A Flexible Statistical Framework for Estimating Excess Mortality
Open accessPreprint — published as 10.1097/ede.0000000000001445
Abstract
Quantifying the impact of natural disasters or epidemics is critical for guiding policy decisions and interventions. When the effects of an event are long-lasting and difficult to detect in the short term, the accumulated effects can be devastating. Mortality is one of the most reliably measured health outcomes, partly due to its unambiguous definition. As a result, excess mortality estimates are an increasingly effective approach for quantifying the effect of an event. However, the fact that indirect effects are often characterized by small, but enduring, increases in mortality rates present a statistical challenge. This is compounded by sources of variability introduced by demographic changes, secular trends, seasonal and day of the week effects, and natural variation. Here we present a model that accounts for these sources of variability and characterizes concerning increases in mortality rates with smooth functions of time that provide statistical power. The model permits discontinuities in the smooth functions to model sudden increases due to direct effects. We implement a flexible estimation approach that permits both surveillance of concerning increases in mortality rates and careful characterization of the effect of a past event. We demonstrate our tools’ utility by estimating excess mortality after hurricanes in the United States and Puerto Rico. We use Hurricane Maria as a case study to show appealing properties that are unique to our method compared to current approaches. Finally, we show the flexibility of our approach by detecting and quantifying the 2014 Chikungunya outbreak in Puerto Rico and the COVID-19 pandemic in the United States. We make our tools available through the excessmort R package available from https://cran.r-project.org/web/packages/excessmort/ .
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The over-dispersed Poisson model used to estimate expected deaths has been validated in simulation and when using historical data.
“The model has been validated in simulation and when using historical data.23”
From Excess deaths associated with covid-19 pandemic in 2020: age and sex disaggregated time series analysis in 29 high income countries (2021) · cites this paper as [23]
Estimates of expected deaths should ideally account for recent temporal trends in mortality that are likely to have occurred in the absence of the pandemic.
“However, estimates of expected deaths should ideally account for recent temporal trends in mortality that are likely to have occurred in the absence of the pandemic.”
From Excess deaths associated with covid-19 pandemic in 2020: age and sex disaggregated time series analysis in 29 high income countries (2021) · cites this paper as [23]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
(2005) · Springer eBooks · doi:10.1007/0-387-27255-0_9 · Cited by 20,333 · see citing works
Ziegel, E. (2002) · Technometrics · doi:10.1198/004017002320256422 · Cited by 19,149 · cited 1×
“The uncertainty associated with count data, such as daily deaths totals, are typically accounted for with Poisson models [12, 13].”
BackgroundInternational Statistical Classification of Diseases and Related Health Problems
Eisfeld, J. (2014) · TSQ Transgender Studies Quarterly · doi:10.1215/23289252-2399740 · Cited by 3,570 · see citing works
Gill, J. (2001) · doi:10.1198/tech.2001.s641 · Cited by 2,963 · cited 1×
“The uncertainty associated with count data, such as daily deaths totals, are typically accounted for with Poisson models [12, 13].”
BackgroundKarlinsky, A., & Kobak, D. (2021) · eLife · doi:10.7554/elife.69336 · Cited by 772 · see citing works
Reconstruction of New Orleans after Hurricane Katrina: A research perspective
Kates, R. W., Colten, C. E., Laska, S. B., & Leatherman, S. P. (2006) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.0605726103 · Cited by 731 · cited 1× · see citing works
“5 billion investment was made to construct stronger levees [28, 29, 30].”
