key: cord-0720150-j8p970p8
authors: Pavelka, Martin; Van-Zandvoort, Kevin; Abbott, Sam; Sherratt, Katharine; Majdan, Marek; Jaruka, Pavol; Kraj, Marek; Flasche, Stefan; Funk, Sebastian
title: The impact of population-wide rapid antigen testing on SARS-CoV-2 prevalence in Slovakia
date: 2021-05-07
journal: Science
DOI: 10.1126/science.abf9648
sha: caeb154873dcf3faf9a7853049ac9a774e299f85
doc_id: 720150
cord_uid: j8p970p8

Slovakia conducted multiple rounds of population-wide rapid antigen testing for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in late 2020, combined with a period of additional contact restrictions. Observed prevalence decreased by 58% (95% confidence interval: 57 to 58%) within 1 week in the 45 counties that were subject to two rounds of mass testing, an estimate that remained robust when adjusting for multiple potential confounders. Adjusting for epidemic growth of 4.4% (1.1 to 6.9%) per day preceding the mass testing campaign, the estimated decrease in prevalence compared with a scenario of unmitigated growth was 70% (67 to 73%). Modeling indicated that this decrease could not be explained solely by infection control measures but required the addition of the isolation and quarantine of household members of those testing positive.

identifying as infected those who are no longer infectious (8) .

By contrast, rapid antigen tests are cheap and can be quickly produced in large quantities, offering results on site in 15 to 30 min without the need for a laboratory. They are less sensitive in detecting infections with low viral load that are less likely to transmit, but can detect over 70% of likely infectious cases. A recent observational study estimated the sensitivity of lateral flow devices in detecting infectious individuals to be as high as 83 to 91% (9) . This makes mass testing a viable part of the portfolio of nonpharmaceutical interventions (10, 11) .

In October and November 2020, Slovakia used rapid antigen tests in a campaign that targeted the whole population to identify infectious cases at scale, rapidly reduce transmission, and thus allow easing of lockdown measures (12) . A pilot took place between 23 and 25 October in the four most affected counties, followed by a round of national mass testing on 31 October and 1 November (round 1). High prevalence counties were again targeted with a subsequent round of testing on 7 and 8 November (round 2) (Fig. 1 ).

In total, 5,276,832 SD-Biosensor Standard Q rapid antigen tests were conducted by trained medical personnel during the mass testing campaigns, with 65% of the respective populations tested in the pilot, 66% in mass testing round 1 and 62% in round 2. This corresponded to 87, 83, and 84% of the age-eligible population (10 to 65 years and older adults in employment) in each round, respectively. It does not include residents who were quarantining at the time of the campaign or the 534,300 tests that were conducted on medical, military, and governmental personnel who were not included in geographical county data.

A total of 50,466 participants tested positive, indicating the presence of currently infectious SARS-CoV-2. The proportion of positive tests was 3.91% (range across counties: 3.12 to 4.84%) in the pilot, 1.01% (range: 0.13 to 3.22%) in round 1, and 0.62% (range: 0.28 to 1.65%) in round 2 (Fig. 2, C and D) .

The potential for large numbers of falsepositive tests has been a point of criticism for mass testing campaigns. Although multiple studies have found high specificity for the Biosensor test kit, they are not sufficiently powered to exclude specificity levels that at a population level would yield an overwhelming amount of false positives (13) . From the low test-positive rates in some counties, we estimate with 95% certainty that the specificity of the SD Biosensor Standard Q antigen test exceeded 99.85%, and the occurrence false positives was therefore not of major concern in this study.

The counties with the highest prevalence were found in the Northern part of the country, whereas the two main Slovakian cities of Bratislava and Košice had some of the lowest observed prevalences (Fig. 1C) . Reflecting this pattern, we found that high county-level prevalence was associated with a younger average population age and a lower population density ( fig. S8 ). Given that prevalence varied at a much smaller than county scale (14) , such associations may be clearer at the individual or community level, as observed in other countries.

In the four counties where the pilot was conducted, observed infection prevalence decreased by 56% [95% confidence interval (CI): 54 to 58%] between the pilot and round 1 of the mass testing campaign and a further 60% (95% CI: 56 to 63%) between rounds 1 and 2, totaling a decrease of 82% (95% CI: 81 to 83%) over 2 weeks. There was little heterogeneity between counties ( Fig. 2B) .

