Progress tracking for autism care

The progress your raw scores are hiding.

Two therapists can score the same child differently, and the same child scores differently at home than in clinic. So a rising line might be real progress, or just a change in who measured and where. Throughline estimates those effects and removes them, so clinicians see the child's underlying progress instead of measurement noise.

See a live case →What it catches
not realRawDeconfounded
The shaded gap is the confound: apparent gains that vanish once rater leniency and setting are removed.

Where standard tracking breaks down

Deconfounding, in one line

It means removing the part of a score that comes from who rated the child and where, so what is left reflects the child's actual ability. A raw score blends all three. Throughline pulls them apart.

Standard progress tracking

  • Shows the raw assessment score over time.
  • A new, more lenient rater can read as progress.
  • A child tested at home instead of clinic can read as change.
  • Hard to separate a real gain from a moving ruler.

Throughline

  • Estimates each rater's and setting's offset from consensus.
  • Places every score on one common scale.
  • Plots the corrected line right next to the raw one.
  • Flags the gains that vanish once the confounds come out.

Raters disagree more than people expect. Inter-rater agreement on VB-MAPP runs around an ICC of 0.63, so close to a third of a score is the rater, not the child. The correction is established psychometrics. Putting it in a clinician's hands, live, is what is new.

Artifact progress

Raw scores climb while the deconfounded line stays flat. Gains that are really a more lenient rater or an easier setting.

Real regression

A decline that survives correction. Real loss of skill, not one rough afternoon at home.

On track

Gains that hold up after deconfounding. Progress a clinician can act on with confidence.