Knowledge and attitudes regarding HIV/AIDS and its prevention among Ghanaian women

The survey and sampling design

To perform an analytic cross-sectional study, this study used data (Women in reproductive age 15-49 years dataset) from the Ghana Multiple Indicator Cluster Survey (MICS) 2017/18. From October 2017 to January 2018, the Ghana Statistical Service gathered data with the assistance of collaborators including the Ministries of Health, Education, Sanitation and Water Resources, Gender, Children and Social Protection, Ghana Health Service, and Ghana Education Service. Under the Statistics for Results Facility – Catalytic Fund, UNICEF provided technical help, while UNICEF, KOICA, UNDP, USAID, and the World Bank provided financial assistance (SRF-CF).

The sample frame was the 2010 Ghanaian Population and Housing Census (PHC). All women between the ages of 15 and 49 years who were either inhabitants or guests who slept in selected residences the night before the survey (14,374) were included.

Ethical consideration

The studies were conducted using MICS data that was made publicly available. Institutions that ordered, funded, or oversaw the surveys were responsible for ethical procedures. Each participant's verbal consent was gained, and minors aged 15-17 years were individually interviewed after adult approval was secured in advance from their parents or caretakers. All participants were made aware of the voluntary nature of their involvement, as well as the confidentiality and anonymity of their data. Respondents were also advised of their right to refuse to answer any or all of the questions, as well as their freedom to end the interview at any moment.

Data analysis

The data was analyzed using SPSS version 20. (IBM Corp., 2011, and NY). Tables with frequencies and percentages were used to present the results of categorical variables including sample characteristics. The link between the dependent and independent variables was determined using the chi-square test. A binary logistic regression model was used to identify the predictive determinants of HIV knowledge and attitude levels. And results were given as exponentiated B coefficients or adjusted odds ratios (AOR) with matching confident intervals (CI). A p-value of 0.05 was used to determine the research's statistical significance.

Sample description

Table 1 describes the women included in this study, most (40.6%) of them were within the youth age group (15-24 years), and the majority (78.5%) had history of attending school. In terms of the type of residence more than half of them (51.2%) were from rural areas, most (13.8%) from the Ashanti region, and the majority (38.2%) were of the Akan tribe. About 43.5% of them were married and the majority (91.1%) were without functional difficulties. In terms of media exposure most (88.0%) were not at all into new paper reading, and only 28.6% of them were listening to the radio every day, majority of them never used computers or tablets and never used the internet (Table 2).

Knowledge on HIV prevention

Table 3 reports the HIV prevention knowledge variables that respondents were assessed on. Their responses to these questions were recorded as 1 for correct answer and 0 for incorrect answer and composite scored. Those with composite scores of 5 and below were categorized as having a poor knowledge level regarding HIV prevention and those composite scores of 6 and above were classified as having a good knowledge level regarding HIV prevention. At the end of the categorization the majority (81.9%) had a good knowledge level regarding HIV prevention and the remaining (18.1%) had a poor knowledge level. Further analysis revealed that all under-studied independents variables of the respondents were significantly associated with the dependent variable (HIV knowledge level) (P < 0.001) as presented in Table 5.

Attitude towards HIV

Table 4 reports the HIV attitude variables that women were assessed on. Their responses to these questions were recorded as 1 for correct answer and 0 for incorrect answer and composite scored. Those with composite scores of 3 and below were categorized as having poor attitude levels towards HIV and those composite scores of 4 and above were classified as having good attitude levels towards HIV. At the end of the categorization the majority (89.5%) had a poor attitude level towards HIV and the remaining (10.5%) had a good attitude level towards HIV. Further analysis revealed that all under-studied independents variables of the respondents were significantly associated with the dependent variable (attitude level towards HIV) (P < 0.05) as presented in Table 5.

Predictors of good HIV prevention knowledge

All understudied independent variables (demographic characteristics and media exposure) of the respondents were significantly associated with the dependent variable (HIV knowledge level) and were therefore included in the binary logistics model. The model revealed the following factors as predictors of good knowledge on HIV prevention: age, history of education, region of orientation, ethnicity, health insurance status, wealth index, radio use, computer or tablet use, and internet use (Table 6).

Relative to the respondents’ age group of 15-24 years, the likelihood of a good HIV knowledge level was 17% more for women of the age group of 25–34 years (AOR = 1.17, 95%, C.I. = 1.02–1.36). Those with no educational history were 54% less likely to have good HIV prevention knowledge when compared to those with educational attendance history (AOR = 0.46, 95%, C.I. = 0.40–0.52).

In terms of regional prediction, a good HIV knowledge level occurred 76% more for women from the Upper East region relative to women from the Western region (AOR = 1.76, 95%, C.I. = 1.31–2.35). Meanwhile, women from the Ashanti region, Brong Ahafo Region, and Northern region were less likely to have good HIV prevention knowledge when compared to those from the Western region. Women from Ashanti region were 27% less likely (AOR = 0.73, 95%, C.I. = 0.58–0.93), women from Brong Ahafo region were 35% less likely (AOR = 0.65, 95%, C.I. = 0.51–0.82), and women from Northern region were 28% less likely (AOR = 0.72, 95%, C.I. = 0.56–0.93).

