Downloaded from http://bmjopen.bmj.com/ on November 30, 2015 - Published by group.bmj.com Open Access Research Incentivising safe sex: a randomised trial of conditional cash transfers for HIV and sexually transmitted infection prevention in rural Tanzania 102459 Damien de Walque,1 William H Dow,2 Rose Nathan,3 Ramadhani Abdul,4 Faraji Abilahi,4 Erick Gong,5 Zachary Isdahl,6 Julian Jamison,7 Boniphace Jullu,4 Suneeta Krishnan,8 Albert Majura,4 Edward Miguel,6 Jeanne Moncada,9 Sally Mtenga,3 Mathew Alexander Mwanyangala,4 Laura Packel,10 Julius Schachter,9 Kizito Shirima,3 Carol A Medlin11 To cite: de Walque D, Dow ABSTRACT ARTICLE SUMMARY WH, Nathan R, et al. Objective: The authors evaluated the use of Incentivising safe sex: conditional cash transfers as an HIV and sexually Article focus a randomised trial of transmitted infection prevention strategy to incentivise conditional cash transfers for - Existing prevention strategies have had a limited safe sex. impact on the trajectory of the HIV/AIDS HIV and sexually transmitted infection prevention in rural Design: An unblinded, individually randomised and epidemic. Tanzania. BMJ Open 2012;2: controlled trial. - Conditional cash transfers have been used e000747. doi:10.1136/ Setting: 10 villages within the Kilombero/Ulanga successfully in a variety of settings to promote bmjopen-2011-000747 districts of the Ifakara Health and Demographic activities that are beneficial to the participants, Surveillance System in rural south-west Tanzania. such as school participation or health check-ups Participants: The authors enrolled 2399 participants, for children. < Prepublication history and aged 18e30 years, including adult spouses. - This trial asks whether conditional cash transfers additional materials for this Interventions: Participants were randomly assigned can be used to prevent people from engaging in paper are available online. To activities that are harmful to themselves and to either a control arm (n¼1124) or one of two view these files please visit others, such as unsafe sex. the journal online (http:// intervention arms: low-value conditional cash transfer bmjopen.bmj.com). (eligible for $10 per testing round, n¼660) and high- Key messages value conditional cash transfer (eligible for $20 per - We designed and evaluated a novel intervention Received 9 December 2011 testing round, n¼615). The authors tested participants that tests for risky sexual behaviour repeatedly Accepted 19 December 2011 every 4 months over a 12-month period for the over short time intervals, reinforcing learning presence of common sexually transmitted infections. This final article is available about safer behaviour with cash transfer incen- In the intervention arms, conditional cash transfer for use under the terms of tives conditional on testing negative for a set of payments were tied to negative sexually transmitted the Creative Commons curable sexually transmitted infections (STIs). infection test results. Anyone testing positive for Attribution Non-Commercial - After 12 months, the results from the adjusted a sexually transmitted infection was offered free 2.0 Licence; see model showed a significant reduction in the http://bmjopen.bmj.com treatment, and all received counselling. combined point prevalence of the four curable Main outcome measures: The primary study end STIs tested every 4 months by nucleic acid point was combined prevalence of the four sexually amplification tests in the group that was eligible transmitted infections, which were tested and reported for the $20 payments, but no such reduction was to subjects every 4 months: Chlamydia trachomatis, found for the group receiving the $10 payments. Neisseria gonorrhoeae, Trichomonas vaginalis and - The results suggest that conditional cash trans- Mycoplasma genitalium. The authors also tested for fers used to incentivise safer sexual practices are HIV, herpes simplex virus 2 and syphilis at baseline a potentially promising new tool in HIV and STIs and month 12. prevention. Additional larger study would be Results: At the end of the 12-month period, for the useful to clarify the effect size, to calibrate the combined prevalence of any of the four sexually size of the incentive and to determine whether transmitted infections, which were tested and reported the intervention can be delivered cost effectively. every 4 months (C trachomatis, N gonorrhoeae, T For numbered affiliations see vaginalis and M genitalium), unadjusted RR for the end of article. high-value conditional cash transfer arm compared to transfer arm compared to the low-value conditional controls was 0.80 (95% CI 0.54 to 1.06) and the cash transfer arm was 0.76 (95% CI 0.49 to 1.03) and Correspondence to adjusted RR was 0.73 (95% CI 0.47 to 0.99). the adjusted RR was 0.69 (95% CI 0.45 to 0.92). No Dr Damien de Walque; Unadjusted RR for the high-value conditional cash harm was reported. ddewalque@worldbank.org de Walque D, Dow WH, Nathan R, et al. BMJ Open 2012;2:e000747. doi:10.1136/bmjopen-2011-000747 1 Downloaded from http://bmjopen.bmj.com/ on November 30, 2015 - Published by group.bmj.com Incentivising safe sex: CCT for HIV/STI prevention in Tanzania In the context of the staggering social, economic and ARTICLE SUMMARY human costs of the AIDS epidemic in sub-Saharan Strengths and limitations of this study Africa, it is perhaps not as great a leap as it would first - This paper reports the results of a novel approach for HIV and appear to apply the logic of conditional cash transfers to STI prevention. the private arena of human sexuality with the aim of - Our study methodology is rigorous, and the results are likely to incentivising safer sexual practices among high-risk advance a global conversation about economic approaches to populations. Numerous studies have documented the HIV/STI prevention. responsiveness of sexual behaviour to incentives, such as - Our main outcome measure is the combined point prevalence sex workers willing to forego condoms when clients pay of four STIs repeatedly tested by nucleic acid amplification extra,8 and increases in transactional sex in the face of tests over the course of the year and which have been household financial difficulties.9 Economic theory incontrovertibly linked to risky sexual activity. These biological outcomes, however, cannot be used to infer the relative suggests several pathways through which risky sexual importance of STI treatment seeking behaviour versus other behaviours could