aUS Department of Health and Human Services/Centers for Disease Control and Prevention, National Center for HIV, STD, and TB Prevention, Atlanta, Georgia, USA
bPopulation Services International, Kenya
cUS Health and Human Services/Centers for Disease Control and Prevention, Office of the Director, Kenya
dNational AIDS and STD Control Programme, Kenya
eLiverpool VCT, Treatment and Care, Nairobi, Kenya
fLiverpool School of Tropical Medicine, Liverpool, UK.
Received 4 April, 2008
Revised 2 July, 2008
Accepted 9 July, 2008
Correspondence to Elizabeth Marum, PhD, US Department of Health and Human Services/Centers for Disease Control and Prevention, National Center for HIV, STD, and TB Prevention, 1600 Clifton Road MS E-04, Atlanta, GA 30333, USA. E-mail: firstname.lastname@example.org
Client-initiated voluntary counseling and HIV-testing (VCT) targets asymptomatic adults who want to know their HIV status and is being scaled up in high-prevalence countries, including Kenya, with an estimated prevalence of 7% [1,2]. Mass media communications have been an effective public health tool to promote vasectomy, family planning, condoms, smoking cessation, and use of insecticide-treated bed nets [3–7]. We analyze the impact of a professionally designed mass media campaign with four phases on the uptake of VCT in Kenya between early 2001 and mid-2005.
In 2000, there were three sites in Kenya where quality-assured VCT services were provided on a regular basis, and few Kenyans perceived benefits to knowing HIV status [8–10]. The Government of Kenya made a commitment to the rapid scale-up of VCT  and procured test kits through a World Bank loan . Plans were developed by a national taskforce to establish VCT services in health facilities and to set up stand-alone VCT centers in cities and larger towns . As a result, 585 VCT sites had been registered by June of 2005 .
To increase public demand for VCT, a professionally designed mass media campaign was launched. A simple, recognizable logo was used on all advertising and print materials. Signboards with the logo were provided to all registered VCT sites meeting quality assurance standards. There were four phases to the campaign, which included radio, television, posters, flyers, and signage (billboards, street signs). Dates given are for the intensive, multimedia phases of the campaign.
Phase 1: May–September 2002
The above-mentioned phase was intended to build knowledge of and confidence in VCT services, create links between consumers and VCT centers, and launch the VCT logo. The target audience was 15–39-year-olds residing in urban and periurban communities. Commonly asked questions, many taken from formative research, were posed, with the suggestion to ‘discuss this question at a VCT center near you.’ The questions included specific reference to the possibility of testing HIV positive.
Phase 2: August 2002–January 2003
A second phase targeted urban youth (15–24 years), with the message to ‘get in control of their life’ by knowing their HIV status. An upbeat, ‘lifestyle’ approach was used. Popular entertainers stated they had been tested themselves but did not reveal their results. No direct mention of the possibility of testing HIV positive was made. The campaign used a slang term ‘chanuka’ meaning ‘get smart.’
Phase 3: July–Dec 2003
A third phase was targeted at urban/periurban young couples, with the objective of establishing a norm to know each other's HIV status during key life events. Taking advantage of the popular phrase ‘chanuka,’ this phase used the slogan ‘chanukeni pamoja’ meaning ‘get smart together.’ The campaign featured celebrity couples, each representing a different life event (dating, about to get married, having the first baby).
Phase 4: January–April 2005
The fourth phase overtly discussed HIV and AIDS. It targeted male family decision makers and established couples. The advertisements portrayed ordinary Kenyans, such as tea plantation workers, a mechanic, and market women marketed to appeal to adults as parents, spouses, and family role models. The characters made statements such as ‘I took an HIV test because I love my family’ and ‘My family knows I'm HIV positive.’
Registered VCT sites complete a national data form for each client. Information on demographics, reasons for seeking testing, sources of information on VCT, and test results are recorded. Client records from 131 sites are included. Data represent 22% of the registered sites in Kenya, including 55% of all registered stand-alone sites, and an estimated 38% of clients tested nationally . Forty-eight sites (37%) were in urban areas. They originate from sites in 29 of the 70 districts, spread geographically throughout eight (out of nine) provinces in Kenya.
Client records were analyzed using SAS v9.1 (SAS Institute Inc., Cary, North Carolina, USA), after being entered into an Epi Info 2002 database (Centers for Disease Control, Atlanta, Georgia, USA). First, descriptive statistics by month of report was used to calculate the uptake of VCT and client characteristics. Twenty variables on the national data form were considered. Second, log binomial regression analysis was used to evaluate trends in prevalence rates between 2001 and 2005 . Each variable presented in Table 1 was considered against the selected baseline for comparison. All confidence intervals (CIs) are 95% confidence intervals.
