Abstract
While overall HIV prevalence in Malawi has decreased, it is still high in the Southern region of the country. Behavioral prevention activities are crucial to continue the reduction in HIV prevalence. Behavior change is influenced by many factors. Previous work indicates knowledge about HIV transmission, self-efficacy to protect oneself from exposure, and accurate risk perception of one’s susceptibility all impact sexual behavior. The current study looks at the effects of a behavior change communication program in Malawi called the BRIDGE II Project on psychosocial and behavioral variables. The program sought to address barriers to individual action and confront societal norms related to sexual risk behavior through a mix of community-based activities and mass media messages delivered through local radio stations. Using cohort data (n = 594), we examined the effect of BRIDGE exposure on three variables that affect HIV behaviors: knowledge, self-efficacy, and risk perception, as well as two behavioral outcomes: HIV testing and condom use at last sex. Data were collected at baseline and for a midterm evaluation. Regression analyses showed exposure to BRIDGE was significantly associated with knowledge level (β = 0.20, p < .001) and self-efficacy (β = 0.35, p < .001) at midterm when controlling for baseline scores, but not risk perception. Psychosocial variables did not show a significant relationship to either behavioral outcome. However, program exposure was a significant predictor of both HIV testing in the past year (OR = 1.40, p < .001) and condom use at last sex (OR = 1.26, p < .05). This study suggests such a communication intervention may play an important role in not only affecting HIV-related behaviors themselves, but also critical factors that affect HIV behaviors, including knowledge and self-efficacy. It is recommended that communication efforts around HIV risk reduction be increased.
Keywords: prevention, Malawi, communication, HIV testing, condom use
Introduction
HIV incidence among Malawian adults dropped by at least half since 2002 (UNAIDS, 2012). While prevalence declined among the general population, females, urban residents, and those in the Southern region are at higher risk (Malawi Country Report, 2012; NSO and ICF, 2011; NSO and ORC, 2005). With no known cure, prevention is paramount to reversing the epidemic.
Knowledge, self-efficacy, and risk perception have shown associations with reductions in sexual risk and HIV treatment-seeking behavior (Baidoobonso, Bauer, Speechley, & Lawson, 2013; Berendes & Rimal, 2011; Chepngeno-Langat, 2013; Odu et al., 2008; Rimal et al., 2009a; Rimal et al., 2009b; Rimal, Limaye, Roberts, Brown, & Mkandawire, 2013). The program of focus here, BRIDGE II, attempted to affect these mediating factors to ultimately promote HIV testing and condom use.
The BRIDGE II Program
The Johns Hopkins University Center for Communication Programs has spearheaded a multi-level HIV prevention intervention program in the southern region of Malawi, where prevalence is highest (NSO and ICF, 2011) known as BRIDGE (Bridge, Redefine, Integrate, Develop, Generate, and Expand). BRIDGE I was a social and behavior change communication program designed to influence perceptions and discussions of HIV/AIDS alongside adoption of improved health behaviors (Rimal et al., 2009a). BRIDGE II seeks to address barriers to individual action and confront societal norms as a means of scaling up prevention activities.
Program activities include community-based participation (e.g., small group discussions, interactive drama, community-wide events, couples’ counseling, community referral, capacity building of local structures to implement activities) and mass media messages delivered through radio. The BRIDGE program does not directly provide testing services, but encourages and provides referrals. The program does provide condoms; however, condoms can also be accessed through other community outlets.
This paper seeks to examine whether a cohort sample in the Southern region was affected by the program at the point of midterm evaluation, specifically regarding the three intermediate individual level factors of interest (HIV knowledge, self-efficacy, and risk perception) and the behaviors of testing and condom use at last sex.
Method
Chi square analyses and t-tests were used to conduct an attrition analysis, while regressions assessed relationships regarding program exposure, intermediate factors, and behavior. Johns Hopkins University and the Malawi National Health Sciences Research Committee provided ethical clearance.
Setting and participants
Participants were adults (18+ years) in 11 districts in southern Malawi. In 2009, we conducted baseline household-level surveys to assess knowledge, attitudes, and behaviors. All participants were informed they could be re-contacted for additional data collection. At the time of writing, two waves have been collected—baseline and midline (November 2011).
Materials and procedures
Malawian interviewers collected data. Questionnaires in Chichewa were pre-tested, responses informed subsequent revisions, and English back-translation confirmed proper concept translation.
