Duration and Area of the Study
The study was conducted in Tigray public hospitals. Tigray is found in the northern part of Ethiopia. Tigray regional state has two comprehensive specialized hospitals, 15 general hospitals, 22 primary hospitals, and 223 health centers. From the 15 general hospitals seven hospitals namely Kahsay Abera, Suhul, Adwa, Adigrat, Mekelle, Lemlem Karl and Alamata are giving MDR TB treatment services. There were 118 registered MDR TB patients in the above seven hospitals of Tigray regional state. The study was conducted from April to June, 2019 in those seven multi drug-resistant tuberculosis treatment center hospitals of Tigray regional state.
Study Design
Hospital-based unmatched case-control study was conducted.
Study Population
Sample Size Determination and Sampling Technique
Sample Size Determination
The required sample size was determined by using the Epi-Info version 7.2.2.12 from the previous study conducted in Ethiopia. The estimated sample size was determined based on the following assumptions: confidence interval of 95% at the power of 80%, with the ratio of 1:2 (case to control). AOR=2.4, 95% CI=95%, Power= 80%, percent controls exposed=16.9, and percent of cases exposed for cases was taken 32.8 from a study conducted in eastern Shoa, Ethiopia [21].
Finally, the sample size was 242. By adding a 5% non-response rate the total sample size was 254 (85 and 169 controls).
Sampling Techniques and Procedures
First, the sample size was allocated proportionally to the seven MDR TB treatment centers of Tigray according to the number of patients registered to take the treatment. Then a frame of MDR-TB patients enrolled in second line drug treatment at MDR treatment centers of Tigray was created using the medical registration number. Finally, a simple random sampling technique was used to select the MDR TB cases. MDR-TB patients enrolled to second line-drug at MDR treatment center hospitals of Tigray taken as cases and simple random sampling technique was used to include 85 study participants from the list of MDR TB patient’s registration book. 169 Controls were also selected by a simple random sampling technique from the TB registration book in similar hospitals.
Data Collection Tools and Procedures
The primary data were collected by face-to-face interviews using pretested structured questionnaires, while secondary data were collected by reviewing participant’s medical charts using checklists for the matching study participants. Seven trained BSc nurses were assigned to conduct an interview with the participants and review the corresponding participant records. Three senior BSc nurse supervisors were assigned to supervise the whole data collection process. The questioner contained socio-demographic related factors, behavioral related factors and clinical related factors. Throughout the data collection process, data collectors were used the N-95 respiratory mask to prevent from acquiring the disease. The data collection period was from April to June 2019.
Data Quality Assurance
Data quality was ensured by giving two days of training for data collectors and supervisors and by providing supervision during the data collection period. First, the questioner was adapted from a published paper in English form [21], then translated into Tigrigna (local language) and back-translated into English to ensure its consistency. Each questionnaire was checked for completeness before leaving the study area. The questionnaire was pretested on 5 % of other patients who did not participated in the study for completeness and appropriateness to the local context. Based on the findings the necessary amendments have done. Every questionnaire was checked by the principal investigator on the spot.
Data Processing and Analysis
The data was coded and entered into Epi data manager version 4.4.3.1 then exported to SPSS version 20 for further statistical analysis.
Descriptive statistics such as frequencies, percentages, median, inter-quartile range were computed. Finally, the report was summarized and presented using, texts, tables, and figures. Binary Logistic regression model was used to test the association between independent and dependent variables. All variables at p-value <0.25 in bivariable logistic regression were entered into multivariable logistic regression to determine the association between a set of independent variables with the dependent variable. The odds ratio was estimated at 95% CI to show the strength of an association and statistical significance was declared at p-value <0.05. The model fitness was checked using Hosmer-Lemeshow goodness-of-fit the p- value which was >0.05 which was fitted (0.436). The variance inflation factor (VIF) was used to assess multicollinearity between the independent variables.
Study Variables
The dependent variable was MDRTB (yes or no). The Independent variables included the socio-demographic variables (age, sex, religion, socio economic status, educational status, occupational status, marital status, residence, family size, the number of rooms in the patient’s household ,and number of windows), clinical related variables (HIV status, history of contact with known TB patient, history of contact of with known TB patient, other underlying/chronic disease, number of TB episodes TB, outcome, history of interruption the first line anti TB, directed observed therapy, encountered side effects, category of TB, and duration of firs-line treatment)and behavioral related variables (prison status, alcohol consumption and cigarette smoking). Data on socio-demographic characteristics, behavioral characteristics and some the clinical characteristics were collected through face to face interviews. The remained clinical characteristics were collected by the review of the patient records and registration books.