Brief Description of the Project

CARONTE aims at addressing fundamental open questions on the structure and functioning of competing death causes and their joint impact on overall mortality levels and forecasts, through the creation of innovative bridges between demography and statistics. A key lesson learned from the recent pandemic is that cause–specific mortality patterns do not propagate in isolation, but co–evolve via intricate systems of interrelated patterns that are rooted in the temporally–varying and demographically–multifaceted dependence structures underlying such causes. For example, a cause–specific death rate may decline as a result of improved treatment of a given disease or simply because other causes have grown meanwhile. Unveiling the endogenous and exogenous determinants that explain the interplay between different causes of death and their effects on overall longevity, can unlock unprecedented knowledge to learn, experiment and control current and future mortality dynamics. Despite the availability of recent studies suggesting that a full understanding of modern mortality trends necessarily requires a finer–scale analysis of cause–specific mortality, the complexity underlying the functional, compositional and discrete nature of multivariate cause–specific mortality processes still hinders progress in the field.

Motivated by the above gap, project CARONTE bridges demography and recent advancements in functional data analysis, compositional data analysis, network models and discrete choice models, to develop a unique statistical modeling framework which can learn the complex systems of graphical dependencies behind causes of death and unveil their combined effects on overall mortality. This will allow to:

  • Explain the dynamic co–evolution of causes of death patterns, while understanding how variations in the incidence of one cause, or groups of causes, affect current and future dynamics of the other causes.
  • Explain trends in overall mortality, along with their differences across countries and cohorts, also in terms of causes of death co–evolution.
  • Study and forecasting the impact of potential mortality shocks (e.g., covid–19) and health policies on the composition, dependence and dynamics of causes of death patterns.
These outputs will push forward the research frontier by creating cutting–edge statistical models and demographic theories that may be used to assist local and national institutions in allocating resources for health care and retirement schemes, by also taking into account how specific policies addressing a certain cause of death can affect, directly or indirectly, other causes of death.

To address these goals, CARONTE will analyze several databases through an interdisciplinary approach which combines the leading expertise in demography, public health and statistics of the three Universities which are part of the project, namely Bocconi University (PI: Daniele Durante), University of Padova (co–PI: Stefano Mazzuco) and University of Rome Tor Vergata (co–PI: Marco Stefanucci). Hence, the Post–Docs hired will conduct research in a vibrant environment which promotes top research, while facilitating effective interdisciplinary collaborations and cross–fertilization across data–related fields. The PI and co–PIs of CARONTE have fruitfully collaborated in the past years within the PRIN–MIUR 2017 Grant: “Unfolding the SEcrets of LongEvity: Current Trends and future prospects" [SELECT] [Start: 08/2019 — End: 08/2023]. As witnessed by the success of SELECT in terms of publications, junior recruiting and organization of national/international workshops, CARONTE can be expected to yield similarly–successful outcomes.

Team Members

[PI] Daniele Durante, AP of Statistics, Bocconi University.

[co–PI] Stefano Mazzuco, Professor of Demography, University of Padova.

