Study area
The study was conducted in two zones (Mekelle and Eastern zones) of Tigray Region, Northern Ethiopia. Mekelle is the Regional capital which is located 778km from Addis-Ababa. The region has seven administrative zones: comprising a total of 53 Weredas (districts) and 673 Tibias (sub-districts). The population is above 5.3 million which is about 6% of the total population of Ethiopia and the rate of urbanization is about 20% per year. Economically active age group accounts for about 51.8 % of the total population, 95.5% of the population are Orthodox Christians, 4.1% are Muslims and 0.4% are Catholics (18). There are 712 health posts, 214 health centers and 38 hospitals (1 specialized hospital, 15 general hospitals, 22 primary hospitals) in the region. There are also well-organized community-based structures that support the health system at the grass-root level (health extension workers and women development army). The private health sector also plays a significant role in relieving the burden of the patient load from the public facilities with strong policy support (19, 20). Currently, almost all health facilities are providing directly observed short-course tuberculosis treatment by trained health workers (TB focal persons). The program was introduced in all hospitals, health centers and in most of the health posts for the past years (19).
Study design, population, period and inclusion
A cross-sectional survey was conducted to investigate the effect of total delay of PTB patients on clinical severity and risk of transmission in health facilities of two zones of Tigray Region, Northern Ethiopia. Study participants were bacteriologically positive or x-ray positive pulmonary tuberculosis new cases during one year period (i.e. October 2018 to October 2019) in the health facilities of the two zones of the Tigray region. PTB patients whose age is 18 years and above , with no previous history of TB, non-MDR-TB cases and who are diagnosed and initiated treatment within the study settings were included in the study.
Sample size
In calculating the sample size that enable us to estimate the effect of delay on clinical severity and risk of transmission, we considered 95% confidence interval, 80% power, equal number of exposed and non-exposed (delay was exposer). The calculated sample size for the objective was 850 and we have considered 425 TB patients as non-exposed and the remaining 425 as exposed.
Recruitment of study participants
Out of the seven zones of the region, two zones (i.e. Eastern zone and Mekelle zone) were selected because of their proximity to the research Institute where the TB culture is performed, and logistical reasons that emanates from the need to have frequent visit to the study sites. Furthermore, the only functional specialized hospital is located in Mekelle. Within the selected zones, there were 26 health centers and 6 hospitals that reported an average monthly cases of 3 and above in the years 2016/2017 and 2017/18 (20). Hence, of the 26 health centers and 6 hospitals, 16 were selected by lottery method and five hospitals were included in the study excluding one hospital as it become for ophthalmological service only. The planned sample size was proportionally allocated to each health facility based on the average number of TB cases reported by the facility in the prior two years. Study participants within each facility were consecutively enrolled until the allocated sample size was attained.
Data collection
The data collection tool was drafted from previously tested instruments and from reviewed literature (12, 21, 22) and then modified to fit to the context of the research interest. Study participants were identified during their treatment start and invited for consultation and enrolment. The samples for laboratory investigations were collected before treatment initiation. Data were collected on socio-demographic and lifestyle factors, clinical and diagnosis related factors, outcome measurement using Bandim TB score, and measurement of risk of transmission as described below.
Socio-demographic and lifestyle data: Background data collected from study participants include gender, age, residence, occupation, marital status, education, family size, income as indebt (poor), income=expense, and saving; smoking history as smoker and non-smoker, and alcohol intake (Yes, No).
Clinical and diagnostics data: Clinical examination and interviews of patients were carried out at recruitment or inclusion. During this time Symptoms (respiratory symptoms, constitutional symptoms, or both), smear status (as positive and negative), pulmonary cavity by chest x-ray (yes/no), and patient delay duration by (<30 days, 31–60 days, and >60 days) were collected. Respiratory symptoms included dyspnea, hemoptysis, cough, or expectoration while constitutional symptoms included malaise, loss of appetite, weight loss, fatigue, fever, chills, chest pain or night sweats.
Smear microscopy of all the study participants was done at three GeneXpert sites of the study area. Three slides per individual were prepared and examined by two laboratory technologists independently. X-ray of the study participants was done in five of the study hospitals who have X-ray machine. These results were examined by a medical doctors specialized in radiology
Bandim TB score: Clinical severity was assessed by the Bandim TB score (8). The Bandim TB score is a newly developed tool to assess clinical status of patients with TB. Its value is expressed as a numeric index based on cough, hemoptysis, dyspnoea, chest pain, night sweating, anemic conjunctivae, tachycardia, positive lung auscultation, increased temperature, body mass index (BMI) and middle-upper arm circumference (MUAC). A Bandim TB scores ≥8 correlates with mortality and lower TB score correlates with favorable outcomes, with cure and treatment success (23, 24). In these studies and in other settings (8) severe TB was defined as TB score ≥8, which has a strong predictive capacity for mortality.
Scoring and severity classes: The score was based on 11 clinical variables (i.e. cough, hemoptysis, dyspnoea, chest pain and night sweating, anemic conjunctivae, tachycardia, positive lung auscultation, increased temperature, body mass index (BMI<16 and BMI<18) and middle-upper arm circumference (MUAC<200mm and MUAC<220mm). Each of the 11 variables is scored 0/1 point; the total score was generated resulting the maximum score of 13 points. For all variables examined and used in the score, missing values were coded as zero. The patients were classified into two severity classes as TB score <8 as having less severe and TB score having >8 as having a higher severity class.
Measurement of risk of transmission and definitions
In this study sputum smear positivity and presence of pulmonary cavity were used as proxy measures for the risk of TB transmission (22, 25, 26). Patients who were smear-positive and had cavitation or smear-positive but had no lung cavitations or smear-negative but had lung cavitation was considered as having risk of transmission. Whereas those patients who were smear-negative and had no lung cavitations were considered as having no risk of transmission. Lung cavitations were defined as the presence of at least one cavity on a chest radiograph taken during the episode of TB (25). Total delay is the time period from the onset of TB symptoms to first start of anti-TB treatment and classified as <30 days, 31-60 days and >60 days.
Data quality and management
The questionnaire was pretested before the actual data collection. During sample collection for laboratory investigation, patients were advised to submit thick sputum rather than saliva and the sputum sample was collected as per standards (27). Before smearing, specimens were assessed macroscopically for quality and recorded as purulent with whitish color or bloody. The study numbers or codes were covered with wrap around stickers before microscopy. As part of internal quality control (IQC), all the positive slides and randomly selected 50% of negative slides were re-examined by a different laboratory technologist.
Data analysis
Data was computerized and analyzed using SPSS version 21. For categorical variables, univariate analysis was performed by using frequencies and percentages whereas mean or standard deviation was employed for continuous variables. Normality was checked for those continuous variables and the total treatment initiation delay was not normally distributed. Hence, Mann-Whitney U and Kruskal Wali's tests were used to compare median delay across the clinical spectrum of the study participants. Initially bivariate logistic regressions was used to model association of pre-specified characteristics and the odds of delayed treatment initiation and then those variables with p-value of at most 0.2 were entered into the multivariate model. A comparison of the proportion of smear-positive tuberculosis and the presence of a pulmonary cavity (risk of transmission) with the duration of patient delay was modeled using a multivariate logistic regression model. Similarly, the association between clinical severity and the total delay was modeled using multivariate logistic regression. In both logistic regression models the goodness of fit was checked using Hosmer and Lemeshow test. In all analysis any result with the p-value less than 5% was considered as being statistically significant.