A cross-sectional retrospective study was conducted, analysing quantitative and descriptive secondary data. The study design was used for population-based surveys and to assess the prevalence of diseases in clinic-based samples.²¹

The study setting was the City of Johannesburg in Gauteng province, South Africa. This included data from all government hospitals and clinics within the City of Johannesburg Metropolitan District.

The study population included everyone who accessed healthcare services in Johannesburg. The sample was restricted to those who accessed TB services during our study period (January 2018 – December 2021) in the Johannesburg district.

The study used de-identified routine TB secondary data from Johannesburg facilities operating before and during the COVID-19 pandemic, including individuals who accessed TB services, showed symptoms, or tested positive or negative between January 2018 and January 2022. Data outside this time frame or from outside Johannesburg were excluded.

Descriptive analysis was undertaken to categorise patients into children (< 10 years), young adolescents (10–14 years), older adolescents (15–19 years), young adults (20–24 years), adults (25–49 years) and older adults (≥ 50 years). Additional demographic information, including district within the Johannesburg Metropolitan Municipality, sex, race and year of diagnosis, was also collected. Patient diagnoses were classified as follows: abscess, anaemia, cough, COVID-19, drenching night sweats (DNS), haemoptysis, HIV, immunodeficiency, weight loss, lower respiratory tract infections, lymphoma, lymphadenopathy, malignancy, malaria, malnutrition, meningitis, noncommunicable diseases (NCDs), pregnancy, pleural effusion, pneumonia, sepsis and TB.

Information on demographics and diagnoses was summarised by age groups using proportions for categorical variables. Descriptive analysis was done to characterise the study setting, sample and provide an overall description of the data. A period-over-period chart was used to illustrate changes in TB screening data across the study months or period using timeframe analysis. This period over chart shows the relationship between the TB data across months and highlights important trends. Data was represented on a line graph to show the trend of screening tests over the study period. Descriptive statistical analysis was used to outline the characteristics of the dataset by summarising and presenting key features through tables, charts, and graphs of the sampled data, enabling evidence-based conclusions.

The Institutional Research Ethics Committee of the University of Johannesburg granted ethical approval and guided the research process (REC-2920-2024).

The age distribution of participants was summarised using categorical age groups. The majority of individuals (69.5%) fell within the 25–49-year age group, representing the largest proportion of TB service users in the dataset. This was followed by participants aged 50 years and above (15.3%), young adults aged 19–24 years (7.4%), and those with unspecified age information (4.2%). Smaller proportions were observed among older adolescents aged 15–18 years (1.6%), children under 10 years (1.5%) and young adolescents aged 11–14 years (0.6%).

Because age was captured in categories rather than as continuous values, measures of central tendency, such as mean and standard deviation, are not appropriate. Instead, the distribution is best characterised by proportions across the predefined age groups. Visual inspection of the age-group frequency distribution indicates that the data are left-skewed, with the highest concentration of cases among adults aged 25–49 years.

Males represented 59.2% of the sample, while females accounted for 37.9%; 3.0% had unspecified sex information. Racial identifiers were largely unavailable, with 99.8% of participants recorded as ‘unknown’. This limits meaningful interpretation of race-related patterns.

Across the four study years, the highest number of TB cases was recorded in 2018 (32.5%), followed by 2019 (27.8%). Service utilisation declined sharply during the COVID-19 pandemic, with 20.3% of cases recorded in 2020 and 19.4% in 2021. This decline corresponds with the onset of national COVID-19 restrictions and disruptions to routine healthcare services.

Tuberculosis service utilisation varied across the seven administrative regions of the City of Johannesburg. Region D reported the highest proportion of cases (35.0%), followed by Region F (16.9%) and Regions B and G (11.6% each), while Region C reported the fewest cases (5.6%). While these differences may reflect variation in population density, facility distribution and service accessibility, the relative decline in TB patient attendance following the onset of COVID-19 was consistent across all regions, ranging between 38% and 42%. This uniformity suggests that the disruption to TB services was widespread and systemic, affecting service utilisation across the metropolitan area regardless of regional differences.

