Research method
Cohort Study
10 studies
A cohort study follows people over time from an exposure or a disease-free baseline to later events — multimorbidity, incident depression, a cardiovascular subtype, cancer death. Investigators count person-time, estimate rates or hazard ratios, and adjust for measured confounders. Unlike a randomised intention-to-treat analysis, assignment to 'exposed' is not under the investigator's control, so residual confounding remains on the table even after a well-fitted Cox model.
Researchers reach for cohorts when the question is whether an exposure tracks later disease in people who were event-free at the start. It answers 'who got what, and how fast, over follow-up?' Its main limitation is causal: a healthy-lifestyle HR below 1 is not a trial of a lifestyle programme, a population attributable fraction is not a tested prevention package, and a drug–adverse-event signal can still be confounding by indication.
Studies
10
Findings
6
6 supporting · 0 challenging · 0 qualifying citations
Open tensions
2
Currently
What we know
- A multinational prospective cohort linked a healthy lifestyle index to later disease over a median 11.0 years, during which 3,244 people developed multimorbidity. Higher HLI was inversely associated with CVD (HR 0.77 per 3 units) and type 2 diabetes (HR 0.67), and more weakly with cancer (HR 0.89). Observational lifestyle scores remain open to residual confounding.
- Diet-quality scores in 15,093 Spanish adults free of depression at baseline showed a threshold-like rather than strictly linear association with new physician-diagnosed depression. For the PDP score, highest versus lowest quintile HRs were about 0.78 using baseline diet and 0.74 using updated diet. That is still an observational association, not a feeding trial.
- The 45 and Up Study accumulated 1.35 million person-years and 27,511 major CVD events. Current smoking raised risks across IHD/AMI, stroke and heart failure, and especially PAD (RR about 5); former smokers sat in between. Subtype risks are not identical, and residual confounding is still possible despite adjustment.
- Rotterdam Study analyses estimated that modifiable cardiovascular risks, related diseases and education could still account for roughly one quarter to one third of dementia cases, a potential-prevention fraction that had not declined between cohorts spanning about two decades. Those PAR estimates assume risk-factor causality and eliminability; they are not a trial of a named prevention package.
- A UK primary-care cohort of newly diagnosed depression found 87.7% received antidepressants (citalopram, fluoxetine and amitriptyline most often), and 8 of 11 common drugs were linked to significantly higher falls versus non-use over 5 years. Observational signals can reflect confounding by indication; absolute risks still need to be weighed against benefit.
Largest unresolved question
Follow-up of thousands of event-free adults is a different design from a 74-person telemedicine case series, even if both get filed as cohorts. The HLI, diet-depression, smoking-CVD and Rotterdam papers estimate rates or attributable fractions over years. The WeChat paper describes outcomes after care already delivered, with no control group, so it cannot show that monitoring caused recovery.
Procedure
How it works
The main stages of the method. Specific protocols and analysis choices vary between studies.
Start from an event-free baseline
People enter while free of the later outcome — no multimorbidity yet, no incident depression, no cardiovascular event. The clock starts at that baseline, not at diagnosis of the exposure itself.
Record an exposure the investigator does not assign
Lifestyle index, diet-quality score, smoking, antidepressant use, or another exposure is observed. Unlike a randomised trial, who is 'exposed' is not under the investigator's control.
Follow person-time to later events
Deaths, incident disease, or a subtype are counted over years of follow-up. A 74-person case series filed as a cohort is not the same design as thousands of event-free adults.
Estimate rates or hazard ratios
Cox or similar models adjust for measured confounders. Residual confounding, confounding by indication, and a population attributable fraction that assumes risks could be eliminated all remain on the table.
Do not read the HR as a trial of a programme
A cohort answers who got what, and how fast, over follow-up. A healthy-lifestyle HR below 1 is not a trial of a lifestyle programme, and a drug–adverse-event signal can still be confounding by indication.
Common misconceptions
An HR of 0.67 for type 2 diabetes means a healthy-lifestyle programme was shown to prevent diabetes.
That figure is an observational contrast per 3 units of a lifestyle index over a median 11.0 years, with 3,244 multimorbidity events in the same cohort. It is not a randomised lifestyle trial.
