Fundamental Lack of Information in Observed Disease and Hospitalization Data

TL;DR

This paper demonstrates, using SEIR-models, that traditional disease data cannot reveal the true extent of underreporting, implying the need for alternative methods to determine actual infection rates.

Contribution

It provides a theoretical proof that underreporting levels cannot be inferred from standard disease observables alone.

Findings

Traditional observables do not reveal true attack rates.

Underreporting levels are fundamentally unidentifiable from disease data.

Alternative methods are required to estimate true infection levels.

Abstract

We present a proof based on SEIR-models that shows it is impossible to identify the level of under reporting based on traditional observables of the disease dynamics alone. This means that the true attack rate must be determined through other means.