[co–PI] Marco Stefanucci, AP of Statistics, University of Rome Tor Vergata

Research Progress

Publications in peer–reviewed journals and works in progress.
[1] Bondi, L., Pagano, M., Bonetti, M. (2024). The sparsity index in Poisson size-biased sampling: Algorithms for the optimal unbiased estimation from small samples. Statistics & Probability Letters, 214, 110217.
[2] Bonetti, M., Basellini, U., Nigri, A. (2024). The average uneven mortality index: Building on the ‘e-dagger’ measure of lifespan inequality. Demographic Research, 50, 1281–1300.
[3] Pavone, F., Legramanti, S., Durante, D. (2024). Learning and forecasting of age-specific period mortality via B-spline processes with locally-adaptive dynamic coefficients. The Annals of Applied Statistics, 18(3), 1965–1987. [arXiv:2209.12047]
[4] Nigri, A., Levantesi, S., Scognamiglio, S. (2024). Disaggregating death rates of age-groups using deep learning algorithms. Journal of Official Statistics, 4 (2), 262–282.
[5] Depaoli, E.G., Stefanucci, M., Mazzuco, S. (2024). Functional concurrent regression with compositional covariates and its application to the time-varying effect of causes of death on human longevity. The Annals of Applied Statistics, 18(2), 1668–1685.[arXiv:2301.06333]
[6] Castiglione, C., Romanò, G., (2025). Age-dependent analysis of mortality patterns in Italy: a network perspective via dynamic stochastic block models. Statistics for Innovation I, SIS 2025 Proceedings, 271–276.
[7] Romanò, G., Aliverti, E., Durante, D. (2025). Bayesian local clustering of age-period mortality surfaces across multiple countries. arXiv:2504.05240 [under review].
[8] Romanò, G., Castiglione, C., Durante, D. (2025). Dependent stochastic block models for age-indexed sequences of directed causes-of-death networks. arXiv:2510.01806 [under review].
[9] Bernardi, M., Canale, A., Stefanucci, M. (2025). On the degrees of freedom of some lasso procedures. arXiv:2511.21595 [under review].
[10] Pozza, F., Durante, D., Szabo, B. (2026). Skew-symmetric approximations of posterior distributions. Journal of the Royal Statistical Society, Series B (Statistical Methodology) [in press].
[11] Kannankeril Joseph, V.J., Conte Keivabu, R., Muttarak, R., Zagheni, E., Mazzuco, S. (2026). Harvesting effect and extreme temperature-related mortality in Italy. European Journal of Population, 42, 3.
[12] Romanò, G., (2026). Bayesian hierarchical modeling of array-structured demographic data. PhD Thesis, Bocconi University.
[13] Diaconu, V., Zarrulli, V., Mazzuco, S. (2026). A double decomposition of standard deviation below the modal age at death and the role of causes of death. European Journal of Population. [in press].
[14] Romanò, G., Piccarreta, R., Durante, D. (2026+). Bayesian temporal clustering of age-specific causes of death profiles across multiple countries. [in preparation].
[15] Novaro, A., Benussi, D., Aliverti, E., Scarpa, B. Mazzuco, S., (2026+). Evolution of sub-national longevity and causes of death composition using data on Italian provinces. [in preparation].
[16] Stefanucci, M., Bertarelli, G., Mazzuco, S. (2026+). Functional regression of mortality compositions: An application to Italian provinces. [in preparation].
[17] Mazzuco, S., Stefanucci, M. (2026+). Reflections on alternative approaches in causes of death data: cause-specific hazard rates or compositions by cause?. [in preparation].
[18] Stefanucci, M., Bernardi, M., Canale, A. (2026+). Interpretable functional linear regression via local sparsity promoting penalty. [in preparation].
[19] Alaimo di Loro, F., Barone, S., Stefanucci, M. (2026+). Bayesian representation of overlap group lasso. [in preparation].
[20] Bernardi, M., Canale, A., Stefanucci, M. (2026+). Convex clustering with unequal and sparse weights. [in preparation].
Presentations at national and international conferences.
[1] SPS Conference, Bocconi University, Milan, Italy, 6/2024 [Bonetti]
[2–3] European Population Conference 2024, Edinburgh, United Kingdom, 6/2024 [Mazzuco, Stefanucci]
[4] 52nd Meeting of the Italian Statistical Society (SIS), Bari, Italy, 6/2024 [Stefanucci]
[5] International Symposium on Nonparametric Statistics (ISNPS) 2024, Braga, Portugal, 6/2024 [Stefanucci]
[6] ISBA World Meeting 2024, Venezia, Italy, 7/2024 [Romanò]
[7–11] Workshop PRIN CARONTE, Padova, Italy, 7/2024 [Diaconu, Mazzuco, Nigri, Romanò, Stefanucci]
[12] International Biometric Conference, Atlanta, USA, 12/2024 [Bonetti]
[13] CFE-CMStatistics 2024, London, United Kingdom, 12/2024 [Castiglione]
[14] BAYSM 2025 (Online), 4/2025 [Romanò]
[15–16] Population Association of America Annual Meeting, Washington, USA, 4/2025 [Diaconu, Mazzuco]
[17] Early-Career Workshop on Nonparametric Statistics, Roma, Italy, 5/2025 [Romanò]
[18–19] 53rd Meeting of the Italian Statistical Society (SIS), Genova, Italy, 6/2025 [Castiglione, Stefanucci]
[20] BNP14, Los Angeles, USA, 6/2025 [Romanò]
[21] International Workshop on Functional and Operatorial Statistics, Novara, Italy, 6/2025 [Stefanucci]
[22] 30th International Population Conference, Brisbane, Australia, 7/2025 [Zarulli]
[23–30] Climbing Mortality Models II, Misurina, Italy, 8/2025 [Aliverti, Benussi, Castiglione, Gasparin, Mazzuco, Novaro, Stefanucci, Zarulli]
[31] CLADAG 2025, Naples, Italy, 9/2025 [Stefanucci]
[32] GRASPA 2025, Rome, Italy, 9/2025 [Stefanucci]
[33] 2025 World Statistics Congress, The Hague, Netherlands, 10/2025 [Bonetti]
[34] Sixth Meeting of the Multi-Cause Network, Barcelona, Spain, 10/2025 [Castiglione]
[35] CFE-CMStatistics 2025, London, United Kingdom, 12/2025 [Romanò]
[36] International Conference on Statistics and Data Analysis, Seville, Spain, 12/2025 [Stefanucci]
Meetings and workshops organized.
[1] Climbing Mortality Models II [final workshop of CARONTE]. August 27 – 29, 2025 [Misurina, BL, Italy]
[2] CARONTE workshop on advances in causes of death modelling. July 8 – 9, 2024 [University of Padova]
[3] CARONTE kick–off meeting. March 22, 2024 [online]