Study design and setting
A quasi-experimental (pre- and post-test) design, with switching replication, was used for this study. A self-administered structured questionnaire was used to determine the TBIC-related knowledge and practices of nurses in an intervention group and a comparison group. It was administered to both groups before the intervention (baseline, T0), after which the intervention group was exposed to the educational programme. The same cohorts of nurses in both groups were followed up and the questionnaire was again administered 6 months later (T1), as illustrated in Figure 1 below. After the second data collection wave, the comparison group also received the training. Then six months later, final data collection was conducted (T2).
The study was conducted in Ibadan, the capital city of Oyo State, South-West Nigeria. It is the third largest metropolitan area in Nigeria, and the largest by geographical area (3,080 km2) with an estimated 2011 population of about 3,034,206 (density of 985/km2) [29,30]. Oyo State has the third highest TB burden in Nigeria, with 6901 cases reported in 2017 .
Study population and sample
Nurses who work at two secondary health facilities in two local government areas (LGAs), Ring Road State Hospital (Ibadan South-West LGA) and Adeoyo Maternity Teaching Hospital (Ibadan North LGA) constituted the study population: 173 and 217 nurses respectively (total=390), from available administrative data at the study sites. The LGAs, which are non-contiguous, were purposively selected to avoid the effect of contamination. With an expected moderate effect size (ES>0.50<0.80) in the TBIC knowledge of the nurses, a significance level of 5% and power of 80%, the study required at least 32 participants in each group . However, because of the public health and health system considerations of the educational intervention, all available nurses at the study sites were encouraged to participate. One hundred (100) nurses were eventually enrolled at each site (total=200).
The multi-method educational intervention, which took place over a 3-hour period, consisted of didactic lectures in the form of Microsoft PowerPoint presentations prepared from WHO and CDC materials on TBIC; a 14-minute video presentation titled, Implementing TB Infection Control in Outpatient Settings, produced by CDC; as well as sessions for general discussion and practical demonstration [33-35]. A session on hand hygiene was included in the training as this has been recommended by WHO for implementation in the context of general infection control . To serve as reminders, handouts or notes of the lectures were provided to the nurses after the training session. Also, CDC-designed educational materials (signages, posters and stickers on TBIC workplace practices) were conspicuously displayed at the facility after the training. Adjustments in the training time was made to accommodate the nurses’ work schedule in order to train as many of them as possible: 5 training sessions were held for a group of 18-25 nurses each time, including sessions that were conducted in the evening for the nurses on night shift. Generally, the educational materials covered the following topics concerning TB: cause, transmission, symptoms and signs, infectiousness, risk factors and TB infection control measures.
The self-administered questionnaire used for this study had sections on socio-demographic characteristics, and TBIC-related knowledge and practices of the nurses. These scales were adapted from an instrument used by Kanjee et al.  to study TBIC in a high drug-resistance setting in South Africa. Subject matter experts (2 consultant chest physicians and 2 senior TB nurses) examined the study instrument for face validity to ensure that the instrument items were relevant, adequate and appropriate while the content validity was improved upon by the review undertaken by the Sefako Makgatho Health Sciences University Research Ethics Committee. Some of the initial instrument items were deleted and others rephrased based on expert opinions. As shown by Cronbach’s alpha of 0.6 and 0.8 for the knowledge and practice scales respectively, the scales were found to have acceptable internal consistency. Fifteen nurses, who were eventually excluded from the main study, participated in the pilot-tested of the questionnaire at one of the study sites. Their feedback on the clarity of the questions and challenges encountered in responding to them were utilized to rephrase some of the questions. The study instrument was drafted in English and because this is the official language used by the participants and the medium of instruction during their training, there was no need for translation. The knowledge final scale contained 33 items, with each having response options of “true”, “false”, or “I don’t know”. Each correct answer had a score of “1” and an incorrect answer, “0” while “I don’t know” was considered an incorrect answer. The knowledge and practice scales had a maximum possible scores of 33. The TBIC practice scale had 6 items which measured self-reported frequency of adherence to various TBIC practices. It was scored using a 5-point Likert-type scale: “never” (1 point), “rarely” (2), “sometimes” (3), “often” (4), and “always” (5), giving it a maximum possible score of 30.
Two research assistants and a supervisor were recruited and trained for the study. The nurses were informed about the study through their administrative structure. After explaining the purpose of the study to the participants, each of them that consented to take part in the study received a copy of the information leaflet, consent form and study questionnaire. After signing the consent, the self-administered questionnaire was issued out and this was collected after completion in May 2014 (wave 1). Six months after the training was conducted on the intervention group, the questionnaire was again administered on both groups in November 2014 (wave 2). After this, the comparison group received the intervention. At the end of another six months (i.e. in May 2015), the questionnaire was administered again on both groups (wave 3).
The data collected was analyzed using SPSS Statistics version 24. Descriptive statistics in the form of means and frequencies were used to show the socio-demographic characteristics. Independent t-test and chi-square test (χ²) respectively were utilized for comparison of the continuous and categorical variables between the intervention and comparison groups. The knowledge and practice scores of the respondents at different time points were presented as mean percentage scores, and independent t-test was also utilized to demonstrate significant differences between the scores of both groups. In addition, the scores were categorized into “good” and “poor” scores using cut-off points of 80% and 100% for knowledge and practices respectively. The cut-off for good practice score was set at 100% because optimal performance of TBIC measures is essential to minimize the nurses’ risk of contracting TB. The level of statistical significance was set at p<0.05. The effect size (Cohen’s d) was calculated using the 6th month measurement to investigate the magnitude of the change in the knowledge score of the intervention group resulting from the educational programme.
The study was approved by Sefako Makgatho Health Sciences University Research Ethics Committee (MREC/H/271/2013: PG) and Oyo State Ministry of Health Research Ethical Review Committee in Nigeria (AD 13/479/557). Permission was obtained from Oyo State Hospitals Management Board and the management of Adeoyo Maternity Teaching Hospital and Ring Road State Hospital, both in Ibadan, Oyo State, Nigeria. Participation in the study was completely voluntary and measures were taken to ensure privacy and confidentiality of the participants, and informed consent was obtained from each participant. In order to ensure anonymity of the participants, their name, address and other unique identifier were not included in the questionnaire. The participating facilities and individual participant questionnaires were allocated identification numbers, to which the participants could not be linked.