If about a quarter to a third of dementia is 'preventable,' a public-health package has already been proven.
Rotterdam PAR estimates assume the risks are causal and could be eliminated. They motivate action; they do not demonstrate that any specific prevention programme would avert that fraction.
If 8 of 11 common antidepressants are linked to more falls, those drugs should be treated as the established cause.
The UK cohort compared users with non-use among people with newly diagnosed depression. Confounding by indication is built into that contrast: people at higher fall risk may be more likely to receive certain drugs.
Related
Claim ledger
What the evidence shows
Drawn from 10 studies in this library. Mix labels say which citation roles are present; they are not a strength score. Supports means evidence for a finding; Challenges means evidence against a stated position; Qualifies marks scope.
A multinational prospective cohort linked a healthy lifestyle index to later disease over a median 11.0 years, during which 3,244 people developed multimorbidity. Higher HLI was inversely associated with CVD (HR 0.77 per 3 units) and type 2 diabetes (HR 0.67), and more weakly with cancer (HR 0.89). Observational lifestyle scores remain open to residual confounding.
Diet-quality scores in 15,093 Spanish adults free of depression at baseline showed a threshold-like rather than strictly linear association with new physician-diagnosed depression. For the PDP score, highest versus lowest quintile HRs were about 0.78 using baseline diet and 0.74 using updated diet. That is still an observational association, not a feeding trial.
The 45 and Up Study accumulated 1.35 million person-years and 27,511 major CVD events. Current smoking raised risks across IHD/AMI, stroke and heart failure, and especially PAD (RR about 5); former smokers sat in between. Subtype risks are not identical, and residual confounding is still possible despite adjustment.
Rotterdam Study analyses estimated that modifiable cardiovascular risks, related diseases and education could still account for roughly one quarter to one third of dementia cases, a potential-prevention fraction that had not declined between cohorts spanning about two decades. Those PAR estimates assume risk-factor causality and eliminability; they are not a trial of a named prevention package.
A UK primary-care cohort of newly diagnosed depression found 87.7% received antidepressants (citalopram, fluoxetine and amitriptyline most often), and 8 of 11 common drugs were linked to significantly higher falls versus non-use over 5 years. Observational signals can reflect confounding by indication; absolute risks still need to be weighed against benefit.
Not every paper tagged 'cohort' is a large incidence study. A WeChat telemedicine series of 74 home-quarantined COVID-19 patients hospitalised 6 people (8%) after detected deterioration; all 74 recovered. That is a small retrospective care series without a concurrent control, not a prospective exposure–disease cohort like 45 and Up or the diet-depression follow-up.
Debates
Tensions and limits
Some items are genuine disagreements on the same question. Others mark different assays, populations, or outcomes.
Follow-up of thousands of event-free adults is a different design from a 74-person telemedicine case series, even if both get filed as cohorts. The HLI, diet-depression, smoking-CVD and Rotterdam papers estimate rates or attributable fractions over years. The WeChat paper describes outcomes after care already delivered, with no control group, so it cannot show that monitoring caused recovery.
Follow-up of thousands of event-free adults is a different design from a 74-person telemedicine case series, even if both get filed as cohorts. The HLI, diet-depression, smoking-CVD and Rotterdam papers estimate rates or attributable fractions over years. The WeChat paper describes outcomes after care already delivered, with no control group, so it cannot show that monitoring caused recovery.
- Healthy lifestyle and multimorbidity risk
- Diet quality and depression risk
- How broadly does smoking raise heart and vessel disease risk?
