Research method
Time-Series Analysis
4 studies
Time-series analysis relates day-to-day (or season-to-season) changes in an exposure series to changes in a count or index series — hospital admissions, emergency-department visits, or satellite greenness — after controlling for weather, season and other time-varying confounders. The usual unit is a city-day, not a person. A Poisson or similar model can estimate a percent change per 10 μg/m³; it does not identify who was exposed or prove a single causal mechanism on that day.
Air-pollution epidemiologists use time series when they need acute, population-level associations that exploit daily variation rather than long-term address contrasts. The design answers 'on higher-pollution days, do counts of this diagnosis rise after lags and weather adjustment?' Its main limitation is that correlated gases, particle composition, dust-storm flags and lag choices can all move the estimate, and a city-day coefficient is not a personal exposure or a demonstrated biological pathway.
Studies
4
Findings
4
4 supporting · 0 challenging · 0 qualifying citations
Open tensions
2
Currently
What we know
- A 10-year (1995–2004) Nicosia series linked daily hospital admissions to urban PM₁₀ and ozone, separating meteorologically confirmed dust-storm days. Each 10 μg/m³ same-day PM₁₀ was linked to a 0.85% rise in all-cause admissions; lag-1 ozone raised cardiovascular admissions 2.91% per 10 ppb. On 63 dust-storm days, all-cause admissions were 4.8% higher. Respiratory admissions showed no overall PM₁₀ association.
- A seasonal Poisson Medicare analysis (2000–2003) related emergency admissions to PM₂.₅, then meta-regressed effects on particle species-to-mass ratios. Per 10 μg/m³: +1.89% cardiovascular, +2.25% MI and +2.07% respiratory admissions. Composition modified toxicity, so mass alone was incomplete. The series does not include people under 65.
- Pooled daily ED visits from 14 hospitals in seven Canadian cities linked CO, NO₂, O₃, SO₂, PM₁₀ and PM₂.₅ at lags 0–2 days (and 3-hour averages) to cardiac and respiratory diagnoses. CO and NO₂ showed the most consistent cardiac associations (warm-season CO: +5.2% MI/angina per 0.7 ppm). Ozone tracked asthma and COPD (COPD +6.2% per 18.4 ppb in the warm season). Warm-season PM₁₀ and PM₂.₅ were strongly linked to asthma visits. Three-hour same-day averages showed no consistent signal. Edmonton supplied ~70% of visits.
- The same method label also covers satellite NDVI time series. Random-forest maps trained on >300,000 classified orthophoto pixels plus NDVI found 2018 early wilting over about 21,524 km² (10.8% of Central European forest). Precipitation (negative) and temperature (positive) anomalies ranked among the top drivers. Wilted patches stayed less green in spring 2019 (NDVI gap +0.015 versus unaffected forest). The models explained 35–42% of deviance.
Largest unresolved question
City hospital series do not agree on which pollutant or lag carries the acute signal. Nicosia finds same-day PM₁₀ for all-cause admissions (0.85% per 10 μg/m³) and lag-1 ozone for cardiovascular admissions (2.91% per 10 ppb) but no overall PM₁₀–respiratory link. Canadian EDs find CO/NO₂ most consistent for cardiac diagnoses and ozone for asthma/COPD, with 3-hour averages null and Edmonton dominating the sample. Medicare finds PM₂.₅ mass plus composition for CVD, MI and respiratory admissions among those 65+. Treating 'time series show pollution raises admissions' as one coefficient hides those splits.
Common misconceptions
A same-day percent increase in admissions means pollution caused those visits through a known day-level biological mechanism.
The estimate is a city-day association after weather adjustment. Nicosia cannot separate particle composition from mass or measure individual exposure; Canadian gases are correlated so a single causal pollutant is not identified; Medicare composition meta-regression still does not isolate one species.
If respiratory admissions have no overall PM₁₀ association, PM does not affect lungs in that city.
Nicosia's null is for respiratory admissions versus PM₁₀ mass, while all-cause admissions still rose 0.85% per 10 μg/m³ and dust-storm days were 4.8% higher. Canadian warm-season PM₁₀/PM₂.₅ tracked asthma visits, and Medicare PM₂.₅ was linked to +2.07% respiratory admissions. Outcome definition, age group and particle metric all change the result.
An NDVI wilting map of 10.8% of forest is a census of tree death.
The product does not separate species mortality from reversible leaf browning, and 35–42% of deviance explained leaves many local factors unmeasured. Spring-2019 greenness remaining lower (NDVI gap +0.015) is persistence of a spectral signal, not a mortality count.
