Does condom social marketing improve health outcomes and increase usage and equitable access?

Abstract

Condom social marketing (CSM) has increased condom supplies, broadened commercial markets for condoms and introduced marketing innovations in developing countries. Yet rigorous and reliable evidence of the impact on condom usage and disease prevention is limited, as is evidence of the impact on equity of access to condoms for poor populations, women and people living with HIV. One strand of research on CSM reports mostly on output (e.g. sales and processes) and market growth; but these have been found to be highly unreliable measures of condom usage. Another strand of research reports primarily on changes in sexual behaviour, attitude or condom usage, using survey data. While random sampling is rare, these studies often use representative samples, which provide some measure of validity. There have been attempts to improve the reliability or results to good effect, but challenges remain for researchers, scholars and donors, including the need to supplement output data with measures of behaviour change, use rigorous designs which are built into programmes a priori, report on equity measures, report on potential harms of CSM programmes, and encourage external and systematic reviews.

Résumé

Le marketing social du préservatif a accru l'approvisionnement en préservatifs, élargi les marchés commerciaux du préservatif et introduit des innovations commerciales dans les pays en développement. Pourtant, les données rigoureuses et dignes de foi concernant son impact sur l'utilisation du préservatif et la prévention des maladies sont limitées, ainsi que sur l'équité d'accès aux préservatifs des populations pauvres, des femmes et des personnes vivant avec le VIH. Une série de recherches sur le marketing social du préservatif portent principalement sur les résultats (par exemple les ventes et les processus) et la croissance du marché, mais il s'agit là de mesures très peu fiables de l'utilisation du préservatif. Une autre série de recherches concernent essentiellement les changements dans le comportement sexuel, l'attitude à l'égard du préservatif ou son utilisation, à l'aide de données d'enquêtes. Si l'échantillonnage aléatoire est rare, ces études ont souvent recours à des échantillons représentatifs, qui offrent un certain degré de validité. On a tenté avec succès d'améliorer la fiabilité ou les résultats, mais des défis demeurent pour les chercheurs, les spécialistes et les donateurs. Il est notamment nécessaire : de compléter les données sur les résultats avec des mesures du changement comportemental ; d'utiliser des configurations rigoureuses incorporées préalablement dans les programmes ; de faire rapport sur les mesures d'équité ; de notifier les dommages potentiels des programmes de marketing social du préservatif ; et d'encourager les évaluations externes et systématiques.

Resumen

El mercadeo social del condón (MSC) ha aumentado los suministros de condones, ampliado los mercados comerciales para los condones e introducido innovaciones de marketing en los países en desarrollo. No obstante, existe poca evidencia rigurosa y confiable del impacto del uso de condones y la prevención de enfermedades, así como del impacto en la equidad de acceso a los condones para los sectores pobres de la población, mujeres y personas que viven con VIH. Una línea de investigación sobre el MSC informa principalmente sobre los resultados (p. ej., ventas y procesos) y el crecimiento del mercado; pero se ha encontrado que éstas son medidas poco fidedignas del uso del condón. Otra línea de investigación informa principalmente sobre los cambios en comportamiento sexual, actitud o uso del condóm, utilizando los datos de encuestas. Aunque el muestreo aleatorio es raro, estos estudios a menudo usan muestras representativas, que ofrecen algún grado de validez. Ha habido intentos de mejorar la fiabilidad o los resultados, pero los investigadores, eruditos y donantes aún afrontan retos, como la necesidad de suplementar los datos sobre los resultados con medidas de los cambios de comportamiento; utlizar diseños rigurosos que se incorporen en los programas a priori; informar sobre las medidas de equidad; informar sobre los posibles daños de los programas de MSC; y promover revisiones externas y sistemáticas.

