Skip to main content

Our take —

The study, available as a preprint and thus not yet peer-reviewed, sought to describe the COVID-19-related mortality disparity among Native Americans in the US. The study found a standardized mortality ratio of 2.77 compared to white populations, and this was even higher in some states, with South Dakota having a mortality ratio of 9.7 as compared to the white population. They found that the standardized mortality ratio was highly correlated with cthe perent of Native Americans living on reservations. The study had many limitations due to its ecological study design, including use of data collected as far back as 2014, and potential underreporting of Native American race/ethnicity. Regardless, results show a high level of disparity in Native American mortality from COVID-19 compared to other racial/ethnic populations in the US.

Study design

Cross-Sectional, Ecological

Study population and setting

The study sought to describe risk factors for COVID-19 infection and related mortality among Native American/American Indian communities in the US. COVID-19-related death counts from the National Center for Health Statistics from January 1, 2020, through January 16, 2021 were used. Midyear population estimates of 2019 were drawn from the US Census Bureau data from 10 states: Arizona, California, Oklahoma, New Mexico, Washington, New York, South Dakota, Minnesota, Utah, and Mississippi. The American Community Survey (ACS) and the Behavioral Risk Factor Surveillance System (BRFSS) were used to estimate potential risk factors that may impact transmission and mortality. For the ACS, the analysis used 2014 – 2018 data to estimate the type of health insurance, income-poverty ratio, and household living arrangements. They also extracted data on frontline worker status using data from 2018. They used the BRFSS from 2011 to 2018 to estimate smoking status and health conditions including asthma, chronic obstructive pulmonary disease (COPD), kidney disease, cancer, heart disease, diabetes, and obesity. More recent ACS or BRFSS versions were not yet available. Finally, they utilized data from MultiState, which generates a rating of open-ness during the pandemic based on state policies and capacity/industry restrictions. They categorized race as non-Latino Native American (including American Indian and Alaskan Native), non-Latino white, non-Latino Black, and Latino, and using the My Tribal Area tool, integrated 2014 – 2018 ACS estimates of Native Americans living on- vs. off-reservation. They generated standardized mortality ratios compared to the 3 other racial categories overall and by state. They disaggregated this based on reservation living status, occupation, and chronic health conditions and behavioral risk factors, generating correlation estimates for each.

Summary of Main Findings

In this study, 2,789 COVID-19-related deaths were estimated from January 1, 2020 to January 16, 2021 among Native Americans. They estimated a crude death rate of 1.63 times that among the US white population, and a standardized mortality ratio of 2.77. This was greater than the standardized mortality ratio within the Black population (1.64) and the Latino population (1.81). Stratifying by state, they found geographic differences as well, with South Dakota having the highest standardized mortality ratio at 9.7 compared to the state’s White population, while California had the lowest at 1.6 times the mortality to the state’s White population. The standardized mortality ratios for the 10 states were correlated with increasing percentages of Native Americans living on reservations (correlation = 0.8). In their sociodemographic and behavioral correlations, they found the income-poverty ratio was highly negatively correlated with the standardized mortality ratio (-0.86).

Study Strengths

The study made use of a wide range of data to describe the health disparities impacting Native Americans, an often underreported population. They disaggregated by meaningful variables indicative of structural risks of disease, such as living on a reservation which may impact access to health services, and living in multigenerational or crowded households, and having insurance. They also examined individual-level factors, such as clinical risks through COPD and diabetes.

Limitations

The study’s primary limitation was that they used many different data sources which may have different reporting guidelines and criteria. Therefore, these results paint an overall picture of Native American health and health disparities, but do not generate individual-level estimates of risk factors and are limited to standardization by age and place alone.. They also limited their analysis to individuals reporting Native American as their only race, which likely underreports the true number of Native American people in the US. This standardization does not reflect differences in the underlying clinical health between white and Native American populations likely due to differences in access to health services and clinical care, and may be biased. They also used data from prior years going as far back as 2014 which may not reflect more recent trends in disease and social factors.

Value added

This is a large study of Native American people in the US, reflecting the health disparities they face compared to white and other racial/ethnic groups.

Our take —

This ecological study on the impact of SARS-CoV-2 mitigation measures in the US found that county-level case and death daily growth rates declined following the implementation of mask mandates and increased following restaurant reopenings. However, due to several key limitations in the study design and statistical analyses, results from this study should be interpreted very cautiously.

Study design

Ecological

Study population and setting

The study used county-level data on SARS-CoV-2 cases and deaths as well as the implementation dates of mask mandates and on-premise dining reopenings to evaluate the impact of masks mandates and restaurant reopenings on subsequent SARS-CoV-2 spread and mortality in the US. Dates of policy implementation were extracted from state and county websites, while case and death data were extracted from state and county public health department websites. Primary outcomes included daily growth rates for cases and deaths, which were estimated as the change in log cases or deaths from the previous day.

