Contrast Card Generator

Corrected age, on a slider

Designing the "how early was your baby?" control taught us something worth saying out loud: corrected age only ever adjusts downward, and the easiest honest way to ask about prematurity is to ask for the number the hospital already gave you.

Mixed evidenceUpdated Aug 2, 2026For parents, caregivers

Combines established science with practitioner guidance and our design opinion.

In one line

Designing the “how early was your baby?” control taught us something worth saying out loud: corrected age only ever adjusts downward, and the easiest honest way to ask about prematurity is to ask for the number the hospital already gave you.

Key points

  • The card is sized to what the eye can resolve, and in the first months that tracks corrected age, not birthday age. For a baby born six weeks early, using the calendar age designs a pattern with detail finer than they can currently see. That is the one sizing error that makes a card unreadable rather than merely easy.
  • Corrected age = actual age − how early they arrived. A baby born at 34 weeks is six weeks “behind” the 40-week term reference, so at four months old their corrected age is about two and a half months.
  • It is a subtraction, never an addition. A baby born after term does not get age added on. Corrected age exists to avoid holding an early baby to a timetable they never started; there is no equivalent adjustment in the other direction. Our first slider design had a “born late” half that moved the handle and changed nothing — we kept it, because parents of post-term babies will look for it, and made the readout say plainly that no correction applies.
  • The unit you ask in matters more than the control you ask with. We started by trying to slide days before or after term, centred on zero. It never worked: the two directions have wildly different real ranges (post-term stops around +2 weeks; the preterm side reaches back to around 22 weeks), so a centred zero forces the two halves of the slider onto different scales, where the same drag means nine days on one side and one day on the other.
  • The fix was to change the question, not the widget. Slide gestational age at birth, 22 to 42 weeks — the number written on the discharge summary, the number a parent says out loud. One honest linear scale, a notch at 40 weeks for term, and the tool does the translating: “born 6 weeks early · corrected age 1 month.”
  • Corrected age is not forever. It is generally applied through roughly the first two years, after which the difference stops being meaningful for most children.

What building the control actually taught us

The slider had to be finished before some of this was obvious. Three things came out of the build rather than the research:

  • A day is not always a day. The age slider steps by one day for the first three months, by one week up to a year, and by one month after that. That is not a UI flourish; it is what the acuity table looks like when you plot it. A single day genuinely changes the right card for a three-week-old and genuinely changes nothing for a two-year-old, so a track with one step size everywhere is either uselessly coarse at the start or absurdly long at the end. If you have ever wondered why parents of newborns count in weeks and parents of toddlers count in years, this is the same fact seen from the other side.
  • Ask for the number people were given, not the number you need. We need “weeks premature”. Nobody is told that. They are told “she was born at 34 weeks”. Asking for gestational age and doing the subtraction ourselves removed an arithmetic step from the most stressed audience the site has — and removed a whole class of quiet error, because a parent who mis-subtracts produces a card that is confidently the wrong size.
  • The honest answer to “born late?” is a visible nothing. The 40–42 week end of the track moves the handle and leaves the corrected age exactly where it was. That looks broken unless you say why, so the readout states it: corrected age = actual age, because correction only ever adjusts for babies born early. A tool that silently ignores an input teaches its user not to trust it.
  • We stopped asking for a date of birth entirely. Once the age is a slider, the date of birth is not needed for anything — so the generator no longer has a field for it, no longer stores one, and structurally cannot put one in a shared link. There is an optional helper that converts a birth date into an age and then throws the date away. The most private way to hold a piece of personal data is to never take it in the first place.

The evidence

  • [established science / practitioner guidance] Correcting for prematurity when assessing early development, and doing so for roughly the first two years, is standard clinical practice; the correction subtracts the weeks of prematurity from chronological age. See docs/02-research/03-prematurity-and-corrected-age.md for the sourced summary.
  • [established science] Visual acuity in the first months improves rapidly and predictably enough to be worth designing to — see docs/02-research/01-infant-visual-development.md. Which is exactly why getting the age wrong by six weeks matters at two months old and matters much less at two years old.
  • [established science / mixed] Preterm infants measured at term-equivalent age do not track identically to term newborns — some studies find slightly higher grating acuity, others find delays where there is a complication. Corrected age is the sound default, not a prediction about an individual child.
  • [our design opinion] That gestational age at birth is the better thing to put under the user’s thumb than days-from-term is a design judgement, based on it being the number families are actually told and on it needing no scale trickery. So is the choice of step sizes. No study was involved in either.

Practical takeaways

  • If your baby arrived early, enter the gestational age at birth in weeks (the “born at 34 weeks” number). The tool works out the corrected age and sizes the card from that.
  • Watch the corrected-age readout move as you set it — the gap between calendar age and corrected age is the thing that changes the card.
  • If your baby arrived at or after 40 weeks, there is nothing to correct. That is not the tool ignoring you; it is how corrected age works.
  • Beyond about two years, use calendar age. Beyond about three, vision is close enough to adult that the age question stops driving the design at all.
  • You never have to type a birth date. If it is easier to think in dates than in weeks, the helper will convert one for you and then forget it.

Caveats & disclaimer

Not a medical device and not a developmental assessment. Corrected age here is used only to choose a pattern size, not to judge whether a child is “on track” — that is a conversation for your paediatric or ophthalmology team. Individual variation is large, and babies born prematurely should be followed by clinicians for retinopathy of prematurity (ROP) and other vision matters regardless of anything this tool produces.