Social marketing is a commonly used strategy in global health. Social marketing programmes may sell subsidized products through commercial sector outlets, distribute appropriately priced products, deliver health services through social franchises and promote behaviours not dependent upon a product or service. We aimed to review evidence of the effectiveness of social marketing in low- and middle-income countries, focusing on major areas of investment in global health: HIV, reproductive health, child survival, malaria and tuberculosis. We searched PubMed, PsycInfo and ProQuest, using search terms linking social marketing and health outcomes for studies published from 1995 to 2013. Eligible studies used experimental or quasi-experimental designs to measure outcomes of behavioural factors, health behaviours and/or health outcomes in each health area. Studies were analysed by effect estimates and for application of social marketing benchmark criteria. After reviewing 18 974 records, 125 studies met inclusion criteria. Across health areas, 81 studies reported on changes in behavioural factors, 97 studies reported on changes in behaviour and 42 studies reported on health outcomes. The greatest number of studies focused on HIV outcomes (n = 45) and took place in sub-Saharan Africa (n = 67). Most studies used quasi-experimental designs and reported mixed results. Child survival had proportionately the greatest number of studies using experimental designs, reporting health outcomes, and reporting positive, statistically significant results. Most programmes used a range of methods to promote behaviour change. Programmes with positive, statistically significant findings were more likely to apply audience insights and cost-benefit analyses to motivate behaviour change. Key evidence gaps were found in voluntary medical male circumcision and childhood pneumonia. Social marketing can influence health behaviours and health outcomes in global health; however evaluations assessing health outcomes remain comparatively limited. Global health investments are needed to (i) fill evidence gaps, (ii) strengthen evaluation rigour and (iii) expand effective social marketing approaches.

Needle-syringe programs (NSP) have been effective in reducing HIV and hepatitis C (HCV) infection among people who inject drugs (PWID). Achieving sustainable reductions in these blood-borne infections requires addressing structural factors so PWID can legally access NSP services. Systematic literature searches collected information on NSP coverage and changes in HIV or HCV infection prevalence or incidence at the population level. Included studies had to document biomarkers (HIV or HCV) coupled with structural-level NSP, defined by a minimum 50 % coverage of PWID and distribution of 10 or more needles/syringe per PWID per year. Fifteen studies reported structural-level NSP and changes in HIV or HCV infection prevalence/incidence. Nine reported decreases in HIV prevalence, six in HCV infection prevalence, and three reported decreases in HIV incidence. The results support NSP as a structural-level intervention to reduce population-level infection and implementation of NSP for prevention and treatment of HIV and HCV infection.

Objective There are multiple reasons why females who inject drugs may be more likely to become infected with HIV than males who inject drugs. Where this is the case, special HIV prevention programs for females would be needed. Design International systematic review and meta-analysis of studies across 14 countries. Methods Countries with high seroprevalence (>20%) HIV epidemics among persons who inject drugs (PWID) were identified from the Reference Group to the UN on HIV and Injecting Drug Use. Systematic literature reviews collected data on HIV prevalence by gender for these countries. Non-parametric and parametric tests along with meta-analytic techniques examined heterogeneity and differences in odds ratios (OR) across studies. Results Data were abstracted from 117 studies in 14 countries; total sample size N=128,745. The mean weighted OR for HIV prevalence among females to males was 1.18 [95% CI 1.10–1.26], with high heterogeneity among studies (I2 = 70.7%). There was a Gaussian distribution of the log ORs across studies in the sample. Conclusion There was a significantly higher HIV prevalence among females compared to males who inject drugs in high seroprevalence settings, but the effect size is extremely modest. The high level of heterogeneity and the Gaussian distribution suggest multiple causes of differences in HIV prevalence between females and males, with a specific difference determined by local factors. Greater understanding of factors that may protect females from HIV infection may provide insights into more effective HIV prevention for both females and males who inject drugs.

Background: Persons who inject drugs (PWID) are at an elevated risk for human immunodeficiency virus (HIV) and hepatitis C virus (HCV) infection. In many high-income countries, needle and syringe exchange programs (NSP) have been associated with reductions in blood-borne infections. However, we do not have a good understanding of the effectiveness of NSP in low/middle-income and transitional-economy countries.

BackgroundInjecting drug use continues to be a primary driver of HIV epidemics in many parts of the world. Many people who inject drugs (PWID) are sexually active, so it is possible that high-seroprevalence HIV epidemics among PWID may initiate self-sustaining heterosexual transmission epidemics.MethodsFourteen countries that had experienced high seroprevalence (<20%) HIV epidemics among PWID and had reliable data for injection drug use (IDU) and heterosexual cases of HIV or AIDS were identified. Graphs of newly reported HIV or AIDS cases among PWID and heterosexuals were constructed to identify temporal relationships between the two types of epidemics. The year in which newly reported cases among heterosexuals surpassed newly reported cases among PWID, aspects of the epidemic curves, and epidemic case histories were analyzed to assess whether it was “plausible” or “highly unlikely” that the HIV epidemic among PWID might have initiated the heterosexual epidemic in each country.ResultsTransitions have occurred in 11 of the 14 countries. Two types of temporal relationships between IDU and heterosexual HIV epidemics were identified, rapid high incidence transitions vs. delayed, low incidence transitions. In six countries it appears “plausible” that the IDU epidemic initiated a heterosexual epidemic, and in five countries it appears “highly unlikely” that the IDU epidemic initiated a heterosexual epidemic. A rapid decline in incidence among PWID after the peak year of new cases and national income were the best predictors of the “highly unlikely” initiation of a heterosexual epidemic.DiscussionTransitions from IDU concentrated epidemics to heterosexual epidemics are common in countries with high seroprevalence among PWID though there are distinct types of transitions. Interventions to immediately reduce HIV incidence among PWID may reduce the likelihood that an IDU epidemic may initiate a heterosexual epidemic.