Data processing and analysis

Our methodological approach is situated at the intersection of grounded theory [22] and thematic narrative analysis [23] and triangulates between three primary analytical components: (1) analysis of quantifiable characteristics of the narratives (“quantitative attributes”); (2) thematic qualitative data analysis, focusing on thematically-related text segments and memoing for emergent themes; and (3) a narrative-based approach, focusing on plot summary and thematic keywords.

Our methods were developed to enable cross-national and temporal analysis and have three main advantages: (i) they ground the analysis in three distinct, though intersecting, dimensions of the data; (ii) allow triangulation; and (iii) facilitate the generation and validation of interpretive hypotheses.

We addressed the three primary analytical components as follows:

1. We quantified discrete components of each narrative, double-entering them into Qualtrics research software (Qualtrics, Provo, UT); in cases of discrepancy, consensus was reached via dialogue and/or adjudication by a third team member. Examples of these quantitative attributes of the narratives included whether a narrative includes a focus on prevention, infection, and/or life post-infection or whether it includes sexual transmission of HIV. Data were downloaded to Excel files; descriptive statistics were calculated in SAS software, Version 9.4 (SAS Institute Inc., Cary, NC).
2. Narratives (overwhelmingly handwritten) were transcribed into English or French and entered verbatim into MAXQDA 12 qualitative data analysis software (VERBI Software, 1989–2016), where they were labelled with reference to a codebook of 54 codes, including, for example, condom, abstinence, and fidelity/infidelity.
3. Each narrative was then summarized and independently double-coded with up to six out of 44 keywords, which included abstinence, condoms, and fidelity/infidelity. Discrepancies in coding were resolved through dialogue. The narrative summaries served as an aide-memoire and data-management tool, allowing us to easily identify common story arcs or to rapidly situate segments of data within the context of the overall plot or argument. The summaries also function in combination with thematic coding, allowing us to isolate both individual text segments related to a specific theme, for example, condoms, and those narratives in which condoms are a central theme.

Our qualitative analysis addresses both narratives coded with prevention-related keywords (3. above) and segments of other texts thematically coded with prevention themes (2. above). Existing thematic coding across a range of deductive (e.g. “poverty”) and inductive themes (e.g. “negative reactions to HIV diagnosis”) was complemented with fine-coding specific to themes related to the prevention of sexual transmission of HIV. Examples of fine codes include “barrier to condom: sexual violence” and “facilitator of abstinence: agency”.

This study, comprising the secondary analysis of existing data, was approved by Emory University Institutional Review Board. The narratives were provided for analysis by Global Dialogues (www.globaldialogues.org; info@globaldialogues.org). We cite them verbatim. Country names are abbreviated as follows: SZ–Swaziland; KY–Kenya; NG–Nigeria; BF–Burkina Faso; and SN–Senegal. Excerpts are identified by the country, contest year, gender, age, and geographic location of the author. The study’s limitations, related to its distinctive data source, are described at the end of the Discussion section.

Prevalence of prevention strategies

The prevalence of HIV prevention in the narratives differs by prevention strategy, by country, and over time. Of the ABC, condoms appear most prominently (mentioned in 33% of the 1,343 narratives and a central theme in 14%), followed by abstinence (mentioned in 19% and a central theme in 4%), and fidelity, infidelity and/or partner reduction (mentioned in 13% of the narratives and a central theme in 11%). Condoms referenced in the sample are almost exclusively male condoms; a female condom is mentioned in five narratives, in only one of which (a male-authored 2005 Kenyan narrative) it is used in a sexual encounter. Characters are tested for HIV in over half (n = 812, or 60%) of the 1,343 narratives, though this often serves to provide a turning point in the plot rather than being represented as a prevention strategy. Among the narratives in which testing occurs, 743 (92%) include positive tests, while 121 (15%) include negative tests; testing leads to characters taking action to prevent HIV transmission in 5% (n = 38) of these narratives.