BackgroundMortality in Puerto Rico after Hurricane Maria
Kishore, N., Marqués, D., Mahmud, A. S., Kiang, M. V., Rodriguez, I., Fuller, A., Ebner, P. J., Sorensen, C. J., Racy, F. D. C. J., Lemery, J. M., et al. (2018) · New England Journal of Medicine · doi:10.1056/nejmsa1803972 · Cited by 706 · cited 1× · see citing works
“On May 29, 2018, three days after a study was published describing a survey of 3,299 households and reporting a much higher estimate of the death count received worldwide media coverage [8], the government finally made the data public and acknowledged the possibility of a higher excess death count of 1,427 [9] based o…”
BackgroundSelf-citationWeinberger, D. M., Chen, J., Cohen, T., Crawford, F. W., Mostashari, F., Olson, D., Pitzer, V. E., Reich, N. G., Russi, M., Simonsen, L., et al. (2020) · JAMA Internal Medicine · doi:10.1001/jamainternmed.2020.3391 · Cited by 485 · see citing works
Foundations of Linear and Generalized Linear Models
Agresti, A. (2015) · CERN Document Server (European Organization for Nuclear Research) · Cited by 473 · see citing works
Islam, N., Shkolnikov, V. M., Acosta, R. J., Klimkin, I., Kawachi, I., Irizarry, R. A., Alicandro, G., Khunti, K., Yates, T., Jdanov, D. A., et al. (2021) · BMJ · doi:10.1136/bmj.n1137 · Cited by 468 · see citing works
Self-citationSocioeconomic status determines COVID-19 incidence and related mortality in Santiago, Chile
Mena, G. E., Martinez, P. P., Mahmud, A. S., Marquet, P. A., Buckee, C. O., & Santillana, M. (2021) · Science · doi:10.1126/science.abg5298 · Cited by 431 · see citing works
A Statistical Algorithm for the Early Detection of Outbreaks of Infectious Disease
Farrington, C. P., Andrews, N. J., Beale, A. D., & Catchpole, M. A. (1996) · Journal of the Royal Statistical Society. Series A (Statistics in Society) · doi:10.2307/2983331 · Cited by 417 · cited 2× · see citing works
“1 Model-based approach provides informative parameterizations Extending the work developed for monitoring outbreaks [14, 15, 16, 17] we modeled death counts with the following mixed model Yt | εt ∼ Poisson {μt[ 1 + f(t) ]εt} for t = 1, .”
MethodRossen, L. M., Branum, A. M., Ahmad, F. B., Sutton, P., & Anderson, R. N. (2020) · MMWR Morbidity and Mortality Weekly Report · doi:10.15585/mmwr.mm6942e2 · Cited by 373 · see citing works
Mortality Displacement of Heat-Related Deaths
Hajat, S., Armstrong, B., Gouveia, N., & Wilkinson, P. (2005) · Epidemiology · doi:10.1097/01.ede.0000164559.41092.2a · Cited by 370 · cited 1× · see citing works
“However, we also noted a decrease in cumulative deaths for several months after the February 20 consistent with a harvesting effect [21, 22].”
BackgroundDushoff, J. G., Plotkin, J. B., Viboud, C. G., Earn, D. J. D., & Simonsen, L. (2005) · American Journal of Epidemiology · doi:10.1093/aje/kwj024 · Cited by 257 · see citing works
Santos‐Burgoa, C., Sandberg, J., Suarez, E. L., Goldman-Hawes, A., Zeger, S. L., Garcia-Meza, A., Pérez, C. M., Estrada‐Merly, N., Colón‐Ramos, U., Nazario, C. M., et al. (2018) · The Lancet Planetary Health · doi:10.1016/s2542-5196(18)30209-2 · Cited by 191 · cited 1× · see citing works
“On August 2018 the official death toll was revised to 2,975 based on a study commissioned by the government of Puerto Rico [10].”
BackgroundAn improved algorithm for outbreak detection in multiple surveillance systems
Noufaily, A., Enki, D. G., Farrington, P., Garthwaite, P. H., Andrews, N. J., & Charlett, A. (2012) · Statistics in Medicine · doi:10.1002/sim.5595 · Cited by 179 · cited 2× · see citing works
“1 Model-based approach provides informative parameterizations Extending the work developed for monitoring outbreaks [14, 15, 16, 17] we modeled death counts with the following mixed model Yt | εt ∼ Poisson {μt[ 1 + f(t) ]εt} for t = 1, .”
MethodThe World Mortality Dataset: Tracking excess mortality across countries during the COVID-19 pandemicPreprint
Karlinsky, A., & Kobak, D. (2021) · medRxiv · doi:10.1101/2021.01.27.21250604 · Cited by 161 · see citing works
Monitoring Count Time Series in R : Aberration Detection in Public Health Surveillance
Salmon, M., Schumacher, D., & Höhle, M. (2016) · Journal of Statistical Software · doi:10.18637/jss.v070.i10 · Cited by 126 · cited 2× · see citing works
“1 Model-based approach provides informative parameterizations Extending the work developed for monitoring outbreaks [14, 15, 16, 17] we modeled death counts with the following mixed model Yt | εt ∼ Poisson {μt[ 1 + f(t) ]εt} for t = 1, .”
MethodInfectious diseases that pose specific challenges after natural disasters: a review.
Ligon, B. (2006) · Seminars in Pediatric Infectious Diseases · doi:10.1053/j.spid.2006.01.002 · Cited by 124 · cited 1× · see citing works
“Because deaths due to bacterial infections have been reported to increase after natural disasters in locations with poor public health infrastructure [26, 27], we defined an index based on ICD10 codes between A00 and A79 as bacterial infections.”