Finally, a multivariate Poisson regression model was used to determine how much of any observed increase in VCT utilization may be related to the mass media campaign or any observed decrease due to events such as test-kit stock outs . As the number of sites was increasing at the same time as the client flow, VCT clients per site per month were modeled and relevant variables in the national VCT data form (see Table 1) were considered in the development of the model. The model used the natural logarithm of the number of sites reporting per month for an offset and accounted for overdispersion. To account for the trend in the changing number of sites over time, the final model presented is a function of the natural logarithm of the number of sites reporting per month (as an offset) and as a quadratic function of the reporting month number. Indicator variables accounting for the timing of the campaigns (including 3 months past their completion) and of stock outs of test kits were included in this model. The coefficients of the indicator variables can be interpreted as the percentage change in client load per site per month (i.e., changes beyond the slope of uptake over time) versus the baseline. All coefficients presented have 95% CIs.
In total, 381 160 client records were analyzed. Table 1 shows VCT utilization and client characteristics over the period. These reveal a significant increase [prevalence ratio = 1.067, CI 1.064, 1.071, P < .0001] in VCT uptake among young people under age 25, particularly in urban areas [rural trend significantly less than the urban (P < .0001)]. There is a significant trend over time for fewer women to attend VCT sites (prevalence ratio = −1.3% per year, CI −1.06, −1.56, P < .0001, Table 1).
When asked how they learned of the service, most VCT clients provided multiple responses. Clients reported messages from healthcare workers as well as increased exposure to information through the mass media multiple channels. Radio and billboards were the most widely cited in both rural and urban areas.
Between 2001 and mid-2005, there was a linear increase in the number of new sites in the database with an average of 1.44 new sites per month (correlation coefficient = 0.98). Declines reflect sites that ‘graduated’ to independent data management, removing them from the database. Figure 1 shows an exponential increase in utilization of VCT. The figure also shows client flow per site per month in order to control for the increase in the number of sites over the time period. It shows pulses of increased VCT uptake following the first and fourth phases of the campaign (indicated by shaded bands) but a linear increase following the second and third phases.
The number of clients and the number of new sites were closely correlated. Poisson regression resulted in a well fitted model revealing an increase in utilization attributable to the first phase of the campaign of 28.5% (CIs 15.9%, 42.5%); and the fourth by 45.2% (CI 28.4%, 64.1%). Conversely, the second phase had a negative impact (−11.2%; CI −19.6%, −2%) and the third no impact (−7.1%; CI −15.9%, 2.6%). In March and April 2003, inadequate supplies of HIV test kits had a significant adverse impact on VCT utilization (−15.2%; CI −27.1%, −0.5%) but other disruptions in supply in 2004 did not.
Data presented in this study are from wide range of site types and geographical areas in Kenya and therefore likely to be representative of overall trends. The increases in VCT utilization in stand-alone sites in urban areas are also reflected in health centers in rural areas in our data. A subset of 30 health centers and stand-alone sites reported data throughout the period, and the results are consistent in this subset. Available data from nonparticipating sites indicate similar trends , though multivariate analysis was not possible on these data. We present a retrospective analysis of observational data collected during a period of multiple changes, including the expansion of antiretroviral therapy (ART) provision. Our data were not designed to look at the impact of ART on the uptake of VCT in asymptomatic individuals. Thus, even after multivariate analysis, it remains difficult to draw firm conclusions on the impact of the campaigns alone. Our data show that the campaign led to an increase in VCT utilization in participating sites over and above the increase in service provision. The first and final phases of the campaign, which made overt mention of possibly testing HIV positive, had significantly more impact than the understated, ‘lifestyles’ approach used in the second and third phases. The overlap in timings between phases one and two of the campaign makes it difficult to differentiate a negative effect of phase 2 from a lack of sustained response to phase 1, a challenge that has been observed in other settings .
A number of other factors emerged of interest to policy makers and programmers in other countries considering mass media promotion of VCT. First, the cost of the mass media campaign accounted for 10% of the total cost of the VCT program in Kenya . Second, the increase seen in 2005 of client flow per site per month (Fig. 1) shows that the sites had capacity to absorb more clients, and the system was not yet saturated. Mass media channels are more accessible in urban areas, which saw a greater impact of the campaign, raising the possibility that the campaigns exacerbated inequities of access to VCT. Thus, whereas expansion in well served urban areas may no longer be justified in the Kenyan setting, rural areas remain underserved. Finally, anecdotal evidence suggests that the professional nature of the advertising campaign had unanticipated benefits. It contributed to a sense among the public and health workers that VCT was a legitimate, well supported new service. Stigma around HIV testing appears to have declined considerably, and going for a test became normalized, evidenced not only by data presented here, but also by the crowds who gather openly where mobile VCT services are offered.
Ambitious targets have been set internationally for the expansion of HIV services. Meeting these goals will require an exponential increase in the number of persons tested. We have demonstrated that the mass media promotion of VCT, along with a readily identifiable logo, has contributed to the expansion of the number of VCT sites and the utilization of VCT services in Kenya. Our evidence suggests that a professionally designed mass media campaign could be a beneficial component in achieving international goals for HIV prevention and treatment elsewhere in Africa.
