Simulation-based inference of the time-dependent reproduction number from temporally aggregated and under-reported disease incidence time series data

Philosophical Transactions of the Royal Society A Mathematical Physical and Engineering Sciences, Journal Year: 2025, Volume and Issue: 383(2293)

Published: April 2, 2025

During infectious disease outbreaks, the time-dependent reproduction number ( R t ) can be estimated to monitor pathogen transmission. In previous work, we developed a simulation-based method for estimating from temporally aggregated incidence data (e.g. weekly case reports). While that approach is straightforward use, it assumes implicitly all cases are reported and computation slow when applied large datasets. this article, extend our develop computationally efficient in real-time accounting both temporal aggregation of under-reporting (with fixed reporting probability per case). Using simulated data, show failing consider stochastic lead inappropriately precise estimates, including scenarios which true value lies outside inferred credible intervals more often than expected. We then apply 2018 2020 Ebola outbreak Democratic Republic Congo (DRC), again exploring effects under-reporting. Finally, how extended account variations reporting. Given information about level reporting, framework used estimate during future outbreaks with under-reported data. This article part theme issue ‘Uncertainty quantification healthcare biological systems (Part 2)’.

Language: Английский

---

estimateR: an R package to estimate and monitor the effective reproductive number

BMC Bioinformatics, Journal Year: 2023, Volume and Issue: 24(1)

Published: Aug. 11, 2023

Abstract Background Accurate estimation of the effective reproductive number ( $$R_e$$ Re ) epidemic outbreaks is central relevance to public health policy and decision making. We present estimateR, an R package for through time from delayed observations infection events. Such include confirmed cases, hospitalizations or deaths. The implements methodology Huisman et al. but modularizes procedure allow easy implementation new alternatives currently available methods. Users can tailor their analyses according particular use case by choosing among implemented options. Results estimateR allows users estimate outbreak based on observed hospitalization, death any other type event documenting past infections, in a fast timely fashion. validated with simulation study: yielded estimates comparable alternative publicly methods while being around two orders magnitude faster. then applied empirical case-confirmation incidence data COVID-19 nine countries dengue fever Brazil; parallel, already (i) SARS-CoV-2 measurements wastewater (ii) study influenza transmission clinical studies. In summary, this provides flexible various diseases datasets. Conclusions modular extendable tool designed surveillance retrospective investigation. It extends method developed makes it variety pathogens, scenarios, observation types. Estimates obtained be interpreted directly used inform more complex models (e.g. forecasting) value .

Language: Английский

---

Estimating the epidemic reproduction number from temporally aggregated incidence data: A statistical modelling approach and software tool

PLoS Computational Biology, Journal Year: 2023, Volume and Issue: 19(8), P. e1011439 - e1011439

Published: Aug. 28, 2023

The time-varying reproduction number (Rt) is an important measure of epidemic transmissibility that directly informs policy decisions and the optimisation control measures. EpiEstim a widely used opensource software tool uses case incidence serial interval (SI, time between symptoms in their infector) to estimate Rt real-time. SI distribution must be provided at same temporal resolution, which can limit applicability other similar methods, e.g. for contexts where window reporting longer than mean SI. In R package, we implement expectation-maximisation algorithm reconstruct daily from temporally aggregated data, then estimated. We assess validity our method using extensive simulation study apply it COVID-19 influenza data. For all datasets, influence intra-weekly variability reported data was mitigated by weekly estimated on sliding windows reconstructed strongly correlated with estimates original revealed well scenarios regardless aggregation presence weekend effects, were more successful recovering true value those obtained These results show this novel allows successfully recovered simple approach very few requirements. Additionally, removing administrative noise when are reconstructed, accuracy improved.

Language: Английский

---

Nowcasting and forecasting the 2022 U.S. mpox outbreak: Support for public health decision making and lessons learned

Epidemics, Journal Year: 2024, Volume and Issue: 47, P. 100755 - 100755

Published: March 2, 2024

In June of 2022, the U.S. Centers for Disease Control and Prevention (CDC) Mpox Response wanted timely answers to important epidemiological questions which can now be answered more effectively through infectious disease modeling. Infectious models have shown valuable tools decision making during outbreaks; however, model complexity often makes communicating results limitations makers difficult. We performed nowcasting forecasting 2022 mpox outbreak in United States using R package EpiNow2. generated nowcasts/forecasts at national level, by Census region, jurisdictions reporting greatest number cases. Modeling were shared situational awareness within CDC publicly on website. retrospectively evaluated forecast predictions four key phases (early, exponential growth, peak, decline) three metrics, weighted interval score, mean absolute error, prediction coverage. compared performance EpiNow2 with a naïve Bayesian generalized linear (GLM). The had less probabilistic error than GLM every phase except early phase. share our experiences an existing tool nowcasting/forecasting highlight areas improvement development future tools. also reflect lessons learned regarding data quality issues adapting modeling different audiences.