be reduced by a conditional cash behaviour changes, such as increased condom use or reducing transfer programme that conditions payment on nega- riskiness of partners. tive sexually transmitted infections (STIs) tests. Standard - The results reported in this study are limited to a 12-month theory predicts that the incentives could operate by experiment and cannot address the sustainability of improve- raising the implicit price of unsafe sex (risking losing the ments in STI outcomes over a longer period, particularly after conditional cash transfer) or by bringing the rewards of the conditional cash transfers have been discontinued. risk avoidance much closer to the present (eg, a condi- tional cash transfer within weeks may be more powerful for some people than the spectre of developing AIDS Conclusions: Conditional cash transfers used to incentivise safer many years in the future) or both. If the conditional cash sexual practices are a potentially promising new tool in HIV and transfer was sufficiently large, then this higher income sexually transmitted infections prevention. Additional larger study could also relieve economic pressures on young women would be useful to clarify the effect size, to calibrate the size of the to engage in transactional sex; but even if incentives incentive and to determine whether the intervention can be delivered were small, recent behavioural economics research cost effectively. suggests that regular reminders of this new frame for Trial registration number: NCT00922038 ClinicalTrials.gov. viewing sexual behaviour could still ‘nudge’ individuals to overcome inertia and extricate themselves from unduly risky sexual relationships.10 In Malawi, small financial incentives have already been shown to increase the uptake of HIV testing and counselling.11 In the only INTRODUCTION prior study similar to ours, a follow-on Malawi interven- Innovative solutions for AIDS prevention are desperately tion promised a single cash reward in 1 year’s time for needed. The Joint United Nations Programme on HIV/ individuals who remained HIV negative, but this design AIDS reported that five people are infected for every two had no measurable effect on HIV status.12 By contrast, placed on treatment, and, in 2009, approximately 2.8 we used the above theory to design and evaluate a novel million people were newly infected.1 Large-scale behav- intervention that tests for risky sexual behaviour iour change interventions aimed at promoting safer repeatedly over shorter time intervals, reinforcing sexual practices have proven less effective and more learning about safer behaviour with conditional cash unreliable at stemming the tide of the epidemic than transfer incentives each time. hoped.2 3 It has been far more difficult than was first anticipated to persuade high-risk populations to adopt safer sexual behaviours and practices that serve their METHODS longer term interests. Trial design Conditional cash transfer programmes have become an This study is an unblinded, individually randomised and increasingly popular approach for incentivising socially controlled trial. It has three separate armsda control desirable behavioural change.4 The principle of condi- arm with an allocation ratio of 50% and two intervention tionalitydmaking payments contingent, for example, on arms (low-value conditional cash transfer and high-value a minimal level of schooling attendance or preventive conditional cash transfer), with an allocation ratio of care useddistinguishes conditional cash transfer 25% each. No important changes to methods were programmes from more traditional means tested social implemented after trial commencement. programmes. The evaluation of conditional cash transfer programmes has shown that they can be effective at Participants raising consumption, education and preventive health- Inclusion criteria consisted of males and females, aged care,5 as well as actual health outcomes.6 Similarly, 18e30 years (and spouses starting at age 16 years and ‘contingency management’ approaches have shown potentially older than 30 years), residing in one of 10 important substance abuse reductions by conditioning study villages within the Kilombero/Ulanga districts of rewards on negative tests for drug or alcohol use.7 the Ifakara Health and Demographic Surveillance 2 de Walque D, Dow WH, Nathan R, et al. BMJ Open 2012;2:e000747. doi:10.1136/bmjopen-2011-000747 Downloaded from http://bmjopen.bmj.com/ on November 30, 2015 - Published by group.bmj.com Incentivising safe sex: CCT for HIV/STI prevention in Tanzania System13 in south-west Tanzania. The villages consisted (regardless of arm) was offered counselling and free STI of eight rural villages and two semi-urban neighbour- treatment (for self and partners) through health facili- hoods in Ifakara town, with participants evenly distrib- ties of the District Ministry of Health serving the uted across the villages. On average across the 10 research communities. Individual pre-test and post-test villages, approximately 20% of the 18e30 old residents counselling was provided to study enrollees at each were enrolled in the study. There were three exclusion testing interval, following Tanzania national testing criteria: being pregnant at the time of registration, guidelines. In addition, monthly group counselling having the intention to permanently migrate out of the sessions emphasising relationship skills training adapted Ifakara Health and Demographic Surveillance System from a subset of the Stepping Stones curriculum18 were area within the next year and unwillingness to partici- also made available to all study participants in all villages pate if assigned to the control arm. HIV-positives were but were not mandatory. eligible for enrolment. Outcomes Interventions The biological markers used in the study were selected The intervention arm was divided into two subarmsd both due to their likely prevalence levels in the study a low-value conditional cash transfer arm eligible for up population and due to their status within the epidemi- to $30 over the course of the study (10 000 Tanzanian ological literature as reasonable proxies for risky sexual shillings or approximately $10 per testing round) and behaviour. The primary outcome measure, as defined in a high-value conditional cash transfer arm eligible for up the study protocol, is the round-specific combined point to $60 (20 000 Tanzanian shillings or approximately $20 prevalence of the four STIs that