Associations between mask mandates and restaurant reopenings with the primary study outcomes were estimated using weighted least squares regression, and were assessed 20 days within implementation as well as 21-40 days, 41-60 days, etc. through 100 days following implementation. Associations between mitigation measures and outcomes in the pre-implementations periods were also examined. Models included adjustment for bar closures, stay-at-home orders, gathering size limitations, daily testing rates, county fixed effects, and day.

Summary of Main Findings

This study found that mask mandates, which were implemented in 75% of US counties, were associated with lower daily case and death growth rates (0.5 and 0.7 percentage point daily decrease within 1 to 20 days after implementation, respectively; p<0.001) following their passage. While counties allowing for reopening of on-premises dining (98% of all counties) did not experience an initial rise in cases within the first 40 days, they did experience a significant increase in both the daily growth rates of cases by 41 days and deaths by 61 days after reopening (p<0.001 for both).

Study Strengths

Data used for this study were gathered directly from primary sources (i.e., local government websites).

Limitations

There were several major limitations to this study. First, there were numerous SARS-CoV-2 mitigation policies implemented over the same calendar timeframes as mask mandates and restaurant restrictions/reopenings, and it is unclear why the impacts of only two of a large number of county-level mitigation measures were assessed in this study. Relatedly, while the authors controlled for several mitigation measures that may have confounded associations between mask mandates and restaurant restrictions with daily changes in cases and deaths, it is unclear whether all mitigation measures implemented within counties were actually included in the analysis and the extent to which those mitigation measures included temporally overlapped with one another: substantial overlap (i.e., statistical collinearity) would likely impact meaningful estimation of exposure effects. Third, cases and death rates change over time due to infectious disease dynamics, which were not accounted for in the analysis. Fourth, and perhaps most critically, authors controlled for daily testing rates in their adjusted models. Testing rates are arguably a proxy for case rates, one of the two primary study outcomes, thus potentially rendering effect estimates from the case rate models uninterpretable. Lastly, the primary outcomes were based on daily reported changes in cases and deaths, which are likely subject to extreme variations due to reporting/surveillance biases.

Value added

This study contributes to a large body of ecological studies examining the impacts of SARS-CoV-2 mitigation measures on virus associated cases and deaths.

Our take —

In a study of US health departments from June to July 2020, higher COVID-19 caseloads per health department staff member correlated with poorer timeliness of COVID-19 case investigations and lower yield in the number of close contacts elicited from each COVID-19 case investigation. Given the inclusion of data from only 56 health departments and narrow data collection period, results may not reflect contact tracing performance outcomes of health departments not included in this study or at different stages of the SARS-CoV-2 pandemic.

Study design

Ecological

Study population and setting

Between June 25 and July 31, 2020, investigators calculated COVID-19 contact tracing performance for 56 CDC-funded health departments in the United States. These indicators included average caseload per investigator (ratio of confirmed or probable COVID-19 cases to number of case investigators), average contact tracing load per investigator (ratio of elicited close contacts from case investigations to number of case investigators), case investigation timeliness (proportion of confirmed or probably COVID-19 cases contacted for a case interview within 24 hours of reporting to the health department), contact tracing timeliness (proportion of close contacts elicited from case investigations notified of exposure within 24 hours), and contact tracing yield (ratio of close contacts elicited from a case investigation to number of index COVID-19 cases interviewed). Contact tracing performance metrics were compared using the Spearman correlation coefficient (r).

Summary of Main Findings

During the reporting period of approximately one month, health departments reported a median caseload per investigator of 31 (range: 1–96 cases) and a median of 29 elicited close contacts requiring follow-up per investigator (range: 0.5–200). Among cases prioritized for interview, a median of 1.15 close contacts were elicited per case, and 42/53 (79%) health departments elicited fewer than two close contacts per case. A median of 57% of COVID-19 cases were interviewed within 24 hours of initial case reporting to the health department, and a median of 55% of close contacts were notified of their exposure within 24 hours of initial interview with the index case. Twelve (25%) health departments reported reaching fewer than 32% of elicited close contacts within a 24-hour period. Investigator caseload and case investigation timeliness (correlation coefficient: –0.68) and contact tracing yield (correlation coefficient: –0.60), respectively, were inversely correlated. Correlations were similar across health departments with different staffing models (i.e., allocating different staff members to case investigations and contact tracing, using the same staff members for case investigations and contact tracing, or a mix).

Study Strengths

The study used health system capacity metrics as indicators of contact tracing performance. Indicator correlations were also compared across health departments that allocated staff to case investigations and contact tracing efforts in different ways, which supports the consistency of reported correlations across different contact tracing models.

