Contact behaviour before, during and after the COVID-19 pandemic in the Netherlands: evidence from contact surveys in 2016-2017 and 2020-2023 (preprint)

Background: The first wave of the COVID-19 pandemic in 2020 was largely mitigated by reducing contacts in the general population. In 2022 most contact-reducing measures were lifted. Aim We assess whether the population has reverted to pre-pandemic contact behaviour and how this would affect the transmission potential of a newly emerging pathogen. Methods The PIENTER Corona study was held every 2-6 months in the Netherlands from April 2020, as a follow-up on the 2016-2017 PIENTER3 study. In both studies, participants (ages 1-85) reported the number and age group of all face-to-face persons contacted on the previous day. The contact behaviour during and after the COVID-19 pandemic was compared to the pre-pandemic baseline. Results We found an average of 15.2 (13.3-16.9, 95% CI) community contacts per person per day in the post-pandemic period, which is 14% lower than the baseline value of 17.6 (16.3-18.9). Children have the highest number of contacts as before the pandemic. Mainly adults aged 20-59 have not reverted to their pre-pandemic behaviour, possibly because this age group works more often from home. Although the number of contacts is structurally lower compared to the pre-pandemic period, the effect on the potential spread of a newly emerging respiratory pathogen is limited if all age groups were equally susceptible. If younger age groups were less susceptible, as observed during the first COVID-19 wave, the transmission potential as well as the required control effort would be lower. Conclusion Continuous monitoring of contacts is needed to be prepared for a future pandemic.

Social contact patterns following the COVID-19 pandemic: a snapshot of post-pandemic behaviour from the CoMix study (preprint)

The COVID-19 pandemic led to unprecedented changes in behaviour. To estimate if these persisted a final new round of the CoMix survey was conducted in four countries at a time when all societal restrictions had been lifted for several months. We conducted a survey on a nationally representative sample in the UK, Netherlands (NL), Belgium (BE), and Switzerland (CH). Participants were asked about their contacts and behaviours on the previous day. We calculated contact matrices and compared the contact levels to a pre-pandemic baseline to estimate R0. Data collection occurred from 17 November to 7 December 2022. 7,477 participants were recruited. Some were asked to undertake the survey on behalf of their children. Only 14.4% of all participants reported wearing a facemask on the previous day, varying between 6.7% in NL to 17.8% in CH. Self-reported vaccination rates in adults were similar for each country at around 86%. Trimmed mean recorded contacts were highest in NL with 9.9 (95% confidence interval [CI] 9.0 to 10.8) contacts per person per day and lowest in CH at 6.0 (95% CI 5.4 to 6.6). The number of contacts at home were similar between the countries. Contacts at work were lowest in the UK (1.4 contacts per person per day) and highest in NL at 2.8 contacts per person per day. Other contacts were also lower in the UK at 1.6 per person per day (95% CI 1.4 to 1.9) and highest in NL at 3.4 recorded per person per day (95% CI 4.0 to 4.0). Using the next-generation approach suggests that R0 for a close-contact disease would be roughly half pre-pandemic levels in the UK, 80% in NL and intermediate in the other two countries. The pandemic appears to have resulted in lasting changes in contact patterns that would be expected to have an impact on the epidemiology of many different pathogens. Further post-pandemic surveys are necessary to confirm this finding.

Projecting COVID-19 intensive care admissions in the Netherlands for policy advice: February 2020 to January 2021 (preprint)

0.IntroductionModel projections of COVID-19 incidence into the future help policy makers about decisions to implement or lift control measures. During 2020, policy makers in the Netherlands were informed on a weekly basis with short-term projections of COVID-19 intensive care unit (ICU) admissions. Here we present the model and the procedure by which it was updated. Methodsthe projections were produced using an age-structured transmission model. A consistent, incremental update procedure that integrated all new surveillance and hospital data was conducted weekly. First, up-to-date estimates for most parameter values were obtained through re-analysis of all data sources. Then, estimates were made for changes in the age-specific contact rates in response to policy changes. Finally, a piecewise constant transmission rate was estimated by fitting the model to reported daily ICU admissions, with a change point analysis guided by Akaikes Information Criterion. ResultsThe model and update procedure allowed us to make mostly accurate weekly projections, accounting for recent and future policy changes, and to adapt the estimated effectiveness of the policy changes based only on the natural accumulation of incoming data. DiscussionThe model incorporates basic epidemiological principles and most model parameters were estimated per data source. Therefore, it had potential to be adapted to a more complex epidemiological situation, as it would develop after 2020.

