Intervention time series analysis of voluntary, counselling and testing on HIV infections in West African sub-region: the case of Ghana

ABSTRACT

In this paper, intervention time series models were developed to examine the effectiveness of the voluntary counselling and testing (VCT) programme in the northern and southern sectors of Ghana. Pre-intervention data of HIV reported cases in the northern and southern sectors were first modelled as Box–Jenkins univariate time series. Second, the adopted models from the pre-intervention data were extended to include the intervention variable. The intervention variable was coded as zero for the pre-intervention period (1 January 1996–31 December 2002) and one for the post-intervention period (1 January 2003–31 December 2007). The models developed were applied to the entire data for the two sectors to estimate the effect of the VCT programme. Our findings indicate that the VCT programme was found to be associated with detection of 20 and 40 new HIV infections per 100,000 persons per month in the northern and southern sectors (p < .10), respectively. The VCT programme in Ghana, like most West African nations, has insignificant impact. Intervention time series models can be used to reliably examine the impact of the VCT programme. The impact of the VCT programme is minimal and we therefore recommend that the National AIDS Control Programme and other stakeholders re-double their efforts to maximise the impact of the programme.

KEYWORDS:

• Box–Jenkins model
• HIV
• impact
• intervention time series
• NACP
• unit root
• VCT programme

Acknowledgements

We are grateful to GSMF, NACP and Ghana AID Commission, especially, Kwadwo Asante (NACP) and Professor Amo-Sekyi (Chairman, GAC) for their assistance during the data collection in Ghana. We also want to express our sincere thanks to the regional coordinators of NACP and their staff for their immerse support and cooperation given to us during data compilation. PAS conception and design of the study; acquisition of the data; formulation of the analysis plan; interpretation of data; drafting of the manuscript and incorporating co-author feedback. JC and UT: reviewed and edited it critically for important intellectual content.

Disclosure statement

No potential conflict of interest was reported by the authors.

Ethics approval

Only de-identified data were analysed to preserve patient confidentiality. Approval was given by Edith Cowan University ethics committee and GAC and NACP to carry out the study.