Midline Research Report
BBC World Service Trust
The full 69-page paper outlines a midline survey conducted to measure the awareness and reach of the 3-year STOP HIV campaign, which the BBC World Service Trust (WST) launched in April 2005 in an effort to raise awareness and change behaviour around HIV and AIDS among young Nigerians aged 15-24. Funded by the United Kingdom (UK) Department for International Development (DFID) and carried out in partnership with the Nigerian Society for Family Health (SFH), the project included 2 weekly phone-in talk shows ("Flava" and "Ya Take Ne?"), radio and television public service announcements (PSAs), a weekly TV drama ("Wetin Dey"), and film outputs (in partnership with the Nigerian video film industry, HIV- and AIDS-related messages were incorporated into the storylines of dozens of films). In addition to assessing the effectiveness of each of these media outputs, the midline survey measured changes in knowledge, attitudes, and self-reported behaviours (practices) regarding: HIV and AIDS, sexually transmitted infections (STIs), condoms, HIV testing, and stigmatisation and discrimination. It also looked at media consumption habits and preferences, gender equality, and perceptions towards issues such as abstinence and faithfulness.
Prior to the broadcast of the programme outputs, the BBC WST's Research and Learning Group conducted a quantitative baseline study of the intended audience. This study of 6,004 active media consumers (those who have either watched television or listened to the radio within the last month) aged between 15 and 24 years old was conducted in October 2005 across 6 states in Nigeria: Lagos, Abuja, Kano, Cross Rivers, Kaduna, and Enugu. It served as a benchmark measurement of knowledge, attitudes, and practice (KAP) in relation to HIV and AIDS and STIs against which the midline and endline surveys can be compared. In March 2007, the first of 2 midline studies was conducted. It was a survey of active radio listeners (those who have listened to the radio within the last 4 weeks). The second study, carried out in July 2007, was a survey of 2,412 regular media consumers (those who have either watched television or listened to the radio within the last 4 weeks) across the same 6 survey states. This report outlines key findings from the second midline survey, and highlights the main learning points for the STOP HIV campaign.
Multi-stage, proportional-to-size random probability sampling was used, with specific survey locations selected to be proportionally reflective of the urban and rural composition of the 6 states. Households were selected using a random walk method with qualifying respondents selected randomly from each chosen household using a Kish grid. The questionnaire was translated into the 3 major Nigerian languages - Hausa, Ibo, and Yoruba - and administered using locally trained interviewers. Male respondents were interviewed by male interviewers and female respondents by female interviewers.
Campaign Awareness and Reach
Overall awareness of the STOP HIV campaign was high in the research states; 84% of media consumers were aware of the campaign. By analysing sample contact data, 97% of respondents who qualified for the survey against demographic criteria were found to be media consumers, implying that approximately 8 in 10 young Nigerians were aware of the STOP HIV campaign. Campaign awareness was higher in the south (hereafter, defined as Lagos, Enugu, and Cross Rivers) compared with the northern states (hereafter, defined as Kano, Kaduna, and Abuja). Campaign awareness increased to 91% when including the radio dramas of the BBC WST's local partner SFH.
Eighty percent of respondents had actually seen or heard at least one of the campaign outputs. This was approximately equivalent to 5 million young Nigerians in the 6 research states. Researchers estimate that, if this level of audience reach were consistent across Nigeria, the STOP HIV campaign could have a potential audience of 20 million out of the estimated 27.1 million young Nigerians within the age range of 15-24 years.
Radio spots reached the widest audience (70% in northern states and 60% in southern states). This was followed by TV spots (39%), Ya Take Ne (19%), Wetin Dey (10%), and Flava (6%). TV spots attracted bigger audiences in the south (54%, compared to 25% in northern states), while radio outputs had a strong reach across Nigeria. These trends reflect general, albeit changing, media consumption patterns in Nigeria - television consumption is higher in the south whilst radio consumption is generally higher in the north. The results also suggest that the Trust's multi-platform and multi-language campaign strategy is being successful and outputs are being consumed according to local language needs and media access conditions.