Effectiveness of a COVID-19 contact tracing app in a simulation model with indirect and informal contact tracing (preprint)

During the COVID-19 pandemic, contact tracing was used to identify individuals who had been in contact with a confirmed case so that these contacted individuals could be tested and quarantined to prevent further spread of the SARS-CoV-2 virus. Many countries developed mobile apps to find these contacted individuals faster. We evaluate the epidemiological effectiveness of the Dutch app CoronaMelder, where we measure effectiveness as the reduction of the reproduction number R. To this end, we use a simulation model of SARS-CoV-2 spread and contact tracing, informed by data collected during the study period (December 2020 - March 2021) in the Netherlands. We show that the tracing app caused a clear but small reduction of the reproduction number, and the magnitude of the effect was found to be robust in sensitivity analyses. The app could have been more effective if more people had used it, and if time intervals between symptom onset and reporting of contacts would have been shorter. The model used is novel as it accounts for the clustered nature of social networks and as it accounts for cases informally alerting their contacts directly after symptom onset, without involvement of health authorities or a tracing app.

Wastewater-based reproduction numbers and projections of COVID-19 cases in multiple cities in Japan, 2022 (preprint)

Background: Wastewater surveillance has expanded globally to monitor the spread of infectious diseases. An inherent challenge is substantial noise and bias in wastewater data due to their sampling and quantification process, leading to the limited applicability of wastewater surveillance as a monitoring tool and the difficulty. Aim: In this study, we present an analytical framework for capturing the growth trend of circulating infections from wastewater data and conducting scenario analyses to guide policy decisions. Methods: We developed a mathematical model for translating the observed SARS-CoV-2 viral load in wastewater into effective reproduction numbers. We used an extended Kalman filter to infer underlying transmissions by smoothing out observational noise. We also illustrated the impact of different countermeasures such as expanded vaccinations and non-pharmaceutical interventions on the projected number of cases using three study areas in Japan as an example. Results: Our analyses showed an adequate fit to the data, regardless of study area and virus quantification method, and the estimated reproduction numbers derived from wastewater data were consistent with notification-based reproduction numbers. Our projections showed that a 10-20% increase in vaccination coverage or a 10% reduction in contact rate may suffice to initiate a declining trend in study areas. Conclusion: Our study demonstrates how wastewater data can be used to track reproduction numbers and perform scenario modelling to inform policy decisions. The proposed framework complements conventional clinical surveillance, especially when reliable and timely epidemiological data are not available.

Contact patterns of frail and non-frail elderly persons in the Netherlands during the COVID-19 pandemic (preprint)

Background During the COVID-19 pandemic social distancing measures were imposed to protect the population from exposure, especially elderly and frail persons who have the highest risk for severe outcomes. These restrictions greatly reduced contacts in the general population, but little is known about behaviour changes among elderly and frail persons themselves. Our aim was to quantify how COVID-19 measures affected contact behaviour of elderly and how this differed between frail and non-frail elderly. Methods In 2021 a contact survey was carried out among persons aged 70 years and older in the Netherlands. A random sample of persons per age group (70-74, 75-79, 80-84, 85-89, 90+) and gender was invited to participate, either during a period with stringent (April 2021) or moderate (October 2021) measures. Participants provided general information on themselves including their frailty, and reported characteristics of all persons with whom they had face-to-face contact on a given day, over the course of a full week. Results In total 720 community-dwelling elderly persons were included (overall response rate of 15%), who reported 16,505 contacts. During the survey period with moderate measures, non-frail participants had significantly more contacts outside their household than frail participants. Especially for women, frailty was a more informative predictor for number of contacts than age. During the survey period with stringent measures, frail and non-frail participants had significantly lower numbers of contacts compared to the survey period with moderate measures. The reduction of number of contacts was largest for the eldest non-frail participants. As they likely interact closely with highly aged and highly frail persons, this reduction of number of contacts indirectly protects frail elderly from SARS-CoV-2 exposure. Conclusions The results of this study reveal that social distancing measures during the COVID-19 pandemic differentially affected the contact patterns of frail and non-frail elderly. The reduction of contacts may have led to direct protection of elderly persons in general but also to indirect protection of frail elderly.

