Data source, study design and sampling procedures
The current study used data taken from the 2015-16 Malawi Demographic and Health Survey (MDHS). The 2015-16 MDHS sample was selected using a two-stage cluster sampling design and produced a nationally representative samples. The census sampling frame is considered as a complete list of all the census standard enumeration areas (SEAs). Thus, in the first stage, 850 SEAs (i.e. 173 SEAs in urban areas and 677 SEAs in rural areas) were selected with probability proportional to the SEA size. In the course of the second stage, a fixed number of 30 and 33 households per urban rural cluster/SEA, respectively, were selected with an equal probability systematic selection criterion. All women and men of reproductive age 15–49 years and 15–54 years respectively, who were either permanent residence of the selected households or visitors who stayed in the household the night prior to the data collection were eligible for the interviews. The MDHS selected a total of 27,516 households, of which 24,562 women and 7,478 men were successfully interviewed for the response rate of 97.7% and 94.6% respectively. Data were collected on socio-demographic characteristics and major health indicators, including knowledge, attitudes, and behaviors related to other health issues such as injections, smoking, fistula, tuberculosis HIV/ acquired immune deficiency syndrome (AIDS), and non-communicable diseases (NCDs). The datasets for women and men were explored and after excluding respondents with missing data, a total of 28,862 respondents (6,937 men and 21,925 women) were included in our analysis.
Variables
Dependent variable
The dependent variable considered in this study was correct and adequate knowledge regarding the mode of TB transmission. This variable was created from the following 6 questions to evaluate the correct knowledge regarding mode of TB transmission among adult male and female.
-
TB is spread from person to person through the air when coughing or sneezing?
-
TB can be transmitted by sharing utensils?
-
TB can be transmitted through food? Q4: TB can be transmitted by touching a person with TB?
-
TB can be transmitted through sexual contact?
-
TB can be transmitted through mosquito bites?
For the purpose of this study, the response to Q1 “Through air when coughing or sneezing” was used to measure the knowledge about the mode of TB transmission. The responses from ‘Q2’ to ‘Q6’ were regarded as misconceptions. However, individuals who responded ‘yes’ to the Q1 and responded ‘no’ to the other questions were recorded to have correct knowledge.
Independent variables
The present study considered the following covariates as independent variables; sex of the respondents, age of the respondents, educational level, wealth index, religion, occupation, marital status, amount of media exposure, perception about TB cure, perception about keeping secret when family member gets TB, place of residence, geographical religion, and ethnicity. These variables were selected after a thoroughly review of literature [11][12][34][35]. The covariates were categorized as follows: sex of the respondents (male/ female), age of the respondents in years (< 25/ 25–34/ 35–44/≥45), educational level (no formal education/ primary/ secondary or high), wealth index (poorest/ poorer/ middle/ richer/ richest), religion (Roman catholic/ Church of Central African Presbyteria/ Anglican/ Seventh Day Adventist/ Baptist/ other Christian/ Muslim/ No religion/other), occupation (not working/ professional or technical or managerial/ clerical or sales or services/ agricultural employee/ skilled manual/ unskilled manual), marital status (never in union/ currently in union / formerly in union), amount of media exposure (0/ 1/ 2/ 3), Tuberculosis can be cured (no/ yes), keep secret when family member gets TB (no/ yes), place of residence (urban/ rural), geographical religion (northern/ central /southern), and ethnicity (Chewa/ Tumbuka/ Lomwe/ Tonga/ Yao/ Sena/ Nkhonde/ Ngoni/ Mang'anja/ Nyanja/ Other). In this study, the amount of media exposure was specifically composed from three domains namely watching television, frequency of reading a newspaper, and or listening to radio at least once a week. The MDHS asked respondents the following questions; (1) Do you read a newspaper or magazine at least once a week, less than once a week or not at all? (2) Do you listen to the radio at least once a week, less than once a week or not at all? And (3) Do you watch television at least once a week, less than once a week or not at all? [36]. Each mass media related variables was recoded to binary level such that the correct answer was coded 1 while incorrect answer was coded 0. Amount of media exposure was organized by summing up the reported frequency of each media if an activity was performed at least once a week with 0 recorded as the least possible score and 3 recorded as the highest possible score. Wealth index is defined as a composite measure of a household’s cumulative living standard and was created using easy-to-collect data on a household’s ownership of selected assets. These assets included televisions, bicycles, materials used for constructing the house, access to safe drinking water, sanitation facilities, and other characteristics of a household. Through a statistical technique named principal component analysis (PCA), placed the household into quintiles [37].
Statistical analysis
We conducted our analyses while taking into account the complex design of the survey (i.e. weighting, clustering, and stratification). First, the baseline statistics were presented as frequency and weighted percentage. Second, the bivariate analyses using Chi-Square test were performed to explore the distribution of the selected characteristics according to the correct knowledge about the mode of TB transmission among adult women and men. Third, using the generalized estimating equation (GEE) logistic regression, the multivariable analyses were performed to investigate the strength of associations between the selected factors and correct knowledge about TB transmission. GEE models were used to account for the correlated responses within the DHS data. Both unadjusted and adjusted odds ratios (aORs) and 95% confidence intervals (CIs) with their p-values were presented. p-values < 0.05 were considered as statistical significant. All analyses were performed using SAS software for Windows, version 9.4 (SAS Institute, Cary, NC, USA).
Ethical Considerations
The 2015–2016 MDHS was implemented by the National Statistical Office (NSO) and the Community Health Sciences Unit (CHSU). The protocols and procedures that were developed for data collection were reviewed and approved by the ICF Macro Institutional Review Board (ICF Macro IRB) and the Malawi National Health Science Research Committee. The MDHS 2015–2016 complied with all requirements of the US Department of Health and Human Services’ the 45 Code of Federal Regulations 46 (45 CFR 46), Protection of Human Subjects [38]. Before this study was conducted, the authors sought permission from the MEASURE DHS for use of the data beyond the primary purpose by which data were collected. At the beginning of each interview, informed consent (both written and oral consent) was obtained from all eligible participants. Furthermore, a parent or guardian provided consent prior to the participation by a child or anybody below the age of 18 years [38].