were regularly testedd per testing round). Those amounts were determined C trachomatis, N gonorrhoeae, T vaginalis and M genitaliumd based on focus-group discussions in neighbouring at months 4, 8 and 12. This measure of combined point villages conducted before the intervention started, prevalence was constructed at study design to ensure balancing sufficient incentive levels against concerns sufficient power to detect differences in the control and about scalability and potential coercion. All participants treatment groups in response to the conditional cash were tested for STIs at baseline and then every 4 months transfer intervention. For logistical reasons, M genitalium for 1 year. Participants in the two intervention arms were testing was not conducted at baseline. We also tested for eligible to receive conditional cash transfer incentive HIV, herpes simplex virus 2 and syphilis at baseline and payments if they tested negative for curable STIs at the month 12. 4-, 8- and 12-month testing rounds. STIs tested at all All STI testing was conducted by the Ifakara Health these incentivised rounds were Chlamydia trachomatis, Institute microbiology laboratory in Ifakara. All test Neisseria gonorrhoeae, Trichomonas vaginalis and Myco- results were available within 7e10 days and were plasma genitalium, which are transmitted through returned to participants the following week. Ten per unprotected sexual contact and therefore serve as cent of all samples, and all positives, were sent to the a proxy for risky sexual behaviour as well as vulnerability University of California Chlamydia Laboratory for to HIV infection.14e17 confirmation analysis (quality control). Individuals in the conditional cash transfer arms were Specimens for chlamydia, gonorrhoea, trichomonas not eligible for the cash award at the 4-, 8- and 12-month and M genitalium were collected by a self-administered testing rounds if they tested positive for any of the vaginal swab for women. Men provided a ‘first-catch following: C trachomatis, N gonorrhoeae and T vaginalis. urine’ (about 20e30 ml) sample. Specimen collection Those converting from negative at baseline to positive at among women was always observed by a nurse at the 12 months for syphilis or herpes simplex virus 2 were testing station. For men, the specialised receptacle used also ineligible to receive the 12-month conditional cash to collect a urine sample was provided only after drop- transfer. HIV testing was conducted at baseline and ping off personal belongings upon checking into the month 12, but payments were not conditioned on those testing section of the study station. Men were asked to results because of local ethical sensitivities. M genitalium urinate into the study receptacle in the vicinity of the results did not affect conditional cash transfer eligibility study station. Detection used GenProbe Aptima (GenP- because there is some uncertainty around transmission robe Inc, San Diego, California, USA) nucleic acid pathways; however, it was included in the combined amplification tests. prevalence measure used as primary outcome to increase To test for HIV, herpes simplex virus 2 and syphilis, statistical power. Individuals in the intervention arms a single venous blood sample of approximately 5e10 ml testing positive for any of the conditioned curable STIs was collected from each participant at baseline and did not receive the conditional cash transfer but were month 12. For herpes simplex virus 2, we used the Focus eligible to continue as a study participant in subsequent HerpeSelect HSV-2 ELISA IgG assay (Focus Technolo- rounds after having been treated and cured of the gies, Cypress, California, USA) to detect serum anti- infection. Individuals in the control arm were not bodies. Treponema pallidum was identified using rapid eligible for conditional cash transfer, but all other study plasma reagin with reactive tests confirmed by T pallidum procedures were identical between the control and particle agglutination assay. Active syphilis was defined as intervention arms. Anyone testing positive for a STI rapid plasma reagin+/T pallidum particle agglutination de Walque D, Dow WH, Nathan R, et al. BMJ Open 2012;2:e000747. doi:10.1136/bmjopen-2011-000747 3 Downloaded from http://bmjopen.bmj.com/ on November 30, 2015 - Published by group.bmj.com Incentivising safe sex: CCT for HIV/STI prevention in Tanzania assay+. For HIV, we used a series of three rapid tests for lation with limited formal education. Participants were screening results (SD Bioline HIV-1/2 3.0; Standard not blinded to arm assignment since awareness of their Diagnostics, INC., Kyonggi-do, Korea), confirmation of eligibility for the conditional cash transfer was a critical positives (DetermineÒ HIV-1/2; Inverness Medical component of the intervention. Japan Co, Ltd, Tokyo, Japan), and tie-breaking (Uni- Spousal pairs were assigned the same intervention arm Gold HIV, Trinity Biotech plc. Bray, Ireland). and the protocol prescribed for randomisation to occur after both spouses had enrolled. Sample size Early study planning had initially assumed a sample size Statistical methods of 3000, which would have provided improved power for Each individual was coded as per their initial rando- gender subgroup analysis in our main comparisons, but mised assignment as per an intent-to-treat design. due to logistical fieldwork constraints the recruited However, individuals who were not present at any given sample size was reduced to approximately 2400. We round were treated as missing and dropped from the present here the ex-post power calculations at this actual analysis for that round due to lack of outcome data. We recruited sample size and actual infection rates, based report sample means at baseline to verify the balance on a comparison of combined STI prevalence rates of across the three study arms. Unadjusted outcomes at the 12% between two equal-sized study arms for a single three follow-up rounds are reported using RRs, that is, post-treatment measurement of proportions controlling the probability of being positive for any STI in the for one baseline measure and assuming a two-sided intervention arm, divided by the probability of being alternative hypothesis. We calculated that a total sample positive for any STI in the control arm. RRs are calcu- size of 2400 individuals would be sufficient to provide at lated from logistic regressions using the margins and least 90% power to detect a one-third