Using percentiles in the interpretation of Patient-Reported Outcomes Measurement Information System scores: Guidelines for autism.
Compare PROMIS scores to the new autism percentiles, not the general ones, before you call sleep or anxiety severe.
01Research in Context
What this study did
Irwin et al. (2022) built the first autism-only percentile charts for four PROMIS parent-proxy scales.
Parents of autistic children rated sleep, anxiety, mobility, and pain. The team placed each score on a curve that shows how the child compares only to other autistic kids, not to the general public.
What they found
Sleep and anxiety scores that look "severe" against U.S. norms are often average inside the autism group.
Using the new percentiles keeps you from calling a typical autism profile a clinical crisis.
How this fits with other research
Prabhakaran et al. (2024) push the same idea further. They warn that Indian families who sleep in the same bed may get high CSHQ scores that do not mean pathology. Together the papers say: know your reference group and your culture.
Schwichtenberg et al. (2013) seems to disagree. They saw only mild sleep trouble in preschoolers who had an autistic sibling but no diagnosis. The clash clears up when you notice they studied non-autistic kids; Julia’s percentiles are for diagnosed autism, a group with naturally higher sleep needs.
Chen et al. (2011) give the opposite advice for PEP-3: use raw scores, not percentiles, to track growth. The two papers pair nicely—use raw numbers to watch one child over time, use Julia’s percentiles to judge if that child stands out right now.
Why it matters
Next time a parent says "My child’s sleep score is in the 95th percentile," ask "Which curve?" Switch the computer view from the general-U.S. line to the autism line before you decide whether to write a sleep treatment goal. You may save the family weeks of needless intervention.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Open the PROMIS score report, toggle to the autism norm, and re-check if the red-flagged sleep score is still extreme.
02At a glance
03Original abstract
The objectives of this study were to (1) demonstrate the application of percentiles to advance the interpretation of patient-reported outcomes and (2) establish autism-specific percentiles for four Patient-Reported Outcomes Measurement Information System (PROMIS) measures. PROMIS measures were completed by parents of autistic children and adolescents ages 5-17 years as part of two studies (n = 939 parents in the first study and n = 406 parents in the second study). Data from the first study were used to develop autism-specific percentiles for PROMIS parent-proxy sleep disturbance, sleep-related impairment, fatigue, and anxiety. Previously established United States general population percentiles were applied to interpret PROMIS scores in both studies. Results of logistic regression models showed that parent-reported material hardship was associated with scoring in the moderate-severe range (defined as ≥75th percentile in the general population) on all four PROMIS measures (odds ratios 1.7-2.2). In the second study, the percentage of children with severe scores (defined as ≥95th percentile in the general population) was 30% for anxiety, 25% for sleep disturbance, and 17% for sleep-related impairment, indicating a high burden of these problems among autistic children. Few children had scores at or above the autism-specific 95th percentile on these measures (3%-4%), indicating that their scores were similar to other autistic children. The general population and condition-specific percentiles provide two complementary reference points to aid interpretation of PROMIS scores, including corresponding severity categories that are comparable across different PROMIS measures.
Autism research : official journal of the International Society for Autism Research, 2022 · doi:10.1002/aur.2833