**4. Discussion**

This is the first study in Zimbabwe to assess the impact of COVID-19 on TB and HIV services in selected health facilities in Harare, the capital city. In summary, despite attempts by the NTP, the NAP and other parts of the health sector to prepare for COVID-19 [15,27], there was a large negative impact on almost all aspects of TB case detection, diagnosis and treatment, as well as HIV testing. The referral of HIV-positive persons to ART was also affected but only to a small degree. A health service already under strain was hit hard and brought almost to its knees in the first 12 months of the COVID-19 pandemic [27].

With respect to TB program activities, the numbers of people presenting to health facilities with presumptive PTB declined considerably over the 12-month COVID-19 period. These findings were similar to those described in the early months of COVID-19 in clinics in Tehran, Iran [28] and Nigeria [29], where lockdown restrictions, transportation difficulties and community fear of health facilities were thought to hinder health facility access. There was a corresponding decline in numbers diagnosed and registered with TB, in line with reports from other countries where the decreases in TB case notifications in the early months of COVID-19 compared with previous years were 48% in clinics in China [30], 48% in clinics in Brazil [31] and 56% in India [32].

A modeling analysis at the start of the pandemic suggested that a 3-month suspension of TB services due to COVID-19 lockdown followed by ten months restoration back to normal would cause over five years an additional 1.2 million TB cases in India, 25,000 additional TB cases in Kenya and 4000 additional cases in Ukraine, mainly as a result of the accumulation of undetected TB during lockdown [33]. Unfortunately, in Zimbabwe, there was no restoration of services back to normal. On the contrary, from October 2020 to February 2021, there was a steady decline in numbers of persons with presumptive PTB and registered TB, coinciding with the expiry of cartridges for Xpert MTB/RIF assays. Sputum specimens were not collected for investigation, and patients with TB-related respiratory symptoms stayed away. The one encouraging finding was an overall increase in bacteriological positivity in the COVID-19 period compared with the previous year. There are several possible explanations that include: (i) when laboratories were functioning and reagents were available, standard laboratory operating procedures remained intact; (ii) there was selective self-referral of only those who were really sick with TB; and (iii) patients may not have had access to community TB testing services as a result of COVID-19, and therefore those with TB came to Harare city health facilities for investigation.

TB treatment success rates declined during the COVID-19 pandemic. However, there was some improvement in the second half of the year, and particularly in January and February 2021, when outcomes were better than at the corresponding times before COVID-19. The probable reason for better outcomes at this time was, as discussed earlier, a series of health care worker strikes towards the end of the pre-COVID-19 period as a result of heightened economic decline. This disrupted routine follow-up of TB patients on treatment, resulting in a high number of patients not being evaluated.

With the onset of the COVID-19 pandemic, we were concerned that COVID-19 restrictions would hinder patients collecting anti-TB medications and compromise drug adherence. However, the longer periods given between anti-TB drug refills and their synchronization with ART medication refills made things easier for patients and partly mitigated this challenge. The large proportion of patients "not evaluated" was a huge challenge, compromising the program's ability to record successful treatment outcomes. This was partly due to the environmental health technicians responsible for following up patients in the community being repurposed to COVID-19 activities. Concerted efforts, however, were made to tackle this issue, and through human resources support from PEP-FAR (the President's Emergency Plan for AIDS Relief), treatment success rates considerably improved in the last two months.

With respect to HIV services, there was a significant decrease in numbers of people presenting for HIV testing, although this did improve during the latter half of the COVID-19 period. There were two main reasons for the large decline. First, the cessation of VMMC services would have resulted in a large decline in males being HIV tested during this time. Second, in April 2020, during the national lockdown, the MOHCC issued a directive for all community-based HIV testing to be temporarily halted to minimize the risk of community health workers and community members from contracting COVID-19. The improvement in the latter half of the COVID-19 period may have been due to the strong support from PEPFAR partners.

The overall reductions in HIV testing were similar to what has been observed in Europe [34], the USA [35] and Africa [36]. It is likely, though, that we have overestimated the COVID-19 impact on HIV testing because of Zimbabwe adopting a more targeted approach, promoting HIV self-testing and focusing on index HIV testing before the onslaught of COVID-19. The increase in HIV positivity observed in the COVID-19 period in our study is probably a result of this more targeted approach to identifying clients more at risk of HIV. While referrals to ART decreased in the COVID-19 period, it was encouraging to see that they were still maintained at over 90%.