Key Words:

• condoms
• social marketing
• evidence-based medicine
• equity and access

Condom social marketing (CSM) involves the promotion and sale of condoms using commercial marketing techniques (e.g. advertising) and distribution methods, at a price that is cheaper than the private sector. CSM programmes have helped to increase supplies of condoms in developing countries and broadened commercial markets for condoms.^1–6 Creative use of mass media and consumer-focused methods by CSM organisations has introduced innovative condom messages and packaging (for example, Papo⁷ and Purdy⁸) and opened up advertising channels which previously prohibited condom advertising. CSM organisations have also collaborated with governments to improve public sector condoms, including packaging and promotion (e.g. in Kenya).⁷

In general, social marketing has been described as a highly evidence-based discipline and credited with introducing monitoring and evaluation to some not-for-profit and public sector programming.³⁹ Among international development donors there has been a groundswell of support for CSM in the past decade.^10–13 Yet there are indications that, despite its successes, the evidence base on CSM is limited with respect to rigour and reliability.

This paper presents a brief state of the evidence on CSM based on published literature reviews; an analysis of themes in the CSM literature; and recommendations for the research agenda from the perspective of evidence-based practice.

Quality of the evidence based on literature reviews

Systematic and literature reviews assessing the impact of CSM on condom usage have found very low levels of rigour, few examples of randomisation and limited use of controls or matched comparison groups.^13–17 From an evidence-based perspective, this means conclusions drawn from the evidence may have limited reliability. While all forms of evidence have value, the best way to measure effectiveness of an intervention, establish causality and ensure it does more good than harm is through rigorous study designs that include random sampling and experimental methods. Without randomisation there is a greater risk that the effects of an intervention will be overstated or misinterpreted,¹⁸ and it is difficult to determine if the results of a study are generalisable to the larger population or in other settings or contexts.¹⁹²⁰

Literature reviews published in 2001, 2007 and 2010 found a limited number of evaluations of equity of access and reported that evidence of equity for poor or vulnerable populations through such programmes is weak.²²¹²² For example, in 2007 Patouillard²¹ used relatively loose inclusion criteria (e.g. not just randomised trials) and found only six CSM studies that included equity measures.

Also on the issue of equity, many CSM programmes have been shown to be more successful in increasing condom uptake among men than women and among urban rather than rural populations.⁸²³ A recent systematic review and meta-analysis of behavioural HIV prevention interventions aimed at HIV-positive people could not find any relevant CSM studies or, therefore, any data on whether they are helping to reduce the transmission of HIV. This was in part because most studies did not target HIV-positive people or did not assess serostatus.²⁵ In countries with generalised epidemics, CSM programmes which target the general population may or may not reach HIV-positive people, depending on local epidemiology of HIV infection and the promotion methods used. However, targeting only HIV-positive people could have negative consequences, for example by further stigmatising condoms as something associated only with disease prevention, rather than health, sex and other positive aspects of condoms. Much remains to be done and learned in this regard.

Indicators of impact used to measure condom social marketing programmes

The reviews referenced above concluded that there is a paucity of rigorous research on CSM. Among evaluation studies of CSM programmes published in peer-reviewed journals, there is a distinct strand which reports mostly on output (e.g. sales), process (e.g. promotional strategies), awareness or market size or growth^{13,15–17,25,26} and another strand which reports primarily on changes in attitude or behaviour in relation to condoms, based primarily on survey data from non-experimental studies.²⁷²⁸

Regarding the first strand of studies, analyses have found that sales are highly unreliable as measures of condom usage. For example, this was the conclusion of Meekers and Van Rossem in their analysis of inconsistencies between data on condom use and sales.²⁹ Also, reporting of output may indicate a shift in where people get their condoms, rather than an absolute increase in uptake of condoms.¹ Sales or output data may be logical measures where a primary aim of a programme is financial sustainability (i.e. weaning itself off donor funding),^30–32 but they are not a valid proxy for usage. Couple years of protection (CYP) is another measure sometimes used as a proxy for condom usage. CYP is the amount of contraception necessary to protect one couple for one year, calculated based in part on how many units of a method of contraception were distributed during that period. Thus, CYP is largely a measure of output rather than usage.³ Several other drawbacks of CYP have also been pointed out in the literature.³³³⁴

