Infection age model links epidemic spread to insurer reserve forecasting

The COVID-19 pandemic exposed a fundamental weakness in how health insurers prepare for the future. Actuaries conventionally set aside reserves by examining past claims, assuming that individual illnesses are largely independent of one another and that claim patterns change slowly. During an epidemic, none of these assumptions hold: one person’s infection drives the next, case numbers can surge within weeks, and the cost of treating a patient climbs steeply as the illness moves from initial diagnosis to hospitalisation and intensive care. A new study published in the KeAi journal Risk Sciences proposes a framework designed to close this gap between epidemiology and actuarial practice.

The research, conducted by a team from Tsinghua University, Meta and Middle Tennessee State University, introduces “infection age” — the time elapsed since a person became infected — as a central variable that shapes both how infectious an individual is and how much their care is likely to cost. “Policyholders with different infection ages will produce different medical claims,” explains corresponding author Runhuan Feng, a professor at Tsinghua University’s School of Economics and Management. “The concept of infection age introduces heterogeneity in individual infectivity into the population, and at a micro level this is highly useful for actuaries performing loss reserving.”

Methodologically, the team borrowed a familiar tool from property and casualty insurance: the loss triangle, a table that organises claims by when they occurred and how they develop over time. By building an “infection triangle” that records how disease unfolds across cohorts of newly infected policyholders, and a companion “payment triangle” that converts this into projected medical costs, the researchers created a structure that aligns the epidemiological timeline with the actuarial one. This alignment makes it possible to estimate the healthcare liabilities of individuals who are already infected but still require ongoing treatment — a quantity the authors term “infected but not removed”.

A key finding of the analysis is that epidemiological quantities, in particular the effective reproduction number, directly shape how reserves evolve, whether premiums are adequate, and what the ultimate cost ratio will be. Reserve levels can rise rapidly during the growth phase of an outbreak and then stabilise or decline as infections fall, a pattern that differs markedly from traditional insurance products. The study also shows that late-stage, severe cases requiring hospitalisation or intensive care dominate total liabilities, even though early-stage cases account for much of the claim frequency.

Using a numerical illustration inspired by COVID-19 dynamics, the authors demonstrate how the framework can be used for stress testing, allowing insurers to see how changes in transmission intensity or treatment severity feed through to reserve requirements. “Our work establishes a formal linkage between epidemiological modelling and financial risk quantification,” says Feng. “Sudden shifts in the reproduction number, which are already monitored closely by public health surveillance systems, could trigger parametric catastrophe insurance or signal the need for reinsurance.”

The authors hope the approach gives insurers, public health institutions and regulators a systematic way to bring real-time outbreak data into reserve estimation and strengthen the financial resilience of healthcare systems during global health emergencies. They caution, however, that the current illustration uses synthetic data and a simplified setting, and that future work should calibrate the framework to real-world epidemic surveillance and insurance claims records as such information becomes available. The study is a computational modelling exercise rather than an analysis of actual claims, so its practical performance in live insurance portfolios remains to be demonstrated.

Subject of Research: Actuarial modelling of epidemic healthcare costs using infection age

Article Title: Tracking "infection age" helps health insurers forecast the cost of an epidemic

Article References: Tracking "infection age" helps health insurers forecast the cost of an epidemic. Original publication (Bibliographic details unavailable.)

Image Credits: AI Generated

DOI: Not provided

Keywords: actuarial science, epidemic modelling, infection age, health insurance, loss reserving, reproduction number, COVID-19, risk sciences, stress testing, healthcare liabilities, Tracking, infection