Also, respondents’ ethnicity predicted their good HIV knowledge level. Relative to women from the Akan tribe, women from the Mole Dagbani tribe were 23% less likely to have a good HIV knowledge level (AOR = 0.77, 95%, C.I. = 0.64–0.94), women from the Grusi tribe were 32% less likely to have good HIV knowledge (AOR = 0.68, 95%, C.I. = 0.52–0.89), women from the Mande tribe were 48% less likely to have good HIV knowledge (AOR = 0.52, 95%, C.I. = 0.28–0.97) and lastly women from other tribes not specified in this current study were 27% less likely to have a good HIV knowledge level (AOR = 0.73, 95%, C.I. = 0.59–0.89).

Those without health insurance were 10% less likely to have good HIV knowledge as compared to those with health insurance (AOR = 0.90, 95%, C.I. = 0.81–0.99). Also, those with no disability were 22% more likely to have good HIV knowledge as compared to those with a disability (AOR = 1.22, 95%, C.I. = 1.04–1.44).

The increased wealth index status of the women positively predicted HIV good knowledge, women of the second wealth index were 29% more likely to have good HIV knowledge when compared to those of the poorest wealth index (AOR =1.29, 95%, C.I. = 1.11–1.48). Also, women of the middle wealth index were 53% more likely to have good HIV knowledge when compared to those of the poorest wealth index (AOR = 1.53, 95%, C.I. = 1.30–1.80). Again, women of the fourth wealth index were 96% more likely to have good HIV knowledge when compared to those of the poorest wealth index (AOR = 1.96, 95%, C.I. = 1.62–2.37). Finally, women of the richest wealth index were 130% more likely to have good HIV knowledge when compared to those of the poorest wealth index (AOR =2.30, 95%, C.I. = 1.83–2.89).

Finally, women's exposure to media predicted their HIV knowledge. Women who spent less than a week listening to the radio were 25% more likely to have good HIV knowledge when compared to those who did not listen to the radio at all (AOR = 1.25, 95, C.I. = 1.08–1.45). Also, women with at least once a week listening to the radio were 62% more likely to have good HIV knowledge when compared to those who did not listen to the radio at all (AOR = 1.62, 95, C.I. = 1.40–1.89). Women who spent almost every day listening to the radio were 46% more likely to have good HIV knowledge as compared to those who did not listen to the radio at all (AOR = 1.46, 95, C.I. = 1.28–1.67). With computer and internet use, those who did not ever use of computer or tablet were 34% less likely to have good HIV knowledge when compared to those who had used computers or tablets (AOR = 0.64, 95%, C.I. = 0.49–0.83). Those who had never used the internet were 45% less likely to have good HIV knowledge as compared to those who had ever used computers or tablets (AOR = 0.55, 95%, C.I. = 0.43–0.71).

Predictors of a good attitude towards HIV

All of the respondents’ understudied independent variables (demographic characteristics and media exposure) were significantly associated with the dependent variable (attitude level towards HIV) and were therefore included in the binary logistics model. The model revealed the following factors as predictors of a good attitude level towards HIV: age, ever attended school, residence type, region of orientation, health insurance status, wealth index, media exposure, and marital status.

Women aged 35 years and above were 30% more likely to have a good attitude toward HIV as compared to women of the age group 15-24 years (AOR = 1.30, C.I. = 1.08–1.57). A good attitude toward HIV was 20% less likely among those without a history of education attendance as compare to those with a history of attending education (AOR = 0.80, 95, C.I. = 0.66–0.98). Rural women were 36% less likely to have a good attitude towards HIV when compared to those from urban residences (AOR = 74, 95%, C.I. = 0.63–0.88). Those with the ‘single’ marital status were 31% more likely to have a good attitude toward HIV as compared to those married (AOR = 1.31, 95%, C.I. = 1.11–1.55). With regard to regional prediction, all the studied regions positively predicted good attitudes toward HIV relative to the Western region.

Regarding health, women with no health insurance were 24% less likely to have a good attitude towards HIV as compared to those with health insurance (AOR = 0.76, 95%, C.I. = 0.66–0.87). Considering the wealth index, women of the middle wealth index were 34% more likely to have a good attitude towards HIV as compared to those of the poorest wealth index (AOR = 1.34, 95%, C.I. = 1.03–1.72). Women of the fourth wealth index were 44% more likely to have a good attitude towards HIV as compared to those of the poorest wealth index (AOR = 1.44, 95%, C.I. = 1.11–1.87). Lastly, women of the richest wealth index were 110% more likely to have a good attitude towards HIV as compared to those of the poorest wealth index (AOR = 2.10, 95%, C.I. = 1.60–2.77).

Finally, media exposure predicted a good attitude towards HIV, those who read newspapers or magazines almost every day were 138% more likely to have a good attitude toward HIV as compared to those who don’t read at all (AOR = 2.38, 95%, C.I. = 1.41–4.03). Those who did not ever use computers or tablets were 33% less likely to have a good attitude toward HIV as compare to those with ever use (AOR = 0.77, 95%, C.I. = 0.63–0.95).