- WeChat monitoring of home COVID patients
Study Role Design N Population Outcome Healthy lifestyle and multimorbidity risk Supports CohortEPIC multinational prospective cohort; healthy lifestyle index N=291778 · 64% women; median follow-up 11.0 years EPIC adults free of cancer, CVD, and T2D at baseline across European centres Incident cancer, CVD, T2D, and subsequent multimorbidity by lifestyle index Diet quality and depression risk Supports CohortSUN Project prospective cohort; diet scores as time-updated exposures N=15093 · Spanish adults free of depression at baseline Spanish university-graduate adults without baseline physician-diagnosed depression Incident physician-diagnosed depression (HRs by MDS, PDP, AHEI-2010 adherence) How broadly does smoking raise heart and vessel disease risk? Supports Cohort45 and Up Study; current and past vs never smokers for fatal/non-fatal CVD subtypes N=188167 · Analysis dataset 188,167 participants; 27,511 major CVD events over 1.35 million person-years Australian adults in the 45 and Up Study Hospitalisation or death from major CVD subtypes by smoking status WeChat monitoring of home COVID patients Supports OtherWeChat telemedicine monitoring programme for home-quarantined COVID-19 patients (not randomised) N=74 · 6 (8%) hospitalised after detected deterioration; all 74 recovered Home-quarantined patients with COVID-19 Detection of deterioration requiring hospital admission
PaperFren reads this as a limit on how far one study travels — different assays, populations, or outcomes — not a forced fight between papers.
Protective lifestyle HRs and harmful drug HRs are not interchangeable clinical instructions. Diet-quality HRs around 0.74–0.78 and HLI HRs of 0.67–0.77 do not prove that prescribing a diet would prevent depression or diabetes, while extra falls on named antidepressants may mark who is prescribed them rather than a pure drug effect. Norwegian cancer-survival SES gradients even became non-significant after stage and smoking were in the model.
Protective lifestyle HRs and harmful drug HRs are not interchangeable clinical instructions. Diet-quality HRs around 0.74–0.78 and HLI HRs of 0.67–0.77 do not prove that prescribing a diet would prevent depression or diabetes, while extra falls on named antidepressants may mark who is prescribed them rather than a pure drug effect. Norwegian cancer-survival SES gradients even became non-significant after stage and smoking were in the model.
- Diet quality and depression risk
- Healthy lifestyle and multimorbidity risk
- Antidepressant adverse events cohort
- Does socioeconomic status change cancer survival?
Study Role Design N Population Outcome Diet quality and depression risk Supports CohortSUN Project prospective cohort; diet scores as time-updated exposures N=15093 · Spanish adults free of depression at baseline Spanish university-graduate adults without baseline physician-diagnosed depression Incident physician-diagnosed depression (HRs by MDS, PDP, AHEI-2010 adherence) Healthy lifestyle and multimorbidity risk Supports CohortEPIC multinational prospective cohort; healthy lifestyle index N=291778 · 64% women; median follow-up 11.0 years EPIC adults free of cancer, CVD, and T2D at baseline across European centres Incident cancer, CVD, T2D, and subsequent multimorbidity by lifestyle index Antidepressant adverse events cohort Supports CohortUK primary-care database; newly diagnosed depression followed for adverse events N=238963 · Final eligible cohort; 87.7% received antidepressants during follow-up Adults aged 20–64 with newly diagnosed depression in UK primary care Adverse outcomes (including falls) by antidepressant class/drug vs non-use Does socioeconomic status change cancer survival? Supports CohortNOWAC; education and household income vs cancer survival with stage and smoking adjustment N=3849 · 3,849 incident invasive cancers with SES data from a 91,814-woman study population Norwegian women with incident primary invasive cancer in NOWAC Cancer survival by education and household income
PaperFren reads this as a limit on how far one study travels — different assays, populations, or outcomes — not a forced fight between papers.
Timeline
How understanding moved
Study years are when the paper was published. Evidence edits are dated changes to this page's claims. Explanations are when PaperFren added a Discovery — not a claim that the science happened that day.
2026
Concept page published
Cohort Study
Change log
What changed
Dated edits to this page's evidence: studies added or removed from a claim, claims added or withdrawn, and new explanations tagged here. Coverage edits are distinct from reviewed research changes and corrections. Dates record PaperFren updates; study publication dates appear in the timeline.
- Coverage updated · Concept page published
Papers
10 studies in this library bear on Cohort Study, ordered by citations.
- Night shifts and type 2 diabetes risk
Longer rotating night-shift work associated with higher type 2 diabetes risk in two large female nurse cohorts.
- COVID-19 death risk and comorbidities
In 31,461 US adults with COVID-19, older age, male sex, Black race, and several comorbidities including heart disease predicted higher mortality.