intervention- nlcom post-estimation commands in the Stata V.12 statis- related reduction in STI point prevalence (significant at tical software package. We further report adjusted the 5% level) in both intervention arms combined. This outcomes using RRs to account for residual variation sample size would also retain at least 80% power to across arms after randomisation. Adjustments have been detect a reduction in a single intervention arm (eg, the made for standard socioeconomic variables, such as high-value conditional cash transfer arm) compared to gender, education, age, marital status, income, socio- the control arm, and if the prevalence was assumed to be economic status, subvillage and baseline STI status. Age as high as 20%, then this power rises to over 90%. and income are continuous variables, while the other Subgroup analysis by gender would not be powered at adjustment variables are categorical. We cluster SEs both the 80% level for our main comparison of the high-value at the household and at subvillage levels, accounting for conditional cash transfer arm against controls assuming the possible correlation within couples and the variation a 12% prevalence level, although it would retain in selection probability at that subvillage level. We approximately 90% power when comparing the present a subgroup analysis by gender. We used Stata combined arms against the control arm assuming 20% V.12.1 (Stata Corp) for statistical analysis. prevalence. RESULTS Randomisation Participant flow Individual-level randomisation took place at the study A total of 5370 individuals were randomly selected from station after baseline interview and testing, with partici- the Ifakara Health and Demographic Surveillance System pants selecting coloured balls from an opaque bag. The sample (figure 1). Eight hundred and sixty-four (16.1%) randomisation took place in public view and in two of those individuals were not found, six (0.1%) had died stages with participants first randomly selecting one of and 344 (6.4%) had migrated. Fieldworkers assessed for four balls to determine their allocation to the interven- eligibility 4156 individuals: 173 (4.2%) did not meet the tion or the control arm. In order to study potential peer- inclusion criteria, among them 35 (0.8%) were not in the effects, in randomly selected subvillages, the probably of study age range and 138 women (3.3%) were currently selection in the intervention arm was 75% (three balls pregnant. Of those eligible, 133 (3.3%) explicitly refused out of four) and in the other subvillages, it was 25% (one to participate in the study and 168 (4.2%) declined for ball out of four); based on the distribution of partici- other reasons. All others (3682) were given an invitation pants across subvillages, we thus expected 48% of the to come to a study station the following week: 2409 overall sample to be randomised into the control arm. (65.4%) registered for the study and were randomised Participants randomised into the intervention arm were into one of the three study arms, while 1273 did not further invited to choose one of two balls from a second come to the study station for registration. bag determining in which of the two intervention arms Of the 2409 registered participants, 1124 (46.7%) were (low-value conditional cash transfers and high-value randomly allocated to the control arm. Among the conditional cash transfers) they would be allocated. participants, 1285 were randomly selected, in a first These highly transparent procedures were deemed stage, to one of the two conditional cash transfer arms: necessary for acceptability of randomisation in a popu- 615 (25.5%) were randomly assigned in the high-value 4 de Walque D, Dow WH, Nathan R, et al. BMJ Open 2012;2:e000747. doi:10.1136/bmjopen-2011-000747 Downloaded from http://bmjopen.bmj.com/ on November 30, 2015 - Published by group.bmj.com Incentivising safe sex: CCT for HIV/STI prevention in Tanzania Figure 1 Participant flow diagram. conditional cash transfer arm and 660 (27.4%) in the line were more likely to be lost to follow-up, despite the low-value conditional cash transfer arm. Ten (0.4%) fact the participants were clearly told that HIV status individuals assigned to the intervention arms were would not affect eligibility for conditional cash transfers. intentionally dropped from the analysis since they failed to be further randomised in one of the two subarms. In Recruitment the control arm, 967 were tested and interviewed at Recruitment and baseline data collection took place round 2 (attrition 14%), 983 (attrition 12.5%) at round from 10 February to 9 April 2009. The second, third and 3 and 1039 (attrition 7.6%) at round 4. In the high-value fourth rounds of interviews and testing took place from conditional cash transfer arm, 570 were tested and 9 June to 15 August 2009, 29 September to 5 December interviewed at round 2 (attrition 7.3%), 567 (attrition 2009 and 16 February to 1 May 2010, respectively. The 7.8%) at round 3 and 585 (attrition 4.9%) at round 4. In conditional cash transfer intervention was stopped after the low-value conditional cash transfer arm, 568 were 1 year, following the protocol. tested and interviewed at round 2 (attrition 13.9%), 567 (attrition 14.1%) at round 3 and 618 (attrition 6.4%) at Process round 4. Overall, attrition was lower at round 4 because The intervention was well accepted and accessed by the the field team made extensive additional effort to study participants as indicated in the participant flow contact and interview attriters. Symptomatic individuals and the low attrition numbers. Furthermore, study in all study arms were particularly encouraged to come participants randomised into the conditional cash to the study station in order to receive free STI treat- transfer arms declared that the financial incentives ment. Attrition was not predicted by any of the baseline motivated them to modify their behaviour. In the high- STI results, except that HIV-positive individuals at base- value conditional cash transfer arm, 317 (59.0%) de Walque D, Dow WH, Nathan R, et al. BMJ Open 2012;2:e000747. doi:10.1136/bmjopen-2011-000747 5 Downloaded from http://bmjopen.bmj.com/ on November 30, 2015 - Published by group.bmj.com Incentivising safe sex: CCT for HIV/STI prevention in Tanzania declared that the money motivated them ‘very much’ to the respective rounds were included in the analysis change their behaviour and 67 (12.5%) stated that it (refer to the sample sizes in