Limitations

The study’s unit of analysis was the health department, rather than the individual case investigator or contact tracer; this prevented investigators from attributing contact tracing performance outcomes to unmeasured differences in staff member characteristics (e.g., individual caseload, training, years of experience) beyond the contextual factors (i.e., average caseload per investigator) reported in this study. Furthermore, the reported correlations in this study may be entirely spurious; these unadjusted estimates do not control or account for other factors, like SARS-CoV-2 incidence and laboratory capacity, that could confound the relationship between health department characteristics and contact tracing performance indicators. Because data from only 56 health departments were included in this study, the results may not be generalizable to health departments that did not participate. Similarly, because data were collected in a narrow timeframe (June–July, 2020), these results may not reflect contact tracing performance at different stages in the SARS-CoV-2 pandemic with varying health department staffing capacity and contact tracing demands.

Value added

This is among the first studies to correlate health workforce capacity and health systems indicators to performance outcomes of COVID-19 contact tracing efforts.

Our take —

In a large survey of over 300,000 US residents aged 13 and older between June 3 and July 27, 2020, 85% of respondents said they were “very likely” to wear a mask while grocery shopping, while only 40% were “very likely” to wear a mask while visiting family and friends. In an analysis conducted at the state level, the authors estimated that an increase of 10% in self-reported mask wearing by this measure was associated with a more than threefold increase in the odds of epidemic control, as defined by an estimated SARS-CoV-2 reproduction number below 1. Although these results support the consensus that mask use is an effective component of transmission control, they should be interpreted cautiously for multiple reasons including survey representativeness, crude measures of exposure and outcome, and likely unmeasured confounding.

Study design

Cross-sectional; Ecological

Study population and setting

This study related US state-level self-reported mask use to transmission control, as measured by an estimate of the reproduction number (Rt) of SARS-CoV-2. Mask use was assessed with a survey delivered by the SurveyMonkey.com online platform and answered by 378,207 individuals 13 years or older between June 3 and July 27, 2020. Results were analyzed as unweighted data and by weighting for age, race, sex, education, geography, and political affiliation to reflect the composition of the U.S. population. Respondents were asked how likely they were to wear a mask “while grocery shopping” or “while visiting with family or friends in their homes,” on a four-point scale ranging from “very likely to “not at all.” A binary classification of mask-wearing was created, defined as responding “very likely” to both questions. Individual mask use data were then aggregated at the state-level each week: this was the primary exposure measurement of mask community use used in subsequent models. Logistic regression models were fit to aggregated state-level weekly estimates of Rt that were dichotomized as either below or above 1 (epidemic slowing vs. epidemic growing). Models were adjusted for several possible confounding variables, including physical distancing (defined by state-level weekly time spent at home relative to a baseline period, measured with aggregated mobility data from Google), state population density, proportion of non-white respondents, and a linear time trend.

Summary of Main Findings

A high proportion (84.7%) of respondents reported that they would be very likely to wear masks at the grocery store, while only 40.2% reported they would be very likely to wear masks while visiting family and friends; 39.8% answered “very likely” to both questions. Self-reported mask use increased linearly with age and was higher among women, nonwhite people, and people with lower income. Mask use varied considerably by geographic region, and was highest on both coasts, along the southern border, and in urban areas. In multivariable logistic regression, self-reported mask use was associated with transmission control (defined as Rt <1): a 10% increase in mask use had an estimated odds ratio for epidemic control of 3.53 (95% CI: 2.03 to 6.43). Results were broadly similar, though attenuated in some instances, under other assumptions including alternative Rt estimates, an alternative definition of mask wearing, dichotomizing Rt at different thresholds, and using self-reported community contacts rather than mobility data. A separate analysis found no association between state-level mask mandates and changes in self-reported mask use.

Study Strengths

This study employed self-reported mask-wearing behavior, rather than mask policies, as the exposure of interest. The sample size was large and was weighted to match the distribution of some US demographic variables. The authors performed a range of sensitivity analyses.

Limitations

Self-reported mask-wearing behavior is subject to bias; for example, respondents may have provided answers in line with perceived social desirability, and this may have occurred differently across geographic regions. The survey was administered via a web platform, and thus respondents are more likely to have internet access than the broader U.S. population, and may have been non-representative in other ways. Moreover, those who responded to the survey may have systematically differed from those who did not respond, which is particularly concerning given that the response rate was only 11%. The measure of dichotomous mask-wearing was crude and may have ignored meaningful gradations in behavior. The outcome measure was similarly crude, and more problematic, since the time-varying reproduction number at the state level is determined by a heterogeneous array of factors, many of which relate to geographically specific transmission dynamics. It is unlikely that the potential confounding variables included in the model (including a linear time trend) adequately accounted for determinants of Rt that may also be related to self-reported mask use. Lastly, self-reported mask use may have been affected by characteristics of local epidemics that were not entirely accounted for by adjusting for prior peak Rt (e.g., test positivity rate, local hospital capacity, local social distancing policies).

