Spatial distribution of pulmonary tuberculosis patients with delays to seeking evaluation in five referral hospitals of Uganda and the associated predictors.

Abstract

Background: Decreasing tuberculosis (TB) transmission requires interventions targeting communities with greatest need. Understanding spatial distribution and predictors of pulmonary TB patient’s (PTB) delays seeking evaluation are key to this success. Design/Methods: This was cross-sectional secondary data analysis study utilizing the East African Public Health Laboratory Network (EAPHLN) study data. The EAPHLN study evaluated GeneXpert MTB/RIF performance in five referral hospitals in Uganda. In this study, we described spatial distribution of PTB patient delays (≥30 days) to seeking evaluation at sub-county level using Moran’s I. Seeking evaluation was defined as going for formal health care for their symptoms. We performed bivariate and multivariate logistic regression adjusting by hospital to identify predictors for delays. Results: Adult PTB patients enrolled into the study were 473, median age 34; IQR; 16, 195 (41%) had delays to seeking evaluation. Sub-county clustering of PTB patients was seen in four referral hospitals (Global Moran’s I; 0.20- 0.57, p<0.0001). Two referral hospitals had clustering of delays (Global Moran’s I; 0.42 and 0.31, P values; 0.003 and 002). Being married (OR 1.7, 95%CI; 0.98-3.30, P=0.04) and having noticeable weight loss (OR 1.7, 95%CI 0.02-3.02, P=0.04) were associated to delayed seeking of care while, having night sweats (OR 0.6, 95% CI; 0.44-1.52, P=0.03) was associated to seeking care early. Conclusions: Areas of large concentration of PTB patients and individual factors were not associated with delays. Areas of high PTB patient concentration and their predictors suggest increasing targeted screening by Village Health Teams, potentially reducing TB transmission and infection.