Opening up safely: public health system requirements for ongoing COVID-19 management based on evaluation of Australia’s surveillance system performance

Countries recognised as having the best initial responses to the coronavirus disease 2019 (COVID-19) pandemic include Singapore, Taiwan, South Korea, Australia, New Zealand, Canada, Iceland, UAE, Germany, and Greece. These countries were largely successful in minimising morbidity and mortality prior to the availability of vaccines [1, 2]. Key to Australia’s success was early identification of community transmission through surveillance, control of outbreaks through non-pharmaceutical public health measures restricting mobility and social interaction, and effective management of international borders which reduced the risk of subsequent re-introductions [3, 4]. In addition to reducing health impacts, these measures have limited the number and duration of restrictions on community mobility and interaction. Consequently, the social and economic burdens of the pandemic were lower in Australia than in many other countries [5].

In the initial phases of the global pandemic, reliance on nationwide lockdowns in Australia was necessitated by delayed closure of international borders [6], which allowed community transmission to establish itself throughout many urban settings, and by limited surveillance, testing, and public health contact tracing capacity for detection and control of such transmission [7]. From November 2020 to June 2021, with improved capacity and capability of public health systems, and a reduction in the number of introductions from international arrivals, Australia was able to effectively identify and control outbreaks with a combination of measures, which sometimes included short lockdowns.

Since late 2020, the epidemiology and thus control of COVID-19 has changed due to two critical and opposing developments. The first is the emergence of novel variants of severe acute respiratory syndrome coronavirus 2 (SARS-COV-2) with higher transmissibility [8]. The second is the availability of several vaccines against SARS-CoV-2 [9], many of which have been in use since early 2021, 1 year after the disease began circulating globally.

It is unclear as yet whether herd immunity can be achieved with the current vaccination programme, as current vaccines are less effective at preventing transmission than preventing serious disease [10, 11]. Furthermore, until transmission is controlled globally, we are likely to continue to see the emergence of new variants, which may have higher transmission potential and increased severity of disease, including in the vaccinated and in vaccine-ineligible. Despite high coverage in wealthy nations [12], achieving high global vaccine coverage in most low- and middle-income countries will take years [13‐15]. There is therefore a strong imperative to understand the continuing public health system requirements for detecting and controlling community transmission, including of variants of concern (VoCs), once planned vaccine programmes are completed and other public health control measures are lifted.

All confirmed cases of SARS-CoV-2 detected in the community in Australia in the period from 1 November 2020 to 30 June 2021 were linked to discrete outbreaks seeded from outside the country, and during this time Australia also had high rates of community testing. Hence, no cases of unknown origin were observed, meeting the criteria for considering that ongoing community transmission of SARS-CoV-2 in Australia was eliminated during this period [16]. All recorded outbreaks in this period have been linked through epidemiological investigation and genomic sequencing to international arrivals. Several resulted in local transmission in the community, all of which were subsequently contained, aside from outbreaks in Victoria and NSW (New South Wales) that commenced late in the study period and led to ongoing community transmission from July 2021 onwards.

The study period (1 November 2020 to 30 June 2021) therefore provides an ideal context in which to evaluate surveillance system performance as community transmission was controlled and case detection was close to complete [17]. The aim of this paper is therefore to define critical public health requirements for the ongoing management of COVID-19, including surveillance system requirements to support detection and control. This information informs current and future contexts of widespread vaccination, including approaches to international borders, and the emergence of VoCs. To address this aim, we:

This paper describes a study evaluating the implementation of the COVID-19 surveillance system in Australia utilising publically available testing, case, and outbreak data. We applied the StaRI checklist for reporting of implementation studies when drafting our paper to ensure an accepted and standardised methodology and reporting of our study [18].

Once vaccine coverage rates are maximised, the priority for surveillance shifts from detection of SARS-CoV-2 to detection of SARS-CoV-2 VoCs associated with increased severity of disease. This surveillance will require ongoing investment to maintain current community-based disease surveillance systems and implement strategies for screening of returned travellers on a larger scale. Such surveillance systems will need to be linked to greatly increased genomic sequencing capacity if novel VoCs are to be detected early in transmission. Given Australia’s limited genomic sequencing capacity, there is an urgent need for investment in scaling up and decentralising, including through novel technologies capable of large volume throughput screening to detect mutations of concern.