Value added

Many prior ecological studies of mask-wearing effectiveness relied on mask mandates and policies; this is one of the few to measure self-reported mask use at a large scale.

Our take —

This study used sewage samples from Alameda and Marin Counties in California, USA, to demonstrate that population-level wastewater sampling can be used to identify SARS-CoV-2 variants circulating in the local population. The authors presented a methodology that can be used to generate genomic data from wastewater samples and showed that the majority of variants detected were also available in publicly available sequences from clinical patient samples. These findings suggest that wastewater sampling may be an effective tool for detecting circulating SARS-CoV-2 variants even in the absence of, or prior to, sequencing of patient-derived samples.

Study design

Ecological

Study population and setting

This study included data from 27 samples from wastewater treatment facilities from Alameda and Marin Counties in Northern California, USA. These samples, all 24-hour 1-liter composite samples of raw sewage (meaning sewage samples were collected every 15 minutes over the course of 24 hours and then combined), were collected between May 19 and July 15, 2020. The goal of the study was not only to detect SARS-CoV-2 in wastewater, but to assemble whole SARS-CoV-2 virus genomes and to compare observed variants to those found in publicly available sequences from clinical patient samples in California.

Summary of Main Findings

The authors found multiple SARS-CoV-2 genotypes in sewage known to be present in local communities. Assembled whole genomes from wastewater samples matched published genomes from clinical samples. In addition, the majority of individual variants (even those at low frequencies) detected within wastewater samples were also observed in patient-derived genomes from California, and that variants found in two or more samples were 2.3 times more likely to be observed in California or US patient-derived genomes than variants observed only once. This suggests that the variation observed in wastewater samples is representative of the SARS-CoV-2 variation circulating in the sampled population — individuals in Alameda and Marin counties, California. The authors also found evidence of new variants not yet found in patient-derived samples from California and suggested that these variants may be detected in clinical patient samples in the near future. On a technical side, the authors addressed the difficulties inherent to isolating virus in wastewater samples, and reported that both the RNA extraction method and the initial concentration of SARS-CoV-2 in the sample have a significant effect on genome recovery and coverage.

Study Strengths

This study demonstrated that wastewater samples can be used not only to detect the presence of SARS-CoV-2 in the community, but also to detect distinct genotypes circulating in the local population. The authors presented a comparison of RNA extraction and sequencing methods that may help to inform future wastewater sequencing studies, and their findings suggest that wastewater sampling may be an effective tool for surveying SARS-CoV-2 variants at a population scale.

Limitations

The authors provided some evidence that common variants occur at similar frequencies in wastewater samples but, in the absence of phylogenetic context or additional technical replicates, it is difficult to evaluate if variant frequencies in wastewater samples can be used to track the spread and abundance of particular viral lineages.

Value added

This study demonstrated that sequencing of wastewater samples can detect SARS-CoV-2 and identify multiple genotypes known to be circulating in local communities.

Our take —

A national study in Israel found that SARS-CoV-2 test positivity and incidence increased across pediatric and adult age groups following school reopenings in May 2020; increases in SARS-CoV-2 hospitalizations and deaths, however, were not observed until later in the summer, coinciding with the lifting of other restrictions on large social gatherings. This study had major limitations, including the use of aggregate national data and short time intervals in analysis, restricted to just three weeks following school reopenings.

Study design

Ecological

Study population and setting

On May 3, 2020, schools in Israel were partially reopened following nearly two months of closure (since March 14). By May 17, schools throughout Israel were completely reopened through the end of the academic year (June 19 for secondary schools and June 30 for elementary schools). While Israel imposed standard public health mitigation measures within schools upon reopening (i.e., mask use, maintaining physical distance from others), they did not employ cohort-based mitigation measures (e.g., keeping the same students and teachers together in the same classroom) in contrast to other countries. Using age-disaggregated COVID-19 incidence, hospitalization, and mortality estimates, the authors investigated the impact of partial and complete school reopenings on SARS-CoV-2 transmission dynamics, comparing changes in SARS-CoV-2 test positivity, incidence (7-day average of new cases per 100,000 persons), hospitalizations, and deaths in school-aged children/adolescents (ages 19 and under) and the general adult population (ages 20 and above) in the week before and the two to three weeks following school reopenings.