- Healthy lifestyle and multimorbidity risk
Healthier lifestyle scores tracked lower CVD and diabetes risk and lower multimorbidity transitions.
- Mediterranean diet and depression risk
Overall Mediterranean diet assignment did not significantly reduce depression; a nuts arm suggested benefit in diabetes subgroup analyses.
- How much dementia could prevention still avert?
In Rotterdam cohorts, about one quarter to one third of dementia cases were potentially attributable to modifiable risks—and that share did not shrink over two decades.
- How broadly does smoking raise heart and vessel disease risk?
In a large Australian cohort, current smoking raised risk across nearly all CVD subtypes—especially peripheral arterial disease—and quitting lowered risk.
- Diet quality and depression risk
In the SUN cohort, higher adherence to Mediterranean, pro-vegetarian, and AHEI-2010 diet scores tracked with lower risk of incident depression over about 8.5 years.
- WeChat monitoring of home COVID patients
WeChat-based telemedicine monitored 74 home-quarantined COVID-19 patients; 8% needed admission and all recovered.
Show 2 more studiesShow fewer studies
- Antidepressant adverse events cohort
In 238,963 adults aged 20–64 with depression, most used antidepressants and several common agents associated with higher fall rates versus non-use.
- Does socioeconomic status change cancer survival?
In Norwegian women, lower education and income predicted worse cancer survival until stage and smoking were accounted for.
Compare studies
Select 2–10 studies. Design and N are labels, not a ranking.
Nothing selected yet.
Questions
What is still open
Follow-up of thousands of event-free adults is a different design from a 74-person telemedicine case series, even if both get filed as cohorts. The HLI, diet-depression, smoking-CVD and Rotterdam papers estimate rates or attributable fractions over years. The WeChat paper describes outcomes after care already delivered, with no control group, so it cannot show that monitoring caused recovery.
Protective lifestyle HRs and harmful drug HRs are not interchangeable clinical instructions. Diet-quality HRs around 0.74–0.78 and HLI HRs of 0.67–0.77 do not prove that prescribing a diet would prevent depression or diabetes, while extra falls on named antidepressants may mark who is prescribed them rather than a pure drug effect. Norwegian cancer-survival SES gradients even became non-significant after stage and smoking were in the model.
Ask PaperFren about Cohort Study
Study this conceptflashcards and short-answer questions
Contrast the HLI CVD HR of 0.77 per 3 units with what a randomised depression-treatment trial would have to show. What can the cohort claim, and what can it not?
The cohort can claim that, over a median 11.0 years with 3,244 multimorbidity events, higher lifestyle scores tracked lower CVD and T2D hazards (HR 0.77 and 0.67) and a weaker cancer association (HR 0.89). It cannot claim that assigning a lifestyle programme would produce those HRs, because exposure was not randomised.
Why is a PAD relative risk of about 5 in current smokers a stronger descriptive finding than the WeChat paper's 8% hospitalisation rate as evidence that a method 'prevents deterioration'?
45 and Up compared current and former smokers with never smokers across 1.35 million person-years and 27,511 major CVD events, so the PAD RR describes an exposure–rate gradient. The WeChat series followed 74 already-infected patients with no concurrent control; all recovered, but that does not identify monitoring as the cause.
A student says socioeconomic cancer-survival inequality 'disappeared, so SES does not matter.' What did the Norwegian Women and Cancer analysis actually do?
An overall negative SES–survival gradient was present. After adjustment for stage and pre-diagnosis smoking, education and income differences became non-significant. That is a decomposition of an observational gradient, not proof that treatment is equitable or that SES never mattered — smoking and stage carried much of the association.
How should a student read PDP depression HRs of about 0.78 and 0.74 next to the antidepressant-falls cohort without turning either into a prescription?
Both are observational. Better diet-quality quintiles predicted less incident depression in 15,093 Spanish adults, with a threshold-like dose–response, but diet was not randomised. Extra falls on 8 of 11 antidepressants versus non-use may still be confounding by indication. Cohort models organise follow-up; they do not by themselves name the right clinical action.