients aged between 15 through 40 by

January 2005 who received at least one counselling consultation per each of the

study years (2005 to 2009) and were not reported dead during the study period.

This was to ensure that valid comparisons and conclusions are made.

Clients who did not remain alive until the end of 2009 were excluded from

the study. Also excluded were clients stopped from taking HAART, lost to follow-

up or were not retained in care during the study period.

10

Sample Size and Sample Selection

Stratified sampling methodology using HAART status as the stratifying

variable was used. For each stratum an equal number of records was selected

using simple random sampling technique.

A precision of 10% was targeted in all parameter estimations at a 95%

confidence level with a significance level of 0.05. Assuming a design effect due to

stratification of 2; 80% file recovery rate (analogous to response rate) in case of

need for data verification, and the proportion of clients at risk of sexual risky

behaviour i.e. p, of 50%, the sample size per stratum was determined to be 170.

That is, 170 records for clients on HAART and 170 records for HAART-naïve

clients were randomly selected resulting in a total of 340 clients.

Data Management and Analysis

Data analysis was done using Stata 11 [8] statistical software after random

sampling using Microsoft Excel [9]. For each client, his or her sexual risk index

was computed for the period before the cut-off quarter (base quarter) in the case

of the treatment group; and before 15th June 2007 for the control group. The

sexual risk index was also computed for the period after the base date/quarter. For

a given client the difference between the two sexual risk indexes was computed.

That is, the sexual risk index after the base date/quarter minus the sexual risk

index before the base date/quarter.

11

For each participant, a decreased or constant risk at zero level (the desired

outcome) was assigned a value of 1 while an increased or constant non-zero risk

was assigned a value of 0 (the undesired outcome). Binary logistic regression was

used to adjust for the effect of covariates such as sex, age-group and number of

counselling sessions on the relationship between HAART status and change in

sexual risk. Complex sample analysis using weights was performed with a weight

of 20 for non-HAART clients and a weight of 5 for clients on HAART.

Ethical Considerations

This study was reviewed and evaluated by the relevant bodies for ethical

and scientific merit. Approvals were obtained from the Faculty of Health Sciences

Research and Ethics Committee; TASO Institutional Review Committee; and from

the Uganda National Council for Science and Technology (UNCST).

Permission was granted by TASO to access and use their routine

healthcare data. To ensure confidentiality, no personal identifiers such as names

and registration numbers of clients were included in the report. Only the

researchers directly involved in this study had access to the data and only for the

purpose of this study.

Funding of this study was part of the Fellowship offered to Francis

Wasagami by MEASURE Evaluation. Francis has worked with TASO Uganda

since 2002. Prof Andy Beke was Francis’ supervisor and is lecturer in the School

of Health Systems and Public Health at the University of Pretoria.

12

Results

Table 1 shows the distribution of the sample by sex and HAART status. The

mean age at the beginning of study period for clients on HAART was 33.1 years

(95% CI: 32.3 to 33.8) while that of the HAART-naïve was 32.1 years (95% CI:

31.3 to 32.8). The rank-sum test showed that clients in both groups were similar in

terms of the baseline sexual risk index (p= 0.1605).

After adjusting for sex, education level, religion, TASO branch, alcohol use,

number of children alive, baseline sexual risk index, age, number of counselling

session in the study period, duration in care, occupation and residence, HAART

was not a significant predictor of change in risky sexual behaviour (OR: 0.976;

95% CI: 0.63 to 1.511; p=0.912). Instead, significant co-factors were: baseline risk

(adjusted OR for a 0.0001 change in baseline risk: 1.603; 95% CI: 1.181 to 2.175;

p=0.003); age of the client (adjusted OR: 1.11; 95% CI: 1.048 to 1.176; p<0.0001);

being in a monogamous marriage (adjusted OR: 0.487; 95% CI: 0.263 to 0.9;

p=0.022); being in a polygamous marriage (adjusted OR: 0.325; 95% CI: 0.13 to

0.811; p=0.016); and having never been in marriage (adjusted OR: 0.159; 95% CI:

0.031 to 0.833; p=0.03).

A rank-sum test did not reveal a significant difference in two-year risk at

baseline for the two study groups (p= 0.295). A similar test did not show a

significant difference between the two groups in terms of the two-year sexual risk

index at the end of the study period (p= 0.88)

13

The estimated percentage of HAART initiated clients that experienced a

reduced risk or constant zero-risk sexual behaviour was 60% (95% CI: 52.59% to

67.41%) while that for HAART-naïve clients was 61% (95% CI: 53.19% to

67.98%).

The test for equality of proportions of clients that experienced a reduced

risky sexual behaviour among the treatment and comparison groups showed that

the two proportions were not statistically different at 5% level of significance (p=

0.9117).