Summary of Main Findings

In the 21-27 days following school reopenings, SARS-CoV-2 test positivity increased significantly across adult age groups but not among children <10 years (Rate Ratio [before vs. after reopening]: 1.46, 95% CI: 0.85–2.51) or 10-19 years (RR 0.93, 95% CI: 0.65–1.34). Accounting for the number of SARS-CoV-2 tests performed, COVID-19 incidence increased significantly in adults and children <10 years (Incidence Rate Ratio [IRR]: 2.2, 95% CI: 1.56–3.11) and 10-19 years (IRR 1.29, 95% CI: 0.94–1.76) in the two and three weeks following school reopenings. By July 31, children 19 years and younger accounted for nearly one-third (29%) of SARS-CoV-2 infections in Israel, but this increase in SARS-CoV-2 test positivity and incidence grew most dramatically in children and adolescents (<20 years) after schools closed permanently for the academic year (June 19 and 30 for secondary and elementary schools, respectively). COVID-19 hospitalizations and deaths remained relatively stable in the 49-day period following school reopenings, increasing significantly in the 21-28 days and 28-34 days, respectively, following easing of restrictions on attending large social gatherings (on June 12) (hospitalizations: IRR 3.95, 95% CI: 3.2–4.8; mortality: IRR 4.0, 95% CI: 1.9–8.3). SARS-CoV-2 test positivity and 7-day average COVID-19 incidence increased significantly across pediatric and adult age groups after June 12, when restrictions on attending large social gatherings were eased.

Study Strengths

This study adjusted for changes in testing rates over time, which may have reduced impacts of surveillance biases introduced by changing test guidelines over calendar time. Israeli surveillance protocols following school reopenings, including testing of asymptomatic household contacts of school-attending children infected with COVID-19, likely resulted in more accurate estimates of SARS-CoV-2 burdens among school-aged children, compared to symptomatic testing only. Lastly, the study’s use of multiple SARS-CoV-2 transmission metrics, from test positivity to mortality, provided a detailed picture of the COVID-19 epidemic in the context of school reopenings.

Limitations

This study had several key limitations. First, trends in COVID-19 test positivity, incidence, hospitalizations, and deaths were only compared between the seven days before school were reopened and the two to three weeks after schools were reopened; this narrow time interval prohibited assessment of potential downstream impacts (i.e., COVID-19 hospitalizations and deaths) that could occur following multiple generations of SARS-CoV-2 transmission. Moreover, it may take several weeks or longer before outbreaks or transmission in schools to occur; if such transmission events occurred, they would not have been captured in this study. Second, because school reopenings occurred within days of other restrictions on social gatherings being lifted, the independent effect of school reopenings on SARS-CoV-2 transmission cannot be disentangled from other policy changes. Third, because children infected with SARS-CoV-2 are more likely than adults to present asymptomatically, estimates of SARS-CoV-2 infections in pediatric age groups may be underestimated. Fourth, given the author’s exclusive use of national data, the estimates generated in their analyses do not account for background subnational differences in COVID-19 incidence, hospitalizations, or deaths. Fifth, the absence of any information on control measures implemented in schools during reopenings, or how these control measures may have differed across schools, makes the results difficult to interpret or to generalize to other settings. Lastly, because schools in Israel were reopened at a time when community SARS-CoV-2 transmission had declined substantially since March 2020, the generalizability of these results may be limited to settings with low COVID-19 burdens at the time of school reopenings.

Value added

This is among the first studies globally to estimate impacts of school reopenings on national age-specific SARS-CoV-2 transmission dynamics.

Our take —

In this study, the authors compared COVID-19 trends among US counties with large universities (20,000+ students) that reopened with in-person classes, those with large universities that reopened with virtual instruction, and those without large universities. Comparing the 21-day periods before and after the start of classes, counties with in-person instruction exhibited substantially larger increases in COVID-19 incidence and SARS-CoV-2 test positivity relative to other counties. However, the use of data aggregated to the county level, inattention to COVID-19 mitigation measures implemented during the study period, and handling of university instructional format as fixed over time limit the attribution of these trends in county-level COVID-19 incidence to university re-openings alone.

Study design

Ecological

Study population and setting

The investigators aggregated US county-level COVID-19 incidence between July and September 2020. Counties were classified into three groups based on the presence or absence of large (20,000+ enrolled students) not-for-profit universities and the instructional modality (i.e., in-person or virtual classes) used in the first days of the academic year. County-level COVID-19 incidence, SARS-CoV-2 test positivity rates, and “hotspot” status (defined by both absolute numbers of new cases and recent trends in incidence) were compared across counties (1) with universities offering in-person instruction, (2) with universities offering virtual instruction, and (3) without large universities. Investigators also conducted a separate analysis, in which counties with universities offering in-person classes were matched with non-university counties based on population size, the % of urban residents, and proximity (i.e., within 500 miles of each other).

