Effects of the Gama Cuulu Radio Serial Drama on HIV-Related Behavior Change in Zambia

Abstract

The Gama Cuulu radio serial drama is written and produced in Zambia's Southern Province. It promotes behavior change and service use to prevent HIV transmission. The authors evaluated the effects of Gama Cuulu on intermediate outcomes (e.g., perceived norms), as well as number of sexual partners, condom use, and HIV testing in the past year among adults between 18 and 49 years of age. The authors used a pretest/posttest assessment with a comparison group design, with Southern Province as the intervention area and Western Province as the comparison area. Approximately 1,500 in-person interviews were conducted in both provinces in 2006 (pretest), 2007, and 2008. Regression models included terms for province, time, and the interaction of the two. Outcomes improved in both provinces (e.g., by 2008, 37.6% of participants in Southern Province and 28.3% participants in Western Province tested for HIV in the past year). Pretest-to-posttest changes in condom use (from 20.2% to 29.4% in Southern Province) and 5 intermediate outcomes were significantly different in the 2 provinces. However, changes in condom use were not associated with listening to Gama Cuulu and changes in other outcomes were similar in both provinces. Weak intervention effects might be attributable to implementation challenges or the saturation of HIV programs in Zambia.

Acknowledgments

This study was funded by the U.S. Centers for Disease Control and Prevention through the President's Emergency Plan for AIDS Relief. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.

The authors gratefully acknowledge the program staff, without whom the Gama Cuulu serial drama would not have existed. The authors also acknowledge support from program champions and consultants throughout the life of the project, including Marc Bulterys, Christine Galavotti, Carmen Villar, Janet Moore, Pamela Bachanas, Nicola Harford, Joseph Petraglia, Katina Pappas-DeLuca, Carl Kendal, Thomas Prag, Neil Shaw, and Gordon Adam.

Notes

¹We use italics to distinguish the MARCH strategy or program from the nongovernmental organization named MARCH.

^a Actual numbers of cases and do not sum to 4,538 because of missing data. Percentages and chi-square test statistics are based on weighted data.

^b p < .05.

²In article text, variable names are in italics for clarity.

Note. For “number characters know,” scores ranged from 1 to 5. For “identification with,” scores ranged from 1 (characters not at all like me and never understand characters) to 5 (characters totally like me and always understand characters).

^a Actual numbers of cases and do not sum to 4,538 because of missing data. Percentages and chi-square test statistics are based on weighted data.

^b Chi-square analyses were not performed because categories differed in 2007 and 2008.

^c p < .05.

^a Only intermediate variables with significant differences between the two provinces are presented. Appendix includes a list of intermediate constructs for which there were no statistically significant differences.

^b Number of cases for analysis. Numbers are actual, and percentages and test statistics based on weighted data.

^c Number who believed many/all in community believed girls who are abused bring shame on the family, etc.

^a For brevity and to draw attention to the effects of Gama Cuulu, we do not present adjusted odds ratios or beta coefficients for the control variables retained in the final models. A list of the variables in each model follows:

	“Testing, what people do” included age and time living in area
	“Girls bring shame” included member of Tonga tribe, educational attainment, and time living in area
	“Women report, care” included exposed to other interventions
	“Community discusses HIV” included exposed to other interventions
	“Community has solutions” included employment status and educational attainment
	“Two or more partners, last year” included sex, marital status, educational attainment
	“Condom use, last sex” included sex, age, partner type (boyfriend/girlfriend/other vs. spouse/live-in partner), educational attainment, read paper at least weekly
	“HIV testing last year” included sex and member of Tonga tribe

^b p < .05.

This article is not subject to US copyright law.