Traffic and chronic disease in Helsinki's districts

I wanted to know whether people who live near motorways get more lung disease. In Helsinki you can get much closer to an answer than almost anywhere in Europe — and still not all the way. This is what the open data actually supports, and where it stops.

The finding that shaped everything else: Helsinki does not publish an asthma figure per district. The open file gives only a seven-disease composite, in which asthma is one unweighted seventh. So nothing on this page is a measure of asthma, and I have tried not to dress it up as one.

What is actually published

The City of Helsinki publishes an age-standardised kansantauti-indeksi (chronic-disease index) for each of its 33 reported districts (peruspiiri). It is built from Kela entitlements to special medicine reimbursement, and it is the unweighted mean of seven sub-indices: asthma, diabetes, rheumatoid arthritis, psychoses, coronary artery disease, heart failure and hypertension. Each is indexed against the city average (Helsinki = 100).

The seven are averaged with equal weight, so asthma contributes about 14% of the number regardless of how common it is. An asthma-only series exists — the city commissioned these figures from Kela as a custom statistics order — but it is not open data.

Map of Helsinki's 33 districts coloured by chronic-disease index, with motorway and trunk roads drawn at a width proportional to measured daily traffic. — Chronic-disease index by district, 2023 (Helsinki = 100; blue below, red above). Road width is proportional to vehicles per day measured at Fintraffic counters. Hatched = Östersundom, for which no index is published.

Measuring the traffic, not just the geometry

The obvious exposure measure — the share of a district's land within 300 m of a motorway — turns out to be close to useless. It cannot tell a forest beside a motorway from a roadside neighbourhood, and it treats a slip road like a six-lane artery.

So instead each road is weighted by what actually drives on it. Finland's road authority runs automatic counters (TMS/LAM) that are open through Digitraffic, including raw per-vehicle records. 72 of them sit in the Helsinki region; the busiest carries about 118,000 vehicles a day. Exposure then becomes vehicle-kilometres per day per square kilometre of district.

Exposure measure	Spearman ρ	p	n
Traffic density (vehicle-km/day/km²)	+0.34	0.054	33
Share of land within 300 m of a motorway	+0.09	0.62	33

Weighting by measured traffic roughly quadruples the association. That is the methodological point of the exercise, and it is stable: averaging three weekdays of counts instead of one changes ρ from +0.338 to +0.339.

Scatter plot of road-traffic density against the chronic-disease index for 33 Helsinki districts, Spearman rho +0.34. — Each dot is a district; marker area scales with population. Unadjusted for income, age or smoking.

Explore it

Open the interactive map full screen · Download the district table (CSV)

What this is not

Not an asthma result. Six of the seven component diseases are not respiratory. Three of them — the cardiovascular ones — do have a real air-pollution pathway, which is part of why the correlation is not zero. It still cannot be read as a respiratory signal.
Not a causal estimate. 33 ecological units, unadjusted for income, education, smoking or housing age, all of which track both motorway proximity and morbidity.
Not population-weighted. Exposure is per unit of land, not per resident. The two highest-exposure districts, Länsi-Pakila and Itä-Pakila, are among the healthiest — they are quiet residential areas that happen to sit between two trunk roads. The correlation is carried by the inner and north-eastern districts.
Not a time series. The city states explicitly that its different vintages are not comparable.
p = 0.054 is a weak signal, and I am reporting it as one.

If you wanted a real answer

The exposure side is the easy half here: the Finnish Meteorological Institute's ENFUSER model gives hourly NO₂ and PM2.5 at 10–15 m resolution across the Helsinki region, free. The binding constraint is the outcome side, and the credible route is register data with residential coordinates under a Findata permit — not district averages.

For a causal question, the stronger design is a natural experiment: Stockholm's congestion charge, Germany's staggered low-emission zones, or London's ULEZ, where children's lung-function growth improved measurably while the earlier, weaker LEZ showed nothing.

Data sources

Districts — City of Helsinki WFS, avoindata:Maavesi_peruspiirit (land only). CC BY 4.0
Chronic-disease index — "Helsingin sairastavuusindeksi", Kaupunkitieto / City of Helsinki, from Kela and Statistics Finland, via Helsinki Region Infoshare. CC BY 4.0
Roads — OpenStreetMap contributors via Overpass API. ODbL
Traffic — Fintraffic / Digitraffic TMS (LAM) counters. CC BY 4.0