Summary of Main Findings

Counties with universities offering in-person instruction (n = 79) experienced a substantial average increase (56.2%) in COVID-19 incidence comparing the 21-day periods before and after classes started. In contrast, counties with universities offering virtual instruction (n = 22) and counties without large universities (n = 3,009) experienced COVID-19 incidence reductions of 17.9% and 5.9%, respectively, over the same time period. Mean SARS-CoV-2 test positivity followed similar patterns, with increases of 1.1% in counties with universities offering in-person classes and reductions of 1.8% and 0.6%, respectively, in counties with universities offering virtual classes and non-university counties. Although detection of COVID-19 hotspots increased across counties during the study period, the increase was substantially larger in counties with in-person university instruction (30.4%) relative to that observed in counties where universities offered virtual classes (9.1%) and non-university counties (1.5%), respectively. Results were roughly comparable in a sensitivity analysis where 68 counties with universities offering in-person classes were compared to 68 counties with non-university counties: in-person university counties were associated with an increase of 10.6 cases per 100,000 residents relative to non-university counties.

Study Strengths

The investigators used a difference-in-differences method, which can be useful in comparing groups differentially exposed to a non-randomized policy. The investigators also conducted a sensitivity analysis through which counties were matched on demographic and geographic factors, demonstrating the robustness of their findings.

Limitations

Counties without and with large universities, offering either in-person or virtual learning in summer 2020, may have differed in unmeasured ways (e.g., local mask mandates, political affiliation, presence or absence of large congregate residences) that affected COVID-19 outcomes; this likelihood limits the ability to attribute observed changes in incidence to university re-openings. Additionally, data aggregated to the county level weakens the plausibility of university re-openings causing these secular changes because the source of COVID-19 infections (i.e., university campuses or in community settings) was not ascertained. Because the investigators treated instructional format (in-person or virtual) as a static, non-time-varying measure in their analysis, effect estimates may have been biased due to misclassification of counties with universities switching instructional modalities (i.e., from in-person to virtual or vice versa) during the study period. Counties with multiple universities were assigned the instructional format of the university with the earliest start date, which may have led to misclassification. Finally, the selection of 20,000 enrolled students as the population size threshold for counties with large universities limits inferences that can be drawn about the association between university re-openings and county-level COVID-19 transmission in settings with smaller (<20,000) universities. Sensitivity analyses using smaller thresholds for classification of counties into university or non-university groups may have permitted comparison of COVID-19 incidence patterns across counties with universities of variable population sizes.

Value added

This is the first study to compare secular trends in COVID-19 incidence and transmission dynamics at the county level and to attribute these changes to university re-openings in late summer 2020.

Our take —

In this study of 10 countries during the spring of 2020, the authors found no evidence that countries implementing “more restrictive” NPIs such as stay-at-home orders and business closures reduced their daily growth in new COVID-19 cases more than Sweden and South Korea, who were categorized as having “less restrictive” NPIs. However, the analysis suffers from several methodological problems that undermine confidence in the results, and should not be regarded as providing much evidence in support of the conclusion that lockdown was ineffective.

Study design

Ecological

Study population and setting

This study estimated the association between “more restrictive” (vs. “less restrictive) non-pharmaceutical interventions (NPIs) and growth rates in new COVID-19 cases in 10 countries from February through the beginning of April 2020. The authors categorized 8 countries (England, France, Germany, Iran, Italy, the Netherlands, Spain and the United States) as having “more restrictive” NPIs (such as stay-at-home orders and business closures), and compared them to 2 countries (South Korea and Sweden) categorized as having “less restrictive” NPIs. Data at the subnational level (e.g., US states) were used in model estimation. Linear models were fit estimating the daily change in the natural log of confirmed COVID-19 cases; there were indicator variables for a total of 51 NPIs across countries, and fixed effects for each subnational unit and day of the week. Changes in case definitions were modeled with an indicator variable. Models were fit separately for each pairwise comparison (n=16) of countries with “more” and “less” restrictive NPIs; each policy coefficient in “more restrictive” countries was added together, and summed policy coefficients from “less restrictive” were subtracted to obtain an overall estimate of the marginal effect of imposing more restrictive NPIs.

Summary of Main Findings

The average daily growth rate in new cases across the 10 countries prior to NPI implementation was 0.32; the rate in South Korea was 0.25 and the rate in Sweden was 0.33. The estimated association between all combined NPIs on the growth rate was statistically significant in all countries, with the exception of Spain, ranging from -0.10 (95% CI: -0.06 to -0.15) in England to -0.33 (95% CI: -0.09 to -0.57) in South Korea. None of the 16 pairwise comparisons between countries with “more restrictive” NPIs and Sweden or South Korea resulted in a statistically significant association indicating a benefit for more restrictive NPIs.

Study Strengths

Data were analyzed at the subnational level.