Biomedical prevention methods are rarely mentioned or depicted in the narratives. Prior to 2014, male circumcision is represented as a traditional practice that is a mode of blood route transmission. VMMC for prevention purposes is ambiguously referenced in only one narrative, a 2014 text from Swaziland by a 19-year-old rural male author in which a male character, after having unprotected sex, questions the protection his circumcision offers. Only three narratives, all by 2014 male Kenyan authors, reference PEP in their narratives. These depict female-to-male transmission risk in which male characters access PEP to prevent transmission. PrEP and TasP are absent from the sample.

Individual versus structural factors

While there are no representations of structural or combination prevention per se, representations differ in their emphasis on characters’ individual responsibility to protect themselves from HIV via behavioural strategies versus acknowledgement of the social and economic factors that increase vulnerability. Isolated narratives depict family members (NG 2008, M 18 R) or community organizations (BF 2008, M 22 R) providing economic support that allows female characters to avoid HIV risk associated with transactional sex, pointing to community-generated structural prevention efforts. However, the majority of representations of HIV prevention in the sample depict behavioural methods (i.e. abstinence, fidelity, condoms) for primary prevention.

Narratives emphasize individual responsibility for HIV prevention in 1 out of 3 narratives that include abstinence, fidelity/infidelity, partner reduction, or condoms. In these narratives, authors depict characters’ unwillingness or inability to put into practice these methods due to overwhelming lust, lack of perceived risk, or lack of knowledge about HIV infection. These narratives are most prominent in the Nigerian and least prominent in the Burkinabe sample.

One in 5 of the narratives highlight the social and economic vulnerabilities that place characters, overwhelmingly female, at risk of contracting HIV. For example, 1 in 3 narratives that include scenarios of condom use or non-use depict male characters refusing to wear condoms, threatening to withdraw economic support from female partners, or using sexual violence to obtain sex without a condom. Economic vulnerability, sexual coercion and violence, and male control of preventive practices (e.g. condom use) are among the primary obstacles to female characters successfully preventing HIV, and are most prominent in the 2014 sample. Narratives representing structural barriers are most common in the Swazi and Kenyan samples. While less blaming than narratives that solely emphasize individual responsibility for prevention, these narratives may reinforce a sense of tragedy and disempowerment in the face of HIV as narrators and characters lament the circumstances that facilitate HIV infection and inhibit prevention but rarely offer tangible ways to address them. One in 15 narratives, most prominently from Kenya and in the 2014 sample, acknowledge social and economic drivers of HIV risk while still emphasizing individuals’ responsibility to use HIV behavioural prevention strategies. Counter to the aims of structural and combination prevention, these narratives are particularly disempowering; for example, in an albeit extreme case, a narrative in which a girl is violently raped begins with the message to abstain from premarital sex (KY 2005, F 12 U).

Gender

More narratives feature female protagonists over time, reaching 75% of narratives in 2014, compared to 44% with male protagonists. While more narratives depict male characters becoming HIV-positive in earlier years (72% in 1997 compared to 49% in 2014), the opposite is true for female characters (64% in 1997 vs. 74% in 2014).

The gender of protagonist differs by prevention strategy: the majority of the narratives in which condoms are a central theme depict male protagonists, while female protagonists predominate in narratives that include the abstinence theme. The overarching plotline and narrator commentary in the personal responsibility narratives reinforce moralistic messages that condemn active female sexuality, idealize female virginity, and connect HIV with immoral sexual behaviour; personal choices, particularly girls’ and women’s personal choices, are highlighted in these narratives as the driver of HIV risk.

