Thinking · Evidence

Air quality is an access problem too. The respiratory story is the half that gets counted.

Abhijith Magal · 8 September 2026 · 9 min read

A health programme's register records a missed visit. It does not record why. When the reason is a week of air that nobody should have been walking through, the entry still reads the same as a patient who lost interest, and at the end of the quarter the programme reports a retention problem it cannot explain.

Air quality is written about almost entirely as a respiratory story, and on that story the evidence is deep. This page is about the other story, the one almost nobody writes: what a severe episode does to access for everything else. The argument holds anywhere air turns bad for weeks at a time, from wildfire smoke to winter inversions. India is the worked example because its bands and its emergency plan are public, precise and tied to forecasts.

What does the evidence on air pollution and healthcare actually measure?

Mostly, who comes in. The large literature on short-term exposure counts hospital admissions, emergency attendances and outpatient visits for conditions pollution worsens, and it finds them rising with concentration. A time-series study in Xi'an covering 1,340,791 outpatient visits across six clinic departments is typical in design: it asks which departments see more patients as pollutants rise, and it found associations even at moderate pollution levels.

That is the right question for a hospital planning beds. It is the wrong question for a programme trying to keep a cohort in care, because it cannot see the visits that never happened. Where researchers have looked at behaviour, they have found people do change what they do when warned. A study of English data found that air pollution warnings reduced asthma admissions among children by 8 per cent, evidence of avoidance behaviour. Whether the same warnings, or the same air, also keep people away from care they need for something else is a question the literature has barely asked.

We count who bad air sends to hospital. Almost nobody counts who it keeps from the rest of their care.

Why would bad air reduce care for unrelated conditions?

Because access runs on journeys, and a severe episode taxes every journey. Four mechanisms are plausible, and each one is testable in a programme's own records.

• Patients deferOlder patients, people with heart or lung disease and parents of young children are told to limit exertion outdoors. A routine review for diabetes or hypertension is exactly the trip that gets postponed.
• Field work stopsCommunity health workers and outreach teams walk, cycle and wait outside. When the advisory says to avoid prolonged exposure, the home visit is the first thing to go, often by the worker's own sound judgement.
• Facilities re-prioritiseA surge in respiratory cases pulls staff and beds. Elective procedures and non-urgent outpatient work are what a stretched facility moves.
• Emergency orders reach the dayWhere authorities run graded plans, higher stages restrict movement and activity in ways that fall on the same journeys a programme depends on.

None of these shows up as a climate event in programme data. Each shows up as a lower count: fewer reviews, fewer visits, fewer refills collected. That is the misattribution pattern set out for heat and flood in the climate risk hiding in your access data. Air is the same failure under a different stressor.

What do India's AQI bands mean for a programme?

India's National Air Quality Index, launched in 2014, reports the worst of up to eight pollutants as a single number from 0 to 500 in six bands, each with a published health statement. The table sets the official bands and statements beside the Delhi-NCR emergency stages, and adds a planning column. That last column is our reading of what a programme should prepare for. It is a design prompt, not an official advisory.

Two things in that table are worth more than the rest. The first is how low the national line sits against the global one. The 24-hour PM2.5 value at the top of India's "Satisfactory" band, 60 µg/m³, is four times the WHO's 24-hour guideline of 15. A programme outside India reading these bands should translate them, not transplant them. The second is the note in the CAQM schedule that stages are invoked on forecast, in advance of the AQI reaching the projected level. The emergency system already runs ahead of the air. Most programmes run behind it.

What should a programme change before the season?

Use the forecast the authorities already use, and act on it the way they do. That is the whole move, and it redirects systems a programme already has rather than asking for new ones.

Tie the programme calendar to the published stages. When a Stage I or II forecast lands, bring forward the refills and reviews due in the following fortnight, so the cohort has what it needs before travel gets hard. Move routine follow-up for high-risk patients to phone or messaging for the duration, and write down that the change was air-driven.

Give the deferral a code. Add one field to the visit register, "deferred: air quality", and use it every time. It costs a health worker two seconds, and it turns an unexplained quarterly dip into a counted climate event. Without it, the programme's own data will keep telling the funder a retention story that is not true.

Close the loop when the stage lifts. The damage from a bad week is not the week. It is the patients who were deferred and never re-contacted. Set a window, a week or two after revocation, in which every deferred patient is called back. That window is the difference between a pause and a loss to follow-up, and it maps directly onto the Continuous condition of climate-access.

What is still unmeasured, and who should measure it?

The size of the effect. There is good evidence that bad air sends people to hospital and good evidence that people change behaviour when warned. There is very little, and in India almost nothing we could find, measuring how much routine care for unrelated conditions falls during a severe episode and how much of it comes back. That is a measurement gap, not a finding, and this page will not dress it up as one.

Programmes are well placed to close it, because they hold the one dataset the research lacks: named cohorts with scheduled visits. A deferral code, run through one season across a few sites, would produce the first honest count of what a bad-air month costs in missed care. The Delhi-NCR season is weeks away. The forecast will be public before the air arrives. The rest is whether the register is ready to record what happens. The wider case for reading climate through access is set out in climate change is a healthcare story, and the terms are defined in the glossary.

Questions worth asking after this

What is the WHO guideline for PM2.5?

The WHO's 2021 Global Air Quality Guidelines set 5 micrograms per cubic metre as the annual mean and 15 micrograms per cubic metre as the 24-hour mean for fine particulate matter, PM2.5. The 2005 values were 10 and 25.

Does air pollution reduce healthcare use for other conditions?

It is plausible and largely unmeasured. Research consistently finds more respiratory and cardiac visits and admissions as pollution rises, and some evidence shows people change behaviour when warned. Very little research measures whether routine care for unrelated conditions falls during severe episodes, which is a gap programme data could help close.

What are the GRAP stages in Delhi-NCR?

Under the Commission for Air Quality Management schedule revised on 21 November 2025, Stage I applies at AQI 201 to 300, Stage II at 301 to 400, Stage III at 401 to 450, and Stage IV, "Severe +", above 450. Stages are invoked in advance on the basis of forecasts from IMD and IITM.

Abhijith Magal

Founder, Syntropy Earth. Nine years across two global pharmaceutical multinationals in patient access and commercial roles, with health-equity work alongside the WHO-Foundation and UNICEF. He works on climate-access: where climate disruption breaks the link between patients and care. More about Abhijith →

This page describes published standards and evidence. It is not medical advice. Patients should follow the advice of their own clinicians and local health authorities during pollution episodes.

Last updated: 24 September 2026