tables 1 and 2). The motivated them ‘somewhat’. In the low-value conditional reductions in sample size from the unadjusted (table 2) cash transfer arm, those numbers are 194 (37.4%) for to the adjusted analysis (table 3) were from 2105 to 2077 ‘very much’ and 107 (20.6%) for ‘somewhat’. at round 2, from 2117 to 2092 at round 3 and from 2242 to 2211 at round 4 due to missing data on covariates in Baseline data the logistic regression model (table 2 results are similar Table 1 describes the baseline characteristics of the when using the smaller samples from table 3). participants by study arm. The prevalence of the six STIs tested at baseline was distributed similarly across arms. Outcomes and estimation Participants were also similar according to gender and Table 2 presents the unadjusted RR ratios compared to education. However, individuals in the two intervention the control group. At months 4, 8 and 12 when the arms had slightly lower self-reported socioeconomic outcome is the combined point prevalence of the four status, and individuals in the low-value conditional cash curable STIs tested every 4 months by nucleic acid transfer arm also had a higher income. amplification tests (columns 1e3), the RRs are not We verified that there was no deviation from protocol statistically different at the 5% significance level. At that could have led to differential secondary spousal month 12, the number of positives was 57 (9.7%) in the enrolment across arms: 604 out of the 2399 participants high-value conditional cash transfer arm, while it was 79 were spouses who joined the study after their spouse was (12.8%) in the low-value conditional cash transfer arm initially invited. They were distributed as follows: 279 out and 126 (12.1%) in the control group. At month 12, this of 1124 (24.8%) in the control arm, 156 (25.4%) out of unadjusted analysis estimated a reduction in the RR of 615 in the high-value cash arm and 169 (25.6%) out of those four curable STIs for the high-value conditional 660 in the low-value cash arm. Tests for statistical cash transfer arm of 20% (95% CI 6% increase to 46% differences with the control arm yielded p values of reduction). The RRs were also not statistically different 0.673 for the high-value conditional cash transfer arm at the 5% significance level in column 4 for the combi- and 0.742 for the low-value conditional cash transfer nation of syphilis prevalence and new cases of HIV and arm, so differences across the three study arms in the herpes simplex virus 2. Those three STIs were detected percentage of spouses joining the study are minimal and by serology performed only at baseline and round 4. For not statistically significant. the combined point prevalence of chlamydia, gonor- rhoea, trichomonas, M genitalium at month 12, the Numbers analysed unadjusted RRs are not statistically different than 1 at Except for the 10 (0.4%) individuals who failed to be 5% significance level when men and women are assigned to either the high or low-value conditional cash considered separately (columns 5 and 6). At month 12, transfer arm, all participants tested and interviewed at for the combined point prevalence of the four curable Table 1 Summary statistics at baseline, by arm (1) (2) (3) Variables Control High-value CCT Low-value CCT Female 561 (49.9%) 314 (51.1%) 329 (49.9%) Age 27.2 (5.6) 27.6 (5.4) 27.6 (5.7) Education None 139 (12.4%) 70 (11.4%) 79 (12.0%) Primary 863 (76.8%) 482 (78.4%) 660 (78.3%) Secondary 122 (10.9%) 63 (10.2%) 64 (9.7%) Married 842 (75.0%) 474 (77.1%) 476 (72.7%) Low SES 582 (51.8%) 344 (55.9%) 377 (57.2%) Yearly income 239 311 (425 091) 257 017 (531 370) 283 218 (534 399) Chlamydia 21 (1.9%) 15 (2.4%) 16 (2.4%) Gonorrhoea 8 (0.7%) 8 (1.3%) 6 (0.9%) Trichomonas 130 (11.6%) 88 (14.3%) 79 (12.0%) Herpes simplex virus 2 380 (33.9%) 226 (36.8%) 225 (34.2%) Syphilis 17 (1.5%) 8 (1.3%) 15 (2.3%) HIV 41 (3.7%) 17 (2.8%) 27 (4.1%) N 1124 615 660 Data are represented in means (SD) or numbers (%). Yearly income in Tanzanian Shillings (Tsh). At baseline, 1000 Tsh ¼ approximately 1US$. Low SES corresponds to the lowest two ranks on a self-reported socioeconomic status scale from 1 to 7. CCT, conditional cash transfer. 6 de Walque D, Dow WH, Nathan R, et al. BMJ Open 2012;2:e000747. doi:10.1136/bmjopen-2011-000747 Downloaded from http://bmjopen.bmj.com/ on November 30, 2015 - Published by group.bmj.com Incentivising safe sex: CCT for HIV/STI prevention in Tanzania STIs tested by nucleic acid amplification tests, unad- STIs tested by NAAT* justed RR for the high-value conditional cash transfer Month 12 combined prevalence of four arm compared to the low-value conditional cash transfer 0.85 (0.55 to 1.15) 0.94 (0.62 to 1.25) arm was 0.76 (95% CI 0.49 to 1.03) and was 0.56 (95% CI women only 0.26 to 0.87) for men only. 163 (14.3%) 79 (15.2%) 39 (12.9%) 45 (14.2%) Table 3 presents results from adjusted regressions. Adjustments have been made for gender, education, age, 1137 marital status, income, socioeconomic status, subvillage (6) and baseline STI status. At months 4 and 8 (columns 1 and 2), the combined prevalence of the four STIs tested STIs tested by NAAT* Month 12 combined by nucleic acid amplification tests is shown to have RRs prevalence of four 0.70 (0.34 to 1.07) 1.25 (0.73 to 1.77) lower than 1 for the two conditional cash transfer arms compared to the control arm, but not significantly so. However, at month 12 (column 3) for the combined 34 (11.3%) men only 99 (9.0%) 47 (9.0%) 18 (6.4%) prevalence of the STIs tested by nucleic acid amplifica- tion tests, the adjusted model estimated a 27% reduction 1105 in the RRs for the high-value conditional cash transfer (5) arm compared to the control arm (95% CI 1% to 53% serology at baseline Month 12 combined reduction), while the RR is not statistically different from measure for three 1.09 (0.76 to 1.43) 1.03 (0.71 to 1.35) STIs detected by 1 for the low-value conditional cash transfer arm. At and month 12y month 12, for the three STIs detected by serology 232 (10.4%) 104 (10.0%) (without having been tested at months 4 and 8), the RR 64 (11.0%) 64 (10.4%) for the low-value conditional cash transfer arm is 0.82 (column 4) but is not significantly lower than the control 2241 arm (95% CI 0.60 to 1.03). In a subgroup analysis by (4) gender (columns 5 and 6), for the four STIs tested by STIs tested by NAAT* nucleic acid amplification tests, the RRs for the high- Month 12 combined CCT, conditional cash transfer; NAAT, nucleic