Effectiveness of prevention strategies

Condoms, abstinence, fidelity and, to a lesser extent, partner reduction are often framed as effective HIV prevention methods via narrator commentary but are rarely depicted within a plotline as successfully preventing HIV transmission. For example, although narrator commentary cites abstinence as an effective HIV prevention method (specifically for unmarried youth) in 1 in 4 narratives that include the abstinence theme, fewer than half of these depict characters successfully protecting themselves via abstinence. For male characters, the most common barriers to abstinence are peer pressure and overwhelming sexual desire. Female characters, while facing these individual-level barriers as well, are also depicted as pressured, coerced, or forced into sexual intercourse by male partners, 1 in 5 of whom are in an economically advantageous position.

Condoms are depicted as not used (i.e. “unprotected sex”) approximately four times more often than they are used. When the young authors depict condom use, half of the time they portray characters becoming infected with HIV despite having used condoms. While these narratives occur in all countries, they are particularly prominent in the Nigerian sample where condom use results in HIV transmission four times more often than it prevents HIV transmission. As with representations of abstinence, the most common barriers to condom use for male characters are individual-level (specifically, lack of perceived vulnerability to HIV infection or loss of control due to lust), whereas female characters are most often challenged by partner refusal that is exacerbated by economic vulnerability, or by sexual violence.

Fidelity is frequently represented as ineffective; for example, 3 out of 4 characters practicing fidelity become infected. In these cases, infections occur primarily as a result of a partner’s inability or refusal to remain sexually exclusive or due to infection via blood. Male characters are depicted as facing similar challenges to fidelity as to abstinence (e.g. lack of perceived vulnerability to HIV, sexual desire for and access to female partners). Female infidelity is most often linked to the economic incentive of having multiple partners.

In all three cases in which PEP is referenced, it is highlighted as an effective prevention strategy in the case of risky sex or female-on-male sexual violence. Only male characters, two out of three of whom are depicted as sexually assaulted by women, use PEP; despite the much higher proportion of male-on-female sexual violence in the 2014 sample (57 female victims versus 7 male victims), no narratives mention PEP in the case of rape of female characters.

Temporal characteristics

Representations of HIV prevention in the narratives evolve over time. Where the 1997 sample focuses on the basic facts of transmission and prevention, the 2005 sample is distinctive for providing contextualized representations of successful prevention. The 2008 sample focuses more on testing as a gateway to treatment, while the 2014 sample sees a decline in prevention and increased preoccupation with the circumstances of infection.

Mentions of condoms and fidelity decline over time, while abstinence is more consistently present. The treatment of abstinence as a central theme peaks in 2005. In comparison with other years, the 2005 sample is distinctive in including a higher proportion of representations of successful abstinence, in which the combination of individual agency, self-control and supportive peers, partners and parents contribute to characters being able to choose to delay or avoid sex. 2014 narratives, in contrast, place less emphasis on the social facilitators of abstinence and depict greater challenges in terms of coercive relationships, while still underscoring the importance of individual agency in the choice to remain abstinent despite an unfavourable environment.

In 1997 (when data come only from Senegal and Burkina Faso), representations of condom use are rare but condoms are mentioned in over half of the narratives, often in combination with abstinence and partner reduction as the young authors underscore the need to prevent HIV in the absence of treatment. The use of humour to promote condom use, distinctive in the 2005 sample, disappears in subsequent years. In addition, the 2005 sample includes more characters, primarily female, negotiating condom use and refusing to have sex without one. Over time, condoms are mentioned and represented less often but more favourably. Representations of condoms as ineffective, either via descriptions of condoms being misused, breaking or simply failing to prevent HIV transmission, which recur in 2005 narratives from Nigeria in particular, become virtually absent in 2014. In 2008 and 2014, a handful of narratives (n = 6) depict the use of condoms to prevent transmission within serodiscordant couples.

Cross-national characteristics

The prominence of narratives featuring individual responsibility and/or structural vulnerability differs by country. While blame and moralizing around HIV prevention that emphasize individual responsibility are found in all countries, they are most common in the Nigerian and, to a lesser extent, the Kenyan sample. Reference to structural vulnerability, particularly in Swazi narratives, tempers the blame authors ascribe to characters for their inability to prevent HIV.