Limitations

The timing of NPI implementation across countries was not randomly allocated; countries enacted policies in response to characteristics of the epidemic that were highly likely to have affected subsequent confirmed case counts. This endogeneity could have severely biased effect estimates and undermines confidence in the study results. Additionally, the sample size of countries was small (10), and did not include many countries for which subnational NPI and case count data are available. The comparison group of countries with “less restrictive” NPIs only included Sweden and South Korea, which may not serve as an appropriate counterfactual proxy. South Korea, for example, enacted a broad array of intensive interventions including a robust testing and contact tracing effort, which employed elements such as video surveillance, mobile phone location data, and credit card monitoring. Many important determinants of epidemic growth that vary across countries (e.g., population density, age structure, number of residents in congregate facilities, mask use, pre-intervention contact patterns) were not included and may have confounded estimates. Reliance on reported case counts is subject to ascertainment bias: testing, and growth in testing capacity, was not uniform across countries during the time period (e.g. South Korea). Additionally, no attention was paid to appropriate lags between policy implementation and outcomes; case growth rates are unlikely to have responded immediately to any policy implementation.

Value added

This study does not provide strong evidence in support of its conclusions.

Our take —

This study found no evidence that in-person school reopenings in the US through October of 2020 caused an increase in COVID-19 hospitalizations at the county level, especially in counties where hospitalization rates were lower to start with (below the 75th percentile). These results were consistent when the authors used several different statistical methods and data sources. In counties with higher baseline hospitalization rates, some results showed an association between more in-person reopening and subsequent COVID-19 hospitalization, but the results were inconclusive and inconsistent across different model specifications. Limitations include the crude classification of school reopening that was subject to disagreement across sources; also, hospitalization data may not have adequately represented trends among the uninsured. Nonetheless, these results provide evidence that school reopenings have not been broad drivers of increased SARS-CoV-2 transmission at the community level.

Study design

Ecological

Study population and setting

This study estimated the county-level effects of fall term school reopening on COVID-19 hospitalization in the US through October, 2020. The authors used a difference-in-differences approach, whereby trends in hospitalizations over time were compared between counties with different reopening modalities. A supplementary analysis used teacher bargaining power as an instrumental variable, as it was hypothesized to affect COVID-19 hospitalization only through its effect on school reopening. The degree of reopening in a county was defined as the proportion of students allowed to attend in-person classes at a given time, which could take on three values at the district level (0 for fully remote, 0.5 for hybrid, and 1 for fully in-person); alternately, counties with any in-person learning were compared to those without. Three sources of school reopening data were employed, separately and in combination: Burbio, a private company that collected data from the websites of 1,200 mostly large school districts (~9% of all US school districts), aggregated to the county level, and imputed results from nearby districts for the least populous counties; MCH, a private company that collected data by telephoning nearly all school districts in the US (8,283 districts were included out of >13,000); and Education Week, a trade publication that collected data on 907 of the largest school districts (~7% of all districts). The outcome was defined as hospitalization with a diagnosis of COVID-19 or COVID-19-related symptoms. The primary source of information on COVID-19 hospitalization was medical claims data from Change Healthcare, which processes 55% of all medical claims in the US; sensitivity analyses were conducted using facility-level data from the US Department of Health and Human Services. Teacher unionization data, used in the instrumental variable analysis, were obtained from the US Department of Education. Potential confounding variables that were considered included state-level transmission control policies and college/university reopenings; county-level fixed effects and time effects were also included in models.

Summary of Main Findings

Across data sources, the proportion of districts that initially opened either fully in-person or with a hybrid approach ranged from 51% to 74%. The results indicate that school reopenings during the late summer and autumn of 2020 did not increase COVID-19 hospitalizations in counties that had lower hospitalization rates from March to July 2020 before reopening (below the 75th percentile). Results were similar when using different analytical approaches (e.g., using different data sources, treating reopening as continuous or as binary with propensity score matching, using teacher bargaining power as an instrumental variable, adjustment for college/university reopenings and state non-pharmaceutical interventions). For counties with higher hospitalization rates before reopening (>36-44 per 100,000 per week), some model variants showed increased hospitalization in counties with in-person learning, but results were inconsistent across methods. Because hospitalization trends in counties with and without in-person learning were not parallel before reopening, the primary results presented were from propensity-matched models and instrumental variable analysis.

Study Strengths

The outcome variable of COVID-19 hospitalization is less subject to ascertainment bias than outcomes related to SARS-CoV-2 test positivity (due to variability in testing policy, availability, and accuracy). The authors performed multiple variants of their analysis using different data sources and statistical methods, providing several robustness checks on their results and focusing their conclusions on those results that were robust across approaches. The instrumental variable approach was a reasonable and unique approach to address possible confounding, and its limitations were discussed. There was generally careful framing of the results with sources of uncertainty highlighted.