Wetin Dey Audience Engagement
The researchers make some specific observations about the TV drama Wetin Dey, in particular, noting that it was building a loyal and engaged audience. After 3 months of weekly broadcasting on the national TV channel NTA, 20% of NTA's audience had watched Wetin Dey; this is equivalent to 16% of TV viewers and 10% of the total sample. Half of Wetin Dey's audience watched every week, and more than half spontaneously recalled the 5 main characters in the drama. There were high levels of character engagement among the audience, further reinforced by qualitative research findings, supporting the drama's use of multi-character storytelling.
When asked about the messages learned from the drama, over half (57%) mentioned testing for the HIV virus that causes AIDS, 37% mentioned "using condoms", and 29% mentioned "reducing stigma around HIV and AIDS". However, the research showed that characters, and their associated storylines, were interpreted in different ways by different segments of the audience. This implies that a uniform quantitative "effect" would be unlikely and difficult to measure; more subtle, qualitative measures of success are required.
Levels of Project Exposure
The evaluation examined whether there is a relationship between exposure to the STOP HIV campaign and the magnitude of behaviour change. After controlling for gender, age, education, and location, a positive relationship was identified between HIV- and AIDS-related knowledge, attitudes, and behaviour and exposure to the STOP HIV campaign. Analysis also revealed that the STOP HIV campaign contributed to positive trends in HIV testing, willingness to talk more with others about sex, condom use at last sex, stigma towards people living with HIV, and general awareness of STIs, HIV, and AIDS.
Recognising that audience members could be exposed to STOP HIV Campaign outputs at varying levels of intensity, measures of campaign exposure were introduced. These amounts were threefold and classified as: High, Medium/Low, and No Exposure. Allocated membership to the appropriate exposure group was made using a combination of recency and frequency of watching or listening to campaign outputs.
Below is a summary of some of the findings as relevant to the formal project success indicators.
Indicator 1: Increased percentage of young Nigerians with comprehensive correct knowledge of HIV and AIDS and prevention methods:
Awareness of HIV significantly1 increased from 91% at baseline to 94% at midline. AIDS awareness remained stable, at around 95% between baseline and 96% in the midline. Among those respondents who were highly exposed to the campaign, there is near-total awareness of HIV (98%) and AIDS (99%). This is significantly higher than HIV and AIDS awareness at baseline (91% and 95%, respectively) and among those with no exposure at the midline (88% and 94%, respectively).
To understand the intended audiences' knowledge of how the HIV virus is transmitted, respondents were asked how someone can contract the HIV virus. This was an open-ended question, and respondents were only probed on whether they could offer any more modes once a response was given. Knowledge of sexual intercourse as a mode of HIV transmission is extremely high and has remained high since the baseline (95% at both baseline and midline). Respondents who were highly exposed to the campaign were significantly more likely to mention blood transfusions (57%) than those with no exposure (42%) or the baseline (49%).
Awareness of risk reduction increased significantly between the baseline survey (87%) and the midline survey (90%). A significantly higher proportion of respondents who had high (93%) or low/medium (91%) exposure to the campaign reported that there is something that a person can do to reduce their risk of contracting HIV, compared to the proportion with no exposure (82%) or the proportion of respondents making this claim at the baseline (87%). Significantly more high exposure respondents mentioned abstinence as a means of risk reduction (68%) compared to the medium/low exposed (60%), unexposed (59%), and baseline (63%).
Awareness of local HIV testing centres significantly increased from 40% at the baseline to 52% at the midline, and respondents reporting that it is possible to get an HIV test in their area significantly increased with exposure to the STOP HIV campaign. Respondents reporting wanting to be tested increased significantly (39% at baseline, 47% midline). Significantly more respondents who were highly exposed (52%) reported wanting to be tested than those with low/medium exposure (46%), no exposure (42%), and baseline (39%). The proportion of respondents reporting having actually taken an HIV test statistically increased from the baseline (10%) to the midline (14%).