MethodSantos‐Lozada, A. R., & Howard, J. T. (2018) · JAMA · doi:10.1001/jama.2018.10929 · Cited by 103 · see citing works
Causes of Excess Deaths in Puerto Rico After Hurricane Maria: A Time-Series Estimation
Cruz‐Cano, R., & Mead, E. L. (2019) · American Journal of Public Health · doi:10.2105/ajph.2019.305015 · Cited by 91 · see citing works
Count data regression charts for the monitoring of surveillance time series
Höhle, M., & Paul, M. (2008) · Computational Statistics & Data Analysis · doi:10.1016/j.csda.2008.02.015 · Cited by 82 · cited 2× · see citing works
“1 Model-based approach provides informative parameterizations Extending the work developed for monitoring outbreaks [14, 15, 16, 17] we modeled death counts with the following mixed model Yt | εt ∼ Poisson {μt[ 1 + f(t) ]εt} for t = 1, .”
Method10th Anniversary Review: Natural disasters and their long-term impacts on the health of communities
Cook, A. G., Watson, J., Buynder, P. G. V., Robertson, A. G., & Weinstein, P. (2008) · Journal of Environmental Monitoring · doi:10.1039/b713256p · Cited by 61 · cited 1× · see citing works
“Because deaths due to bacterial infections have been reported to increase after natural disasters in locations with poor public health infrastructure [26, 27], we defined an index based on ICD10 codes between A00 and A79 as bacterial infections.”
MethodSharp, T. M., Ryff, K., Alvarado, L. I., et al. (2016) · Journal of Infectious Diseases · doi:10.1093/infdis/jiw245 · Cited by 44 · cited 1× · see citing works
“These period coincide with the 2014-2015 Chikungunya outbreak [18, 19] (Figure 1B).”
BackgroundExcess Mortality and Causes Associated with Chikungunya, Puerto Rico, 2014–2015
Freitas, A. R. R., Donalísio, M. R., & Alarcón‐Elbal, P. M. (2018) · Emerging infectious diseases · doi:10.3201/eid2412.170639 · Cited by 42 · see citing works
Hsu, C. H., Cruz-López, F., Torres, D. V., et al. (2019) · PLoS Neglected Tropical Diseases · doi:10.1371/journal.pntd.0007084 · Cited by 35 · cited 1× · see citing works
“These period coincide with the 2014-2015 Chikungunya outbreak [18, 19] (Figure 1B).”
BackgroundModeling excess deaths after a natural disaster with application to Hurricane Maria
Rivera, R., & Rolke, W. (2019) · Statistics in Medicine · doi:10.1002/sim.8314 · Cited by 35 · see citing works
The Causes and Consequences of Puerto Rico's Declining Population
Abel, J. R., & Deitz, R. (2014) · RePEc: Research Papers in Economics · Cited by 29 · see citing works
Sandberg, J., Santos‐Burgoa, C., Roess, A., et al. (2019) · Epidemiology · doi:10.1097/ede.0000000000000970 · Cited by 28 · see citing works
Estimating the death toll of Hurricane Maria
Rivera, R., & Rolke, W. (2018) · Significance · doi:10.1111/j.1740-9713.2018.01102.x · Cited by 26 · see citing works
Flagged first, then by citations
Cited by
Sorted by citations
Wu, X., Nethery, R. C., Sabath, M. B., Braun, D., & Dominici, F. (2020) · Science Advances · doi:10.1126/sciadv.abd4049 · Cited by 1,976 · see citing works
“Analyses using county excess deaths as the outcome (23)”
MethodIslam, N., Shkolnikov, V. M., Acosta, R. J., Klimkin, I., Kawachi, I., Irizarry, R. A., Alicandro, G., Khunti, K., Yates, T., Jdanov, D. A., et al. (2021) · BMJ · doi:10.1136/bmj.n1137 · Cited by 468 · see citing works
Self-citationSocioeconomic status determines COVID-19 incidence and related mortality in Santiago, Chile
Mena, G. E., Martinez, P. P., Mahmud, A. S., Marquet, P. A., Buckee, C. O., & Santillana, M. (2021) · Science · doi:10.1126/science.abg5298 · Cited by 431 · see citing works
“Excess deaths, defined as the difference between observed and expected deaths, can provide a measure of the actual impact of the pandemic in mortality by quantifying direct and indirect deaths related to COVID-19 (7, 8, 16).”
