Disease surveillance · Early warning
An outbreak detectedbefore anyone calls it one.
Kinara OS is designed to surface earlier epidemic warning signals by correlating governed information across health, logistics, agriculture, and maritime systems. No new data collection. No shared databases. Every join authorised and logged.
The problem
The signals exist. The connection does not.
Every documented epidemic has a precursor signal that appears in data other than the health system: in food prices, in transport patterns, in port arrivals. That data exists. It is collected daily. It sits in separate systems with separate stewards.
Traditional surveillance waits for the health system to see enough cases to raise a flag. By then, the outbreak is already in the community. Kinara OS is designed to correlate the signal across all four domains before the health system has a cluster to report.
Traditional surveillance is single-domain.
Health information systems track clinic visits and lab results. They cannot see the food security shock that preceded the cluster by two weeks, or the movement pattern that explains its geography. The signals exist. The connection does not.
The data collection is already done.
Kinara OS does not require a new surveillance programme. The maritime, agriculture, logistics and health records that flow through the system every day as part of normal operations are the same records the early warning query runs against. The outbreak signal is a query, not a new data source.
Every join is governed and attributed.
Correlating health data with logistics or agriculture data is a cross-domain query. Kinara Core checks the lawful basis, executes with minimum disclosure, and writes the audit entry before the result is returned. The health steward can see exactly what was asked and what was disclosed.
The hypothesis: earlier warning.
In documented outbreak patterns, the maritime-to-health signal chain spans 14 to 28 days. If cross-domain correlation can surface a warning at Day 21 instead of Day 28 to 35, that is 7 to 14 days earlier for the health steward to act. The summer 2027 pilot is designed to test whether this holds in a real-world setting.
How it works
Follow one outbreak signal across four domains.
Maritime
Day 0 from first signal
Vessel from an affected region clears port customs. Crew manifest recorded. No flag raised. No single domain has context.
The maritime domain records every vessel arrival, crew list, and port of origin. In isolation, this is routine data. Its significance only emerges in combination with what follows.
Agriculture
Day 14 from first signal
A food security shock is recorded across three inland districts: lower yields, disrupted supply. Logistics domain sees unusual inbound demand.
Malnutrition and immune suppression precede disease spikes by two to four weeks in documented outbreak patterns. An agriculture signal this size, from these districts, is a known precursor. No health system sees it in time without a cross-domain query.
Logistics
Day 18 from first signal
Unusual movement pattern flagged on inland routes: higher-than-baseline traffic from the districts, no corresponding economic event.
Movement data from the logistics domain shows population displacement before health facilities register it. Field workers travelling between districts, informal traders shifting routes. These appear in transport records first.
Health
Day 21 from first signal
Clinic visit cluster rises sharply across the same three districts. Multiple facilities. Similar presenting symptoms. No cross-district coordination.
Individual clinics see their own queue. Without a cross-facility signal, each treats a local spike as a local issue. The pattern is invisible inside any single health facility's records.
Kinara Core
Day 21 from first signal
Cross-domain pattern raised. Four correlated signals across maritime, agriculture, logistics and health. Early warning issued to the health steward.
Kinara Core correlates the four signals under governance. Each join is authorised, the minimum data is disclosed, and the alert is attributed in the audit log. The health steward receives the warning with the evidence chain attached. No raw data from any other domain is disclosed.
Applications
What this enables across domains.
Epidemic preparedness
Evaluate whether cross-domain signals can identify emerging health threats earlier than clinic-only surveillance.
Food security and disease correlation
Match agriculture yield shocks to downstream health outcomes. Identify at-risk populations before malnutrition presents clinically.
Vector and transmission mapping
Use logistics movement data to model disease transmission routes. Identify which transport corridors correlate with spread patterns.
Port health screening
Cross-reference maritime crew manifests and port-of-origin data with active outbreak locations to inform port health decisions.
Supply chain readiness
When a health alert is raised, the logistics domain already holds warehouse stock levels and vehicle availability. Response dispatch begins before the formal request is filed.
Data governance
No domain surrenders custody to produce the warning.
The maritime steward does not share its crew manifests with the health system. The agriculture steward does not share its yield records with the logistics system. Each domain retains full custody of its own data throughout.
Kinara Core asks a governed question across the boundary, receives a minimum-disclosure answer, and issues the alert. Every step is in the audit log. The health steward can see exactly what was asked, what domains were queried, and what minimum data was used to produce the signal.
Audit entry · Early warning query
Query raised by
Kinara Core · Pattern engine
Domains queried
Maritime, Agriculture, Logistics, Health
Data disclosed
Minimum: aggregate signals only
Raw records disclosed
None
Lawful basis checked
Yes · Policy ref 04-EW
Alert issued to
Health steward only
Audit entry
Immutable · Cannot be altered
Request a demonstration
We will run the cross-domain query live.
The infrastructure is deployed and available for demonstration using synthetic data. A briefing is 45 minutes: we walk through the signal chain, show the governance layer authorising the join, and display the audit log in real time.