Age-specific severity of SARS-CoV-2 in February 2020 - June 2021 in the Netherlands (preprint)

Background: Severity of SARS-CoV-2 infection may vary over time. Here, we estimate age-specific risks of hospitalization, ICU admission and death given infection in the Netherlands from February 2020 - June 2021. Methods: A nationwide longitudinal serology study was used to estimate numbers of infections in three epidemic periods (February 2020 - June 2020, July 2020 - February 2021, March 2021 - June 2021). We accounted for reinfections and, as vaccination started in January 2021, breakthrough infections among vaccinated persons. Severity estimates were inferred by combining numbers of infections with aligned numbers of hospitalizations and ICU admissions from a national hospital-based registry, and aligned numbers of deaths based on national excess all-cause mortality estimates. Results: In each period there was a nearly consistent pattern of accelerating, almost exponential, increase in severity of infection with age. The rate of increase with age was highest for death and lowest for hospitalization. In the first period, the overall risk of hospitalization, ICU admission and death were 1.5% (95%-confidence interval [CI] 1.3-1.8%), 0.36% (95%-CI: 0.31-0.42%) and 1.2% (95%-CI: 1.0-1.4), respectively. The risk of hospitalization was higher in the following periods, while the risk of ICU admission remained stable. The risk of death decreased over time, with a substantial drop among [≥]70-years-olds in February 2021 - June 2021. Conclusion: The accelerating increase in severity of SARS-CoV-2 with age remained intact during the first three epidemic periods in the Netherlands. The substantial drop in risk of death among elderly in the third period coincided with the introduction of COVID-19 vaccination.

Dynamics of non-household contacts during the COVID-19 pandemic in 2020 and 2021 in the Netherlands (preprint)

The COVID-19 pandemic was in 2020 and 2021 for a large part mitigated by reducing contacts in the general population. To monitor how these contacts changed over the course of the pandemic in the Netherlands, a longitudinal survey was conducted where participants reported on their at-risk contacts every two weeks, as part of the European CoMix survey. The survey included 1659 participants from April to August 2020 and 2514 participants from December 2020 to September 2021. We categorized the number of unique contacted persons excluding household members, reported per participant per day into six activity levels, defined as 0, 1, 2, 3-4, 5-9 and 10 or more reported contacts. After correcting for age, vaccination status, risk status for severe outcome of infection, and frequency of participation, activity levels increased over time, coinciding with relaxation of COVID-19 control measures.

Inferring time-varying generation time, serial interval and incubation period distributions for COVID-19 (preprint)

The generation time distribution, reflecting the time between successive infections in transmission chains, is a key epidemiological parameter for describing COVID-19 transmission dynamics. However, because exact infection times are rarely known, it is often approximated by the serial interval distribution. This approximation holds under the assumption that infectors and infectees share the same incubation period distribution, which may not always be true. We investigated incubation period and serial interval distributions in data on 2989 confirmed cases in China in January-February 2020, and developed an inferential framework to estimate the generation time distribution that accounts for variation over time due to changes in epidemiology, sampling biases and public health and social measures. We identified substantial reductions over time in the serial interval and generation time distributions. Our proposed method provides more reliable estimation of the temporal variation in the generation time distribution, improving assessment of transmission dynamics.

