Study design and population

A health facility based cross sectional study was conducted from April1, 2016 to January 30, 2017. The study population consisted of all forms of newly diagnosed adult TB patients who were registered to the selected health facilities during the study period. Sample size was determined using single population proportion formula as follows. N = [(Zα/2)² ∗ p ∗ (1-p)] /d² where: N = total sample size, Zα/2 at 95% CI is 1.96, d marginal error 0.05, and P is the proportion (47.8%) of TB patients who started treatment after 60 days (median total delay) of initial onset of signs and symptoms [17]. The calculated sample size was 384. The study was carried out in 15 randomly selected public health facilities that provide directly observed treatment short course (DOTS) in the study area. Total sample size was proportionately allocated for the 15 randomly selected health facilities based on their previous one year estimated number of registered TB patients. The proportions of participant TB patients were as follows: 35 from one referral hospital, 18 from one general hospital and 337 from 13 health centers. Thus, TB patients who fulfilled the inclusion criteria were included consecutively from the above mentioned public health facilities until the calculated sample size was attained.

New patients (≥18 years old) diagnosed, based on the national TB guideline [21], with TB of all forms and who had commenced treatment for less than 15 days at the time of data collection were included in the study. TB patients transferred to the selected health facilities for continuation of treatment were included if the transfer was within 15 days of starting treatment. Double counting of study participants was avoided through attachment of reminder note with referral sheet. However, TB patients who were relapsed or treatment failed, critically ill, unable to communicate, refused to participate and to give consent for human immunodeficiency virus (HIV) screening were excluded. TB patients who were temporary residents of the study area (living for < 6 months) were also excluded from the study.