The unattributable burden of disease

Non-communicable diseases remain the leading cause of mortality and disability worldwide and account for most health loss in high-income countries, including Italy.1 Cardiovascular diseases, cancers, diabetes, and chronic respiratory diseases are largely driven by modifiable behavioural, metabolic, and environmental exposures.2 Characterising the distribution of these risk factors and quantifying their contribution to population health is central to modern epidemiology and public health. Over the past three decades, the Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) has provided a comprehensive framework for estimating health loss attributable to diseases, injuries, and their risk factors across countries, and over time. By systematically linking risk exposures to health outcomes, GBD analyses have informed prevention priorities and become an essential tool for understanding population health trends.3 In The Lancet Public Health, the GBD 2023 Italy Risk Factors Collaborators4 present a detailed assessment of the burden of disease attributable to risk factors in Italy between 1990 and 2023, with estimates disaggregated by sex and macro-region. The study highlights several well-established patterns. Exposure to behavioural risk factors included in the analysis has declined overall, with smoking substantially falling among males, yet remaining the leading contributor to disease burden. Among females, metabolic risks account for the largest share of attributable health loss, with particular reference to high systolic blood pressure and BMI. The analysis also reveals persistent geographical and socioeconomic disparities, with less favourable trends in several southern regions.4,5 Yet, one of the most interesting findings from the study4 is that the share of disability-adjusted life-years (DALYs) not attributable to risk factors included in the GBD framework has increased over time, exceeding 60% in Italy in 2023. As described by the authors, this unattributable burden refers to health loss that cannot be assigned to the risk factors currently modelled within the GBD comparative risk assessment framework and should be interpreted as a model-dependent residual. The observed changes in the unattributable burden might reflect, in part, real improvements in major modifiable risk factors, alongside demographic changes such as population ageing, and progress in prevention and treatment. However, they might also arise from conditions for which causal links with specific risk factors remain poorly characterised, from interactions between multiple risk factors, or from complex causal pathways that are difficult to incorporate into comparative risk assessment models. This expanding share of disease burden not captured by current risk attribution models raises a fundamental question: are we merely facing the limits of current attribution frameworks, or are we also overlooking key determinants of population health? Risk attribution has been a powerful framework for understanding disease causation and guiding prevention. Yet, as health systems become more effective at addressing well-established risks, the remaining burden of disease could increasingly reflect complex and interacting influences that are less easily accounted for by traditional risk-factor paradigms. In high-income settings such as Italy, these influences might extend beyond individual risk exposures. Social inequalities, urban environments, food systems, occupational conditions, and other structural determinants shape both exposure to risks, and vulnerability to disease across the life course. These upstream determinants rarely operate in isolation and are not easily incorporated into analytical frameworks primarily designed to attribute health loss to selected individual risk factors.6 Recognising these limitations does not diminish the value of the GBD framework. On the contrary, GBD estimates remain essential for identifying major drivers of disease burden, their trends, and highlighting opportunities for prevention. For example, tobacco use remains the single largest preventable contributor to disease burden among Italian men, underscoring the need for sustained tobacco control policies.7,8 Similarly, the growing contribution of metabolic risks points to the importance of policies that improve food environments, promote physical activity, and strengthen preventive care.8,9 Beyond describing patterns of disease burden, GBD estimates can also inform the evaluation of prevention policies, including national strategies such as the Italian National Prevention Plan. However, their usefulness for monitoring progress ultimately depends on the availability and quality of the underlying exposure data used to inform the models. These estimates rely partly on population surveys and surveillance systems monitoring behavioural risk factors, whose coverage and consistency can vary across regions. In addition, the availability of large longitudinal population cohorts remains more limited in Italy, than in some other high-income countries, which might further constrain the precision with which long-term risk–disease relationships and trends are assessed. As epidemiological methods become increasingly sophisticated, the challenge for public health research might lie not only in measuring risk factors more precisely, but also in identifying and better characterising determinants and pathways that remain insufficiently captured in current analytical frameworks. In this sense, the growing share of unattributable burden of disease invites a simple but important question: are we missing something?