Related
Claim ledger
What the evidence shows
Drawn from 4 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 10-year (1995–2004) Nicosia series linked daily hospital admissions to urban PM₁₀ and ozone, separating meteorologically confirmed dust-storm days. Each 10 μg/m³ same-day PM₁₀ was linked to a 0.85% rise in all-cause admissions; lag-1 ozone raised cardiovascular admissions 2.91% per 10 ppb. On 63 dust-storm days, all-cause admissions were 4.8% higher. Respiratory admissions showed no overall PM₁₀ association.
A seasonal Poisson Medicare analysis (2000–2003) related emergency admissions to PM₂.₅, then meta-regressed effects on particle species-to-mass ratios. Per 10 μg/m³: +1.89% cardiovascular, +2.25% MI and +2.07% respiratory admissions. Composition modified toxicity, so mass alone was incomplete. The series does not include people under 65.
Pooled daily ED visits from 14 hospitals in seven Canadian cities linked CO, NO₂, O₃, SO₂, PM₁₀ and PM₂.₅ at lags 0–2 days (and 3-hour averages) to cardiac and respiratory diagnoses. CO and NO₂ showed the most consistent cardiac associations (warm-season CO: +5.2% MI/angina per 0.7 ppm). Ozone tracked asthma and COPD (COPD +6.2% per 18.4 ppb in the warm season). Warm-season PM₁₀ and PM₂.₅ were strongly linked to asthma visits. Three-hour same-day averages showed no consistent signal. Edmonton supplied ~70% of visits.
The same method label also covers satellite NDVI time series. Random-forest maps trained on >300,000 classified orthophoto pixels plus NDVI found 2018 early wilting over about 21,524 km² (10.8% of Central European forest). Precipitation (negative) and temperature (positive) anomalies ranked among the top drivers. Wilted patches stayed less green in spring 2019 (NDVI gap +0.015 versus unaffected forest). The models explained 35–42% of deviance.
Debates
Tensions and limits
Some items are genuine disagreements on the same question. Others mark different assays, populations, or outcomes.
City hospital series do not agree on which pollutant or lag carries the acute signal. Nicosia finds same-day PM₁₀ for all-cause admissions (0.85% per 10 μg/m³) and lag-1 ozone for cardiovascular admissions (2.91% per 10 ppb) but no overall PM₁₀–respiratory link. Canadian EDs find CO/NO₂ most consistent for cardiac diagnoses and ozone for asthma/COPD, with 3-hour averages null and Edmonton dominating the sample. Medicare finds PM₂.₅ mass plus composition for CVD, MI and respiratory admissions among those 65+. Treating 'time series show pollution raises admissions' as one coefficient hides those splits.
City hospital series do not agree on which pollutant or lag carries the acute signal. Nicosia finds same-day PM₁₀ for all-cause admissions (0.85% per 10 μg/m³) and lag-1 ozone for cardiovascular admissions (2.91% per 10 ppb) but no overall PM₁₀–respiratory link. Canadian EDs find CO/NO₂ most consistent for cardiac diagnoses and ozone for asthma/COPD, with 3-hour averages null and Edmonton dominating the sample. Medicare finds PM₂.₅ mass plus composition for CVD, MI and respiratory admissions among those 65+. Treating 'time series show pollution raises admissions' as one coefficient hides those splits.
- Dust storms and city air raise Nicosia hospital admissions
- Canadian ED visits track CO, NO2, ozone, and PM
- PM2.5 components and hospital admissions
Study Role Design N Population Outcome Dust storms and city air raise Nicosia hospital admissions Supports Other10-year daily time-series of hospital admissions vs PM10/ozone, including dust-storm days Citywide Nicosia admissions 1995–2004 (population ~270,000); 63 dust-storm days — no single person-level N Hospital admissions in Nicosia, Cyprus All-cause, cardiovascular, and respiratory admissions associated with PM10, ozone, and dust storms Canadian ED visits track CO, NO2, ozone, and PM Supports OtherMulti-city daily time-series of ED visits vs gaseous and particulate pollutants N=400000 · Nearly 400,000 ED visits to 14 hospitals in seven Canadian cities Cardiac and respiratory emergency department visits in seven Canadian cities ED visit associations with CO, NO2, O3, SO2, and PM by season and lag PM2.5 components and hospital admissions Supports OtherSeasonal Poisson time-series of Medicare emergency admissions vs PM2.5 and composition N=685716 · 685,716 CVD admissions across 26 US communities (2000–2003); other causes also analyzed Medicare beneficiaries aged ≥65 in 26 US communities Cause-specific emergency admissions related to PM2.5 mass and species ratios
PaperFren reads this as a limit on how far one study travels — different assays, populations, or outcomes — not a forced fight between papers.
The 2018 forest-wilting paper is an NDVI/climate time series, not a health Poisson model. Early wilting of 10.8% of forest and a spring-2019 NDVI gap of +0.015 answer a drought-impact question. That is a real disagreement of scientific target under the same method heading, not a second admissions estimate.