Empowering narratives

A small minority of narratives depict the successful prevention of HIV in ways that do not perpetuate HIV stigma or reinforce a sense of hopelessness or disempowerment. Narratives with successful prevention feature female empowerment and male respect for female characters’ desires. For example, while fewer than 1 in 25 narratives that include condoms depict successful condom use that prevents HIV, these positive representations are facilitated by supportive peers, community leaders and, for female characters, male partners who accede to their requests for protection. In one narrative, Assetou is propositioned by and initially refuses Germain. Her friends tell her about the risks of unprotected sex and urge her to use a condom. Assetou and Germain decide to have sex with a condom (BF 2008, F 21 U). Underlying these positive representations of HIV prevention lies a lack of stigmatization of sexuality in general and youth sexuality in particular. Sex outside marriage in these scenarios is not a deviant act worthy of punishment via HIV infection, but rather a common experience that carries risk that can be navigated with a menu of prevention strategies, deployed depending on the circumstances. Prevention methods therefore represent opportunities to navigate sexual risk rather than being indicative of personal failure. Gender and economic disparities are acknowledged but counteracted by supportive social environments that make it easier for female characters to exercise agency. These scenarios feature most frequently in the Burkina and Swazi samples.

Prevention strategies

The decline in prominence of HIV prevention themes intersects with increased access to life-extending ART (particularly evident in the 2008 sample) and a resurgence in moralistic preoccupation with the circumstances of infection in 2014 (following lowest levels of blame and hopelessness in 2008). It makes sense that young Africans’ narratives about HIV in the mid-to-late 2000s would include ART, following acceleration of access in the mid-decade. However, it is important to ensure that prevention remains a prominent dimension of young Africans’ thinking about HIV.

While it is encouraging that virulently moralistic condemnations of condoms as ineffective in the 2005 sample–which we associated with PEPFAR priorities and the prominence of the faith-based response [25]–have receded, their declining prominence in the narratives is of concern. Promotion and provision of male and female condoms is part of the DREAMS core package, launched in 2014, along with PrEP and testing [8]; both Kenya and Eswatini are part of the DREAMS partnership. It is, however, essential to ensure that condoms remain a central component of meaning-making around HIV prevention for young Africans in general, not just those at particular risk in higher-prevalence countries.

Warning against dangerous complacency, a 2018 Lancet Commission report [2] argues that the dominance of HIV treatment and biomedical approaches has resulted in the neglect and underfunding of primary prevention, along with efforts to address HIV-related stigma and other social and structural drivers. The shift of emphasis we observed in the narratives may reflect a declining prioritization of general youth prevention in the context of the increasing biomedicalization of the epidemic, as prevention has become less a matter for the general population and more focused on key populations and more clinic-based [26]. While mass media and community-level communication strategies were critical to the HIV response in the early days of the epidemic, they may be less prominent in light of more targeted approaches (summed up in UNAIDS “Know your epidemic, know your response” priority) [27]. In neglecting the general population for prevention messaging in favour of key populations, we exclude both those who are not easily reached by more targeted communication strategies and those who can help catalyze and sustain a supportive and enabling environment. This applies as much to biomedical prevention interventions [28] as to condom promotion.

While evidence supporting the major biomedical prevention strategies was in place before the 2014 scriptwriting contest–VMMC (2005–7), PEP (starting from the early 1990s), PrEP (from 2010) and TasP (2011)–these are virtually absent from the young authors’ narratives. This may reflect limited relevance, availability and/or awareness (depending on the strategy and the country). The countries in our study had baseline levels of male circumcision over 90% with the exception of Western Kenya and of Swaziland (which had a male circumcision prevalence rate of 8.2% in 2006–7 [29] but where VMMCs have increased substantially since then [9]). As only a proportion of our Kenyan texts are from Western Kenya, it makes sense that VMMC should feature most prominently in the Swazi sample, though it is disappointing that it is mentioned in only one 2014 text. We have been unable to locate data on PEP coverage in SSA; Ncube et al. [7] note that its availability is limited even in South Africa. PrEPWatch [30] reports the number of PrEP users per country. While Senegal is reported as having no users, as of June 2019, Kenya has the largest program in SSA and is reported to be the only country in SSA that is aggressively promoting PrEP use [11].