The other strand of research involves reporting on impact in relation to behaviour change, such as condom usage. A recent systematic review concluded that measuring the impact of CSM on behaviour change (e.g. impact on reported condom use in last sexual act and in intentions to use condoms) is relatively rare in the literature.¹⁶¹⁷ Behaviour change studies generally utilise self-report survey data, with constructs such as proportion of the population who ever used a condom or used a condom in their last sex act.³⁵ Some are retrospective, (e.g. Eloundouu-Enyegue et al²⁸) but most have prospective designs (e.g. Meekers et al³⁵), which is a better measure of change, as retrospective studies are at risk of recall bias. Most studies of this type do not use random sampling, but many do use representative samples, which provide a better measure of impact and generalisability than, for example, convenience sampling.³⁶

There is a clear debate about risks of bias related to self-report measures of condom use in general, not just in relation to CSM. For example, a study by Noar, Cole and Carlyle³⁷ cited a number of issues common in studies of self-reported condom use measures, and recommend ways to increase reliability of such measures, such as employing multiple measures of condom use and comparison of results to identify discrepancies and measuring behavioural change at intervals of less than three months. In the CSM literature, Meekers' Zimbabwe study²⁷ measured self-reported condom use among a targeted population at short intervals, thus helping to increase the reliability of the findings. This is an example of how some CSM researchers have worked to improve the reliability of results to good effect.

Challenges for condom social marketing researchers, scholars and donors

To address the limitations within the current evidence base on CSM and build on good practices, organisations, scholars, donors and others should ensure that studies of CSM programmes go beyond reporting only on output data, and include measures of behaviour change. This is a particular challenge for CSM interventions that are implemented at the national or regional level; it is difficult to attribute any change in behaviour to the intervention, as many other factors could influence individuals. Nevertheless, it is a vital step towards improving the evidence base.

Increase rigour

In general, CSM research should include more rigorous designs. Experimental studies are encouraged, but well-designed analytical surveys also help to support the evidence base (e.g. prospective longitudinal studies, such as that used by Meekers in Zimbabwe²⁷), particularly if they involve random sampling of participants.

Randomisation in studies is rare in the HIV prevention literature in general,³⁸ and there are continuing ethical debates about its use (for example, see Lie³⁹). However, there are ethical ways to conduct non-placebo, randomised controlled trials of condom promotion and distribution, as demonstrated by condom use studies from various settings, including the United States,⁴⁰⁴¹ Madagascar,⁴² and Kenya and South Africa.⁴³ Randomisation can also be used in non-experimental studies through random sampling techniques rather than, for example, convenience sampling.³⁶ Conducting randomised trials of population-level interventions (such as national level campaigns) poses methodological and pragmatic challenges.⁴⁴⁴⁵ Nevertheless, researchers and CSM organisations should investigate approaches with the highest possible level of rigour, follow recommendations for increasing reliability of the chosen method (for example, those by Noar et al³⁷) and look to other areas of behavioural research where rigorous methods have been used to good effect.⁴⁴⁴⁵

A better quality of evidence can also be encouraged by ensuring that rigorous evaluation is built into a programme, rather than an add-on or post hoc effort. Patouillard²¹ points out that many interventions are not established as research exercises; thus, evaluation must fit around other priorities. This is no doubt the case in many CSM programmes, particularly those dependent on donor funding, as donor priorities may not always match those of scholars of evidence-based practice. This remains a challenge for all those seeking to improve public health in developing countries.