14

Discussion

The findings of this study are consistent with those of Sarna et al. [3] who

found no statistically significant difference in self-reported sexual activity at

baseline (48%) and at 12 months (58%) (p=0.07) in their Mombasa, Kenya study

despite the design of Sarna et al ’s [3] study (prospective cohort with no control

group) being different from the current study’s design (retrospective cohort with a

control group). Similarity of Sarna et al’s [3] findings may therefore be purely

coincidental.

The findings of the current study however contrast with those of Bunnell et

al. [2] who attributed to HAART initiation a 70% (adjusted risk ratio, 0.3; 95%

confidence interval (CI), 0.2-0.7; p= 0.0017) drop in risky sexual behaviour after

six months of the initiation in their rural Uganda study. The lack of a HAART naïve

control group in the Bunnell et al. [2] study might have led to their contrasting

conclusion.

Findings from the current study also contrast with those established by a

study involving men who have sex with men conducted in Amsterdam [4] where

HAART initiation was associated with increased risky sexual behavior among the

participants. The lack of a HAART naïve control group in this Amsterdam study [4],

like in the Bunnell et al. [2] study, might also have led to this finding that is different

from that established by our study.

15

Another study with findings that contrasted with the results of this study was

that carried out by Peltzer et al. [5] in Uthukela health district in KwaZulu-Natal,

South Africa, which showed a significant reduction in risky sexual behavior after

starting HAART. In this study, Peltzer et al. [5] followed up 735 PLWH at 6 and 12

months after initiation on HAART but they did not do the same for HAART naïve

PLWH. This might have rendered their attribution of the reduction in risky sexual

behavior to HAART less reliable.

In a meta-analysis of 12 controlled trials in the USA, Crepaz et al.[10] found

a significant effect of behavioral interventions on change in sexual behaviour of

PLWH. For those studies that established a change in risky sexual behavior

[2,4,5], the observed change may therefore be attributed to the

behavioural/psychosocial sub-components and other aspects of the respective

HIV/AIDS programmes; as well as to personal and environmental attributes of the

PLWH rather than to HAART initiation.

Emerging Predictors of Change in Risky Sexual Behaviour in TASO

The factors that emerged as statistically significant for change in risky

sexual behaviour were: baseline risk (adjusted OR for a 0.0001 change in baseline

risk: 1.603; 95% CI: 1.181 to 2.175; p=0.003); age of the client (adjusted OR: 1.11;

95% CI: 1.048 to 1.176; p<0.0001); being in a monogamous marriage (adjusted

OR: 0.487; 95% CI: 0.263 to 0.9; p=0.022); being in a polygamous marriage

(adjusted OR: 0.325; 95% CI: 0.13 to 0.811; p=0.016); and having never been in

marriage (adjusted OR: 0.159; 95% CI: 0.031 to 0.833; p=0.03).

16

This study established that baseline sexual risk index (SRI) was positively

associated with reduced risky sexual behaviour. In other words, clients with higher

levels of risky sexual behaviour at baseline were more likely to experience a

reduction in risky sexual behaviour than those with lower baseline sexual risk. It

appears that the psychosocial (counselling) component of TASO HIV/AIDS

programme had a profound effect of causing a reduction in risky sexual behaviour

among the more-at-sexual-risk than among the less-at-sexual-risk clients. This

may have resulted from preventive counselling efforts being directed more towards

clients thought to be more-at-risk of risky sexual behaviour, leaving those

perceived to be less-at-risk to deteriorate in terms of risky sexual behaviour since

counselling has been shown to have an effect on sexual behaviour of HIV positive

persons [11].

Another factor that emerged as significant for reduced risky sexual

behaviour was the age of the client. Age was positively related to reduced risky

sexual behaviour, that is, the older the clients the more likely they were to

experience reduced risky sexual behaviour behaviour. This may be attributed to

reduction in sexual desire with age [12] which may as such put older clients at a

lesser risky sexual behaviour than young clients.

The study also found a statistically significant relationship between being

married and change in sexual risky behaviour. Both the clients in monogamous

and polygamous marriages were less likely to reduce their sexual risky levels

(respective adjusted OR of 0.487 and 0.325). This can be attributed to married

17

people’s high likelihood to engage in sex acts quite often in any given quarter; and

also to a possibility that the married were engaging in risky extra-marital affairs.

This result is consistent with the reports which showed increased risk of HIV in

marriages in Uganda [13].

Finally, the study found that clients who reported that they had never been

married at the time of registration into care in TASO were less likely to reduce their

risky sexual behaviour (adjusted OR: 0.159). This may be attributed to the

possibility of engaging in risky sexual acts during their search for possible

marriage partners and also involvement in sexual activities for pleasure.