Participants presented with a diverse range of diagnoses prompting TB screening. The largest proportion (59.8%) fell under ‘non-specified’ diagnoses, followed by individuals with non-communicable diseases (26.7%). Those already diagnosed with TB accounted for 9.7% of cases, while smaller proportions (< 1%) were associated with conditions such as DNS, pneumonia, pleural effusion and meningitis.

When comparing the pre-COVID-19 and peri-COVID-19 periods, several conditions demonstrated a marked decline, including renal artery stenosis, lymphoma, pneumonia and TB. This reduction may reflect decreased healthcare utilisation, limited access to diagnostic services such as imaging and specialist consultations, and shorter or prioritised clinical encounters during the pandemic. In contrast, an increase in lymphadenopathy was observed, which may be attributable to its role as a clinically detectable sign requiring minimal diagnostic resources, as well as heightened vigilance for infectious or inflammatory conditions during the COVID-19 period.

As only individuals who tested positive for TB were included in this study, these trends are interpreted with caution, as the dataset does not capture patients who screened negative or were not tested.

Figure 1 shows the annual distribution of TB cases recorded in the dataset, with an overall decline of 4891 cases across the study period. Notably, a decrease of approximately 1900 cases is observed between 2018 and 2019, prior to the onset of the COVID-19 pandemic. The reasons for this pre-pandemic decline are not fully clear and may reflect underlying health system challenges, changes in case detection practices, or natural variation in TB reporting trends. A further decline between 2019 and 2020 coincides with the onset of the COVID-19 pandemic and is more plausibly linked to disruptions in healthcare access, including movement restrictions, fear of infection, and the reallocation of healthcare resources.

Data beyond the study period were not available to assess the recovery of TB service utilisation following the COVID-19 pandemic, which represents an important area for future research.

The disruption of TB services observed in this study, including the consistent decline in utilisation across all regions, underscores the need for resilient health systems capable of maintaining essential services during widespread public health emergencies. The uniformity of this decline suggests that future interventions should be implemented at a system-wide level rather than targeting specific geographic areas.

Integrated service delivery models, such as combined TB, HIV and COVID-19 screening, together with multi-month dispensing and community-based medication delivery, may help to sustain continuity of care during similar disruptions. In addition, the observed reduction in facility-based attendance highlights the potential role of telemedicine platforms to support patient follow-up, symptom monitoring and adherence counselling without requiring in-person visits.

The large proportion of ‘non-specified’ diagnoses in the dataset further highlights gaps in routine health information systems. While this limits diagnosis-specific interpretation, it does not substantially affect the observed patterns of service utilisation. However, it does underscore the need for improved clinical documentation and diagnostic coding practices to enhance the value of routine data for service planning, monitoring, and future research.

This study provides valuable insight into the impact of COVID-19 restrictions on TB service utilisation and highlights the broader societal effects of a pandemic. By analysing service utilisation across different lockdown alert levels, the study offers a nuanced understanding of how varying restrictions influenced TB care. It emphasises the importance of out-of-care, community-based interventions that can sustain TB service utilisation during public health emergencies and beyond, particularly for TB patients and those co-infected with other conditions. The descriptive analysis provides insight into the study sample and service utilisation patterns, while the findings underscore the importance of maintaining continuity of care for TB and other chronic conditions during pandemics. In line with existing literature, the results also highlight the influence of social determinants on access to and provision of TB services.

As with other studies relying on secondary data, this research faced limitations related to data availability and quality. Missing or incomplete data constrained the study’s ability to fully explore TB trends and service utilisation patterns during the pandemic. Inconsistencies in identity numbers and constant variables across the pre-COVID-19 and peri-COVID-19 periods limited the capacity to establish associations between variables. Despite these constraints, the study provides meaningful insights into the effects of pandemic restrictions on routine TB services.