Among the 45 counties that were included in round 2 of the mass testing campaign, observed infection prevalence decreased by 58% (95% CI: 57 to 58%) in 1 week. Combining the pilot results with the ones from the two rounds of testing in 45 counties, each round of mass testing was estimated to have reduced observed infection prevalence by 56% (95% CI: 52 to 59%) when adjusted for attendance rates, reproduction number, and prevalence in previous rounds. The estimated reduction between rounds varied considerably by county, from 29% in county Považská Bystrica to 79% in county Medzilaborce, but although heterogeneous showed no regional pattern ( Fig.  2A ). Neither region, attendance rates, prevalence in round 1, nor the estimated growth rate before mass testing showed any significant impact on the observed county-specific reductions.

At the time of round 1 of the mass testing campaign, the incidence of confirmed cases reported through the syndromic surveillance system was rising in nonpilot counties, with an estimated infection growth rate of 4.4% (1.1% to 6.9%) per day. When adjusting for this growth trend, we estimated a self-adjusted prevalence ratio (saPR) of 0.30 (0.27 to 0.33). In the pilot counties, reported infection incidence showed signs of leveling in the week before the mass testing campaign, with an estimated infection growth rate of 1.3% (-7.4 to 7.8%), yielding a respective saPR of 0.31 (0.26 to 0.33).

Because we used the test positivity rate of the subsequent round to estimate the impact of the previous one, we were unable to observe the impact of the last round in each county and hence the full effect of the campaign.

However, we found that the reduction achieved per round of testing was 56% (52 to 59%), indicating that the 41 counties with two rounds of testing likely reduced infection prevalence by 81% (77 to 83%) within 2 weeks and that the four counties included into the pilot testing reduced infection prevalence by 91% (89 to 93%) within 3 weeks.

The observational nature of this study made it difficult to separate the effects of the mass (Top) Description of timing and extent of national contact restriction in Slovakia (color intensity indicates intensity of the measures) and timing and extent of the mass testing campaigns. Open circles and lines in respective colors indicate the start and duration of the contact restrictions, and the blue solid circles indicate the days on which mass testing was conducted, although the highest turnout was usually on the first day. (Left) Box illustrating contact-reducing measures for those testing positive and those who chose not to be tested. (Bottom) SARS-CoV-2 infection incidence as reported by the Slovak Ministry of Health and collected through passive symptom-triggered PCR testing. Using the same color coding as at the top, contact interventions are indicated by horizontal lines, and mass testing campaigns are indicated by vertical lines. Data from the passive surveillance subsequent to the respective first mass testing campaign are omitted to clearly illustrate the trends in infection rates that led up to the mass testing and because mass testing is likely to have changed the sensitivity of the passive surveillance, thereby distorting the observation of infection trends that followed mass testing.

testing campaigns from that of the other nonpharmaceutical interventions introduced over the same period that aimed to reduce contacts and mobility, although much less than during the spring lockdown ( fig. S4 ). Nevertheless, a greater than 50% decline in infection prevalence within 1 week (or 80% in 2 weeks) is noteworthy, particularly while primary schools and workplaces were mostly open. For comparison, a month-long lockdown in November in the UK resulted in just a 30% decrease in prevalence (15) . This, alongside the inability in December to control the rebounding spread of SARS-CoV-2 in Slovakia through even more stringent contact restrictions, indicates that the mass testing campaigns were responsible for a large share of case reduction in the previous months.

To further investigate the relationship between the reduction in prevalence, mass testing, and nonpharmaceutical interventions, we used a microsimulation model for finescale SARS-CoV-2 transmission in a repre-sentative county included in the pilot phase of the mass testing. Among the multiple intervention scenarios tested, only the scenario that assumed a substantial impact of both the additional contact reducing measures and the mass testing campaigns was able to generate reductions in test positivity rates between testing rounds that were similar to those observed (Fig. 3) . Thus, the requirement for quarantine for the whole household after a positive test was essential for the combined effect of mass testing and contact Pavelka observed from the pilot mass testing round to either the first (green) or the second (orange) national round and from the first to the second mass testing round (blue) in the four counties that were included in the pilot. The confidence intervals were estimated using a normal approximation (Wald interval). (C and D) County-level test positivity in the (C) first and (D) second round of mass testing. Gray areas indicate counties that were not part of the second round because their test positivity rate was less than 7 per 1000 and hence have no estimates.

reduction measures. The model predicted a prevalence ratio between the first two testing rounds of 0.30 (0.26 to 0.34) with household quarantine and 0.78 (0.72 to 0.84) without household quarantine.