In summary, while limited relevance or availability accounts for silence around VMMC, PEP and PrEP in most countries in 2014, the silence surrounding TasP in all countries is of concern, given its universal relevance and its potential to benefit stigma reduction, testing, and treatment initiation and adherence. This silence is in keeping with qualitative studies conducted in 2013 in urban communities in South Africa and Zambia [31] and in 2015 among rural HIV-positive and HIV-negative South African men [32], which found no awareness of TasP. Further research is needed to determine to what extent this situation has changed in the intervening period and the extent to which it is attributable to policymakers’ and others’ reluctance to explicitly promote TasP for fear that it may lead to behavioural disinhibition and a rebound in unsafe sexual practices [32]. Continued low awareness of TasP would potentially be depriving SSA of a valuable tool in HIV prevention, treatment and stigma-reduction.

Social and structural drivers

The growing feminization of the epidemic is reflected in an increasing proportion of narratives with female protagonists. Whether the young participants in the scriptwriting competitions observed this phenomenon directly in their communities or indirectly via media representations reflecting epidemiological understandings, they are increasingly propagating and reinforcing a social representation of a feminized epidemic in their narratives. The young authors understandably struggled to reconcile acknowledgement of structural drivers of HIV with individual preventive behaviours, creating disempowering narratives of vulnerability or condemning characters for failing to prevent HIV in the face of often overwhelming structural challenges. In the context of combination prevention, an empowering cultural narrative that models prevention despite structural constraints and avoids stigma or blame is increasingly important. Such a narrative should incorporate biomedical prevention strategies and help build demand for them: as Ahmed et al. [33] suggest, for example, it is critical that PrEP roll-out use a gain-framed and empowering approach.

Social representations can be viewed as a form of social norms in symbolic form. They interact with cultural narratives in a mutually constitutive way (manuscript under review). The data–in Nigeria, in particular–point to a well-established cultural narrative of HIV prevention that takes the form of a moralistic, cautionary tale and that, our 18 years of data suggest, has considerable staying power. Cautionary tales as conceptualized by Moore [24] place primary emphasis on personal responsibility for risk, such that risk becomes a moralizing discourse and allow for individual blame for outcomes. In this way, cautionary tales support the maintenance of stigma by associating HIV infection with immoral behaviours and populations while constructing the disease as preventable via individual conduct [34]. Our analysis of a broader sample (manuscript under review) suggest that the hope-promoting and stigma-reducing influence of the advent of ART access on symbolic representations in the narratives (evident in the 2008 data in particular) may have partially run its course by 2011/2013. The three countries where DHS temporal data on stigma is available show a downturn in accepting attitudes towards people living with HIV among 15–24 year-olds (for Senegal from 2010, for Nigeria from 2013, and for Kenya from 2014) [35]. This points to the need to redouble stigma reduction efforts among youth and to attend to the ways in which HIV communication may be fuelling stigma.

A cultural narrative compatible with combination prevention is, by virtue of its complexity, unlikely to be as compelling as cautionary tales of infection, that seek to capitalize on the deterrent effect of a loss-framed message [36]. The dominant cultural narrative may be resistant to accommodating new information and approaches that lend themselves to less melodramatic and moralistic storylines. Moralistic dimensions of the narratives, particularly in relation to gender and sexuality, reflect power-laden positions and appear least susceptible to change, even in the context of dramatic increase in access to ART. Nonetheless, cross-national differences and the existence of empowering narratives in our sample indicate the potential for alternative framings.

