Self-reported sexually transmitted infections among sexually active men in Ghana

We sourced data from the 6th version of the Ghana Demographic and Health Survey (GDHS). The GDHS is a nationally representative survey that is mostly conducted in 5 year intervals in over 85 low and middle income countries globally. In Ghana, the first round was conducted in 1988, the second one in 1993, third in 1998, fourth in 2003, fifth in 2008 and the most recent one is the 2014 version. The 2014 version covered all the then ten regions (now sixteen) in Ghana. The 2014 GDHS was implemented by the Ghana Statistical Service and the Ghana Health Service with Inner City Fund International giving technical support through MEASURE DHS [11, 12]. The survey gathers information on a variety of topics, including maternal and child health, to track key maternal and child health indicators in low- and middle-income countries. The survey also collects data from men aged 15–59 or 64 years in some countries. The DHS’s sampling technique is a multistage procedure. For example, in the 2014 GDHS the first stage is marked by the collection of clusters made up of enumeration areas drawn up for the 2010 Census of Population and Housing [[11] p317–319]. The clusters were selected from the previous 10 administrative regions in Ghana containing both urban and rural areas. With the second stage households are selected from each cluster. Details of the methodology has been documented elsewhere [11] p317–19]. In the 2014 DHS, 4388 men aged 15 to 59 were interviewed, with 3,498 admitting to having ever had sex. However, for this analysis, 3,051 sexually active men were considered, all of whom had full details on all of the variables of interest.

Stata 14.0 was used to conduct all of the analyses (Stata Corporation, College Station, TX, USA). To analyze our data, we used both descriptive and inferential statistics. We used frequencies and percentages to classify the sample using descriptive statistics. To pick possible variables for the multivariable logistic regression analysis, a Chi-square test was used. The multivariable binary logistic regression model was used to model  variables that had a chi-square test p-value of less than 0.05. Before fitting the final model, the variance inflation factor (VIF) was used to determine multi-collinearity among the independent variables (Mean VIF = 1.45, Minimum = 1.09, Maximum VIF = 2.03), and no evidence of collinearity was found. Adjusted odds ratios (aORs) with 95% confidence intervals (CIs) and p-values were used to present the regression results. Statistical significance was declared at p < 0.05 in all of the analyses. Using the sample weight (v005/1,000,000), the data were weighted to account for the complex sampling structure and non-response. The data was also adjusted using the ‘svy’ command to account for the complex survey nature.

In Table 1, we present the socio-demographic characteristics of the respondents and STI prevalence. Slightly more than 30 % of the respondents were aged 25–34. More than half (52.2%) were in the urban areas, 47.5% were in the middle ecological zone, 74.9% had secondary/tertiary level of education, 47.7% were in the rich wealth index, 57.7% were married and 72.3% were Christians. Almost 60% of the participants (58%) were married. Almost half of the study participants (49%) were of the Akan ethnic group. Almost all the participants (96%) reported being employed whereas 55.9% had their first sex during adolescence. Out of a total of 3051 study participants, 174 representing 6.0% (CI = 4.7–6.8), reported to have contracted STI within 12 months preceding the survey. The chi-square analysis also showed that ecological zone, education, marital status, ethnicity, age at first sex, number of sexual partners in the past 12 month and exposure to television had statistical significance with self-reported STIs.

Table 2 presents adjusted odds ratio at 95% confidence level on the factors associated with STI among sexually active men in Ghana. Compared to men aged 45–59, those aged 25–34 (aOR = 2.96, CI:1.64–5.35), 15–24 (AOR = 2.19, CI: 1.13–4.26) and 35–44 (AOR = 2.29, CI: 1.23–4.24) were more likely to report STI. Respondents who had their first sex when they were 20 years and above had lower odds of reporting STIs compared to those who had their first sex at a younger age (aOR = 0.66, CI: 0.47–0.93). Respondents who had 2 or more sexual partners were more likely to report an STI (aOR = 4.24, CI: 2.52–7.14) as compared to their counterpart who had one or no sexual partner. Men who read newspaper had lower odds (aOR = 0.53, CI: 0.37–0.77) of reporting STI compared to their counterparts who reported that they were not reading newspaper.

The study’s key strength is its use of a nationally representative dataset with a relatively large sample size. In terms of methodology, the DHS survey follows best practices, and this, combined with the use of experienced and well-trained data collectors, resulted in a high response rate. As a result, the findings of this study can be extended to all Ghanaian men who are sexually active. Despite these strengths, the cross-sectional study design prevents causal inference from the findings. Self-reports were used to measure the outcome variable, which were not checked or validated by any medical professional. Since many of the most common STIs are asymptomatic (most notably chlamydia and trichomoniasis in men), self-reporting is often restricted because asymptomatic infections are not recorded. It is also possible that some of the factors we controlled for (e.g., education, wealth index) may have affected symptom recognition or STI in a way that cannot be controlled for without biological testing. In addition, there is also the likelihood of recall and social desirability biases [28]. Furthermore, the type of STI was not specified [12].