Report on equity of access

Many areas of intervention – not just CSM – have limited evidence related to equity of access, and collecting and assessing evidence on equity is complex.⁴⁶ There is a continuing need to examine the socioeconomic distribution of benefits of CSM programmes.²¹ Differing opinions have been expressed in the literature about whether social marketing is the best strategy for reaching hard-to-reach, marginalised or otherwise vulnerable populations,²⁴⁷ and about the characteristics which may make a CSM programme more or less effective in this regard. CSM programmes designed primarily to stimulate manufacturers and other private sector entities to engage in the production, distribution and promotion of condoms (known as the “manufacturer's model”)^30–32 have been shown to increase access to condoms for people who can afford to pay for them, and may ultimately wean people off free or subsidised condoms³¹³² to the benefit of the manufacturer and distributor. There has been limited research on this so far,⁶⁴⁸ but there are suggestions that it:

“…may not have a substantial impact on contraceptive prevalence or other quantifiable public health goals (especially in the poorest countries) as unmet need tends to be found among low-income groups that may not be able to afford commercial prices.” ³⁰

Hence, it is imperative to improve the evidence base by including equity measures in all CSM studies (e.g. socioeconomic baseline and follow-up data in experimental and analytical studies). It is also vital to be as transparent as possible in measuring and acknowledging the limitations of CSM in terms of equity of access for poor or vulnerable populations, and to consider the impact that price may have on access.²⁴⁹ For example, a 2007 study by Population Services International in Trinidad and Tobago included a strong study design, socioeconomic data and measures of the effect of price on the purchase and use of condoms.⁵⁰

Report on potential harms

There have been few studies which assess any measures of harm that may occur as a result of a CSM programme. The exception is an ethnographic study of a CSM campaign in a community in Mozambique, which explains how it clashed with local values and was blamed for an increase in prostitution, “promiscuity” and HIV, and led church officials to ban the use of condoms completely among parishioners.¹¹ An NGO reported that the campaign made it very difficult for them to gain church acceptance of condoms at all. Yet an earlier published evaluation of the CSM intervention by the implementing agency made no mention of negative effects.⁵¹

While this example cannot be generalised to other settings or even time periods (the surveys it reports on were conducted a decade ago), it is a reminder that measuring the effects of an intervention should include measures of both good and harm and is a fundamental aspect of transparency.⁵²⁵³ Measuring harm is facilitated by experimental designs, which can suggest causality, but even analytical and descriptive studies can generate hypotheses related to potential benefits and harms for future research.

Encourage external and systematic reviews

At present, a large proportion of the research on CSM is conducted by organisations involved in CSM, and this should be addressed by encouraging and publishing rigorous external evaluations. External evaluation does not guarantee greater reliability or validity of results, but it does address the fundamental risk of bias.^* The lack of external evaluation may not necessarily be due to a lack of openness on the part of CSM organisations, but may indicate a lack of interest or motivation within the academic community or among funders. This suggests a need for more partnering between CSM organisations and scholars.

In addition, while there are a number of literature and systematic reviews which assess CSM studies,^{2,13,15–17,21,22,24} more varied reviews assessing different aspects of CSM impact could help to improve the evidence base and the understanding of where, when and with which populations CSM works best. Systematic reviews, in particular, are considered the “gold standard” of evidence⁵⁵ as they synthesise the results from multiple studies. Donors such as the UK Department for International Development have commissioned a large number of systematic reviews in recent years to inform their policy and funding decisions (<www.dfid.gov.uk/r4d/SystematicReviewNew.asp>). Reviews of studies which include more than just stand-alone CSM programmes would be useful, such as those which are accompanied by education and skills-building interventions,¹ as well as reviews comparing methods or characteristics of distribution (e.g. public compared to private sector, free compared to subsidised).

Systematic reviews can provide a stronger level of evidence than non-systematic literature reviews, but they too can vary in quality.⁵⁴⁵⁶ For example, they are prone to publication bias, whereby studies with positive results are more likely to be reviewed than negative studies because positive studies are more likely to be published and therefore easily available. Also, due to the difficulties of tracking down unpublished studies, which many CSM evaluations remain, systematic reviews may provide a limited picture of the evidence on a topic. Moreover, they often do not examine or consider how study contexts might affect the delivery, uptake or effectiveness of interventions.⁴⁶ In light of these issues, CSM organisations may consider partnering with academic institutions or researchers to ensure good quality research on CSM in future.

Tehran's first condom shop, Iran, 2005

Notes

* For a discussion of conflict of interest bias in intervention research see, for example, Littell.⁵⁴