“Three people this week found out at my counter that their Medi-Cal was gone. None of them had seen a letter. One walked out without his insulin.”
Contra Costa Early Warning
A community-owned early warning system for HR 1. It watches for warning signs — overdue renewals, paperwork losses, people giving up — so we can act before people lose CalFresh and Medi-Cal they still qualify for.
Official counts tell us who already lost coverage. Early warning tells us while we can still act.
Most coverage loss under HR 1 will come from paperwork, not ineligibility. In Arkansas, more than 95% of people subject to work requirements already met them or should have been exempt. Each indicator below has thresholds set by the community, and it raises its level before losses show up in official counts.
Indicators
Eight readings, each with a threshold the community set. The marker shows where this month falls.
Indicator values are illustrative until data-sharing agreements are in place. Thresholds are proposals for the Steering Committee to adopt.
Community Signals
Neighbors, nurses, pantry volunteers, barbers and caseworkers report what they see. Their stories are the first warning, often months before system data moves.
Converging patterns
Where people get stuck
Share of sample signals naming each barrier. A signal can name more than one.
Coverage Watch
System data from the county and health plans, shown as aggregate counts only. The question every chart answers: did people lose help because they didn't qualify, or because of paperwork?
Automatic renewal rate
Share of Medi-Cal renewals completed from data the county already has, with no form. The community threshold is 65%.
Sample data. Source when live: EHSD monthly renewal report.
Why people lost Medi-Cal
Monthly discontinuances split by reason. Paperwork losses can be prevented; most ineligible losses cannot.
Sample data. Source when live: CalSAWS discontinuance reason codes via EHSD.
Back on within 90 days
People cut off for paperwork can be restored if they return their packet within 90 days. A high rate here means the loss was preventable.
Sample data.
Equity Lens
The same readings, broken out by region, race and ethnicity, language and age, so we can see who is being hit first. Darker cells mean more warning signs.
| Overdue renewals | Paperwork share of losses | Back on in 90 days | Signals this month | |
|---|---|---|---|---|
| Region | ||||
| East County | 1,840 | 74% | 29% | 96 |
| West County | 1,420 | 69% | 36% | 71 |
| Central County | 940 | 63% | 41% | 38 |
| Race and ethnicity | ||||
| Latino | 1,560 | 74% | 29% | 72 |
| Black | 820 | 77% | 26% | 58 |
| White | 980 | 63% | 40% | 41 |
| Asian and Pacific Islander | 560 | 70% | 33% | 22 |
| Multiracial or another group | 280 | 68% | 35% | 12 |
| Language | ||||
| English | 2,310 | 64% | 39% | 118 |
| Spanish | 1,350 | 76% | 28% | 61 |
| Other languages | 540 | 79% | 22% | 26 |
| Age | ||||
| 19–34 | 1,610 | 72% | 33% | 54 |
| 35–54 | 1,580 | 66% | 36% | 87 |
| 55–64 | 1,010 | 70% | 30% | 64 |
“Back on in 90 days”: a lower number is worse, so it is shaded by how far it falls below the county average. Groups under 11 people are never shown. Race and ethnicity come from county records as people report them, shown only as totals.
Mitigation Tracker
When an indicator reaches Warning, these are the responses. Each move is tied to the readings it should bring down.
Response tracker
Renewal Days
Every mitigation in one place: groceries, haircuts with a coverage check, renewals on the spot, address updates and peer navigators.
How it works
Civic Signals is community-maintained and community-owned. It joins two streams, what people report and what the systems count, and turns them into early warnings and a public dashboard.
Signals come in
- Signal form in English and Spanish
- Text message and QR codes at pantries, clinics, barbershops
- 211 referral follow-ups
- Monthly county and health plan counts
Community researchers
- Paid researchers with lived experience review every signal
- Remove names and details that identify anyone
- AI helps group signals; people confirm each pattern
Readings vs. thresholds
- Eight indicators, refreshed monthly
- Thresholds set and changed by the Steering Committee
- Levels: clear, watch, warning, urgent
Alerts and the dashboard
- Warning and urgent alerts to work group leads by email and text
- Public dashboard on civicsignals.net
- Monthly one-page brief for BRIDGE
Data the dashboard needs
Aggregate counts only. No one's case file ever enters Civic Signals.
| Source | What it feeds | How often | Status |
|---|---|---|---|
| Community signal form | Signals, barriers, self-cancellations, relay requests | Continuous | Live |
| EHSD renewal and discontinuance reports | Automatic renewal rate, overdue renewals, reasons for loss, back on in 90 days | Monthly | Needs data-sharing agreement |
| Medi-Cal health plans (CCHP and others) | Renewal reminders through clinics and doctors' offices | Monthly | Needs agreement (state template MEDIL 23-17) |
| Food Bank of Contra Costa and Solano | Volunteer-hour sites, pantry visits, self-cancellation reports | Monthly | In conversation |
| Hospitals and clinics | Uninsured ER visits, application processing days | Quarterly | Proposed |
| 211 Bay Area | Referral outcomes and dead-end numbers | Monthly | Proposed |
Rules that protect people
- Counts, not cases. Only totals leave partner systems.
- Small groups stay hidden. Any cell under 11 people is suppressed.
- No immigration status is ever collected.
- Quotes are shared only with consent, and with identifying details removed.
- The community owns the data. Partners sign a data agreement; the Steering Committee approves every new source.
Technology
Start simple and cheap; grow only when the pilot proves it.
Static site on GitHub Pages · signal form · monthly counts in a shared spreadsheet · thresholds in one editable file
Open-source database · secure upload for partner counts · automatic email and text alerts · Spanish and more languages