Indicator 2: Increased percentage of young Nigerians with correct knowledge of STI symptoms and treatment:
Significantly more respondents with high exposure to the campaign reported being aware of STIs (91%) compared to those at baseline (76%), unexposed at midline (77%), and low/medium exposed (86%) at midline. There were no significant differences between baseline respondents (76%) and those not exposed to the campaign (77%). At the midline, the average number of female and male STI symptoms respondents could name was significantly higher among those with high exposure to the STOP HIV campaign (2 symptoms in both men and women) compared to the baseline and to those with lower levels of exposure to the campaign (1 symptom for men and women at baseline and midline no exposure). Respondents with high exposure to Trust outputs could name significantly more male STI symptoms (2) than those with low/medium exposure (1) or no exposure (1).
Indicator 3: Increased percentage of young Nigerian women and men with the confidence and skills to negotiate equitable partnerships:
The proportion of respondents who have had sex with boy/girlfriends in the last 12 months and who had discussed condom use with their sexual partner(s) remained stable (baseline 81%, midline 81%). There were positive increases across exposure groups; however, these were not statistically significant due to small sub-group sample sizes. The proportion of sexually active respondents reporting feeling confident in convincing a boyfriend or girlfriend to use a condom every time they have sex decreased from baseline (85%) to midline (83%). This difference is not statistically significant.
The proportion of sexually active respondents reporting having used a condom the last time they had sex with their girlfriend or boyfriend increased (64% at the baseline to 69% at the midline). Respondents with high exposure reported higher condom use at last sex (73%) than those at baseline (64%), and low/medium levels of exposure (67%). However, these were not statistically significant, influenced by the small sub-group sample size.
Indicator 4: Increased percentage of young Nigerian men and women with the confidence and skills to negotiate support from their parents and community leaders conducive to their sexual health:
The number of people with whom respondents felt comfortable talking about sex increased. The increase was, however, not statistically significant. Talking with more people about sex is positively associated with exposure to the STOP HIV campaign. There was a significant increase in the average number of people respondents had spoken to about sexual health matters from baseline (1.5) to midline (2.0). Highly exposed respondents reported greater willingness to talk about sex with significantly more people (2.2) compared with baseline (1.5) and low/medium exposure groups (2.0). Significantly more highly exposed respondents reported being very comfortable talking to a wider range of people than was the case at the baseline.
Indicator 5: Increased tolerance and acceptance among young Nigerians towards people living with HIV and AIDS (PLWHA):
Acceptance of people living with HIV and AIDS was assessed by understanding stigma in the public space where one has little to no control (e.g. co-worker in the office, at school), in the personal-public space which "I control" (e.g. people serving me food), and in the familial setting. High levels of exposure to the campaign are associated with significantly lower levels of stigma in the public, private-public, and familial settings. A majority (59%) of the respondents stated they would not stigmatise PLWHA in at least one of the stigma-related scenarios they were asked about during the baseline survey. This significantly rose to 63% during the midline.
Significantly more high and medium/low exposed respondents agreed that a co-worker in the same office who is HIV-positive but healthy can continue his/her normal activities (66% and 66%, respectively) than the baseline (57%). People with high exposure (68%) and medium/low exposure (69%) were significantly more likely to agree to allow a classmate who is HIV-positive to continue his/her normal duties than the baseline (59%). Significantly more highly exposed (69%) and medium/low exposed (68%) respondents agreed that a teacher who is HIV-positive but healthy can continue his/her normal duties compared with the baseline (61%). More people who were highly exposed (68%) to the outputs are likely to accept those community leaders who are living positively than those who were low/medium exposed (64%). Acceptance for the leader is also significantly higher among the exposed groups than at the baseline (59%).
Similarly, higher levels of exposure to the STOP HIV campaign were associated with increased tolerance to HIV in the private-public sphere, such as in a public health environment. Respondents with high exposure and medium/low exposure to campaign formats agreed that a nurse who has HIV but is healthy should be able to continue his/her duties (48% highly, 46% medium/low) - significantly more than those at baseline (30%) and no exposure groups (36%). Significantly more respondents who were highly exposed to the campaign's outputs agreed that a person serving food who has the HIV virus and is healthy should be able to continue their normal activities (42%), compared to baseline (32%) and no-exposure group (37%).