Despite a reduction of more than 50% in test positivity between mass testing campaigns, standard syndromic surveillance did not report a rapid collapse in test-positive cases corresponding to drastic reductions in prevalence. This may be explained by a variety of reasons. Foremost, the national mass testing campaigns are likely to have a major disruptive effect on routine passive syndromic surveillance. Also, the ability of PCR to detect viral RNA well Pavelka continued on next page beyond the infectious period will partially mask a sudden drop in infectious cases. In addition, starting mid-September incidence surveillance has been operating at capacity with long waiting lists for testing and stricter eligibility criteria, which reduced substantially in the period after mass testing and hence may have artificially reduced the observable change in these data. By contrast, data on hospital bed occupancy shows a sudden flattening from mid-November, indicating a sharp decrease in new admissions that is consistent with a sizable reduction in new infections when the mass testing campaigns occurred ( fig. S6 ). Executing a large-scale mass testing campaign comes with several challenges. The need to mobilize sufficient medical personnel to conduct the nasopharyngeal swabs proved to be a major obstacle. Also, the logistics of mobilizing large numbers of assisting army personnel and vast amounts of testing and personal protective equipment (PPE) material proved challenging. Some of the challenges could be overcome by using other rapid antigen tests with similarly high sensitivity but that are also licensed for use with nasal swabs (16, 17) . Nasal swabs can be self-administered and reduce demand on trained personnel and transmission risk in the process of sample collection and can even enable testing at home. Self-administered swabs are also less intrusive and can be better suited for children and mass testing at schools. However, these benefits must be weighed against the potential loss of sensitivity if self-administered swabs are not conducted appropriately (18) . The details of the Slovak mass testing experience need to be studied carefully before considering potential replication elsewhere (19) .

The combination of nationwide restrictions and mass testing with quarantining of household contacts of test positives rapidly reduced the prevalence of infectious residents in Slovakia. Although it was impossible to disentangle the precise contribution of control measures and mass testing, the latter is likely to have had a substantial effect in curbing the pandemic in Slovakia and may provide a valuable tool in future containment of SARS-CoV-2 elsewhere. 6 of 7 

Guidelines for the implementation of non-pharmaceutical interventions against COVID-19

Social Impact of COVID-19-Dros

Socio-economic impact of COVID-19. UNDP; www

An observational study of SARS-CoV-2 infectivity by viral load and demographic factors and the utility lateral flow devices to prevent transmission

Som Zodpovedný; www.somzodpovedny.sk

Ag Rapid Test Device

How Slovakia tested 3.6 million people for COVID-19 in a single weekend

R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing

Data and R code accompanying the manuscript "The impact of population-wide rapid antigen testing on SARS-CoV-2 prevalence in Slovakia

Further, we thank all participants who contributed their time to help curb the pandemic and particularly those who had to quarantine as a result of their or their household's or contact's test result. Funding: This work was supported by a Sir Henry Dale Fellowship jointly funded by the Wellcome Trust and the Royal Society

); and Elrha's Research for Health in Humanitarian Crises (R2HC) Programme funded by the UK Government (DFID), the Wellcome Trust, and the UK National Institute for Health Research (NIHR) (K.V.-Z.). The following funding sources are acknowledged as providing funding for the working group authors

Bill & Melinda Gates Foundation INV-001754

NTD Modelling Consortium OPP1184344

Trust Epidemic Preparedness Coronavirus research program 221303/Z/20/Z: (to C.A.B.P.); EDCTP2 RIA2020EF-2983-CSIGN (to H

The European Union's Horizon 2020 research and innovation program-project EpiPose 101003688

The Global Challenges Research Fund (GCRF) project "RECAP" managed through RCUK and ESRC ES/P010873/1 (to A

MRC MR/N013638/1

The National Institute for Health Research NIHR 16/136/46 (to B

Health Protection Research Unit for Modelling Methodology HPRU-2012-10096 (to

Understanding the dynamics and drivers of the COVID-19 epidemic using real-time outbreak analytics

MR/P014658/1

The UK Public Health Rapid Support Team funded by the United Kingdom Department of Health and Social Care (to

); 206471/Z/17/Z (to O

The following authors were part of the Inštitút Zdravotných Analýz. Each contributed in processing, cleaning, and interpretation of data

Competing interests: M.P. is employed as epidemiologist and public health data analyst at the Slovak Ministry of Health but had no involvement in the design of the mass testing campaigns. M.K. is the current Slovak Minister of Health and was

All other authors declare that they have no conflicts of interest. Data and materials availability: Daily incidence of positive COVID-19 test reports and the results of the mass testing are available through governmental websites (12, 20)

CMMID COVID-19 Working Group Authors List Inštitút Zdravotných Analýz Authors List References

We thank all health care workers, Slovak armed forces, and countless volunteers who helped with the execution of the