The study, led by Vincent Mougin of the Karolinska Institutet together with collaborators including researchers at the Institute for Health Metrics and Evaluation, harnessed the extraordinary statistical infrastructure of the Swedish state. By linking the Total Population Register, the National Patient Register, the Cause of Death Register and the LISA labour market database, the team assembled a cohort of nearly 6.85 million unique individuals diagnosed between 2001 and 2023, contributing more than 196 million person-years of observation. That scale, and the near-complete population coverage of Scandinavian registers, allowed something no previous study had achieved: a harmonised, like-for-like comparison of income-related survival gaps across the entire spectrum of non-communicable disease.

The technical approach was as important as the scale. Rather than comparing raw survival rates — which would be hopelessly confounded, since low-income Swedes tend to be older, and differ in sex distribution and era of diagnosis — the researchers used flexible parametric survival models with regression standardisation. Restricted cubic splines captured the shape of the baseline hazard over time without forcing the data into proportional hazards assumptions, and standardisation produced marginal survival curves: estimates of what survival would look like if everyone in a cohort shared the same income level while keeping their actual age, sex and diagnosis year. Income itself was defined carefully, using five-year lagged, consumption-equivalised household disposable income ranked into quartiles within sex and birth-year strata, a technique designed to strip out short-term income shocks caused by illness itself.

The headline finding is stark in its uniformity. After standardisation, individuals in the bottom income quartile had lower ten-year survival than those in the top quartile for all 100 disease categories, and the confidence interval for the difference excluded zero in every case. For 37 conditions, the absolute gap exceeded ten percentage points. The largest per-person disparities clustered around smoking- and alcohol-related diseases: ten-year survival for other pharynx cancer was 34.2 percent among the poorest versus 58.2 percent among the richest — a 24-point chasm in a country where chemotherapy, surgery and radiotherapy are free at the point of use. Larynx cancer, lip and oral cavity cancer, chronic obstructive pulmonary disease, endocarditis and aortic aneurysm all showed gaps of more than 14 percentage points.

One especially striking pattern emerged in diseases with intermediate prognoses. For conditions where overall ten-year survival hovered between roughly 40 and 60 percent — a group dominated by cardiovascular disease — income acted almost as a switch. Ten years after an ischaemic heart disease diagnosis, the researchers note, a Swede in the top income quartile is more likely to be alive than dead, while someone in the bottom quartile is more likely to be dead than alive. The same grim arithmetic holds for hypertensive heart disease, ischaemic and haemorrhagic stroke, aortic aneurysm, prostate cancer and several others. The authors suggest this may reflect a mix of statistical floor and ceiling effects — survival differences are mechanically compressed where nearly everyone lives or nearly everyone dies — alongside the fact that these intermediate conditions are precisely where treatment quality, adherence and risk-factor management tip outcomes.

To express the gaps in a clinically intuitive metric, the team calculated ten-year restricted mean survival time, effectively the average number of years lived within a decade of diagnosis. Low-income patients with other pharynx cancer lost 1.90 years compared with high-income patients, larynx cancer patients 1.46 years, and COPD patients just over a year. But when the researchers scaled these per-person differences up to the population level, the picture inverted. Rare head and neck cancers, despite their enormous individual gaps, contributed relatively little to the national total because so few people are diagnosed. Instead, the greatest absolute burden came from humdrum, highly prevalent conditions: hypertension alone accounted for roughly 352,000 person-years of survival lost among low-income patients, with atrial fibrillation, stroke, heart failure and urinary tract infections close behind. In total, about 6.2 million person-years of survival were lost within ten years of diagnosis — the difference between what low- and middle-income patients actually experienced and a counterfactual in which they had enjoyed high-income survival trajectories.

Perhaps the most methodologically sophisticated part of the study tackled a deceptively simple question: are these gaps just the familiar income gradient in life expectancy showing through, or does income specifically change what happens after you are diagnosed? The researchers matched each diagnosed patient to five disease-free controls sharing the same birth year, sex and income quartile, then compared survival loss relative to those controls across income groups. For 73 of the 100 disease categories, the poorest patients lost significantly more survival time to their disease than the richest — a difference that persisted after netting out background mortality. This extends a logic familiar from deprivation-stratified relative survival analyses in cancer epidemiology, but applies it across cardiovascular, respiratory, neurological and mental health conditions, suggesting the pattern is not a cancer quirk but a general feature of illness in an unequal society.

Why does income still matter so much in a welfare state? The observational design cannot pin down mechanisms, but the authors sketch two non-exclusive pathways. First, lower income may worsen the course of disease after onset — through later presentation, differences in treatment, adherence to medication, management of comorbidities, and the practical costs of being ill, such as prescription charges, travel and lost earnings. Second, low-income patients may simply enter the register later, at a more advanced stage of disease, since cohort entry here marks first recorded specialist care rather than true disease onset. The breadth of the findings, spanning fast-moving cancers where early detection is everything and slow chronic diseases where long-term management dominates, suggests no single mechanism suffices. Upstream forces compound the problem: smoking and alcohol harm concentrate among the poor, financial insecurity feeds chronic stress, cheaper processed diets and low-control work add cardiovascular risk, and disadvantages accumulate within the same people across a lifetime.

The authors caution that their estimates are probably conservative. Older people living on modest pensions may be misclassified as low-income despite secure circumstances, diluting the measured association, and Sweden’s comparatively low inequality and strong safety nets likely mean the gaps would be wider in countries with higher out-of-pocket costs. Still, income inequality in Sweden has risen faster than in most OECD countries in recent decades, and the life expectancy gap between rich and poor has widened in parallel. For policymakers, the study offers a triage map: intervening on prevalent cardiovascular conditions such as hypertension and ischaemic heart disease would deliver the largest absolute reductions in survival inequality, while focused prevention and early detection for smoking- and alcohol-related cancers — including HPV-driven pharyngeal tumours — would benefit each affected person most. More broadly, the findings are a reminder that universal coverage removes the price tag on a hospital bed but not the many other ways that money buys health.

Subject of Research: Income-related survival disparities across 100 disease categories in Sweden

Article Title: Income-related differences in survival across disease categories in Sweden: a retrospective population-based register study

Article References: Mougin, V., Gray, A., Naghavi, M., Allebeck, P., Kinge, J. M., Wool, E. E., Weaver, N. D., Murray, C. J., & Agardh, E. E. (2026). Income-related differences in survival across disease categories in Sweden: a retrospective population-based register study. The Lancet Regional Health – Europe, 71, Article 101896. https://doi.org/10.1016/j.lanepe.2026.101896

Image Credits: AI Generated

DOI: 10.1016/j.lanepe.2026.101896

Keywords: health inequality, survival analysis, Sweden, income, non-communicable diseases, cardiovascular disease, cancer, COPD, universal healthcare, register study, life expectancy, epidemiology