Most novel incursions of SARS-CoV-2 into the Australian community have been detected through community-based surveillance for symptomatic disease, with the remainder detected through screening of returning overseas travellers and staff linked to quarantine facilities. Community-based symptomatic testing levels in the region of the outbreak were between 1 and 3% of the population tested in the week preceding detection of the outbreak, suggesting this level of testing will allow early detection of most community seeding or amplifying events. Outbreaks were detected, on average, within 5 days of the primary case. All outbreaks in 2021 were due to novel variants of concern. Although outbreaks due to the Delta variant of SARS-CoV-2 appear to be more transmissible and therefore more difficult to control, the surveillance system has still performed well in being able to detect them early.

The Australian national surveillance plan includes under goal 10 a range of indicators for timeliness of detecting individual cases [3]. Given that during the study period, a single case in the community in Australia represented an outbreak, these indicators would provide appropriate guidance against which to compare our findings. However, despite providing definitions of appropriate indicators of timeliness, the strategy does not define goal or targets for these indicators. A preliminary review of other Australian jurisdictions and internationally did not identify specific targets for timeliness either. This is an important gap that must be addressed by policymakers, and our study provides data to inform the development of appropriate and feasible targets in regard to detection of cases and outbreaks.

Wastewater surveillance was of limited utility in the study period. However, in the period following the study, from July to August 2021, there were two outbreaks in which detection of the outbreak was preceded by unexpected positive wastewater surveillance results [43‐45]. Targeted screening of wastewater from high-risk settings (e.g. public housing towers [46]) has also been utilised in recent responses. Given the overall low sensitivity and specificity, detection will rely on community surveillance to rapidly identify cases if it is to be of utility in public health response, investing in maintaining community surveillance through high uptake of testing will continue to be critical.

As our analysis suggests, to have similar performance in detecting variants as Australia has had in detecting novel outbreaks of SARS-Co-V2, sequencing rates of up to 1000-fold magnitude higher than current requirements would be required. Prioritising sequencing to those samples linked to severe disease (hospital/ICU) would require less capacity. However previous research demonstrates that if surveillance were to rely on hospital rather than community-based case detection, the number of cases at the point of outbreak detection would be much higher, making outbreak control much more challenging [35]. A more effective strategy would be to prioritise sequencing to international arrivals, the most likely source of novel VoC introductions in Australia. This would require ongoing comprehensive SARS-CoV-2 testing of all arrivals. Even this alone would necessitate marked scale-up of sequencing capacity if international arrival numbers were to return to pre-COVID-19 levels.

Attempts of other high-income nations to detect SARS-CoV-2 variants highlight current constraints on the capacity of nations to conduct large-scale genomic sequencing. As of February 2021, the USA had sequenced around 96,000 (less than 1%) of its 27 million COVID-19-positive cases [47]. Experts attribute the country’s low sequencing rate to a lack of national coordination, poor surveillance planning, and inadequate investment in sequencing infrastructure and research [47]. In Canada, genomic surveillance has been a key component of the national surveillance strategy, with relatively greater success than the USA [48, 49]. An average of 5% of COVID-19-positive samples are being analysed across the country [50], with a target of sequencing 10% of positive samples [50]. This would bring its efforts closer to the UK, which has been lauded for its world-leading genomic sequencing capability [51]. As of 5 July 2021, the UK has conducted approximately 672,677 sequences out of 5,186,297 infected people (13.0%) in total [52].

However, even the UK’s level of sequencing is much lower than would be needed for Australia to conduct optimal surveillance for VoCs. Given the challenges faced by other developed countries in sequencing a high proportion of samples once the incidence of SARS-CoV-2 increases, it is unlikely that whole genome sequencing would be able to meet such requirements. It is therefore critical to implement genomic surveillance strategies that are effective in early detection of novel VoC introductions or emergence, to develop technologies that can screen for, rapidly and at scale, specific mutations of particular concern (e.g. immune escape mutations). Without such planning and adaptation for future scenarios of increased transmission, not only will it be impossible to sequence enough samples to meet surveillance goals, timeliness of sequencing will also be delayed, in turn delaying the identification of transmission links related to VoCs.

The Omicron variant of SARS-CoV-2 was first identified in late November 2021 globally and soon after designated a variant of concern by WHO [53]. By early December 2021, available data indicated increased transmissibility, including in vaccinated individuals, but impacts on the severity in vaccinated were unclear. The multiple mutations on the spike protein suggest there may be an immune escape and therefore a need for adapted vaccines. Even with Australia’s high levels of genomic sequencing, within 2 weeks of report of the first case, multiple jurisdictions were reporting Omicron cases, many with no known link to international travel or travellers [54]. This is likely to be due to levels of testing of overseas arrivals being around 40%. Retrospective sequencing of SARS-CoV-2-positive samples also identified further cases [55], indicating that higher rates of genomic surveillance may have detected community transmission of this variant earlier. Overall, the early information on the introduction of the Omicron variant supports the conclusions and recommendations made in this paper on the management of future variants based on data prior to its arrival.