Significantly more people in the highly exposed group (52%) and low/medium exposed (53%) agreed that a spouse who is HIV-positive but healthy can continue his/her normal duties cmpared to those who were unexposed (43%). Significantly more respondents with high and medium exposure (71% and 73%) said that they would want the HIV status of a family member who is HIV-positive to remain a secret compared to the baseline (63%) and unexposed group (66% - statistically different to the low/medium exposure group only). Significantly more respondents with high exposure and low/medium exposure (88% and 87%, respectively) reported that they would care for a family member than at the baseline (78%). This suggests that although exposure to the STOP HIV campaign is associated with greater willingness to care for an HIV-positive family member, it is also associated with an increased desire to hide the status of such a family member. This finding suggests that, despite increased tolerance towards people living with HIV/AIDS (PLWHA) at other levels (e.g. public and private-public spaces), stigma in a familial setting may still be a barrier to sustainable behaviour change and is likely driven by the desire to protect family members.
An excerpt from the Conclusions section of the report follows:
"At the mid-point of the campaign, the Trust in Nigeria is showing great promise in delivering HIV and AIDS media outputs to young Nigerian media consumers....The research findings have highlighted the importance and the success of using a multi-platform and multi-language mass media campaign to reach audiences in a heterogeneous country such as Nigeria where, intra alias, language, ethnic and religion vary widely across the country. The research has shown that the Trust's approach in Nigeria has, to date, been able to reach 80% of the target audience - a level of reach that is higher than if the campaign had been based around a single-output or a single media...
...There is strong indication from the research that knowledge, attitudes and behaviour around HIV and AIDS can be changed positively using mass media....Stigma in the public sphere, public-private and familial sphere has also significantly improved over time and across exposure groups. The TV spots have been shown to be particularly effective in this area. The campaign has also faired well in raising awareness to issues which were relatively unknown before the campaign (e.g STIs). However, STI awareness - especially knowledge of STI symptoms in both males and females - is still worryingly low and warrants further targeted communication.
Where the campaign has been less successful is when indicators are already very high (e.g. knowledge that sex is a mode of transmission - already at 95% at baseline and has not increased, self assessed confidence to negotiate and discuss condoms with boyfriends/girlfriends) and when indicators involve the disclosure of an HIV positive family member (a majority would still want this to remain a secret). The latter may take a longer period of time before changes are visible and also warrants further research into the reasons underlying the desire to hide a family member's HIV-positive status. Finally, the campaign has not consistently shifted each of the ABC (Abstinence, Be Faithful and Condoms) indicators and there are still pockets of misconceptions around modes of HIV transmission which need to be addressed.
...The midline study has demonstrated that mass media has the potential to contribute to actual behaviour change, as well as gains in knowledge and shifts in attitudes..."
The recommendations below were given as examples of how findings from the research could provide production- and research-focused insight for future direction of the project.
- Maintain a multi-platform and multi-language mass media campaign to reach the diverse audiences in Nigeria.
- Strive to continue the successfully received health content around STI awareness, willingness to talk to others about sex, knowledge and attitudes towards PLWHA, and condom use at last sex.
- Reinforcement of all basic HIV and AIDS knowledge indicators is strongly recommended.
- Consider using the lack of factual knowledge around HIV and AIDS as a messaging tool kit. For example, developing and executing a 'Basic Facts' drive. Ask the audience if they know the facts i.e. introduce a catchphrase such as 'do you know the 8 fact rule' - the 4 main routes of transmission, the number 1 way to contract HIV is through unprotected sex (5), and there are only 3 ways to prevention HIV infection Abstinence (6) Be faithful (7) Condoms (8).
- Develop consistent condom use messaging to complement the increase in condom use at last sex.
- Continue to address stigma to sustain the change in attitudes towards PLWHA in a public space.
- Challenging misconceptions and reinforcing knowledge of the basic modes of HIV transmission will assist any messaging to address stigma in the public-private space.
- Showcase further examples of PLWHA continuing their normal activities including relationships which include sex.
- Conduct further research into why the status of family members would be hidden.
To inquire about accessing the full report, please see the contact details below.
Email from Emily LeRoux-Rutledge to The Communication Initiative on October 1 2009.
1 Where findings are referred to as significant, this is informed by z and t testing appropriate to the data with the probability level set to p<0.05.