Thinking · Evidence

Three in four deaths are now chronic disease: that changes what a climate shock is.

Abhijith Magal · 3 October 2026 · 8 min read

The NCD deaths and climate story starts with a denominator. WHO's Global Health Estimates, released on 2 October, record that noncommunicable diseases caused 74 per cent of the world's deaths in 2023, against 58 per cent in 2000, when the series began. Of 61 million deaths that year, the ten leading causes took 33 million, and eight of the ten were NCDs. Ischaemic heart disease led, at about 16 per cent of all deaths. Alzheimer's and other dementias ranked fifth, up from 19th in 2000, with deaths tripling to 2.1 million. The age-standardised rate of healthy life lost rose about 20 per cent for depressive disorders and nearly 45 per cent for anxiety disorders between 2019 and 2023.

India sits above the global curve's starting point and inside its direction. WHO's country-level NCD data, last published in its Invisible Numbers portal, put NCDs at 66 per cent of India's deaths in 2019, with cardiovascular disease alone at 28 per cent. The epidemiological transition the GHE describes, communicable diseases falling below half of deaths in low-income countries for the first time in 2023, happened in India years ago.

What a shock hits when deaths are chronic

Hold the 74 per cent against how a disaster is still imagined. The planning picture remains acute: trauma, drowning, waterborne infection, a surge of casualties into an emergency department. All of that still happens. But when three in four deaths come from conditions that are managed rather than cured, through a daily tablet, an insulin cold chain, a monthly review, a dialysis slot, the population a climate event meets is mostly a population in the middle of treatment. The shock's largest health effect runs through the treatment, not past it.

The mechanics are unglamorous. A flood that closes a road for five days does not injure the hypertensive patient; it separates her from a refill, and the NSO has just priced what reconnecting costs, Rs 29,454 for the average rural overnight medical trip, in the distance tax data. A heatwave does not need to touch a heart patient outdoors; heat load plus a missed dose does the work indoors, which is why the chronic-care arm of a heat plan is this season's highest-yield unfunded measure. Displacement multiplies all of it, as I traced in continuity of care under displacement, and heat reaches the medicines themselves before it reaches the patient.

The dementia line deserves its own pause. A cause of death that moved from 19th to 5th describes a population that cannot self-manage a disruption at all: patients who will not ration their own tablets sensibly, will not recognise warning signs, and depend entirely on a carer whose own routine the same event just broke. No Indian disaster protocol I have read treats cognitive impairment as a response category.

Counting treatment days is buildable now

The ordering move here is a measurement move. A health programme, public or CSR-funded, that keeps a patient register already holds everything needed to count treatment-day losses through an event: who was due a refill or review in the window, who received it, and the days to restoration for those who did not. Run routinely, that one metric converts "the flood affected health services" into a number a district officer, a CSR committee and eventually a regulator can act on. It also gives the programme its own performance measure: a continuity design is working when treatment days lost per event falls year on year. The Continuity Cohort Method formalises this way of building programmes around outcomes that outlast the funding cycle; a published version sits on Zenodo at doi.org/10.5281/zenodo.22861246.

This is the denominator slide for any continuity-of-treatment pitch. First the 58-to-74 shift, then India's 66, then the NSO trip cost, then the question that follows from them: when the next event hits this district, who is counting the interrupted treatments? Today the honest answer is nobody, and that is the gap a programme can own.

One limit. The GHE release carries no climate analysis and no India table; India's 66 per cent is WHO's 2019 country figure from a different publication, and the new estimates run to 2023. Before any country number anchors a document, pull India's rows from WHO's Global Health Estimates data portal directly. The depression and anxiety figures measure healthy life lost, not deaths, and the two should never be swapped in a slide.

Common questions

What share of global deaths do NCDs now cause?

74 per cent in 2023, up from 58 per cent in 2000, per WHO's Global Health Estimates released on 2 October 2026. Eight of the ten leading causes of death are NCDs, with ischaemic heart disease first at about 16 per cent of all deaths.

What share of deaths in India are from NCDs?

66 per cent at WHO's last country-level count, published in its NCD data portal for 2019, with cardiovascular disease at 28 per cent of all deaths. Country rows for the new 2023 estimates are in WHO's Global Health Estimates portal and should be pulled directly before quoting.

Why does the NCD share change how climate shocks are understood?

NCDs are managed with daily medicine and routine contact with the health system. A climate event that interrupts supply or visits does its main health damage through broken treatment rather than direct injury, in a population where three of four deaths already run through these conditions.

What should programmes measure after a climate event?

Treatment days lost: enrolled patients whose refill, review or procedure fell in the event window, whether they received it, and days to restoration for those who did not. The metric is countable from an ordinary patient register and turns continuity into a reportable outcome.

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.

This page reads published estimates. It is not clinical guidance; the management of any condition follows national guidelines and the advice of a qualified doctor.

Last updated: 3 October 2026