Sampling
A multi-stage sampling process was used. Stage 1 involved selection and weighting of districts. Stage 2 involved stratification of all traditional authorities (akin to counties in the U.S.) in each district into three groups (low/medium/high exposure), according to the projected program activity level to be implemented. We randomly sampled homes from each exposure level; one person from each household was chosen. At midterm, we stratified all participants into the projected exposure groups and randomly selected 50% from each to be re-interviewed.
Measures
Demographics
Participants reported age, gender, marital status, and education.
Program exposure
Because programming levels changed slightly during implementation, exposure was measured via self-report on four items asking whether or not participants heard of or participated in four campaigns. Participants received a point for each program they heard/participated in.
Knowledge
Assessed with an 11-item scale focused on cure and prevention previously tested by our organization in other African countries. Responses were “Yes”, “No”, and “Don’t know.” Items were scored 1 if correct and 0 if incorrect. “Don’t know” was scored 0 (baseline α= 0.66, midline α = 0.63).
Self-efficacy
Assessed with a nine-item, four-point Likert scale (1 = strongly disagree; 4 = strongly agree) previously tested by our organization. Neither agree/disagree was scored zero. Scale scores were a sum of all items (baseline α = 0.75, midline α= 0.73).
HIV risk perception
A three-item scale determined perception of vulnerability to infection for oneself and family. Responses were a four-point Likert scale (1 = no likelihood at all; 4 = very high likelihood). Neither agree/disagree was scored zero. Scale scores were a sum of all items (baseline α = 0.76, midline α = 0.81).
HIV-related behaviors
Participants were asked: “Have you been tested for HIV in the last 12 months?” to measure testing during the program, and “The last time you had sex, did you and your partner use a condom?” (male or female condom not specified).
Results
We attempted to contact half the baseline sample (906 out of 1812) at midterm but only reached 685 (24.4% attrition). Attrition analyses showed differences for education level (t = 2.72, p<.01), gender (χ2 = 9.06, p < .01) and relationship status (χ2= 7.49, p<.01); those who remained in the sample were more likely to be better educated, female, and married than those who dropped out.
One participant who did not complete baseline and participants with no sexual experience (n = 66) were excluded. Because of our interest in risk perceptions and prevention behaviors, we excluded those who reported being infected (n = 24) at either time point (final n = 594).
Table 1 shows the sample characteristics. The sample had a low education level, a majority was married/cohabitating, and mean age was 27.65 for women and 30.82 for men. Average program exposure level was 0.77 (SD = 0.88) for females and 1.23 (SD = 1.04) for males.
Table 1.
Demgraphic characteristics at baseline.
Female (323) | Male (271) | Total (594) | ||||
---|---|---|---|---|---|---|
n/mean | %/SD | n/mean | %/SD | n/mean | %/SD | |
Age (mean, SD) | 27.65 | 10.31 | 30.82 | 11.92 | 29.09 | 11.17 |
18–24 years | 143 | 44.27 | 105 | 38.75 | 248 | 41.75 |
25–29 years | 85 | 26.32 | 41 | 15.13 | 126 | 21.21 |
30–39 years | 54 | 16.72 | 65 | 23.99 | 119 | 20.03 |
40–49 years | 30 | 9.29 | 41 | 15.13 | 71 | 11.95 |
50+ years | 11 | 3.41 | 19 | 7.01 | 30 | 5.05 |
Education (mean, SD) | 5.49 | 3.38 | 6.49 | 3.33 | 5.95 | 3.39 |
1–4 years | 132 | 40.87 | 82 | 30.26 | 214 | 36.03 |
5–10 years | 166 | 51.39 | 154 | 56.83 | 320 | 53.87 |
11+ years | 25 | 7.74 | 35 | 12.92 | 60 | 10.10 |
Relationship status | ||||||
Married/cohabitating | 255 | 78.95 | 196 | 72.32 | 451 | 75.93 |
Single | 68 | 21.05 | 75 | 27.68 | 143 | 24.07 |
Using logistic regression analyses with the behaviors as outcomes, we entered key demographics (age, gender, education, marital status) and intermediate variable scores (knowledge, risk perception, and self-efficacy as measured at midline) as predictors. Only age and marital status were significant predictors of condom use. Age, gender, and marital status were significant predictors of testing (all p’s< .05). Younger, female, and married respondents were more likely to have tested, and younger, unmarried respondents were more likely to use condoms compared to their counterparts. We used linear regressions to look at whether program exposure was related to each psychosocial intermediate variable, controlling for baseline (Table 2). Exposure was significantly associated with knowledge and self-efficacy (p’s< .01) at midterm.
Table 2.
