The last mile of disaster risk reduction is not a place. It is a person: DRMS 2.0

A warning has not completed its journey because it reached a district, a mobile phone or even a household. It succeeds when the person at risk receives it, understands it, can act on it and can obtain assistance when needed. 

The “last mile” is often described as a geographical problem: how to move warnings from national systems to remote communities, neighbourhoods and households. Geography matters, but the final test is human. Was the person able to become safer? 

Bridging this gap requires shifting the focus from what warning systems deliver to what people can do with the information they receive. The following sections explore three dimensions that are critical to this shift: different experiences of risk, people’s changing locations and the ability to turn warnings into early action. 

A warning is not the same as protection  

Technology has greatly improved our ability to observe hazards, forecast impacts and disseminate warnings. Yet producing risk information and enabling protective action are different achievements. 

From the perspective of the person at risk, the questions are immediate: Am I exposed where I am now? What should I do? Is the recommended action feasible for me? Can I evacuate independently? If I cannot, who knows that I need assistance? 

Figure 1 brings these questions into a single pathway. It shows why dissemination is only one stage: protection depends on connecting authoritative information with the person’s circumstances, action and support.

Different people experience the same hazard differently  

A flood does not create identical risks for everybody in its path. A person who uses a wheelchair may encounter an inaccessible evacuation route; a person who is deaf may not receive an audible warning; a child may be separated from caregivers. The hazard may be the same, but exposure, vulnerability, capacity and support needs are not. 

The Sendai Framework for Disaster Risk Reduction 2015–2030 recognizes persons with disabilities as contributing stakeholders to disaster risk reduction. Yet UNDRR’s 2023 Global Survey, based on 6,342 responses from 132 countries, found limited progress in disability inclusion over the preceding decade. Inclusion therefore needs to be built into preparedness and response architecture, not added after the system has been designed. 

Risk increasingly moves with people  

Many disaster-loss databases and assistance registries are structured around events, administrative jurisdictions or registered addresses. Examples include OpenFEMA’s Registration Intake and Individuals and Households Program (RI-IHP) data, DesInventar Methodology, NOAA NCEI Storm Events Database. These systems are valuable for administration and analysis, but they do not by themselves provide a real-time picture of where a person is when risk changes. People move. Someone registered in one city may be hundreds of kilometres away when a cyclone approaches. Migrant workers, visitors, displaced people and others on the move may have limited knowledge of local hazards or support systems. 

Risk follows the person, not only the registered address. Operational guidance on early warning systems for mobile populations reinforces the need to design for changing locations and circumstances. 

From early warning to early action  

Early Warnings for All provides an important opportunity to advance the transition from warning dissemination to person-centred early action, including the activation of appropriate support systems. Its four interconnected pillars cover risk knowledge; detection, observation and forecasting; warning dissemination and communication; and preparedness to respond. The ambition cannot stop at technological coverage. Protection depends on whether people can translate timely information into feasible action. 

Interoperability is therefore central. The Common Alerting Protocol already provides an international standard for distributing authoritative alerts across multiple channels. The remaining design task is to connect such institutional capabilities with the person’s location, accessibility requirements, support networks and available response pathways. 

From DRMS to DRMS 2.0  

My engagement with this question began with a narrower challenge. The original Disability Disaster Rescue Management System asked whether technology could help identify and locate persons with disabilities during an emergency and connect them to appropriate support. The idea received second prize in the 2022 ICT 4 Inclusion Challenge, a competition organized by Atos and the Deutsche Gesellschaft für Internationale Zusammenarbeit GmbH (GIZ), and was subsequently recognized by the Zero Project as a special-entry Global Winning Project 

That foundation has evolved into DRMS 2.0: a broader, people-centred and inclusive disaster preparedness, early warning, assistance and response coordination architecture. It remains anchored in disability inclusion while extending to older people, children, pregnant women, people with health conditions, visitors and anyone whose circumstances may affect their ability to act safely. 

DRMS 2.0 is location-aware and AI-enabled, but neither location technology nor AI is treated as an end in itself. The architecture is designed to use verified official warnings, connect them with consent-based current location and accessibility context, communicate risk in an understandable form, recommend feasible action, mobilize personal or community assistance and escalate to authorized responders where necessary. 