BackgroundMeasuring the Labor Market at the Onset of the COVID-19 CrisisPreprint
Bartik, A. W., Bertrand, M., Lin, F., et al. (2020) · SSRN Electronic Journal · doi:10.2139/ssrn.3633053 · Cited by 135 · see citing works
Mandated and Voluntary Social Distancing during the COVID-19 Epidemic
Gupta, S., Simon, K. I., & Wing, C. (2020) · Brookings Papers on Economic Activity · doi:10.1353/eca.2020.0011 · Cited by 64 · see citing works
Excess all-cause mortality in the USA and Europe during the COVID-19 pandemic, 2020 and 2021
Rossen, L. M., Nørgaard, S. K., Sutton, P. D., Krause, T. G., Ahmad, F. B., Vestergaard, L. S., Mølbak, K., Anderson, R. N., & Nielsen, J. O. (2022) · Scientific Reports · doi:10.1038/s41598-022-21844-7 · Cited by 46 · see citing works
Soares, G. H., Jamieson, L., Biazevic, M. G. H., & Michel‐Crosato, E. (2021) · Journal of Racial and Ethnic Health Disparities · doi:10.1007/s40615-021-01162-w · Cited by 27 · see citing works
“This study advances the scientific literature by applying a robust statistical approach that estimates the expected number of deaths for a given period while accounting for a recent demographic change in the Indigenous population from Brazil [24].”
BackgroundAcosta, R. J., Patnaik, B., Buckee, C., et al. (2021) · medRxiv · doi:10.1101/2021.08.22.21262432 · Cited by 6 · see citing works
Self-citationIslam, N., López, F. J. G., Jdanov, D. A., Royo‐Bordonada, M. Á., Khunti, K. K., Lewington, S., Lacey, B., White, M. J., Morris, E. J. A., & Zunzunegui, M. V. (2021) · medRxiv · doi:10.1101/2021.11.29.21266617 · Cited by 3 · see citing works
No evidence of substantial underreporting of COVID-19 deaths in Taiwan during 2020
Chen, E. Y., Sachs, M. C., & Dickman, P. W. (2021) · Journal of the Formosan Medical Association · doi:10.1016/j.jfma.2021.03.018 · Cited by 3 · see citing works
“com/ rafalab/excessmort).(3) The statistical analyses were done in R, version 4.”
An ensemble modeling approach for estimating expected patterns in mortality in the United States
Rossen, L. M., Cisewski, J. A., Ahmad, F. B., & Anderson, R. N. (2026) · Epidemiologic Methods · doi:10.1515/em-2025-0035
“Excess mortality statistics have been used to assess the severity of seasonal influenza epidemics, the impact of natural disasters such as hurricanes and heat waves and have been widely used to estimate the impact of the COVID-19 pandemic worldwide [1–13].”
- Authors
- Acosta, R. J., & Irizarry, R. A.
- Affiliations
- Harvard University (US)
- Corresponding author
- Irizarry, Rafael A.
- Venue
- Epidemiology
- Publisher
- openRxiv
- Published
- 2020-06-08
- Volume
- 33
- Issue
- 3
- Pages
- 346-353
- ISSN
- 1044-3983 (ISSN-L)
- PMC
- PMC10200579
- OpenAlex
- W4225809697
- Cited by
- 52
- References
- 18
- Citations by year
- 2026: 7 · 2025: 9 · 2024: 12 · 2023: 13 · 2022: 5 · 2021: 5 · 2020: 1
- FWCI
- 7.04
- Type
- journal article
- Language
- en
- Open access
- green
- doi.org
- Licence
- cc-by
- Locations
- Epidemiology (publishedVersion)
- medRxiv (acceptedVersion, cc-by)
- PubMed (publishedVersion)
- PubMed Central (acceptedVersion)
- Funders
- Centers for Disease Control and Prevention
- National Institute of General Medical Sciences (R35GM131802)
- National Institute of Environmental Health Sciences (T32ES007142)
- Topics
- Health Sciences › Medicine › Epidemiology › Data-Driven Disease Surveillance
- Physical Sciences › Environmental Science › Health, Toxicology and Mutagenesis › Climate Change and Health Impacts
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods in Epidemiology
- Keywords
- excess mortality, natural disasters, hurricanes, Hurricane Maria, chikungunya outbreak, COVID-19 pandemic, secular trends, mortality surveillance
- MeSH terms
- COVID-19, Humans, Puerto Rico, United States, Cyclonic Storms, Pandemics
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