Regression analysis of program exposure on intermediate variables.
Program exposure | ||
---|---|---|
β | 95% CI | |
HIV knowledge | 0.20 | (0.06 – 0.34)** |
Self-efficacy | 0.35 | (0.08–0.62)** |
HIV risk perception | 0.17 | (−0.13–0.46) |
p<.01
Table 3 shows logistic regressions examining whether intermediate variables were related to each behavior, controlling for baseline scores. None of the variables were significant predictors of either behavior. We examined whether program exposure was directly related to the behaviors using logistic regressions. Program exposure was a significant predictor of both testing in the past year (p < .001) and condom use at last sex (p < .05).
Table 3.
Regression analysis of program exposure and intermediate variables on behavioral outcomes.
HIV testing within past year | Condom use at last sex | |||
---|---|---|---|---|
OR | 95% CI | OR | 95% CI | |
HIV knowledge | 1.05 | (0.96–1.16) | 0.98 | (0.87–1.10) |
Self-efficacy | 0.99 | (0.94–1.05) | 1.05 | (0.23–0.97) |
HIV risk perception | 0.98 | (0.93–1.02) | 1.05 | (0.99–1.11) |
Program exposure | 1.40 | (1.16–1.70)*** | 1.26 | (1.03–1.55)* |
Note: OR=odds ratio adjusted for baseline
p<.05
p<.001
Discussion
Results indicated that although the program could be linked with improvements in knowledge and self-efficacy, but not risk perception, these variables were not associated with condom use or HIV testing. There was, however, a direct link between intervention exposure and improvements in condom use and testing.
These findings both verify results from previous studies and raise several questions. Knowledge, self-efficacy, and risk perception are key determinants of health behaviors, according to social cognitive theory (Bandura, 1977). The fact that these variables were not associated with behaviors points to the possibility of imprecise measures. Our findings did show, however, that even after controlling for baseline values, the intervention explained additional variance in midline values. The intervention was able to boost people’s knowledge and self-efficacy beyond their values in the absence of an intervention. This suggests perhaps measurement issues were not strong enough to determine significant findings.
Condom use may be governed more by relationship quality than by individual-level factors such as knowledge and self-efficacy. Similarly, testing may be greatly influenced by social considerations—stigma, support of those with HIV—in addition to individual-level factors. Our other finding—that both testing and condom use were linked with campaign exposure—provides some support for this. Indeed, BRIDGE II went beyond individual-level change and sought to garner community support, change prevailing norms, and mobilize communities. It is likely our exposure variable was tapping into these aspects, and that individual-level factors were not adequate to capture the totality of effects.
Limitations
A limitation is the pre-/post-test design without a control group. With mass media being a large part of implementation, it was difficult to maintain a true control group due to the wide reach. However, this makes it impossible to know if the significant pre/post differences are truly attributable to the program. Also, participants may have engaged in social desirability in their midline self-reports. Finally, there may be fundamental differences in those who engaged in the program versus those who did not, especially for community activities.
It is clear BRIDGE II is linked to positive behavioral changes. The exact pathways of those changes are unclear, however, we are confident efforts to build knowledge, self-efficacy and increase risk perceptions to enable individuals to effect behaviors should continue. Multilevel social and behavioral change communication interventions, such as mass media supported by community interactions provides a strong platform to support behavior change.
Acknowledgments
Thank you to Jane Brown and Madalitso Makwandu for their support on this project.
This research was supported by a grant from the American People through the United States Agency for International Development (USAID). The contents of this report are the responsibility of the Johns Hopkins University Bloomberg School of Public Health Center for Communication Programs and do not necessarily reflect the views of USAID or the United States Government.
Contributor Information
Michelle R. Kaufman, Center for Communication Programs, Johns Hopkins Bloomberg School of Public Health, 111 Market Place, Suite 310, Baltimore, MD, 21202, USA, Ph: +1-410-659-6327, mkaufman@jhsph.edu
Rajiv N. Rimal, Department of Prevention & Community Health, George Washington University, 2175 K St NW, Washington, DC 20037, USA, Ph: +1-202-994-3570, rrimal@gwu.edu
Maria Carrasco, Johns Hopkins University Bloomberg School of Public Health, Baltimore, MD, USA, 615 N. Wolfe Street, Baltimore, MD, 21205, USA, Ph: +1-301-412-0096. mcarrasc@jhsph.edu.
OlaOluwa Fajobi, Center for Communication Programs, Johns Hopkins Bloomberg School of Public Health, 111 Market Place, Suite 310, Baltimore, MD, 21202, USA, Ph: +1-443-794-2768, ofajobi@jhsph.edu.
