Page 143 - SAHCS HIVMed Journal Vol 20 No 1 2019

Page 3 of 7 Original Research

services, reduced risk of HIV infection, reduced risk of other questions any time without consequences. The interviews
sexually transmitted infections [STIs], safety of procedure were conducted in a private environment.
and penile hygiene) and attitudes or perceptions (access to
and costs of MC services, fear of pain, fear of HIV testing, Results
sexual dissatisfaction or satisfaction, promiscuousness, and
stigma). The questionnaire was self-developed based on a Of the total target sample of 176 male participants, 136
literature review of various published studies, such as the consented, with a response rate of 77.2%. The majority were
Zambian Demographic and Health Surveys, as well as aged between 21 and 24 years (50.0%). Of the participants
3
unpublished reports. The 10–15 min interviews were interviewed, 28% were from Northern Technical College,
conducted by trained research assistants who were fluent followed by the Copperbelt University (24.4%) and Zambia
in both English and Bemba, which are the primary local ICT College (12.6%). Most of the respondents reported that
languages. To minimise bias during data collection, we they were not in committed relationships (85.1%), although
ensured age–sex matching, that is, only male research half reported having sexual partners (50.8%). The majority
assistants administered the questionnaire after lecture hours reported identifying with traditionally non-circumcising
in private rooms identified at each of the five campuses. cultural groups (82.1%), such as Bemba, Lozi, Tonga and Ngoni
(Table 1). The overwhelming majority of study participants
identified as Christian (99.2%, n = 133), a religion that does not
Data management generally promote circumcision on religious grounds.
The paper-based questionnaires were physically checked for
completeness and consistency before data were entered into Of the 131 students who responded to a question about their
TM
Epi Info (Centers for Disease Prevention and Control, circumcision status, 82 (62.6%) reported being circumcised.
Atlanta, GA, US; Version 3.5.1). We used Epi Info to conduct The majority (81.5%) of the circumcised students reported
checks on data range, values and consistency. Data were then being single. The vast majority of medically circumcised
exported to SPSS version 22.0 for further statistical analyses. respondents believed that VMMC should be promoted in non-
For clearer interpretation of results, ordinal variables were circumcising communities (98.6%), and 74.5% of uncircumcised
recorded to binary form; responses ranging from ‘very respondents reported that they were considering circumcision
important’ to ‘important’ were recoded to ‘yes’, whereas in the future. Regardless of circumcision status, 97.6% (n = 126)
‘somewhat important’ and ‘not important’ were recoded to of respondents recommended MC as an additional strategy to
‘no’. The resulting data set included both categorical and prevent HIV and other STIs.
continuous variables.
Awareness of VMMC was almost universal, with 97.7%
(n = 133) of the respondents reporting having seen or heard
Statistical analyses messages about it (Table 1). Many respondents ranked
We generated means with their standard deviations to government hospitals and clinics among the top two places
describe continuous variables, and frequency and percentage where one could get circumcised (88.0% and 63.2%,
distributions to describe categorical variables. We used respectively), and the majority knew that the services at those
univariable logistic regressions to assess individual statistical locations were free (96.8%). Most respondents reported that
significance of association between factors and MC status, and they were able to easily access VMMC sites (75.2%, n = 131).
multivariable logistic regression to adjust for other factors. About one in five (22.3%, n = 121) did not know what time
The multivariable logistic regression model was determined during the day these services were provided (Table 1).
in two stages. Firstly, cross tables of the factors associated with
VMMC were done; at this point, chi-square test or Fisher’s With regard to knowledge and attitude towards VMMC,
exact-based p-values, as appropriate, equal to or less than circumcised participants ranked the following as reasons, in
0.10 were deemed indicative of a significant association. order of importance, why they decided to get circumcised:
Secondly, factors found to be significantly associated with (1) it reduces the chances of contracting HIV (90%, n = 80),
VMMC at the first stage were taken into the multivariable (2) it reduces the chances of contracting other STIs (88.8%,
logistic regression model. The Wald chi-square test-based n = 80), (3) it improves penile hygiene (93.2%, n = 74),
p-values equal to or less than 0.05 were deemed indicative of a (4) these services were provided for free (87.3%, n = 79), and
significant association in the multivariable model. (5) facilities were nearer to their residences (87.2%, n = 78)
(Table 2a and Table 2b). Other respondents reported that
Ethical consideration MC reduces the risk of developing penile cancer (90.0%),
while some reported that the services were safe (96.2%) and
This study was conducted after approval from ERES of good quality (90.4%). Nearly two-thirds (n = 80) of students
Converge Research Ethics Board (Reference number ranked enhanced sexual pleasure as an important reason for
00005948) was obtained. A written informed consent was accepting the procedure. Eight exposure variables (such as
obtained after explaining the purpose of the study to the safety of procedure, costs, the quality of MC services,
prospective participants. This was a low-risk single-contact perceptions on promiscuous behaviour, sexual enhancement,
study using a structured questionnaire and participants sexual sensitivity being the ‘same’ regardless of circumcision
were free to withdraw from the study or decline to answer status, reduced chance of HIV and other STIs) met the level of

http://www.sajhivmed.org.za 136 Open Access