There are several strengths to this study. First, the real-time monthly surveillance was embedded within the routine services of the health facilities, and there was cross-checking and validation of the data each month between the country coordinator and the overall study monitoring and evaluation officer. We believe, therefore, that the data were accurate and reflected programmatic practice in the field in Harare. Second, we used two 12-month periods to compare data, thus accounting for seasonal changes that might have affected TB case detection and HIV testing. Third, the conduct and reporting of the study were in line with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines [37].

However, there were some limitations. Our study was limited to health facilities in Harare, and therefore may not be representative of Zimbabwe as a whole. The use of aggregate quantitative data made it impossible to understand the cascade of care for TB case detection and HIV testing during COVID-19. Further mixed-methods research with both quantitative and qualitative analysis amongs<sup>t</sup> TB patients themselves, as has been conducted in Zambia [38], are needed to obtain a better idea of why patient access to health services and health service delivery was so badly affected. We only assessed referrals of HIV-positive persons to ART and did not document the initiation or retention on ART. Previous studies have suggested that ART interruption has been a problem during the COVID-19 pandemic [39]. It would have been interesting to assess this in Zimbabwe, especially as interruption to ART is thought to be one of the most important determinants of HIV-related mortality during the COVID-19 pandemic [40]. Finally, the official monthly reports to WHO of COVID-19 cases and deaths may have underestimated the true burden of COVID-19 in the country. A study on deceased people at the University Teaching Hospital morgue in Lusaka, Zambia, found that just 9% of 70 people, who were confirmed as having SARS-CoV-2 from postmortem nasopharyngeal swabs within 48 h of death, had ever been tested before death [41]. This finding is unlikely to be confined to Zambia, and large numbers of unreported cases and deaths due to COVID-19 are probably occurring in other countries in Africa.

Despite these limitations, there are some important programmatic implications from this study. First, the strengthened monthly surveillance system worked well with both NTP and HIV program directors looking each month at the monthly reports and using the data for decision making and suggesting interventions. The monthly data were also shared with the ten health facilities through the country coordinator, and this enabled the health facilities to implement interventions simultaneously. However, the monthly surveillance and the interventions applied were insufficient to reverse the direct negative effects of COVID-19 on services or the indirect effects such as industrial action and stock-outs of diagnostic tests. Given the challenges both programs faced, it is remarkable that services were able to function at all, so without monthly surveillance, it might have been much worse. The monthly surveillance as conducted in Harare required effort and external support. If this became routine as desired, the NTP would require considerable investment, probably involving electronic case-based reporting systems. Technical support by partners, who have additional human resources to help, would be vital to ge<sup>t</sup> activities and services back to normal.

Second, with COVID-19 likely to become endemic, there is an urgen<sup>t</sup> need to bring TB case detection and treatment outcomes back to pre-COVID-19 levels. This study suggested that better contact tracing of TB patients and heightened attention to reducing the "not evaluated" category were important interventions for TB case detection and TB treatment outcomes. Support from PEPFAR partners also helped to support staff in ascertaining TB treatment outcomes. There are several other suggestions about how TB services could be sustained or improved in the face of the pandemic. These include: better integration and screening of TB and COVID-19 at the health facility and community level; sharing testing algorithms and multiplexing equipment such as GeneXpert platforms within the laboratories; having robust procurement and delivery systems to avoid reagen<sup>t</sup> stock-outs; ensuring effective infection, prevention and control activities within health facilities; having longer 3-month follow-up appointments for patient check-ups and drug collection; providing health information education in health facilities and the community; and mobilizing support networks of TB survivors and TB communities [42–44]. More use of digital platforms for case finding, drug adherence, managemen<sup>t</sup> of adverse drug reactions and training have also been recommended as a way of rapidly restoring TB diagnosis, care and prevention services [45]. The Zimbabwe NTP and MOHCC could consider which, if any, of these innovative approaches might help, and might be feasible and cost-effective to implement.

Third, HIV self-testing, index partner notification and home-based HIV testing services have allowed HIV testing numbers in the COVID-19 era to rebound in some African countries [46–48]. Zimbabwe is already moving in this direction. It is also vital to capture all HIV testing and HIV test results outside of as well as within the health sector in order to keep informed about what is happening and ensure that all HIV-infected persons are able to initiate ART. Zimbabwe again is already implementing such an approach and is tracking on a monthly basis facility and community distribution of HIV test kits, as well as numbers tested for HIV along with their results. Further deployment of electronic data capture systems will be important in this context.

Finally, the resources and funding that have been raised to fight the COVID-19 pandemic have been staggering. The proposed global financing systems that have been proposed as a way of promoting and enabling science and product development in the health sector [49] must be used and leveraged to help fight other diseases, TB and HIV/AIDS included, so that we can stay on track to deliver on the Sustainable Development Goals in 2030.