Anthony Soko, Johns Hopkins University Center for Communication Programs, Malawi, P.O. Box 30782, Lilongwe 3, Malawi, Ph: +265 999 939 795, asoko@malawibridge.org.
Rupali Limaye, Center for Communication Programs, Johns Hopkins Bloomberg School of Public Health, 111 Market Place, Suite 310, Baltimore, MD, 21202, USA, Ph: +1-410-659-6273, rlimaye@jhuccp.org.
Glory Mkandawire, Johns Hopkins University Center for Communication Programs, Malawi, P.O. Box 30782, Lilongwe 3, Malawi, Ph: +265 992 961 826, gmkandawire@malawibridge.org.
References
- Baidoobonso S, Bauer GR, Speechley KN, Lawon E. HIV risk perception and distribution of HIV risk among African, Caribbean and other Black people in a Canadian city: mixed methods results from the BLACCH study. BMC Public Health. 2013;13:184. doi: 10.1186/1471-2458-13-184. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bandura A. Social learning theory. Englewood-Cliffs, NJ: Prentice Hall; 1977. [Google Scholar]
- Berendes S, Rimal RN. Addressing the slow uptake of HIV testing in Malawi: The role of stigma, self-efficacy, and knowledge in the Malawi BRIDGE Project. Journal of the Association of Nurses in AIDS Care. 2011;22:215–228. doi: 10.1016/j.jana.2010.08.005. Epub 2010 Dec 24. [DOI] [PubMed] [Google Scholar]
- Chepngeno-Langat G. Perception of vulnerability to HIV infection among older people in Nairobi, Kenya: a need for intervention. Journal of Biosocial Science. 2013;45:249–266. doi: 10.1017/S0021932012000417. Epub 2012 Jul 13. [DOI] [PubMed] [Google Scholar]
- Joint United Nations Programme on HIV/AIDS. Report on the Global AIDS Epidemic 2012. 2012 Retrieved from http://www.unaids.org/en/media/unaids/contentassets/documents/epidemiology/2012/gr2012/20121120_UNAIDS_Global_Report_2012_en.pdf. [Google Scholar]
- Malawi Country Report. Global AIDS response progress report: Malawi country report for 2010 and 2011. 2012 Retrieved from http://www.unaids.org/en/dataanalysis/knowyourresponse/countryprogressreports/2012countries/ce_MW_Narrative_Report%5B1%5D.pdf. [Google Scholar]
- National Statistical Office (NSO) [Malawi], & ICF Macro. Malawi Demographic and Health Survey 2010. Zomba, Malawi and Calverton, Maryland, USA: NSO and ICF Macro; 2011. Retrieved from http://www.measuredhs.com/pubs/pdf/FR247/FR247.pdf. [Google Scholar]
- National Statistical Office (NSO) [Malawi], and ORC Macro. Malawi Demographic and Health Survey 2004. Calverton, Maryland: NSO and ORC Macro; 2005. Retrieved from http://www.measuredhs.com/pubs/pdf/FR175/FR-175-MW04.pdf. [Google Scholar]
- Odu OO, Asekun-Olarinmoye EO, Bamidele JO, Egbewale BE, Amusan OA, Olowu AO. Knowledge, attitudes to HIV/AIDS and sexual behavior of students in a tertiary institution in south-western Nigeria. The European Journal of Contraception and Reproductive Health Care. 2008;13(1):90–96. doi: 10.1080/13625180701617670. [DOI] [PubMed] [Google Scholar]
- Rimal RN, Böse K, Brown J, Mkandawire G, Folda L. Extending the purview of the risk perception attitude (RPA) framework: Findings from HIV/AIDS prevention research in Malawi. Health Communication. 2009a;24:210–218. doi: 10.1080/10410230902804109. [DOI] [PubMed] [Google Scholar]
- Rimal RN, Brown J, Mkandawire G, Folda L, Bose K, Creel A. Audience segmentation as a social-marketing tool in health promotion: Use of the risk perception attitude framework in HIV prevention in Malawi. American Journal of Public Health. 2009b;99:2224–2229. doi: 10.2105/AJPH.2008.155234. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rimal RN, Limaye RL, Roberts P, Brown J, Mkandawire G. The role of interpersonal discussion in reducing structural disparities and psychosocial deficiencies: experience from the Malawi BRIDGE project. Journal of Communication. 2013;63:51–71. [Google Scholar]