Treatment seeking and antibiotic use for urinary tract infection symptoms in the time of COVID-19 in Tanzania and Uganda

Abstract

Background: There is still little empirical evidence on how the outbreak of coronavirus disease 2019 (COVID-19) and associated regulations may have disrupted care-seeking for non-COVID-19 conditions or affected antibiotic behaviours in low- and middle-income countries (LMICs). We aimed to investigate the differences in treatment-seeking behaviours and antibiotic use for urinary tract infection (UTI)-like symptoms before and during the pandemic at recruitment sites in two East African countries with different COVID-19 control policies: Mbarara, Uganda and Mwanza, Tanzania.

Methods: In this repeated cross-sectional study, we used data from outpatients (pregnant adolescents aged >14 and adults aged >18) with UTI-like symptoms who visited health facilities in Mwanza, Tanzania and Mbarara, Uganda. We assessed the prevalence of self-reported behaviours (delays in care-seeking, providers visited, antibiotics taken) at three different time points, labelled as 'pre-COVID-19 phase' (February 2019 to February 2020), 'COVID-19 phase 1' (March 2020 to April 2020), and 'COVID-19 phase 2' (July 2021 to February 2022).

Results: In both study sites, delays in care-seeking were less common during the pandemic than they were in the pre-COVID phase. Patients in Mwanza, Tanzania had shorter care-seeking pathways during the pandemic compared to before it, but this difference was not observed in Mbarara, Uganda. Health centres were the dominant sources of antibiotics in both settings. Over time, reported antibiotic use for UTI-like symptoms became more common in both settings. During the COVID-19 phases, there was a significant increase in self-reported use of antibiotics like metronidazole (<30% in the pre-COVID-19 phase to 40% in COVID phase 2) and doxycycline (30% in the pre-COVID-19 phase to 55% in COVID phase 2) that were not recommended for treating UTI-like symptoms in the National Treatment Guidelines in Mbarara, Uganda.

Conclusions: There was no clear evidence that patients with UTI-like symptoms attending health care facilities had longer or more complex treatment pathways despite strict government-led interventions related to COVID-19. However, antibiotic use increased over time, including some antibiotics not recommended for treating UTI, which has implications for future antimicrobial resistance.

Figure 1

Timeline of recruitment and COVID-19 restrictions at the two study sites. The shaded area and brown-coloured texts refer to the period of data collection at both study sites.

Figure 2

Treatment-seeking pathways during COVID-19. The chords represent percentages of the total sample, so the wider the circle’s segments, the higher the percentage of patients taking those steps. The outer gray ring distinguishes the different steps (steps 1–3, and last step). The second ring shows the relative share of patients who accessed treatment from that provider overall. Finally, the innermost coloured bars depict the share of patients moving from their current step to another one.

Figure 3

Predicted probabilities of taking ABs anytime in the pathway among patients who tried treating their UTI symptoms before coming to the recruitment clinic. PP – predicted probabilities, CI – confidence interval.

Figure 4

Distribution of patients by the type of antibiotics they took to treat their UTI symptoms. Mwanza, TZ: pre-COVID-19 phase (n = 836); Mwanza, TZ: COVID-19 phase 1 (n = 142); Mwanza, TZ: COVID-19 phase 2 (n = 285); Mbarara, UG: pre-COVID-19 phase (n = 374); Mbarara, UG: COVID-19 phase 1 (n = 109); Mbarara, UG, COVID-19 phase 2 (n = 416). P-values are derived from χ² tests.