In conclusion, this study established that there is no statistically significant

evidence for HAART as being associated with change in sexual behaviour. The

results from the study showed that changes in sexual behaviour after HAART

initiation observed in other studies may have been caused by different

programmatic and environmental factors rather than being on HAART.

18

Acknowledgements

Thanks TASO Management for allowing this study to be conducted in the

organisation. Also acknowledged is the support from TASO staffs who enabled

access to the data and the clients’ files. To Prof Girdler-Brown and Dr. Stephanie

Mullen, thanks for all your technical and statistical guidance.

The study was funded under a Fellowship from the USAID-funded

MEASURE Evaluation Project

19

References

1. Seeley J, Russell S, Khana K, Ezati E, King R, Bunnell R. Sex after ART:

sexual partnerships established by HIV-infected persons taking anti-

retroviral therapy in Eastern Uganda. Cult Health Sex. Oct 2009; 11(7):703-

16. Accessed on 7th March 2011 at

http://www.ncbi.nlm.nih.gov/pubmed/19544115.

2. Bunnell R, Ekwaru JP, Solberg P, Wamai N, Bikaako-Kajura W, Were W.

Changes in sexual behavior and risk of HIV transmission after antiretroviral

therapy and prevention interventions in rural Uganda. AIDS. 2nd January

2006; 20(1):85-92. Accessed on 7th March 2011 at

http://journals.lww.com/aidsonline/Fulltext/2006/01020/Changes_in_sexual_

behavior_and_risk_of_HIV.12.aspx

3. Sarna A, Chersich M, Okal J, Luchters SM, Mandaliya KN, Rutenberg N,

Temmerman M. Changes in sexual risk taking with antiretroviral treatment:

influence of context and gender norms in Mombasa, Kenya. Cult Health

Sex. 2009 Nov; 11(8):783-97.

4. Dukers NHTM, Goudsmit J, de Wit JBF, Prins M, WeverlingGJ, Coutinho

RA. Sexual risk behaviour relates to the virological and immunological

improvements during highly active antiretroviral therapy in HIV-1 infection.

AIDS. 16 February 2001; 15 (3):369-378.

20

5. Peltzer K, Ramlagan S. Safer sexual behaviours after 1 year of

antiretroviral treatment in KwaZulu-Natal, South Africa: a prospective cohort

study. Sex Health. 2010 June; 7(2):135-41.

6. Bell DC, Trevino RA. Modeling HIV Risk. JAIDS. 1 November 1999; 22

(3):280. Accessed on 22 May 2011 at

http://journals.lww.com/jaids/Fulltext/1999/11010/Modeling_HIV_Risk.10.as

px.

7. Uganda Ministry of Health. Uganda HIV Prevention Response and Modes

of Transmission Analysis, March 2009.

8. StataCorp. Stata Statistical Software: Release 11. 2009. College Station,

TX: StataCorp LP.

9. Microsoft Corporation Inc. USA. Microsoft Excel 2007.

10. Crepaz N, Lyles CM, Wolitski RJ, PassinWF, Rama SM, HerbstJH, et al.

Do Prevention Interventions Reduce HIV Risk Behaviors Among People

Living With HIV? A Meta-Analytic Review of Controlled Trials. AIDS. 2006;

20(2):143-157. Accessed on 8th March 2011 at

http://www.medscape.com/viewarticle/521053.

11. Richardson JL, Milam J, McCutchan A, Stoyanoff S, Bolan R, Weiss J,

Kemper C et al. Effect of brief safer-sex counseling by medical providers to

HIV-1 seropositive patients: a multi-clinic assessment. AIDS. 21 May 2004;

18 (8):1179-1186.

12. Delamater JD, Sill M. Sexual desire in later life. Journal of Sex Research.

2005; 42 (2) : 138 – 149.

21

13. Namutebi J. New Vision Newspaper, Uganda. June 13, 2007. Newspaper

article accessed on 18th June 2012 at

http://aidscarewatch.blogspot.com/2007/06/uganda-married-women-at-

higher-risk-of.html.

Tables

Table 1. Distribution of Participants by HAART status and Sex

Sex HAART-naïve HAART Total

Female 146 126 272

Male 24 44 68

Total 170 170 340

Table

Francis WasagamiClick here to download LWW Copyright Transfer and Disclosure Form: copyrightTransferWASAGAMI.pdf

Andy BekeClick here to download LWW Copyright Transfer and Disclosure Form: copyrightTransfer_Andy Beke.pdf

AIDS: Author’s paper submission checklist

Title of paper: ► HAART and Sexual Behaviour: A Comparison of HAART-Initiated

and HAART-Naïve Clients in The AIDS Support Organisation, Uganda

Names of

authors:

► Francis Wasagami; Andy Beke

AUTHORS SHOULD PLEASE ENSURE THAT ALL APPROPRIATE

INFORMATION (EG. CONFLICT OF INTEREST STATEMENTS) ARE

ALSO INCLUDED IN THE TEXT OF THE ARTICLE.