A limitation of this study was that testing rates were calculated for States as a whole and therefore may not reflect the testing rates in the specific area where the outbreak occurred. Rates vary by geographical distance from testing sites. Given all outbreaks except one were detected in major metropolitan areas, which have higher testing rates than rural areas, State-level averages are likely to be an underestimate of testing in urban areas. However, testing rates also vary widely within urban areas based on the socio-economic characteristics of the catchment population. Testing data disaggregated by LGA were not publicly available for most States, but future analysis should be done to refine these estimates if such data becomes available.

Surveillance system requirements must be considered in the context of other public health measures to detect and control COVID-19 transmission. These are vaccine effectiveness, availability and uptake, outbreak response capacity (including effective implementation of social and mobility restrictions), and community engagement and partnership as the cross-cutting enabler of success in all areas. Additional file 4 provides a framework for future non-surveillance-related public health control measures when focus shifts to the management of VoCs.

As referred to previously, the impact of vaccination on transmission, and whether or not other measures are needed, depends in part on the effectiveness of the vaccine. See Additional file 4: Relationship between vaccination and other control measures. As has become apparent, knowledge is rapidly changing and evolving around the effectiveness of available vaccines even for current viral variants, such as the optimal number of doses for the primary schedule and the degree and timing of boosters to address waning of immunity [56]. The impact of vaccines is also related to vaccine availability, which is dependent on production, procurement, and distribution capacity. As seen in Australia in early 2020, limitations in all of these areas were the main barrier to vaccination [57, 58]. Additional file 4 assesses the requirements for ensuring that such barriers are minimised as much as possible in the event a novel vaccine is needed for a VoC. Targets are also critical, and a useful and ethical approach when assessing performance of vaccination programmes is to consider the highest coverage achievable with available vaccines and an effective community support and engagement strategy. Such a strategy would include offering every individual in the community for whom vaccination is not contraindicated their choice of approved vaccine, in a setting and at a time that is feasible and acceptable to them, alongside material support (paid leave, free health care, including primary practitioner support) for those likely to require such support and accessible, practical, and trusted messaging and communication. The target we propose in Additional file 4 is based on this approach. A key consideration is those groups at risk of lower engagement in health interventions, such as socially and economically disadvantaged groups. Past experience with vaccination suggests that the main cause for low uptake, including in these higher risk groups, is likely to be structural barriers to engagement [59, 60]. Therefore, rather than focussing on shifting behaviour in the small minority that are strongly vaccine resistant [61], a more useful and effective approach is to address structural barriers to engagement.

As well as surveillance, critical to Australia’s response has been outbreak response capacity [62], in particular the enumeration and management of cases and contacts efficiently enough to ensure that all transmission chains are controlled once initial seeding is identified through surveillance. As the outbreak evolved, innovations such as secondary contract tracing were needed in order to keep pace with the shortened serial interval of new variants [62]. Further innovations to increase the efficiency of outbreak response, including task shifting and automation where warranted, should be developed now in order to prepare for future scenarios (see Additional file 4). Border controls and social and mobility restrictions were also critical to effective control of SARS-CoV-2, prior to scale-up of testing, surveillance, and outbreak response capacity. These measures may again be critical if a variant emerges that exhibits the characteristics of concern discussed previously. For this reason, continued investments in understanding how these measures can be implemented efficiently and effectively while minimising secondary impacts are needed.

Finally, the most important factor underpinning success in prevention, detection, and control of current and future transmission, including our capacity to open society safely, will be how well governments support all sectors of the community. Investment in applied research on the structural barriers to engagement, and how these need to be addressed, is critical to ensuring success against COVID-19.

Community-based surveillance for disease and routine screening of residents and staff in overseas quarantine facilities achieved early and comprehensive detection of new outbreaks of SARS-CoV-2 in the community in a period when local transmission had been eliminated in Australia and novel introductions through overseas arrivals were limited. Epidemiological investigation to identify transmission sources and link outbreak clusters has been greatly supported by comprehensive genomic sequencing.

Surveillance systems will need to emphasise the detection of VoCs through widespread genomic testing. Other measures to control VoCs that will need to be implemented alongside these surveillance strategies include maintaining and improving the efficiency of outbreak response and developing the capacity to rapidly engineer, manufacture, and distribute variant vaccines at scale, all underpinned by effective community partnerships. Finally, all the measures we have discussed are aimed at mitigating novel variants of concern, the main risk to the future control of COVID-19 in highly vaccinated populations. However, the only way to prevent their emergence is to support high vaccine coverage globally.