Developing a theory informed intervention to improve linkage to treatment for patients diagnosed with tuberculosis using XPERT® MTB/RIF testing in Central and Eastern Uganda

Abstract

Background: In 2019, Uganda notified only 74% of estimated TB cases against an END TB target of 90%. This gap between estimated incident TB cases and those accessing TB treatment is due to a combination of underdiagnosis (people not accessing TB services or not being evaluated for TB when they do) and underreporting (people diagnosed with TB not linked to TB treatment - pretreatment loss to follow up). Previous studies from Uganda have shown that 25-40% of persons diagnosed with TB experience pretreatment loss to follow-up (LFU). We aimed to study factors associated with and reasons for pretreatment loss to follow-up in order to design a theory-informed intervention to improve linkage to treatment for patients diagnosed with TB using Xpert® MTB/RIF testing at selected public health facilities in central and eastern Uganda. Methods: We carried out four sub studies at ten public health facilities across ten districts. We purposively selected health facilities from three levels of the healthcare system involved in TB care (three primary care facilities, four district hospitals and three tertiary referral hospitals). First, we carried out a cross sectional study to describe the magnitude of and patient/health facility factors associated with pretreatment LFU. We then traced patients who experienced pretreatment LFU to determine their vital status. We compared vital status of patients who experienced pre-treatment LFU to that of patients successfully initiated TB treatment to determine the effect of pretreatment LFU on TB associated mortality. We also determined the effect of pre-treatment mortality on overall health facility estimates of TB associated mortality. Next, we carried out a qualitative study based on the Capacity, Opportunity, and Motivation for Behavior (COM-B) model to understand patient and health facility level barriers to and facilitators for linkage to TB treatment. We also identified intervention components from the Behavior Change Wheel that could alleviate identified barriers to and enhance facilitators for linkage to TB treatment. Finally, we conducted a quasiexperimental pre-/post-interventional study at one tertiary referral hospital to test the feasibility and preliminary effectiveness of the chosen intervention components. Data analysis involved descriptive statistics, multivariable logistic regression and Kaplan Meier survival curves (I-II) while thematic analysis was used for the qualitative study. The Pearson chi-squared test was used to compare the proportion of patients linked to TB treatment within two weeks of diagnosis in the intervention period (October to December 2019) using the period June to August 2019 as the comparison period. Results: Sub-study I: Across the ten health facilities, 100/510 (19.6%) patients were not initiated on TB treatment within two weeks of diagnosis (pretreatment LFU). Not having a phone number recorded in the clinic registers (aOR 7.93, 95%CI 3.93-13.05); being HIV-infected (aOR 1.83; 95% CI: 1.09-3.26) and receiving care from a health facility performing more than 12 Xpert tests per day (aOR 4.37, 95%CI 1.69-11.29) were significantly associated with pretreatment LFU. Sub-study II: Out of the 100 patients not linked to TB treatment, we successfully traced 49 patients. In the six months following TB diagnosis, mortality was higher among patients who experienced pretreatment LFU 48.1/1000py vs 22.9/1000py (HR 3.18, 95% CI 1.61 – 6.30). After incorporating pre-treatment deaths among patients who experienced pretreatment LFU, health facility level estimates of TB associated mortality increased from 8.4% (95% CI 6.1%-11.6%) to 10.2% (95% CI 7.7%-13.4%). Sub-study III: At the health facility, lack of knowledge about the proportion of patients not initiated on TB treatment; difficulty communicating sputum results from the laboratory and difficulty tracing patients due to inadequate recording of patient addresses were the main barriers to linkage to TB treatment. At the patient level, long turnaround time for sputum results; lack of transport funds to return to health facilities to collect sputum results and stigma were the main barriers to linkage to TB treatment. The most important facilitators identified were quick access to sputum test results either on the date of first visit (same-day diagnosis) or on the date of first return and availability of TB treatment. We identified education of healthcare workers, restructuring of the service environment to improve sputum results turnaround time and improving communication of test results from the lab to the healthcare workers to the patients as relevant intervention functions to alleviate some of the barriers to and enhance facilitators of TB treatment initiation. Sub-study IV: The median time from sample collection to provision of sputum results improved from 14 hours (IQR 4-47) in the pre-intervention period to 4 hours (IQR 3-6) in the intervention period. The proportion of patients started on treatment within two weeks of diagnosis improved from 58.8% (40/68) before the intervention to 89.1% (49/55) during the intervention (difference 30.3%, 95% CI 16.0% - 44.6%, p<0.01) while the proportion of patients receiving a same-day diagnosis increased from 7.4% (5/68) to 27.3% (14/35) during the intervention (difference 17.6%, 95% CI 6.6% - 33.2%, p<0.01). Conclusion A significant proportion of patients diagnosed with TB experienced pretreatment LFU. These patients were more likely to die from TB than those who were successfully linked to TB treatment. Because the majority of these deaths occurred outside the healthcare system and were not incorporated into health facility level estimates for TB associated mortality, the health impact of TB was inaccurately measured. Barriers to TB treatment initiation existed at both the patient and health facility level and required a multifaceted intervention to be overcome. An intervention that involved educating healthcare workers, restructuring of the work environment and enabling quick and efficient communication between healthcare workers and patients was feasible to implement and showed potential to reduce pretreatment LFU. Further evaluation of this intervention should be considered.