Language: Английский

---

EpiLPS: A fast and flexible Bayesian tool for estimation of the time-varying reproduction number

PLoS Computational Biology, Journal Year: 2022, Volume and Issue: 18(10), P. e1010618 - e1010618

Published: Oct. 10, 2022

In infectious disease epidemiology, the instantaneous reproduction number R t is a time-varying parameter defined as average of secondary infections generated by an infected individual at time t . It therefore crucial epidemiological statistic that assists public health decision makers in management epidemic. We present new Bayesian tool (EpiLPS) for robust estimation number. The proposed methodology smooths epidemic curve and allows to obtain (approximate) point estimates credible intervals id="M2">

Language: Английский

---

Differences between the true reproduction number and the apparent reproduction number of an epidemic time series

Epidemics, Journal Year: 2024, Volume and Issue: 46, P. 100742 - 100742

Published: Jan. 15, 2024

The time-varying reproduction number R(t) measures the of new infections per infectious individual and is closely correlated with time series infection incidence by definition. timings actual are rarely known, analysis epidemics usually relies on data for other outcomes such as symptom onset. A common implicit assumption, when estimating from an epidemic series, that has same relationship these downstream it does incidence. However, this assumption unlikely to be valid given most not perfect proxies Rather they represent convolutions uncertain delay distributions. Here we define apparent number, R

Language: Английский

---

The Role of Seasonal Influenza in Compounding the Outbreak of Infectious Diseases: A Critical Review

Biomedical & Pharmacology Journal, Journal Year: 2024, Volume and Issue: 17(1), P. 1 - 13

Published: March 20, 2024

Infectious diseases continue to pose a persistent threat public health globally. Amidst the array of factors contributing complexity infectious disease outbreaks, role seasonal influenza stands out as significant amplifier. Seasonal influenza, commonly known flu, not only inflicts its burden on communities but also plays crucial in compounding spread and impact other diseases. This review delves into various ways which contributes outbreaks. The outbreak is multifaceted challenge that demands attention from authorities worldwide. Addressing this effect requires holistic approach encompasses vaccination campaigns, strengthened healthcare infrastructure, improved diagnostic capabilities. By understanding mitigating can enhance their resilience responsiveness face evolving threats. Recognizing these dynamics essential for designing effective strategies. implementing comprehensive programs, improving capabilities, enhancing overall preparedness, better navigate complexities outbreaks exacerbated by presence influenza.

Language: Английский

---

Time-varying reproductive number estimation for practical application in structured populations

Epidemiologic Methods, Journal Year: 2025, Volume and Issue: 14(1)

Published: Jan. 1, 2025

Abstract Objectives EpiEstim is a popular statistical framework designed to produce real-time estimates of the time-varying reproductive number, R t ${\mathcal{R}}_{t}$ . However, methods in have not been tested small, non-randomly mixing populations determine if resulting ̂ ${\hat{\mathcal{R}}}_{t}$ are temporally biased. Thus, we evaluate temporal performance when population structure present, and then demonstrate how recover accuracy using an approximation with Methods Following real-world example COVID-19 outbreak small university town, generate simulated case report data from two-population mechanistic model explicit generation interval distribution expression compute true To quantify bias, compare time points estimated fall below critical threshold 1. Results When present but accounted for prematurely incidence aggregated over weeks at later point than daily data, however, does further affect timing differences between data. Last, show it possible correct by lagging subpopulation estimate total Conclusions key parameter used epidemic response. Since can bias near 1, should be prudently applied structured populations.

Language: Английский

---

Time-varying reproduction number estimation: fusing compartmental models with generalized additive models

Journal of The Royal Society Interface, Journal Year: 2025, Volume and Issue: 22(222)

Published: Jan. 1, 2025

The reproduction number, the mean number of secondary cases infected by each primary case, gives an indication effort required to control disease. Beyond well-known basic there are two natural extensions, namely and effective numbers. As behaviour, population immunity viral characteristics can change with time, these numbers vary over time. Real-world data be complex, so in this work we consider a generalized additive model smooth surveillance through explicit incorporation day-of-the-week effects, provide simple measure time-varying growth rate associated data. Converting resulting spline into estimator for both requires assumptions on structure, which here assume compartmental model. calculated based simulated real-world data, compared estimates from already existing tool. derived method estimating is effective, efficient comparable other methods. It provides useful alternative approach, included as part toolbox models, that particularly apt at smoothing out effects surveillance.

Language: Английский

---

Evaluating a novel reproduction number estimation method: a comparative analysis

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: Feb. 13, 2025

This paper presents practical methodologies for determining effective reproduction numbers, R(t), providing valuable insights researchers and public health officials. It proposes multiple simplified approaches estimating R(t) of infectious diseases compares their effectiveness. These include methods based on exponential, fixed (delta), normal, gamma distributions the generation time. The exponential time offer convenience as they rely solely mean number new infections. However, are sensitive to variance distribution: method may underestimate when is small, while overestimate large. normal distribution also risks underestimation, depending growth rate. In contrast, demonstrates greater robustness accuracy across a variety scenarios. A key contribution this work consolidated presentation these estimation methods, along with novel derivation an accurate formula distribution. research offers guidance selecting most appropriate method, emphasizing importance accounting specific characteristics disease's

Language: Английский

---