acid amplification test; STI, sexually transmitted infection. value conditional cash transfer arm are 0.68 for men and prevalence of four 0.80 (0.54 to 1.06) 1.05 (0.75 to 1.35) 0.76 for women. Those two RRs are not significantly different from each other (as confirmed by test of 262 (11.7%) 126 (12.1%) 79 (12.8%) interaction between gender and arm, where an interac- 57 (9.7%) tion term for woman was not significant for either 2242 conditional cash transfer arm (p values 0.648 for high- Robust SEs in parentheses, clustered at both the household and the subvillage levels. (3) value cash transfer arm and 0.391 for low-value cash transfer arm) and is not significantly lower than 1 at the STIs tested by NAAT* 5% level. At month 12, for the combined point preva- Month 8 combined prevalence of four 0.86 (0.60 to 1.12) 0.80 (0.55 to 1.04) lence of the four STIs tested by nucleic acid amplifica- The reference group for the computation of the RRs is the control group. tion tests, adjusted RR for the high-value conditional Table 2 Unadjusted outcomes: RR from logistic regression cash transfer arm compared to the low-value conditional 260 (12.3%) 133 (13.5%) 66 (11.6%) 61 (10.8%) *Chlamydia, gonorrhoea, trichomonas, Mycoplasma genitalium. cash transfer arm was 0.69 (95% CI 0.45 to 0.92) and was 0.52 (95% CI 0.23 to 0.80) for men only. 2117 (2) DISCUSSION STIs tested by NAAT* After 12 months, the adjusted results showed a signifi- Month 4 combined prevalence of four cant reduction in the combined point prevalence of the 1.06 (0.74 to 1.38) 0.97 (0.66 to 1.28) four curable STIs tested every 4 months by nucleic acid yHIV, herpes simplex virus 2, syphilis. amplification tests in the group that was eligible for the 246 (11.7%) 112 (11.6%) High-value CCT 70 (12.3%) Low-value CCT 64 (11.3%) Number positives by study arm $20 quarterly payments, but no such reduction was found for the group receiving the $10 quarterly 2105 payments. Such results were not found at earlier rounds (1) nor for unadjusted results. Furthermore, the impact of the conditional cash transfers did not differ between Number positive High-value CCT Low-value CCT men and women. Control arm Limitation (95% CI) (95% CI) Our main outcome measure is the combined point prevalence of four STIs repeatedly tested by nucleic acid N amplification tests over the course of the year and which de Walque D, Dow WH, Nathan R, et al. BMJ Open 2012;2:e000747. doi:10.1136/bmjopen-2011-000747 7 Downloaded from http://bmjopen.bmj.com/ on November 30, 2015 - Published by group.bmj.com Incentivising safe sex: CCT for HIV/STI prevention in Tanzania have been incontrovertibly linked to risky sexual activity. STIs tested by NAAT* These biological outcomes, however, cannot be used to Month 12 combined prevalence of four infer the relative importance of STI treatment seeking 0.76 (0.46 to 1.07) 0.98 (0.63 to 1.33) behaviour versus other behaviour changes, such as women only increased condom use or reducing riskiness of partners. 159 (14.2%) Furthermore, the lack of a clear result on the combined measure for the three STIs that were detected by 1118 serology only at baseline and month 12 (this measure (6) primarily reflects herpes simplex virus 2 incidence, as HIV and syphilis prevalence were somewhat lower) is STIs tested by NAAT* Month 12 combined puzzling and merits further study. The contrasting prevalence of four 0.68 (0.25 to 1.10) 1.31 (0.73 to 1.89) result with the impact of the high-value conditional cash transfers on the four curable STIs that were tested by nucleic acid amplification tests could point to the men only 98 (9.0%) importance of treatment seeking behaviour rather than safer sexual practices. However, the interpretation of 1093 herpes simplex virus 2 results is complicated by the fact (5) that most transmission occurs via asymptomatic shed- ding by partners who may be otherwise low risk,19 as serology at baseline Month 12 combined measure for three 1.03 (0.74 to 1.32) 0.82 (0.60 to 1.03) well as the fact that it can be transmitted even in the STIs detected by and month 12y context of appropriate condom use.20 Furthermore, this study was not powered to directly examine HIV 227 (10.3%) Results adjusted for gender, education, age, marital status, income, socioeconomic status, subvillage and baseline STIs. conversion, thus implications for HIV prevention remain speculative. 2210 In order to study potential peer-effects, in randomly (4) selected subvillages, the probably of selection in the intervention arm was 75% and in the other subvillages, it STIs tested by NAAT* CCT, conditional cash transfer; NAAT, nucleic acid amplification test; STI, sexually transmitted infection. Month 12 combined was 25%. This might have led to baseline imbalances. prevalence of four 0.73 (0.47 to 0.99) 1.06 (0.75 to 1.38) For this reason, we included subvillage indicator vari- ables in the adjusted models. This might explain some of 257 (11.6%) the differences between the results from the unadjusted and the adjusted models. Finally, the results reported in this study are limited to Robust SEs in parentheses, clustered at both the household and the subvillage levels. 2211 a 12-month experiment and cannot address the (3) sustainability of improvements in STI outcomes over STIs tested by NAAT* a longer period, particularly after the conditional cash Month 8 combined transfers have been discontinued. Nor can they address prevalence of four The reference group for the computation of the RRs is the control group. 0.90 (0.61 to 1.18) 0.85 (0.58 to 1.13) the possibility of adverse consequences to the extent that extrinsic incentives may reduce long-term intrinsic Table 3 Adjusted outcomes: RR from logistic regression 258 (12.3%) *Chlamydia, gonorrhoea, trichomonas, Mycoplasma genitalium. motivation to engage in safe behaviours after incentives are withdrawn. To address these questions, we will follow- 2092 up with study participants 1 year following the end of the (2) intervention study, in the Spring of 2011, to assess whether improved outcomes have been sustained, or STIs tested by NAAT* reversed, in the absence of a positive feedback mecha- Month 4 combined prevalence of four nism in the form of STI testing and conditional cash 0.92 (0.62 to 1.20) 0.94 (0.63 to 1.26) transfers. yHIV, herpes simplex virus 2, syphilis. 