12-3 · Crisis prevention · 2026-11-14

Social Isolation and Suicide Prevention Research

Cohort and simulationSynthetic checks matchedEvaluation data partly established

Assess whether changes in everyday routines can serve as indicators of sustained isolation.

This is a research record, not a help service. If you are in Korea and need to talk to someone now, the suicide prevention counselling line 109 can be reached by phone. Elsewhere, contact your local emergency number.

Research question and target

Use aggregate features for prolonged changes in social connection and count false alerts during ordinary interruptions.

Internal target: At least 85% precision for detecting transitions into sustained isolation
Programme: Crisis prevention
Evidence on this record: Synthetic harness (prevent12)Common synthetic checks

Evaluation data

NHATS Sample Person and Other Person files
R1–R11, 2011–2021
Split rule locked
WESAD, Universität Siegen
ICMI 2018, DOI 10.1145/3242969.3242985
Manifest locked

Evaluation registry · Labels and split definitions (KR)

Synthetic harness · prevent12

Consent-based on-device detection of isolation transitions

Synthetic precision for detecting a sustained transition was 0.978, with a one-sided 95% lower bound of 0.933, against the internal target of 0.85. Whether the threshold stays certified depends on the assumed prevalence, listed below.

Assumed prevalence 0.25: CERTIFIED
Assumed prevalence 0.1: CERTIFIED
Assumed prevalence 0.05: NOT_CERTIFIED
Assumed prevalence 0.02: NOT_CERTIFIED
Recall: 0.978

Limits

Features are computed on the device and leave it only through a fixed allow-list, after local differential privacy.
Synthetic daily-rhythm signals only. No sensor data from real participants has been used.

Synthetic inputs only (synthetic_only). These are virtual metrics on synthetic inputs. No real cohort has been evaluated, clinical validity has not been assessed and no auditor has signed the result. A separate audit program passed 4 of 4 gates.

G1 · Frozen inputs and submission structureG2 · Schema, policy rules and transition completenessG3 · Independent recalculation of thresholds and metricsG4 · Virtual metric verdict and result hash

Audit record (JSON) · Run output (JSON) · 12-3 design specification: on-device isolation transition detection (KR) · 12-1 to 12-4 replay record (KR)

python3 -B -m prevent12.orchestrate --output evidence/prevent12-run.json
python3 -B -m prevent12.audit --run evidence/prevent12-run.json --output evidence/prevent12-audit.json
python3 -B -m unittest discover -s tests -p 'test_prevent12.py' -v

Sources and eligibility

Dartmouth StudentLife
Open paper
A university student sensing and survey study. Reference outcomes for sustained isolation transitions and suicide-related endpoints remain unverified.
WESAD
Public research data
Laboratory-collected wearable measures of stress and affect. Longitudinal isolation data require a separate source.
KOSIS single-person household statistics
Public aggregate data
Aggregated household types provide regional context. Individual isolation or crisis status requires separate measures.

Boundary cases

Distinguish vacations, battery depletion and device non-wear from observed changes
Check for storage of raw location or text fields
Reject inferences drawn from unverified linkage of regional and individual data

Common synthetic checks

Distinguishing missing aggregate observations from changing records · 5/5 matched

Expected and observed values for these cases

What the official call asks

Field validation with community partner organisations, with at least 100 participants at four or more sites, followed by analysis of satisfaction and effect.

Proposal requirements, printed pages 60–62 · All twelve requirements

Interpretation and next evaluation

Living alone, short-term stress, subjective loneliness and sustained isolation are separate endpoints. Calculate precision only against verified transition labels.

Next check: Register the scale, time window and eligibility criteria for isolation. Test missing sensor readings and ordinary interruptions in simulation.

← All twelve projects