AI may assist with interpreting information, providing multilingual and accessible communication, personalization, prioritization and decision support. It should not independently declare a disaster, replace authoritative warning agencies, or remove human responsibility from consequential decisions. 

The present DRMS 2.0 pilot is a controlled proof of concept, not an official emergency service. Its value lies in demonstrating how established capabilities could operate as one continuous person-centred pathway. Figure 2 sets out that pathway and its accountability loop.

Established capabilities integrated around the person  

The functions brought together in DRMS 2.0 are not new or untested in isolation. Hazard intelligence, geospatial targeting, accessible communication, preparedness, community support and authorized emergency response are already established across disaster-management systems and practice. Inclusive multi-hazard early-warning guidance likewise emphasizes accessible communication and participation by people at risk. 

DRMS 2.0 does not claim to invent these components. Its proposed contribution is to integrate them around the person and their current circumstances, so that risk information can lead to appropriate action, assistance and a recorded outcome. The design is informed by the Sendai Framework for Disaster Risk Reduction 2015–2030, Early Warnings for All, the Common Alerting Protocol, UNDRR guidance on inclusive and accessible multi-hazard early-warning systems, and operational guidance on early-warning systems for mobile populations. These frameworks, standards and guidance support the constituent capabilities and principles; the DRMS 2.0 pilots will assess whether their integration performs reliably, inclusively, safely and accountably under real institutional and community conditions. The approach also draws on my professional experience in environmental management, governance, technology-enabled systems and inclusive development. 

Figure 3 distinguishes the evidence base from the innovation and the validation task. The pilots do not need to prove that warning, geolocation, accessible communication or emergency response are possible. They need to determine whether DRMS 2.0 connects them reliably, inclusively, safely and accountably under real institutional and community conditions. 

What the pilots should establish  

The pilot programme should examine whether the integrated system can: 

• match verified hazard information with the person’s current location and exposure accurately; 
• deliver warnings in forms that people can receive and understand; 
• recommend actions that are relevant, timely, safe and feasible; 
• recognize additional support needs and connect the person with appropriate personal or community networks; 
• escalate to the correct authorized responders when necessary; and 
• close the loop through confirmed safety, an accountable handover or a recorded resolution. 

Evaluation should also test consent, minimum necessary data sharing, access controls, cybersecurity, connectivity constraints, language and accessibility performance, institutional interoperability and the risk of false reassurance or inappropriate automation. 

In my own work, DRMS 2.0 has not yet produced population-level evidence that it reduces disaster losses or prevents rescue situations. Those would be impact claims and require field evidence. For the pilot, I propose measuring more immediate indicators of whether the system is functioning as: whether a warning was received and understood; whether the recommended action was feasible; whether requested support arrived; whether escalation reached the correct responder; how long resolution took; and whether the outcome was verified. 

The next frontier is human

We already possess extraordinary capabilities for observing the planet and communicating risk: satellites, AI, GIS platforms, emergency operations centres and growing volumes of data. The next frontier is to connect what institutions know with the individual who needs to act, without creating another isolated emergency application.

That requires governments, disaster management authorities, technology organizations, researchers, humanitarian agencies, organizations of persons with disabilities, communities, and citizens to work together. It also requires consent, privacy, cybersecurity, interoperability, accessible design, and accountable governance from the outset.

The decisive questions are therefore not only: Was the forecast accurate? Was the warning transmitted? They are also: Did the person receive and understand it? Could they act? Did the necessary support arrive? Was responsibility transferred appropriately? Was the person safer because the system existed? 

If evidence shows that people acted earlier and more safely, risk information will have become resilience. If not, the last mile remains unfinished. 

Pramod Dabrase is an environmental scientist and development professional with more than two decades of experience across environmental management, sustainability, governance and inclusive development. Together with a multidisciplinary team, he initiated the Disability Rescue Management System (DRMS), which subsequently evolved into Disability-Disaster Rescue Management System DRMS 2.0, exploring people-centred approaches to disaster preparedness, early warning, assistance, and resilience. He can be contacted at: [email protected]. For more information, visit the DRMS 2.0 platform or access the DRMS 2.0 app.