242 (11.7%) Generalisability While these study results are important in showing that 2077 the idea of using financial incentives can be a useful tool (1) for preventing HIV and STI transmission, it remains an initial study on a limited scale. Even though the study Number positive High-value CCT Low-value CCT site is fairly representative of rural and small town envi- ronments in sub-Saharan Africa, this approach would (95% CI) (95% CI) need to be replicated elsewhere and implemented on a larger scale (in permutations requiring less adminis- trative and laboratory capacity) before it could be N concluded that such conditional cash transfer 8 de Walque D, Dow WH, Nathan R, et al. BMJ Open 2012;2:e000747. doi:10.1136/bmjopen-2011-000747 Downloaded from http://bmjopen.bmj.com/ on November 30, 2015 - Published by group.bmj.com Incentivising safe sex: CCT for HIV/STI prevention in Tanzania programmes offer an efficient, scalable and sustainable amended approval is from 11 February 2010. Tanzania’s HIV prevention strategy. National Institute for Medical Research approved the study 5 February 2009. Interpretation Author affiliations The results indicate that conditional cash transfers based 1 The World Bank, Development Research Group, Washington, District of on negative results of periodic screenings for incident Columbia, USA STIsdan objectively measured marker for risky sexual 2 School of Public Health, University of California, Berkeley, California, USA 3 behaviourdare a potentially useful tool for STI and Ifakara Health Institute, Dar-es-Salaam, Tanzania 4 possibly HIV prevention. The extraordinarily high social Ifakara Health Institute, Ifakara, Tanzania 5 Department of Economics, Middlebury College, Middlebury, Vermont, USA and economic cost of the current HIV and AIDS crisis 6 University of California, Berkeley, California, USA suggests that prevention can be far cheaper than treat- 7 Center for Behavioural Economics, Federal Reserve Bank of Boston, Boston, ment, thus motivating the continued search for innova- Massachusetts, USA 8 tive and effective new prevention approaches, such as Research Triangle Institute International, Research Triangle Park, North conditional cash transfers or other financial incentives. Carolina, USA 9 Department of Laboratory Medicine, University of California, San Francisco, The absence of significant impacts at rounds 2 (month California, USA 4) and 3 (month 8) suggests that the impact of the 10 Global Health Sciences Prevention and Public Health Group, University of conditional cash transfer may take time to materialise, California, San Francisco, California, USA 11 perhaps because it is not easy to extricate oneself from Health Economics and Finance, Global Health Program, The Bill and Melinda complicated sexual relationships, or perhaps because Gates Foundation, Seattle, Washington, USA participants needed time to become accustomed to Acknowledgements We thank Salim Abdulla, Joseph Kambona, Dean Karlan, (and trust) the incentive mechanism. The comparison Flora Kessy, Carol Kolb deWilde, Andrew Mchomvu, Hassan Mshinda, Mead Over, Nancy Padian, Honorathy Urassa and many others for their between the impacts of the conditional cash transfer contributions to the project. intervention in the high-value conditional cash Funding The study was funded by the World Bank Research Committee, the transfer arm to that in the low-value conditional cash Spanish Impact Evaluation Fund and the Knowledge for Change Program transfer arm permits us to better understand at which managed by the World Bank and the William and Flora Hewlett Foundation threshold conditional cash transfers can be effective as through the Population Reference Bureau. The study funders had no role in the an HIV and STI prevention tool. While the results study design, in the collection, analysis and interpretation of data, in the showed a significant reduction in STI incidence in the writing of the report and in the decision to submit the article for publication and researchers were independent from the funders. The findings, arm that was eligible for the $20 conditional cash interpretations and conclusions expressed in this paper are entirely those of transfers every 4 months or up to $60 over 12 months, no the authors. They do not necessarily represent the views of the International such reduction was found for the arm receiving the $10 Bank for Reconstruction and Development/World Bank and its affiliated conditional cash transfers every 4 months or up to $30 organisations or those of the Executive Directors of the World Bank or the over 12 months. This distinction must be interpreted governments they represent. with caution though because assignments were not Competing interests None. masked, hence individuals in the low-value conditional Patient consent The article does not contain personal medical information cash transfer arm could have behaved differently than if about an identifiable living individual. they were to receive the same incentive in the absence of Contributors DdW, WHD, RN and CAM made contributions to each part of the a higher conditional cash transfer arm. Both of these project, planned and designed the study, conducted the analysis, interpreted amounts represent a meaningful proportion of house- the findings and contributed to the manuscript. The Ifakara Health Institute hold income in a country where gross domestic product was the main implementing agency for the project: BJ and FA managed the Ifakara laboratory testing, AM led field operations, MAM facilitated operations, per capita was $440 in 2008, and particularly among our KS programmed the study systems and together with RA managed the study participants who had mean individual annual database and SM was responsible for outreach to participating communities earnings of approximately $250. and health clinics. EG contributed to data analysis, LP conducted in-depth interviews and ZI was project director onsite in Tanzania. From University of California, San Francisco, JM and JS set up the Ifakara Health Institute OTHER INFORMATION laboratory, developed laboratory protocols and were responsible for quality Registration control. SK, JJ and EM as senior investigators have contributed throughout the This randomised control trial is registered at ClinicalTrials. project and are leading subanalyses linked to the main study in their respective fields of expertise. All authors, external and internal, had full access to all the gov, study identifier # NCT00922038. data (including statistical reports and tables) in the study and can take responsibility for the integrity of the data and the