Inferring time-varying generation time, serial interval and incubation period distributions for COVID-19 (preprint)

Background The generation time distribution, reflecting the time between successive infections in transmission chains, is one of the fundamental epidemiological parameters for describing COVID-19 transmission dynamics. However, because exact infection times are rarely known, it is often approximated by the serial interval distribution, reflecting the time between illness onsets of infector and infectee. This approximation holds under the assumption that infectors and infectees share the same incubation period distribution, which may not always be true. Methods We analyzed data on observed incubation period and serial interval distributions in China, during January and February 2020, under different sampling approaches, and developed an inferential framework to estimate the generation time distribution that accounts for variation over time due to changes in epidemiology, sampling biases and public health and social measures. Results We analyzed data on a total of 2989 confirmed cases for COVID-19 during January 1 to February 29, 2020 in Mainland China. During the study period, the empirical forward serial interval decreased from a mean of 8.90 days to 2.68 days. The estimated mean backward incubation period of infectors increased from 3.77 days to 9.61 days, and the mean forward incubation period of infectees also increased from 5.39 days to 7.21 days. The estimated mean forward generation time decreased from 7.27 days (95% confidence interval: 6.42, 8.07) to 4.21 days (95% confidence interval: 3.70, 4.74) days by January 29. We used simulations to examine the sensitivity of our modelling approach to a number of assumptions and alternative dynamics. Conclusions The proposed method can provide more reliable estimation of the temporal variation in the generation time distribution, enabling proper assessment of transmission dynamics.

Inferring the differences in incubation-period and generation-interval distributions of the Delta and Omicron variants of SARS-CoV-2 (preprint)

Estimating the differences in the incubation-period, serial-interval, and generation-interval distributions of SARS-CoV-2 variants is critical to understanding their transmission and control. However, the impact of epidemic dynamics is often neglected in estimating the timing of infection and transmission---for example, when an epidemic is growing exponentially, a cohort of infected individuals who developed symptoms at the same time are more likely to have been infected recently. Here, we re-analyze incubation-period and serial-interval data describing transmissions of the Delta and Omicron variants from the Netherlands at the end of December 2021. Previous analysis of the same data set reported shorter mean observed incubation period (3.2 days vs 4.4 days) and serial interval (3.5 days vs 4.1 days) for the Omicron variant, but the number of infections caused by the Delta variant decreased during this period as the number of Omicron infections increased. When we account for growth-rate differences of two variants during the study period, we estimate similar mean incubation periods (3.8--4.5 days) for both variants but a shorter mean generation interval for the Omicron variant (3.0 days; 95\% CI: 2.7--3.2 days) than for the Delta variant (3.8 days; 95\% CI: 3.7--4.0 days). We further note that the differences in estimated generation intervals may be driven by the "network effect"---higher effective transmissibility of the Omicron variant can cause faster susceptible depletion among contact networks, which in turn prevents late transmission (therefore shortening realized generation intervals). Using up-to-date generation-interval distributions is critical to accurately estimating the reproduction advantage of the Omicron variant.

Dose-response modelling of endemic coronavirus and SARS-CoV-2: human challenge trials reveal the individual variation in susceptibility (preprint)

We propose a mathematical framework to analyze and interpret the outcomes of human challenge trials. We present plausible infection risks with HCoV-229E and SARS-CoV-2 over a wide range of infectious dose, and suggest ways to improve the design of future trials and to translate its outcomes to the general population.

Regional importation and asymmetric within-country spread of SARS-CoV-2 variants of concern in the Netherlands (preprint)

Variants of concern (VOCs) of SARS-CoV-2 have caused resurging waves of infections worldwide. In the Netherlands, Alpha, Beta, Gamma and Delta variants circulated widely between September 2020 and August 2021. To understand how various control measures had impacted the spread of these VOCs, we analyzed 39,844 SARS-CoV-2 genomes collected under the Dutch national surveillance program. We found that all four VOCs were introduced before targeted flight restrictions were imposed on countries where the VOCs first emerged. Importantly, foreign introductions, predominantly from other European countries, continued during these restrictions. Our findings show that flight restrictions had limited effectiveness in deterring VOC introductions due to the strength of regional land travel importation risks. We also found that the Alpha and Delta variants largely circulated more populous regions with international connections after their respective introduction before asymmetric bidirectional transmissions occurred with the rest of the country and the variant dominated infections in the Netherlands. As countries consider scaling down SARS-CoV-2 surveillance efforts in the post-crisis phase of the pandemic, our results highlight that robust surveillance in regions of early spread is important for providing timely information for variant detection and outbreak control.

