A descriptive, cross-sectional survey design was used to assess nurses’ knowledge of and attitudes towards PrEP for the prevention of HIV infection in Tshwane, South Africa.

The study took place in the primary health care (PHC) facilities located in Tshwane sub-district 1, which include Soshanguve, Mabopane, Ga-Rankuwa, Winterveldt and Akasia. From all the health care facilities in Tshwane North district 1, there are approximately 15 PHC clinics, 3 community healthcare centres (CHCs) and 1 midwife obstetric unit (MOU). Primary health care clinics employ approximately 10 nurses each, CHC employs approximately 50 nurses, and the MOU employs approximately 15 nurses. The services provided at these healthcare facilities include family planning, immunisations, integrated management of childhood illnesses (IMCI), emergency care, HIV counselling and testing, antenatal and/or postnatal care, prevention of the mother-to-child transmission of HIV (PMTCT), chronic care, screening and the treatment of sexually transmitted infections (STIs) and PrEP.¹⁷ The researcher intended to include all healthcare facilities in the sub-district but instead collected data from only 10 facilities because some facility managers did not grant permission for data collection.

Using convenience sampling, the researchers asked all the nurses who were at the selected facilities at the time of data collection to take part in the study. This method of sampling was a feasible method because the healthcare workers worked in shifts. Eligibility criteria included healthcare workers (doctors and professional nurses) working in a local healthcare setting (clinic) in sub-district 1, Tshwane, regardless of the years of experience, and ability and willingness to provide informed consent for study participation; and excluded healthcare workers not on duty during the date and time of data collection. All healthcare workers from the selected facilities who met the eligibility criteria were invited to participate in the study. Copies of the informational pamphlet were distributed to a total of 127 potential respondents who were available at the date of data collection for their perusal. A total of 114 (n = 114) questionnaires were completed in full and considered for data analysis because 13 received questionnaires had to be discarded for various reasons, mainly to being inadequately completed. A total of 114 questionnaires were adequately completed and prepared for analysis. Although the study included both doctors and nurses, only nurses participated during data collection.

An anonymous self-administered paper-based questionnaire was used to collect the data. The questionnaire was adapted from the tool used by the Fenway Institute.¹⁰ The questionnaire was available in English and contained close-ended questions. Data collection occurred over a period of 3 months from September 2019 to November 2019. Prior to conducting the study, the questionnaire was pretested on seven professional nurses who were conveniently recruited from a local nongovernmental organisation (NGO) to identify any unclear or ambiguous items in the questionnaire and to verify its accuracy. Data collected from the pretesting of the data collection tool were not included in the data analysis of the main study. No problems were identified in the tool and no changes were made.

The questionnaire was divided into three sections and consisted of 32 questions. It included true-or-false questions and Likert scale questions. Section A focused on demographic information which assessed the age, gender, occupation, marital status, highest level of education, race and religion. Section B assessed the PrEP knowledge of nurses, and section C assessed the attitudes towards PrEP held by nurses. Seven questions were asked to determine the level of knowledge. The total number of correct responses to the seven knowledge questions provided an overall assessment of the respondent’s knowledge of PrEP on a seven-point scale. The results were classified as ‘good’ if they scored 7 out of 7, ‘moderate’ if they scored between 4 and 6, and ‘poor’ if the scores were less than 4 on the scale. Twelve statements were used to elicit respondents’ attitudes. When evaluating the general attitude of respondents towards PrEP, the following criteria were applied: respondents were asked to rate their level of agreement with a series of PrEP-related statements as agree, unsure or disagree; a point value ranging from −1 to +1 was assigned to each response option. The possible score range was −12 to +12. A score of 6 or more showed a positive attitude, while a score of less than 6 showed a negative attitude.

To minimise interruptions with the duties of nurses, the researcher arranged with the facility managers to address the potential respondents during their weekly morning meetings. The research assistant administered the surveys to the respondents after the meetings. The respondents required approximately 15 min – 20 min to complete the questionnaire. The respondents placed their confidential questionnaires in a box before exiting the interview venue. A total of 127 questionnaires were distributed and completed; and only 114 were accepted for analysis because 13 of the received questionnaires had to be discarded for a variety of reasons, including being incomplete.

The data were collected using Epi-Info, which was then exported to STATA for cleaning and analysis. All analyses were performed using STATA version 13 (Stata Corp., College Station, TX, USA). Nonresponses were categorised as missing data, and missing values were omitted from the analysis. The study’s outcome variables were nurses’ knowledge and attitudes regarding PrEP.

For descriptive statistics, continuous variables were summarised using the median with interquartile range. For categorical variables, the frequency and proportion or percentage were used. The Cronbach alpha analysis was applied to verify the reliability and internal consistency of the data collection instrument.¹¹ In this study, the knowledge and attitude was treated as a continuous scale, then calculated to full score, and further given cutoff points. The lower scores indicate poor knowledge and negative attitudes, and the Cronbach’s alpha was 0.741 for knowledge and 0.84 for attitudes. Logistic regression was used to compute odds ratios (ORs) and 95% confidence intervals (CI) for the relationship between the level of knowledge of and the attitude towards PrEP and the respondent characteristics. Finally, the Pearson’s chi-square was applied to assess the relationship between knowledge and attitude. All p-values less than 0.05 were deemed statistically significant. To assess factors associated with knowledge and attitude, which are the two outcomes of the study, we first collapsed knowledge scores (good, moderate, poor) into two categories, which are good and poor, where moderate score was combined to one. The rationale was to have a binary variable for the knowledge score. It was further categorised attitudes into positive attitudes and negative attitudes.