1. DUPLICATE PUBLICATION is not acceptable and includes papers, or letters to

the Editor reporting the same data previously published in any journal. Abstracts of

papers presented at meetings and published in the proceedings of such meetings do

not constitute duplicate publication, but should be disclosed by including a note at the

beginning of the paper, i.e.. "Data presented previously at (state meeting) and

published as abstract in (give reference)". Have you published these data previously?

► No

2. CONFLICT OF INTEREST include financial support from the biomedical

industry or other commercial sources in the form of research grants, bench fees,

consultancy or lecture fees, travelling expenses, payment of registration fees,

consultancy appointments, posts held in the biomedical industry or equipment

manufacturers, stock holdings in the company, free supply of drugs and the like.

These should be stated in relation to each author. Has any of the authors any conflict

of interest? Please state details.

► This study was funded through a Fellowship offered to Francis Wasagami by the USAID

through MEASURE Evaluation Project. Francis Wasagami has worked with The AIDS

Support Organisation (TASO). Prof. Andy Beke was an academic supervisor to Francis

3. CONSENT Please note that patient's, or normal control’s, written consent is

needed not only for full papers, but also for case reports. The written consent needs to

include not only agreement to undergo treatment, or participate in an experiment or an

randomised control trial, but also agreement for anonymised data to be published in a

scientific journal. Was patient’s consent obtained and in what form?

► Because this study used secondary data, no consent was obtained from patients. However all necessary ethical approvals were obtained and the principles strictly adhered to in the

handling and use of these data.

4. ETHICS All studies need to be approved by the local Ethical Committees. Was

your study? Please provide the approval from your local Ethical Committees for any

animal experimentation or human subject studies.

► This study was approved by: University of Pretoria's Faculty of Health Sciences Research and Ethics Committee, The AIDS Support Organisation's Institutional Review Committee and

Uganda National Council for Science and Technology

Author's Checklist

5. AUTHOR’S CONTRIBUTIONS AND APPROVAL OF TEXT Please state

briefly how each of the authors contributed to the study, to data analysis and to the

writing of your paper. Subject to your agreement, we will print this information, if the

paper is accepted for publication. In addition, please confirm that all the authors have

read and approved the text as submitted to AIDS. Justify individual’s contributions

when the author list exceeds 10.

► Francis developed the research protocol, prepared and analysed the data; and wrote the report. Prof Andy Beke conceptualised and finalised the protocol, revised statistical methods,

and revised all stages of drafting of the manuscript

6. STATISTICAL ANALYSIS Kindly please let me know who performed the

statistical analysis of you data.

► The data analysis was done by Francis Wasagami with some guidance from Prof Brendan

Girdler-Brown from School of Health Systems and Public Health, University of Pretoria

7. ARTICLE LENGTH Please acknowledge that you have kept within the word and insert

limits for your submission by ticking the relevant boxes below, and that you have indicated

the word count on the title page of your article

Article length

√ Original papers, 3500 words of text excluding references with no more than five inserts

(figures/tables)

Concise communications, 1800 words of text excluding references with maximum of two

inserts (figures/tables)

Research Letters, 1000 words excluding summary with no more than one insert (figure/table)

Correspondence, 750 words excluding references with no more than one insert (figure/table)

Title page

√ Title no more than 120 characters

Running head, no more than 40 characters

Justification of the number of contributors greater than 10 included in this document

Abstract page

√ Abstract no more than 250 words

√ 5–7 keywords listed

Word count

√ Word count of text (excluding references) included on title page

8. CLINICAL TRIALS AND BEHAVIOURAL EVALUATIONS: Authors reporting

results of randomized controlled trials should include with their submission a complete

checklist from the CONSORT statement, see JAMA 1996; 227:637-639 or

http://www.consort-statement.org. For behavioural and public health evaluations involving

non-randomized designs, authors should include with their submission a complete checklist

from the TREND statement, see Am J Public Health 2004; 94:361-366 or www.trend-

statement.org .

REGISTRATION OF CLINICAL TRIALS: As a condition for publication of a clinical

trial in AIDS, registration of the trial in a public registry is required. Registration of a trial must be at or before the enrollment of participants. The editors of AIDS also do not advocate

one particular registry but require that the registry utilized meet the criteria set out in the

statement of policy of the ICMJE (www.ijcme.org).

Please state that your article includes a clinical trial and that the conditions of submission

above have been met.

► Not Applicable

Other information for the Editor:

►

Name of person completing this form:

► Francis Wasagami

Date: ►26/06/2012