The 2018 forest-wilting paper is an NDVI/climate time series, not a health Poisson model. Early wilting of 10.8% of forest and a spring-2019 NDVI gap of +0.015 answer a drought-impact question. That is a real disagreement of scientific target under the same method heading, not a second admissions estimate.
- 2018 drought wilted about 11% of Central European forests
- Dust storms and city air raise Nicosia hospital admissions
Study Role Design N Population Outcome 2018 drought wilted about 11% of Central European forests Supports OtherRemote-sensing map of 2018 early wilting from >300,000 classified pixels plus NDVI ~21,524 km² (~10.8%) of mapped Central European forest; pixel-based map, not a person/tree cohort N Central European forest canopies in the 2018 extreme drought Spatial extent and climate drivers of early wilting and lingering 2019 greenness loss Dust storms and city air raise Nicosia hospital admissions Supports Other10-year daily time-series of hospital admissions vs PM10/ozone, including dust-storm days Citywide Nicosia admissions 1995–2004 (population ~270,000); 63 dust-storm days — no single person-level N Hospital admissions in Nicosia, Cyprus All-cause, cardiovascular, and respiratory admissions associated with PM10, ozone, and dust storms
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
Time-Series Analysis
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
4 studies in this library bear on Time-Series Analysis, ordered by citations.
- PM2.5 components and hospital admissions
A 10 μg/m³ rise in two-day PM2.5 was linked to about 1.9% more cardiac emergency admissions in 26 US communities.
- Canadian ED visits track CO, NO2, ozone, and PM
In seven Canadian cities, carbon monoxide and NO2 align with cardiac emergency visits, ozone with respiratory visits, and particles with warm-season asthma.
- Dust storms and city air raise Nicosia hospital admissions
Same-day PM10 and lagged ozone increase hospital admissions in Nicosia, and dust-storm days add extra cardiovascular risk.
- 2018 drought wilted about 11% of Central European forests
The 2018 heat-and-drought summer triggered early wilting on about 11% of Central European forest area, worst in Germany and Czechia, with greening still reduced the next spring.
Compare studies
Select 2–10 studies. Design and N are labels, not a ranking.
Nothing selected yet.
Questions
What is still open
City hospital series do not agree on which pollutant or lag carries the acute signal. Nicosia finds same-day PM₁₀ for all-cause admissions (0.85% per 10 μg/m³) and lag-1 ozone for cardiovascular admissions (2.91% per 10 ppb) but no overall PM₁₀–respiratory link. Canadian EDs find CO/NO₂ most consistent for cardiac diagnoses and ozone for asthma/COPD, with 3-hour averages null and Edmonton dominating the sample. Medicare finds PM₂.₅ mass plus composition for CVD, MI and respiratory admissions among those 65+. Treating 'time series show pollution raises admissions' as one coefficient hides those splits.
The 2018 forest-wilting paper is an NDVI/climate time series, not a health Poisson model. Early wilting of 10.8% of forest and a spring-2019 NDVI gap of +0.015 answer a drought-impact question. That is a real disagreement of scientific target under the same method heading, not a second admissions estimate.
Ask PaperFren about Time-Series Analysis
Study this conceptflashcards and short-answer questions
Nicosia reports +0.85% all-cause admissions per 10 μg/m³ same-day PM₁₀, +2.91% cardiovascular admissions per 10 ppb lag-1 ozone, and +4.8% all-cause admissions on 63 dust-storm days. Why is the dust-storm figure more fragile?
It rests on only 63 event days, cannot separate composition from mass, and still does not measure individual exposure or mortality. The PM₁₀ and ozone coefficients use the full 10-year daily series.
Canadian EDs found warm-season CO +5.2% MI/angina per 0.7 ppm and no consistent 3-hour same-day signal. What does the 3-hour null tell you about 'more temporal resolution is always better'?
Three-hour averages did not yield a consistent association in that pooled series. Finer clocks can add noise, reduce event counts per bin, or miss the lag structure (0–2 days) that the daily models used. Edmonton also supplied ~70% of visits, so the daily result is not seven equally weighted cities.
Medicare found +1.89% CVD, +2.25% MI and +2.07% respiratory admissions per 10 μg/m³ PM₂.₅, with composition modifying toxicity. What can that time series not claim?
A single causal species, effects in people under 65, or a personal-exposure mechanism. Mass is an incomplete toxicity metric, but the meta-regression on species-to-mass ratios is still ecological at the city-season level.
Why would it be a mistake to cite the 21,524 km² (10.8%) early-wilting map as evidence that air pollution time series harm forests?
The paper maps a 2018 climate-anomaly wilting response with NDVI and site factors, not a daily PM Poisson model. Precipitation and temperature anomalies ranked among the top drivers. It is a different question under the same method label.