Empirical evidence of transmission over a school-household network for SARS-CoV-2; exploration of transmission pairs stratified by primary and secondary school. (preprint)

Background Children play a key role in the transmission of many infectious diseases. They have many of their close social encounters at home or at school. We hypothesized that most of the transmission of respiratory infections among children occur in these two settings and that transmission patterns can be predicted by a bipartite network of schools and households. Aim and methods To confirm transmission over a school-household network, SARS-CoV-2 transmission pairs in children aged 4-17 years were analyzed by study year and primary/secondary school. Cases with symptom onset between the 1st of March 2021 and the 4th of April 2021 identified by source and contact-tracing in the Netherlands were included. In this period, primary schools were open and secondary school students attended class at least once per week. Within pairs, spatial distance between the postcodes was calculated as the Euclidean distance. Results A total of 4,059 transmission pairs were identified; 51.9% between primary schoolers; 19.6% between primary and secondary schoolers; 28.5% between secondary schoolers. Most (68.5%) of the transmission for children in the same study year occurred at school. In contrast, most of the transmission of children from different study years (64.3%) and most primary-secondary transmission (81.7%) occurred at home. The average spatial distance between infections was 1.2km (median 0.4) for primary school pairs, 1.6km (median 0) for primary-secondary school pairs and 4.1km (median 1.2) for secondary school pairs. Conclusion The results provide evidence of transmission on a bipartite school-household network. Schools play an important role in transmission within study years, and households play an important role in transmission between study years and between primary and secondary schools. Spatial distance between infections in a transmission pair reflects the smaller school catchment area of primary schools versus secondary schools. Many of these observed patterns likely hold for other respiratory pathogens.

Shorter serial intervals in SARS-CoV-2 cases with Omicron variant compared to Delta variant in the Netherlands, 13 - 19 December 2021 (preprint)

The SARS-CoV-2 Omicron variant has a growth advantage over the Delta variant, due to higher transmissibility, immune evasion, or a shorter serial interval. Using S-gene target failure (SGTF) as indication for Omicron, we identify 220 SGTF and 869 non-SGTF serial intervals in the same week. Within households, we find a mean serial interval of 3.4 days for SGTF and 3.9 days for non-SGTF cases. This suggests that the growth advantage of Omicron is partly due to a shorter serial interval.

EpiLPS: a fast and flexible Bayesian tool for near real-time estimation of the time-varying reproduction number (preprint)

In infectious disease epidemiology, the instantaneous reproduction number R ( t ) is a timevarying metric defined as the average number of secondary infections generated by individuals who are infectious at time t . It is therefore a crucial epidemiological parameter that assists public health decision makers in the management of an epidemic. We present a new Bayesian tool for robust estimation of the time-varying reproduction number. The proposed methodology smooths the epidemic curve and allows to obtain (approximate) point estimates and credible envelopes of R ( t ) by employing the renewal equation, using Bayesian P-splines coupled with Laplace approximations of the conditional posterior of the spline vector. Two alternative approaches for inference are presented: (1) an approach based on a maximum a posteriori argument for the model hyperparameters, delivering estimates of R ( t ) in only a few seconds; and (2) an approach based on a MCMC scheme with underlying Langevin dynamics for efficient sampling of the posterior target distribution. Case counts per unit of time are assumed to follow a Negative Binomial distribution to account for potential excess variability in the data that would not be captured by a classic Poisson model. Furthermore, after smoothing the epidemic curve, a “plug-in” estimate of the reproduction number can be obtained from the renewal equation yielding a closed form expression of R ( t ) as a function of the spline parameters. The approach is extremely fast and free of arbitrary smoothing assumptions. EpiLPS is applied on data of SARS-CoV-1 in Hong-Kong (2003), influenza A H1N1 (2009) in the USA and current SARS-CoV-2 pandemic (2020-2021) for Belgium, Portugal, Denmark and France. Author summary The instantaneous reproduction number R ( t ) is a key metric that provides important insights into an epidemic outbreak. We present a flexible Bayesian approach called EpiLPS (Epidemiological modeling with Laplacian-P-splines) for smooth estimation of the epidemic curve and R ( t ). Computational speed and absence of arbitrary assumptions on smoothing makes EpiLPS an interesting tool for near real-time estimation of the reproduction number. An R software package is available ( https://github.com/oswaldogressani ).

