Time series analysis of HIV incidence cases in Ghana : trends, predictions and impact of interventions

Date of Award


Degree Type


Degree Name

Doctor of Philosophy


School of Engineering,


Faculty of Computing, Health and Science

First Advisor

James Cross


The HIV/AIDS epidemic is one of the world’s leading causes of death, particularly in sub-Saharan African nations like Ghana, and threatens socio-economic development in many developing countries. In this thesis Ghanaian HIV data, comprising monthly number of serologically confirmed reported new HIV cases since 1996, was subdivided into northern and southern sectors based on the geographical location of the ten administrative regions. Potential bias in the collection is considered given the strategic location of the two specialist teaching hospital, one in each sector, which receive referrals from the regions.

Time series modelling was applied to the monthly number of new HIV cases in each sector. Moving average of time series analysis of equal weight was applied to determine the trend for cases of incidence of HIV infection in the northern and southern sectors while Box-Jenkins modelling identification and Holt’s (double) exponential smoothing modelling methods were employed to predict of new incidence of HIV cases for both sectors in respect to sex and age groups. The effectiveness of three existing major interventions was examined using intervention modelling whereas cointegration modelling was used to determine the long-run impact of condom utilisation on the incidence cases of HIV infection. The trend analysis and predictions for the next three years reveal a slow increase in the number of new incidence of HIV cases. Although, various interventions have influenced the number of cases of HIV infection, the magnitude of impact fluctuated and declined with time. The analysis of the long-run impact of condom utilisation, on the reduction of new HIV incidence cases, indicates that new cases of infection will actually increase monthly by factor of 0.4 -0.6 for every 1000 condoms issued. These perplexing results may be because issuing of condoms does not ensure usage.

LCSH Subject Headings

HIV-positive persons - Ghana.

HIV (Viruses) – Ghana - Statistics.

AIDS (Disease) – Epidemiology - Mathematical models.

Access Note

Access to this thesis - the full text is restricted to current ECU staff and students by author's request. Email request to library@ecu.edu.au

Access to this thesis is restricted. Please see the Access Note below for access details.