The researcher provided each potential participant with an information sheet stating the purpose of the study and the procedures involved. The researcher provided clarity on any questions raised by the respondents. A written informed consent form was signed and dated by the researcher and the respondent. Participation was voluntary and respondents were allowed to withdraw at any point during the study. There were no negative consequences for nonparticipation or withdrawal from the study. The researcher did not capture any identifying information, maintaining confidentiality throughout the study. The study was conducted in accordance with the Declaration of Helsinki.

Both the Faculty of Health Care Sciences Research Committee and the University Research Ethics Committee gave ethical approval for this study (Ref: /H/182/2019: PG). The Gauteng Provincial Department of Health and the Tshwane District Department of Health gave permission to conduct the study.

The survey tools were distributed to 127 participants but only 114 completed the questionnaire, resulting in a response rate of 89.8%. The sample included nurses from 10 health facilities providing HIV-PrEP services. The median age of the respondents was 47 years (IQR 38–54). There were more females (n = 105; 92.1%) than males (n = 8; 7.0%). Only a small number of respondents, that is, 9.7% (n = 11) were widowed or divorced, and 3.5% (n = 4) lived with their partners. Most respondents (n = 73; 64%) held a diploma, while 36% (n = 41) held a bachelor’s degree (Table 1).

Most respondents’ HIV-PrEP knowledge was moderate (68%) and poor (29%), and 3% had good knowledge (Figure 1). However, about half of the respondents (49.1%) assessed their self-rated PrEP knowledge as good and 35.2% as fair (Table 1).

As indicated in Table 1, most respondents reported being aware of PrEP (94.7%). The responses to specific knowledge questions are presented in Table 2. Most respondents knew that PrEP is intended for use as an HIV prevention tool (97.2%, n = 105), that it is taken orally once daily (90.8%) and that it has been approved for use in the country (88%). More than half of the respondents (53.7%) correctly identified that PrEP is not intended for both treatment and prevention. Few respondents (34.3%) provided a correct response that PrEP is not intended for use by both HIV-negative and HIV-positive individuals. Most respondents (77.8%) correctly identified that PrEP is not currently available to all potential users in PHC facilities, and that individuals on PrEP do not need to be followed up every month for medication side effects and lab toxicities.

Logistic regression indicated that males were 1.53 times more likely than their female counterparts to identify PrEP as an HIV preventive measure (OR = 1.53; 95% CI 0.61–3.85; p = 0.363). The odds of recognising PrEP as an HIV prevention measure was the same for the nurses who had a bachelor’s degree and the diploma qualifications (OR = 1.04; 95% CI 0.55–1.96; p = 0.902). However, the association between both sex and education and knowledge was not statistically significant. Age was not associated with knowledge of PrEP (Table 3).

Overall, HIV-PrEP was viewed negatively by 84.5% of respondents (Figure 2).

Respondents’ attitudes about HIV-PrEP are demonstrated in Table 4. Although the overall attitude was less favourable towards PrEP, the respondents demonstrated positive attitudes regarding providing PrEP education as an essential part of HIV prevention (90.7%), willingness to recommend PrEP to potential users (91.7%), rolling out PrEP in PHC (67.6%) and that using PrEP could empower women who are unable to negotiate condom use (64.8%). Respondents also showed positive attitudes by agreeing that PrEP adherence (34.3% of the respondents) and the dosing regimen (23.3% of the respondents) could be barriers to PrEP use and by disagreeing (66.7% of the respondents) with the statement that it is unethical to give PrEP to healthy people. Negative attitudes were evident in the responses where approximately half of the respondents (53.7%; n = 58) responded that the use of PrEP will lead to increased sexual risk-taking behaviour, and 45.5% (n = 49) responded that PrEP promotes HIV resistance.

Healthcare providers reported negative attitudes regarding PrEP use, holding the view that PrEP could lead to an increase in sexual risk-taking behaviour (53.7%) and HIV resistance (45.4%), and that potential ARV resistance (64.8%) and medication side effects (56.5%) would be potential barriers to PrEP use. In addition, most respondents believed that behavioural interventions were more effective than PrEP (66.7%), which suggests a negative attitude.

The odds of a positive attitude towards PrEP as an HIV preventive measure were 1.92 times higher among males than females (OR = 1.92; 95% CI 0.54–6.83; p = 0.311) and 1.24 times higher among respondents with a bachelor’s degree than respondents with a diploma (OR = 1.24; 95% CI 0.51–3.01; p = 0.636). However, the association between both sex and education and knowledge was not statistically significant. Age was not associated with attitudes towards PrEP (Table 5).