The authors would like to acknowledge persons seeking VCT services, the District Health Management Teams and community and faith-based organizations involved, and the data departments at the Liverpool VCT Center, Nairobi and the US HHS/Centers for Diseases Control and Prevention, Kenya. Data management was supervised by Patrick Kamau and Patrick M. Angala. Direct contributions to the writing of this study were also made by Mary Furnivall of IntraHealth International, Nairobi and Brian Faragher of the Liverpool School of Tropical Medicine. The design and conduct of the scale-up of VCT services in Kenya was supported by the Government of Kenya, the President's Emergency Plan for AIDS Relief (PEPFAR), the Global Fund for AIDS, Tuberculosis and Malaria, the World Bank, the UK Department for International Development, the Japanese International Cooperation Agency, the US Department of Health and Human Services/Centers for Disease Control and Prevention (HHS/CDC), Global AIDS Program, and the US Agency for International Development. The mass media campaign for VCT was conducted by Population Services International (PSI) with funding from HHS/CDC and United States Agency for International Development (USAID) (via Family Health International Cooperative Agreement HRN-A-97-00017-00). The HHS/CDC and Liverpool VCT and Care supported the collection, management, analysis, and interpretation of the data and preparation, review, and approval of the manuscript.
There was no conflict of interest.
Ethics: Data reported in this study were collected routinely for project management at all the participating VCT sites. No additional data for the purpose of this study were collected. These data are based on anonymous and unlinked records.
Author contributions: E.M. conceived and wrote the article. She supervised the CDC Kenya VCT program during the timeframe described. G.M. was involved in the design of the media interventions described. She read and critically appraised drafts of the article. A.H. did the statistical analysis and explained the methods used. He read and critically appraised the article. C.N. was involved in the conception of the mass media campaigns, their execution nationally and read and critically appraised the article. M.T. wrote the article, analyzed data, and was the director of Liverpool VCT, Treatment and Care, Nairobi, that contributed data from 40 sites nationally.
1. Kenya. Ministry of Planning and National Development. Demographic Health Survey 2003. Nairobi: Central Bureau of Statistics; 2004.
2. The Voluntary HIV-1 Counselling and Testing Efficacy Group. Efficacy of voluntary HIV-1 counselling and testing in individuals and couples in Kenya, Tanzania, and Trinidad: a randomised trial. Lancet 2000; 356:103–112.
3. Piotrow PT, Kincaid DL. How should vasectomy be promoted in Guatemala? Stud Fam Plann 1988; 19:248–249.
4. Piotrow PT, Rimon JG, Winnard K, Kincaid DL, Huntington D, Convisser J. Mass media family planning promotion in three Nigerian cities. Stud Fam Plann 1990; 21:265–274.
5. Ramirez AG, McAlister AL. Mass media campaign: a su salud. Prev Med 1989; 17:608–621.
6. Worden JK, Flynn BS, Solomon LJ. Using mass media to prevent cigarette smoking among adolescent girls. Health Educ Q 1996; 23:453–468.
7. Kincaid DL, Merrit AP, Nickerson L, de Castro Buffington S, de Castro MPP, de Castro BM. Impact of a mass media vasectomy promotion campaign in Brazil. Int Fam Plan Perspect 1996; 22:169–175.
8. Arthur GR, Ngatia G, Rachier C, Mutemi R, Odhiambo J, Gilks CF. The role for government health centers in provision of same-day voluntary HIV counseling and testing in Kenya. J Acquir Immune Defic Syndr 2005; 40:329–335.
9. Population Council. HIV/AIDS counselling, testing, care and support services in Nairobi, Kenya. Population Council Report. Nairobi: 1999.
10. National Council for Population and Development. Demographic health survey 1998. Maryland: Marco International, Inc.; 1999.
11. Kenya. Office of the President. Kenya national HIV/AIDS strategic plan 2000–2005. Nairobi: National AIDS Control Council; 2000.
12. World Bank. Project appraisal document on a proposed credit to the Republic of Kenya for the Decentralised Reproductive Health and HIV/AIDS Project. [document on the Internet]. Washington: World Bank; 2000 [updated 2000 November 16; cited 2005 June 25]. Available from: http://go.worldbank.org/PTW6MMQR20
13. Marum E, Taegtmeyer M, Chebet K. Scale-up of voluntary HIV counseling and testing in Kenya. JAMA 2006; 296:859–862.
14. Kenya. Ministry of Health. National Register of VCT Sites in Kenya. Nairobi: National AIDS and STD Control Programme; 2005.
15. Ministry of Health Kenya. AIDS in Kenya: trends, interventions and impact, 7th ed. Nairobi: National AIDS and STD Control Programme; 2005.
16. Spiegelman D, Hertzmark E. Easy SAS calculations for risk or prevalence ratios and differences. Am J Epidemiol 2005; 162:199–200.
17. McCullagh P, Nelder JA. Generalized linear models. New York: Chapman and Hall; 1989.
18. Otwombe KN, Wanyungu J, Nduku K, Taegtmeyer M. Improving national data collection systems from voluntary counselling and testing centres in Kenya. Bull World Health Organ 2007; 85:315–318.
19. Vidanapathirana J, Abramson MJ, Forbes A, Fairley C. Mass media interventions for promoting HIV testing. Cochrane Database Syst Rev 2005; CD004775.
© 2008 Lippincott Williams & Wilkins, Inc.