Thinking · Evidence

When the emergency ends and the access gap does not.

Abhijith Magal · 18 September 2026 · 10 min read

This page is written for funders and programme leads in humanitarian and development health: the people who pay for continuity of care and who are rarely shown whether they got it.

When a flood or cyclone moves people, the response is measured in shelter, water, food and acute injuries over the first weeks. That window closes. The people who were managing a chronic condition before the move are still managing it after, often somewhere with no record of them, and the clinic they left marks them as lost. Nobody is counting how many of them came back into care. That is the gap this page is about, and it is a design problem before it is a research one.

How many people do disasters displace, and for how long?

Tens of millions a year, and a growing number for a long time. The Internal Displacement Monitoring Centre's Global Report on Internal Displacement 2026 recorded 62.2 million internal displacements in 2025. Conflict and violence drove a record 32.3 million of them and, for the first time, overtook disasters, which drove 29.9 million. Disaster displacements fell 35 per cent from the exceptional levels of 2024, but remained 13 per cent above the average of the past decade.

The more important number for health is the stock, not the flow. At the end of 2025, 13.6 million people were still living in displacement caused by disasters, the highest year-end figure on record. Displacement that lasts months is displacement that spans many refill dates.

IDMC flags a warning with its own figures. It saw reduced displacement data availability in 15 per cent of the countries it monitors in 2025, three times the share of 2024. The numbers that do exist are becoming harder to collect.

What does the evidence say about treatment after displacement?

That continuity matters, that displacement breaks it, and that almost nobody has measured by how much. A systematic review of community-based solutions for chronic disease management during natural disasters, published in 2025, found 46 eligible studies. Eighteen named continuity of medication during disasters as a key priority. The authors called for quantitative, longitudinal studies with strong impact data in disaster-prone and lower-income settings, which is another way of saying those studies are scarce.

The clinical case for continuity is not in doubt. A systematic review of patients with diabetes or hypertension found a strong association between higher continuity of care and lower mortality, fewer complications and less use of health services. A review of models of care for hypertension and diabetes in humanitarian crises lists forced displacement among the ways treatment and support are disrupted, alongside health worker attrition, insecurity and broken supply chains. And researchers working with forcibly displaced people living with diabetes and hypertension in Uganda and South Sudan have set out what continuity means in practice: engagement along the care cascade, consistent access to medicines, and diet and lifestyle management.

What none of this provides is a count. How many people in treatment before a disaster displacement are back in treatment three months later is not a number anyone publishes. That is a measurement gap, and naming it is more useful than stretching a citation to cover it.

The origin register records a displaced patient as lost to follow-up. The counting failure and the care failure are the same failure.

Why does the care failure stay invisible?

Because it is recorded as something else. A patient who misses two scheduled visits at the clinic they fled is coded the same way as a patient who stopped caring about treatment. The programme's retention falls, the funder sees attrition, and nobody sees displacement. At the destination, if care is found at all, the patient arrives without a record and is often recorded as a new case.

This is the same misattribution that hides heat and flood inside programme data, set out in the climate risk hiding in your access data. Displacement is its sharpest form, because the person has not just missed a visit. They have left the catchment the programme counts.

What would a programme have to do differently?

Carry the treatment relationship across the move, using systems it already runs. Four changes cover most of the distance.

• Add a statusGive the register a "displaced" status, distinct from "lost to follow-up", triggered by a declared disaster in the catchment. It changes nothing clinically and everything about what the numbers say.
• Supply before the seasonWhere displacement is seasonal and forecastable, move patients on stable regimens to a longer supply before the high-risk months, so the road does not run out before the medicine does.
• Put the record in the patient's handA card or phone record with diagnosis, regimen and last visit lets a destination clinic continue treatment rather than restart it.
• Reconnect on purposeAfter the acute phase, call every displaced patient on the register, and hand over to a destination provider where they have settled.

What should a programme count to know whether it worked?

Two numbers, both built from the register the programme already keeps. The reconnection rate: of patients in active treatment in the affected catchment before the event, the share back in treatment anywhere within a set window, three months being a reasonable first choice. And time to reconnection: the median days between the last visit before displacement and the first visit after.

Neither needs a research grant. Both need the "displaced" status to exist before the event, because nobody can count a category they did not record. A funder that asks for those two numbers after every major disaster in its portfolio would, within a few seasons, hold the dataset the literature is missing. It would also know for the first time whether its continuity spend is buying continuity.

That is where this work is heading: from counting who arrived at a clinic to counting who stayed in care. The pre-landfall version of that count is now a published method: the Continuity Cohort Method fixes a named cohort at the disaster trigger and reads continuity at day twenty-one against the area's own baseline. The four conditions it rests on are defined in what is climate-access, and the argument for reading climate through care is set out in climate change is a healthcare story. The glossary entry on vulnerable populations covers the terms.

Questions worth asking after this

How many people were displaced by disasters in 2025?

IDMC's Global Report on Internal Displacement 2026 recorded 29.9 million disaster displacements in 2025, 13 per cent above the ten-year average, and 13.6 million people still living in displacement caused by disasters at the end of the year.

Is there data on how many displaced people lose chronic disease treatment?

Very little. Reviews of chronic disease care in disasters and humanitarian settings consistently identify medication continuity as a priority and displacement as a cause of treatment disruption, but published counts of how many patients reconnect to care after disaster displacement are scarce.

How can a programme measure continuity of care after displacement?

By adding a displaced status to its register before a disaster, then tracking the reconnection rate, the share of previously active patients back in treatment within a set window, and the median time to reconnection.

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 data and evidence. It is not clinical guidance. Changes to supply duration or treatment should follow national guidelines and clinical judgement.

Last updated: 4 October 2026