The estimated disease burden of acute COVID-19 in the Netherlands in 2020, in disability-adjusted life-years (preprint)

Background. The impact of COVID-19 on population health is recognised as being substantial, yet few studies have attempted to quantify to what extent infection causes mild or moderate symptoms only, requires hospital and/or ICU admission, results in prolonged and chronic illness, or leads to premature death. We aimed to quantify the total disease burden of acute COVID-19 in the Netherlands in 2020 using the disability-adjusted life-years (DALY) measure, and to investigate how burden varies between age-groups and occupations.Methods. Using standard methods and diverse data sources (mandatory notifications, population-level seroprevalence, hospital and ICU admissions, registered COVID-19 deaths, and the literature), we estimated years of life lost (YLL), years lived with disability, DALY and DALY per 100,000 population due to COVID-19, excluding post-acute sequelae, stratified by 5-year age-group and occupation category.Results. The total disease burden due to acute COVID-19 was 286,100 (95% CI:281,700–290,500) DALY, and the per-capita burden was 1640 (95% CI:1620–1670) DALY/100,000, of which 99.4% consisted of YLL. The per-capita burden increased steeply with age, starting from 60–64 years, with relatively little burden estimated for persons under 50 years old.Conclusions. SARS-CoV-2 infection and associated premature mortality was responsible for a considerable direct health burden in the Netherlands, despite extensive public health measures. DALY were much higher than for other high-burden infectious diseases, but lower than estimated for coronary heart disease. These findings are valuable for informing public health decision-makers regarding the expected COVID-19 health burden among population subgroups, and the possible gains from targeted preventative interventions.

The impact of vaccinating adolescents and children on COVID-19 disease outcomes (preprint)

Introduction: Despite the high COVID-19 vaccination coverage among adults, there is concern over a peak in SARS-CoV-2 infections in the coming months. To help ensure that healthcare systems are not overwhelmed in the event of a new wave of SARS-CoV-2 infections, many countries have extended vaccination to adolescents (those aged 12-17 years) and may consider further extending to children aged 5-11 years. However, there is considerable debate about whether or not to vaccinate healthy adolescents and children against SARS-CoV-2 because, while vaccination of children and adolescents may limit transmission from these groups to other, more vulnerable groups, adolescents and children themselves have limited risk of severe disease if infected and may experience adverse events from vaccination. To quantify the benefits of extending COVID-19 vaccination beyond adults we compare daily cases, hospital admissions, and intensive care (IC) admissions for vaccination in adults only, those 12 years and above, and those 5 years and above. Methods and Findings: We developed a deterministic, age-structured susceptible-exposed-infectious-recovered (SEIR) model to simulate disease outcomes (e.g., cases, hospital admissions, IC admissions) under different vaccination scenarios. The model is partitioned into 10-year age bands (0-9, 10-19, ..., 70-79, 80+) and accounts for differences in susceptibility and infectiousness by age group, seasonality in transmission rate, modes of vaccine protection (e.g., infection, transmission), and vaccine characteristics (e.g., vaccine effectiveness). Model parameters are estimated by fitting the model piecewise to daily cases from the Dutch notification database Osiris from 01 January 2020 to 22 June 2021. Forward simulations are performed from 22 June 2021 to 31 March 2022. We performed sensitivity analyses in which vaccine-induced immunity waned. We found that upon relaxation of all non-pharmaceutical control measures a large wave occurred regardless of vaccination strategy. We found overall reductions of 5.7% (4.4%, 6.9%) of cases, 2.0% (0.7%, 3.2%) of hospital admissions, and 1.7% (0.6%, 2.8%) of IC admissions when those 12 years and above were vaccinated compared to vaccinating only adults. When those 5 years and above were vaccinated we observed reductions of 8.7% (7.5%, 9.9%) of cases, 3.2% (2.0%, 4.5%) of hospital admissions, and 2.4% (1.2%, 3.5%) of IC admissions compared to vaccination in adults only. Benefits of extending vaccination were larger within the age groups included in the vaccination program extension than in other age groups. The benefits of vaccinating adolescents and children were smaller if vaccine protection against infection, hospitalization, and transmission (once infected) wanes. Discussion: Our results highlight the benefits of extending COVID-19 vaccination programs beyond adults to reduce infections and severe outcomes in adolescents and children and in the wider population. A reduction of infections in school-aged children/adolescents may have the added benefit of reducing the need for school closures during a new wave. Additional control measures may be required in future to prevent a large wave despite vaccination program extensions. While the results presented here are based on population characteristics and the COVID-19 vaccination program in The Netherlands, they may provide valuable insights for other countries who are considering COVID-19 vaccination program extensions.