accuracy of the data analysis. Protocol DdW and WHD are the guarantors of the study. The study protocol was initially approved by the Provenance and peer review Not commissioned; externally peer reviewed. University of California; Berkeley’s Institutional Review Data sharing statement No additional data available. Board (Committee for Protection of Human Subjects) effective 17 December 2008; approval has been updated numerous times since to reflect protocol amendments, REFERENCES with the latest approval effective 11 October 2011. The 1. UNAIDS. AIDS Epidemic Update: November 2009. Geneva, Switzerland: UNAIDS, 2009. Ifakara Health Institute Institutional Review Board 2. Bertrand JR, O’Reilly K, Denison J, et al. Systematic review of the initially approved the study on 24 July 2008. The latest effectiveness of mass communication programs to change HIV/AIDS- de Walque D, Dow WH, Nathan R, et al. BMJ Open 2012;2:e000747. doi:10.1136/bmjopen-2011-000747 9 Downloaded from http://bmjopen.bmj.com/ on November 30, 2015 - Published by group.bmj.com Incentivising safe sex: CCT for HIV/STI prevention in Tanzania related behaviors in developing countries. Health Educ Res 13. Schellenberg JA, Mukasa O, Abdulla S, et al. The Ifakara demographic 2006;21:567e97. surveillance system, Tanzania. In: Sankoh OA, Kahn K, Mwageni E, 3. Napierala Mavedzenge SM, Doyle AM, Ross DA. HIV Prevention in et al, eds. Population and Health in Developing Countries: Volume 1. Young People in Sub-Saharan Africa: A Systematic Review. J Population, Health and Survival at INDEPTH Sites. Ottawa, Canada: Adolescent Health 2011;49:568e86. International Development Research Centre, 2002:159e64. 4. Fiszbein A, Schady N. Conditional Cash Transfers. Reducing 14. Napierala Mavedzenge S, Weiss HA. Association of Mycoplasma Present and Future Poverty. Washington DC: The World Bank, genitalium and HIV infection: a systematic review and meta-analysis. 2009:361. AIDS 2009;23:611e20. 5. Lagarde M, Haines A, Palmer N. Conditional cash transfers for 15. Niccolai LM, Rowhani-Rahbar A, Jenkins H, et al. Condom improving uptake of health interventions in low- and middle-income effectiveness for prevention of Chlamydia trachomatis infection. Sex countries: a systematic review. JAMA 2007;298:1900e10. Transm Infect 2005;81:323e5. 6. Gertler P. Do conditional cash transfers improve child health? 16. Crosby RA, DiClemente RJ, Wingood GM, et al. Value of consistent Evidence from PROGRESA’s control randomized experiment. Amer condom use: a study of sexually transmitted disease prevention Econ Rev 2004;94:332e41. among african american adolescent females. Am J Public Health 7. Petry NM, Petrakis I, Trevisan L, et al. Contingency management 2003;93:901e2. interventions: from research to practice. Am J Psychiatry 17. Fishbein M, Pequegnat W. Evaluating AIDS prevention interventions 2001;158:694e702. using behavioral and biological outcome measures. Sex Trans Dis 8. Gertler P, Shah M, Bertozzi SM. Risky business. The market for 2000;27:101e10. unprotected sex. J Polit Economy 2005;113:518e50. 18. Jewkes R, Nduna M, Levin J, et al. Impact of stepping stones 9. Robinson J, Yeh E. Transactional sex as a response to risk in on incidence of HIV and HSV-2 and sexual behaviour in rural Western Kenya. AEJ: Applied 2011;3:35e64. South Africa: cluster randomised controlled trial. BMJ 2008; 10. DellaVigna S. Psychology and economics: evidence from the field. 337:a506. J Econ Lit 2009;47:315e72. 19. Koelle DM, Wald A. Herpes simplex virus: the importance of 11. Thornton R. The demand for and impact of learning HIV status. Am asymptomatic shedding. J Antimicrob Chemother 2000;45(Suppl Econ Rev 2008;98:1829e63. T3):1e8. 12. Kohler HP, Thornton R. Conditional Cash Transfers and HIV/AIDS 20. Martin ET, Krantz E, Gottlieb SL, et al. A pooled analysis of the effect Prevention: Unconditionally Promising? World Bank Econ Rev 2011. of condoms in preventing HSV-2 acquisition. Arch Inter Med doi:10.1093/wber/lhr041 2010;170:1233e40. PAGE fraction trail=9.5 10 de Walque D, Dow WH, Nathan R, et al. BMJ Open 2012;2:e000747. doi:10.1136/bmjopen-2011-000747 Downloaded from http://bmjopen.bmj.com/ on November 30, 2015 - Published by group.bmj.com Incentivising safe sex: a randomised trial of conditional cash transfers for HIV and sexually transmitted infection prevention in rural Tanzania Damien de Walque, William H Dow, Rose Nathan, Ramadhani Abdul, Faraji Abilahi, Erick Gong, Zachary Isdahl, Julian Jamison, Boniphace Jullu, Suneeta Krishnan, Albert Majura, Edward Miguel, Jeanne Moncada, Sally Mtenga, Mathew Alexander Mwanyangala, Laura Packel, Julius Schachter, Kizito Shirima and Carol A Medlin BMJ Open 2012 2: doi: 10.1136/bmjopen-2011-000747 Updated information and services can be found at: http://bmjopen.bmj.com/content/2/1/e000747 These include: Supplementary Supplementary material can be found at: Material http://bmjopen.bmj.com/content/suppl/2012/02/09/bmjopen-2011-000 747.DC1.html References This article cites 16 articles, 4 of which you can access for free at: http://bmjopen.bmj.com/content/2/1/e000747#BIBL Open Access This is an open-access article distributed under the terms of the Creative Commons Attribution Non-commercial License, which permits use, distribution, and reproduction in any medium, provided the original work is properly cited, the use is non commercial and is otherwise in compliance with the license. See: http://creativecommons.org/licenses/by-nc/2.0/ and http://creativecommons.org/licenses/by-nc/2.0/legalcode. Email alerting Receive free email alerts when new articles cite this article. Sign up in the service box at the top right corner of the online article. Topic Articles on similar topics can be found in the following collections Collections HIV AIDS (119) Reproductive medicine (44) Infectious diseases (348) Public health (1259) Sexual health (95) Immunology (including allergy) (28) Sexual health (9) Clinical trials (epidemiology) (16) Ophthalmology (5) To request permissions go to: http://group.bmj.com/group/rights-licensing/permissions To order reprints go to: http://journals.bmj.com/cgi/reprintform To subscribe to BMJ go to: http://group.bmj.com/subscribe/ Downloaded from http://bmjopen.bmj.com/ on November 30, 2015 - Published by group.bmj.com Notes To request permissions go to: http://group.bmj.com/group/rights-licensing/permissions To order reprints go to: http://journals.bmj.com/cgi/reprintform To subscribe to BMJ go to: http://group.bmj.com/subscribe/