News Release

Methods and advances in infectious disease burden estimation

Peer-Reviewed Publication

Science China Press

From signals to action: Where infectious disease risk assessment can fail—and how to bridge the gaps

image: 

The pathway extends from signal detection and verification through risk assessment to decision-making and response. Failures may arise from delayed verification, poor characterization of risk, weak linkage to action, or insufficient operational readiness. Corresponding measures include rapid verification, decision-ready and context-sensitive assessment, predefined action triggers, and tested response capacity. Timely data sharing and clear responsibilities support the pathway, while updated evidence feeds back into reassessment and adjustment of response.

view more 

Credit: ©Medicine Plus

Reliable estimates of infectious disease burden are essential for setting prevention priorities, allocating health resources and evaluating interventions. Yet routine surveillance often provides an incomplete picture. Some people do not seek care, undergo testing or have their illness recorded in reporting systems. Different pathogens can cause similar clinical conditions, while a single infection may lead to complications and lasting health consequences. Infectious disease burden is also shaped by transmission within populations and can change rapidly as population immunity, pathogen characteristics and control measures evolve. Vaccination, for example, may protect vaccinated individuals while also reducing transmission and providing indirect protection to others. These features make estimating infectious disease burden and evaluating interventions particularly challenging.

A review published in Medicine Plus, “Methods and advances in infectious disease burden estimation,” brings together the main frameworks and analytical approaches used in this field. Ying Shen, Zhaoming Feng, Quanyi Wang, and Da Huo at the Beijing Center for Disease Prevention and Control, collaborated with Abrar A. Chughtai at UNSW Sydney, Australia, to examine the data requirements, assumptions, applications and limitations of these approaches. The review also discusses developments in multi-pathogen attribution, uncertainty assessment, health economic evaluation and artificial intelligence.

The review distinguishes three components of burden estimation: the outcomes to be measured, the framework used to allocate and attribute burden, and the analytical methods applied to the data. The time perspective is an important choice. An incidence-based approach follows infections acquired during a particular period and estimates their subsequent consequences, including acute illness, complications and long-term sequelae. A prevalence-based approach estimates the health loss experienced by a population during a given period. The former is particularly relevant to assessing the consequences of preventing new infections, while the latter helps characterize current health loss and healthcare needs. Pathogen attribution can also proceed in different directions. Researchers may start with a specific pathogen and trace the health outcomes it causes, including complications and long-term sequelae. Alternatively, they may begin with a clinical outcome, such as pneumonia or diarrhoea, and estimate the proportion attributable to different pathogens.

Estimating infectious disease burden often means working with incomplete records and uncertain causes of illness. The review describes how mortality, hospitalization and syndromic indicators can be used to estimate epidemic-associated excess burden, while multiplier and capture–recapture methods help account for cases missed by surveillance. When diagnostic testing cannot reliably establish the cause of a clinical syndrome, reductions following pathogen-specific vaccination can provide evidence of vaccine-preventable burden. Evaluating interventions may also require transmission dynamic models to capture changes in infection spread and indirect protection. Bayesian approaches offer ways to bring together evidence from different sources and assess uncertainty, particularly when observations are sparse or tests are imperfect.

Pathogen detection does not establish causation. Advances in multiplex testing and metagenomics increasingly allow several pathogens to be detected during the same illness episode. However, a positive result may reflect a causal infection, colonization, prolonged shedding or an incidental finding. Treating every detected organism as a cause of illness can overestimate pathogen-specific burden. The review highlights the importance of combining case-control evidence, longitudinal observations, quantitative microbial measurements and clinical information to improve attribution.

The review emphasizes that no single method is appropriate for every infectious disease or setting. Credible estimates depend on matching methods to the research question, using suitable evidence and making assumptions and uncertainty explicit. Strengthening both analytical methods and surveillance infrastructure can help disease burden research provide more useful evidence for prevention, intervention evaluation and resource allocation.


Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.