Track A · Marina's Mission · Aramco / Social Good
We catch 92% of the blur that needs glasses.
We catch 92% of the blur that needs glasses.
Then we don't lose the kid.
A Texas-compliant eye chart you test on this laptop — lean in and it says step back — plus an agent that follows every failing child until a doctor has looked.
Prescreen, not diagnosis. Nurse-tap is the default; every referral gets a full exam. Closing the loop is what works — Peek RCT: 54% vs 22% attended with reminders. Texas-compliant: 25 TAC §37.23.
25–40
of every ~50 kids who fail a school screening never see an eye doctor
Ly et al. 2023, J AAPOS ↗
92%
of the refractive error that actually reduces vision is caught by the §37.23 referral rule in a held-out backtest — 43% across the whole sample, most already seeing fine with correction
Held-out backtest · n=1,040 vision-reducing ↗
2 in 3
cases of vision loss are correctable with glasses — the exam is the missing step
Full CDC NHANES sample · n=12,981 ↗
Screening station §37.23 · Sloan / HOTV · calibrated optotypes
DSHS protocol Right eye (OD) first, then the left — screen down to 20/20. A line counts as passed only if the child reads more than half of it.
not calibrated
Match a credit card (85.6 mm) so letters render at true physical size.
Right eye (OD)
20/200