The cross-national differences in the narratives, particularly as they relate to sexual morality, blame and hope, are salient. Our analyses extend temporally our cross-national study of symbolic stigma in the 2005 sample [37], with a specific focus on HIV prevention. They demonstrate the consistency of cross-national differences over time, with representations in the Nigerian sample remaining the most stigmatizing, a finding that is statistically significant across a wider range of time points (manuscript under review). Our findings point to the need for HIV communication in Nigeria to address the stigmatizing link between HIV and sexual behaviours represented as immoral. This can potentially be done by elevating empowering narratives that depict a supportive social environment for HIV prevention, possibly drawing on the pragmatic menu of prevention options present in Burkinabe narratives. Young people in Nigeria also need to be encouraged to critically reflect upon the social and economic drivers of women and girls’ vulnerability to HIV, including sexual violence. Kenyan narratives continue to convey distinctively mixed messages, combining blame with acknowledgement of structural barriers. This characteristic, combined with the absence of representations of supportive social environments, points to the need to promote empathy with young Kenyans who are vulnerable to HIV and to disseminate empowering narratives that emphasize the agency of those at risk. There is also a need to go beyond messages of personal responsibility for infection and embrace messages about the collective responsibility to create a supportive social environment for HIV and violence prevention and response. Narratives from lowest prevalence Senegal remain ‘othering’ in their focus on ‘high risk’ groups and migration as a primary risk factor, suggesting a priority need to promote greater awareness of risk perception among Senegalese youth. Narratives from highest prevalence Swaziland, in contrast, continue to reflect the social proximity of the epidemic and awareness of structural factors, while the Burkina Faso sample is characterized by a non-stigmatizing pragmatism. While narratives from Swaziland and Burkina Faso have distinctive strengths from which other countries can learn, it is important to avoid complacency and remain vigilant in relation to stigma. In Swaziland in particular, our data point to a need for greater focus on secondary prevention, while in Burkina Faso, they point to a need to raise awareness of sexual violence, which is largely absent from the narratives.

Across all countries, we recommend increased focus on positively-framed prevention narratives–potentially drawing on examples in the 2005 and Burkina Faso samples–featuring empowered characters (male and female) creatively navigating circumstantial (and, where appropriate, structural) challenges to protect their health and achieve desired outcomes. These narratives can provide alternative role models for young people, cultivate skills development, boost their belief in themselves and their capacities, and align with combination prevention. While these narratives may lack the melodrama of the cautionary tale, their positive emotion can be leveraged in health education and communication materials and curricula to model and build skills around the successful application of prevention strategies.

Limitations

This study has limitations related to its distinctive data source. Data for 1997 are available from only two countries, while those from Nigeria come only from the Igbo-speaking South-East. As contest participants self-select, the data are not representative of the youth populations; participants in the scriptwriting competitions may be better educated and more knowledgeable and motivated about HIV than the general youth population. As a product of the same contest mechanism, however, these biases are likely to be consistent across the five countries hence the country samples, though not representative, are comparable for our purposes. We have little demographic information about individual participants other than their gender, age, country of origin, and type of place of residence. In light of the circumstances in which the texts were written, it is impossible to know which depict lived versus imagined experience. The data are embedded within cultural norms of performance, discourse and persuasion [38], which may be informed by rhetorical considerations specific to the scriptwriting competition, reflecting the young authors’ motivation, for example, to tell what they consider to be a good story and thereby win the contest. The programmatic context of the scriptwriting competitions, including the shift in elicitation and broadening of themes from 2013, may have influenced the consistency of the data in ways that it is impossible for us to gauge. Despite their limitations and in the absence of comparable temporal and cross-national data, the narratives provide rare and distinctive insight into evolving sense-making about HIV prevention among a presumed general population of young Africans from five countries and across a time period of shifting strategies and programmatic priorities in efforts to prevent the sexual transmission of HIV.