Limitations

A primary source of uncertainty is the quality and completeness of the school reopening data. Data from Burbio included <10% of all school districts and imputed reopening policies for 25% of the student population; data from MCH included approximately 60% of all school districts. The data sources displayed generally poor agreement regarding the proportion of districts in each reopening category. The three categories of school reopening also hide considerable variation in the actual educational environment: for example, a district with crowded classrooms and few infection control procedures could be classified as “hybrid” if it offered remote options for some students, while another district with small groups of students taught in strict pods and the majority learning remotely would also receive the “hybrid” designation. Hospitalization data covered approximately half of the population; if those not included (for example, the uninsured) were affected differently by school reopenings, results would be biased. Few potential confounding variables were considered, and those omitted from models (for example, population density, mask use, or local transmission control policies) may have been associated with both reopening and COVID-19 hospitalizations. Finally, the data are ecological in nature and may mask effects on subgroups of concern (for example, staff members or racial minorities).

Value added

This is the most comprehensive study to date, in the US or elsewhere, regarding school reopening and possible effects on SARS-CoV-2 transmission and/or COVID-19 outcomes.

Our take —

While the puzzle is far from solved, this study, available as a preprint and thus not yet peer reviewed, provides evidence that the novel SARS-CoV-2 variant of concern 202012/01 is associated with substantial increased transmissibility. No evidence was found to suggest that the new variant was associated with increased hospitalization or death rates. The study also reported that additional stricter mitigation measures along with high rates of vaccination will be necessary to control transmission due to the estimated increased transmissibility of the new variant and its increasing frequency in the population. Testing and sampling biases may partially explain the observed results, and founder and super spreading events cannot be fully ruled out as explanations for increasing frequency and transmissibility of the new variant.

Study design

Ecological, Modeling/Stimulation

Study population and setting

This study reports on the frequency and transmissibility of SARS-CoV-2 variant of concern (VOC) 202012/01 in England between September 28 and December 1, 2020. The frequency and proportion of the VOC in England was estimated from routine diagnostic PCR data. Diagnostic assays fail to detect the spike glycoprotein (S-gene) in the VOC, and so samples exhibiting S-gene target failure (i.e. S gene drop-out) that are positive for other SARS-CoV-2 genes, can be used as a proxy for VOC in the absence of whole-genome sequencing, and was shown to be 97% specific for the VOC in the UK during the analysis period. After estimating the proportion of all positive samples with S gene drop-out, the authors used a two-strain age and regionally structured transmission model fit to COVID-19 hospitalization and death data from seven regions to explain observed increases in COVID-19 diagnoses and frequency of the VOC during the observation period. Four hypotheses to explain the coincident rise in overall cases and the VOC were evaluated: 1) increased infectiousness of the VOC; 2) immune escape (i.e., reinfection of individuals who had been previously infected with a different variant by the VOC); 3) increased susceptibility to the VOC among children as compared to other variants, and 4) a shorter generation time. The model was also used to assess whether there were differences in odds of hospitalization or relative risk of death due to the VOC. The authors also assessed whether changes in contact patterns over time might explain observed changes in transmission using Google mobility data and age-specific contact data from the CoMix social contact survey. Lastly, the authors projected the course of the epidemic under various control strategies, including lockdowns, school closures, and mass vaccination, through the summer.

Summary of Main Findings

The authors found that the proportion of COVID-19 cases in England due to the VOC has rapidly increased since November. The increasing proportion of VOC cases was also associated with a significant increase in the effective reproductive number, Rt. Of the four hypotheses tested to explain the coincident rise in the VOC and Rt, the observed data were most consistent with the model including increased transmissibility of the VOC relative to pre-existing variants. The authors estimated that the VOC was 56% more transmissible (95% credible interval: 50-74%) than previously observed SARS-CoV-2 variants combined. However, there was no evidence that the VOC led to higher rates of hospitalizations or deaths. Changing contact patterns did not appear to explain the rise in the VOC. Lastly, the authors found that unless schools are closed and vaccinations increased to 2 million per week, the effective reproductive number is unlikely to fall below 1 as a consequence of increased frequency and transmissibility of the VOC.

Study Strengths

This study assessed multiple alternative mechanisms for the rapid rise in COVID-19 cases along with increasing frequency of the VOC in England, including increased transmissibility of the VOC, increased, susceptibility to the VOC among children (which has major implications for school re-openings/closures), and changes in social contact patterns.  Implications of the VOC on vaccination strategies were also explored.

Limitations

Limited information was reported on testing rates or genomic sampling across NHS regions or age-groups over the observation period which may impact interpretation of results. While unlikely, this study does not rule out the possibility that founder and super spreading events with subsequent spread of the VOC to other regions explain the observed data. Of note, there were apparent increases in social contacts among persons <18 years of age immediately prior to the rise in the VOC, which was not addressed in the limitations, despite reported outbreaks among schools over the same time frame. Further, estimates of Rt from contact data appear to fit the data relatively well, except for a brief time period between/during October and November.

Value added

This is the first study to estimate the relative transmissibility of the novel SARS-CoV-2 variant, VOC 202102/01, compared to pre-existing variants and its implications for future transmission dynamics and control.