Does having more long-term illnesses raise the risk of death?
The more long-term conditions a middle-aged or older adult had, the higher their risk of dying over about seven years, and the relative jump in risk was largest in younger adults.
Source
Relationship between multimorbidity, demographic factors and mortality: findings from the UK Biobank cohort
Study at a glance
- Design
- Cohort — Prospective population cohort: self-reported long-term conditions at recruitment (2006-2010) linked to national death records over a median 7 years, analysed with Cox models adjusted for sex, deprivation, smoking, alcohol, physical activity and BMI.
- N
- N=500769 · 500,769 UK Biobank participants with data on long-term conditions linked to mortality status; model Ns varied slightly with missing covariates.
- Population
- Middle-aged and older adults (about 37-73 years) recruited from the general population in England, Scotland and Wales
- Outcome
- All-cause mortality, with cancer and vascular mortality as cause-specific outcomes
Structured fields used in claim comparison tables when every cited study has a complete layer.
What they did
The researchers used UK Biobank, counting how many of 43 long-term conditions (such as diabetes, depression or COPD) each participant reported at recruitment. They followed participants through national death records for a median of 7 years and compared death rates across groups with 0, 1, 2, 3 or 4+ conditions, adjusting for lifestyle and deprivation. They also looked at which types and combinations of conditions mattered most and whether age, sex or socioeconomic status changed the relationship.
What they found
Risk rose step by step with each extra condition: compared with people with none, those with one condition had a hazard ratio of 1.46 and those with four or more had 2.79 for all-cause death. The gradient was steeper for vascular death (HR 3.71 for 4+ conditions) than for cancer death (HR 2.01). Cardiometabolic conditions were linked to all three outcomes. Absolute death rates were highest in older people, but the relative increase was biggest among those aged 37-49, especially men; deprivation did not change the relative effect of multimorbidity.
The limits
What it doesn't show
Conditions were self-reported at baseline and simply counted, with no information on severity, and the count gives each disease equal weight. UK Biobank volunteers are healthier, more often white British and less deprived than the UK population, so effects may be underestimated and may not generalise. As an observational study, residual confounding is likely, and it cannot show that the conditions themselves, rather than associated factors, cause the deaths.
Key terms
- Multimorbidity
- Having two or more long-term health conditions at the same time.
- Hazard ratio (HR)
- How much faster an event such as death occurs in one group compared with a reference group over follow-up; 2 means about twice the rate.
- Dose-response relationship
- A pattern in which the outcome gets steadily more likely as the amount of exposure, here the number of conditions, increases.
- Cardiometabolic conditions
- Conditions of the heart, blood vessels and metabolism, here hypertension, coronary heart disease, peripheral vascular disease, atrial fibrillation, diabetes, heart failure and stroke or TIA.
- Townsend deprivation score
- An area-based measure of socioeconomic deprivation in the UK calculated from census data for a person's postcode.
- Competing risks
- When one type of death prevents another from being observed; here cancer and vascular deaths were modelled separately, censoring the other cause.
Flashcards
0 of 10 answers reviewed
Research intelligence for this paper
See its role on concept claims, tensions it is part of, placement history, and related discoveries.
Quiz yourself
Compared with people with no long-term conditions, roughly how much higher was all-cause mortality in people with four or more?
Common questions
If older people die more often, why do the authors stress younger adults?
Absolute risk was highest in the oldest group, but compared with healthy people of the same age, having four or more conditions multiplied risk more in 37-49 year olds. That suggests middle-aged people with multimorbidity may benefit from targeted prevention.
Why check the results using hospital records?
Self-reports might be wrong. Repeating the analysis with conditions from prior hospital diagnoses gave the same pattern with even larger effects, which strengthens confidence, though hospital records miss conditions that rarely lead to admission.
Does this show deprivation doesn't matter for health?
No. Deprivation was itself a predictor of death, and the most deprived people with 4+ conditions had high absolute mortality; it simply did not change the relative effect of having more conditions.
More on Health equity