Optimal vaccine allocation for COVID-19 in the Netherlands: a data-driven prioritization (preprint)

For the control of COVID-19, vaccination programmes provide a long-term solution. The amount of available vaccines is often limited, and thus it is crucial to determine the allocation strategy. While mathematical modelling approaches have been used to find an optimal distribution of vaccines, there is an excessively large number of possible schemes to be simulated. Here, we propose an algorithm to find a near-optimal allocation scheme given an intervention objective such as minimization of new infections, hospitalizations, or deaths, where multiple vaccines are available. The proposed principle for allocating vaccines is to target subgroups with the largest reduction in the outcome of interest, such as new infections, due to vaccination that fully immunizes a single individual. We express the expected impact of vaccinating each subgroup in terms of the observed incidence of infection and force of infection. The proposed approach is firstly evaluated with a simulated epidemic and then applied to the epidemiological data on COVID-19 in the Netherlands. Our results reveal how the optimal allocation depends on the objective of infection control. In the case of COVID-19, if we wish to minimize deaths, the optimal allocation strategy is not efficient for minimizing other outcomes, such as infections. In simulated epidemics, an allocation strategy optimized for an outcome outperforms other strategies such as the allocation from young to old, from old to young, and at random. Our simulations clarify that the current policy in the Netherlands (i.e., allocation from old to young) was concordant with the allocation scheme that minimizes deaths. The proposed method provides an optimal allocation scheme, given routine surveillance data that reflect ongoing transmissions. The principle of allocation is useful for providing plausible simulation scenarios for complex models, which give a more robust basis to determine intervention strategies.

Testing behaviour and positivity for SARS-CoV-2 infection: Insights from web-based participatory surveillance (preprint)

BackgroundVoluntary testing for SARS-CoV-2 infection is an integral component of an effective response to the COVID-19 pandemic. It is essential to identify populations at a high risk for infection but who are less likely to present for testing. Here, we use internet-based participatory surveillance data from the Netherlands to identify sociodemographic and household factors that are associated with a lower propensity to be tested and, if tested, with a higher risk of a positive test result.MethodsMultivariable analyses using generalised estimating equations for binomial outcomes were conducted to estimate the adjusted odds ratios of testing and of positivity associated with participant and household characteristics.ResultsBased on five months (17 November 2020 to 18 April 2021) of weekly surveys obtained from 12,026 participants, males (adjusted odds ratio for testing (ORt): 0.92; adjusted odds ratio for positivity (ORp): 1.30, age-groups <20 (ORt: 0.89; ORp: 1.27) 50-64 years (ORt: 0.94; ORp: 1.06) and 65+ years (ORt: 0.78; ORp: 1.24), diabetics (ORt: 0.97; ORp: 1.06), and sales/administrative employees (ORt: 0.93; ORp: 1.90) were distinguished as lower propensity/higher positivity factors.ConclusionsThe factors identified using this approach can help identify potential target groups for improving communication and encouraging testing among those with symptoms and thus increase the effectiveness of testing, which is essential for the response to the COVID-19 pandemic and for public health strategies in the longer term.