Modelling a cluster randomised controlled trial to evaluate the impact of a cash incentivised prevention intervention to reduce teenage pregnancy among high school learners in rural KwaZulu-Natal, South Africa.
Date
2013
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
A third of adolescent girls fall pregnant before the age of 20 in South Africa, and
this happens despite contraceptives and condoms being free and mostly accessi-
ble.
This thesis aims to evaluate the impact of a cash conditional incentivised preven-
tion intervention on teenage pregnancy through the use of appropriate statistical
methods that take into account the clustering effect as well as the binary nature
of the response variable. This thesis will focus on the analysis of interim data of
the study which was collected at baseline and first follow up.
Fourteen schools were allocated to two study arms (intervention and control arm),
totalling to 1412 teenage girls who were followed up annually for 3 years. Partic-
ipants in the intervention arm received the conditional cash transfers while those
in the control arm did not. The intervention arm comprised of 704 girls while the
control arm had 708 girls at baseline. The null hypothesis for this study states
that there is no difference in the rate of pregnancy across the study arms. Urine
specimen were collected to test for pregnancy. Baseline data analysis revealed an
overall pregnancy proportion of 3.47% with that of 2.84% and 4.10% respectively
for the intervention and control arm. These findings together with all the other
findings from the applied statistical methods yielded insignificant results thus,
favoring the null hypothesis. Amongst the covariates used, age and grade were
multi-collinear. In all the fitted models, the age variable was statistically signif-
icant (p<0.01) which is indicative that this variable plays an important role in
pregnancy.
From this study, a total of 280 (approximately 21%) girls missed a follow up visit.
No statistical analysis was done to account for the missing data as the study was
analysed at an interim thus, there is a possibility that participants might miss
a particular visit but return for another scheduled visit. Based on the outcome
obtained from the interim data analysis, it is evident that teenage pregnancy oc-
curs regardless of treatment arm thus, we conclude that cash conditional transfers
does not exclusively prevent teenage girls from falling pregnant.
Description
M. Sc. University of KwaZulu-Natal, Pietermaritzburg 2014.
Keywords
Theses--Statistics., Teenage pregnancy--Statistics., Teenage pregnancy--Prevention.