Page 194 - Let them speak Towards a patient reported outcome measure set for

Chapter 7
Most hypotheses regarding correlations between the core PROM set and
associated subscales of legacy questionnaires, including PedsQL, SF-36,
CBCL, ASR, and ABCL, were supported (Pearson’s r ≥ 0.40). Other studies
have also reported strong correlations between PROMIS instruments and
PedsQL and SF-36 subscales in populations with different conditions (66-
68). However, some correlations were unexpectedly low, especially between
PROMIS CAT Anxiety and ABCL Anxious/depressed for caregivers of adults.
A possible explanation is that ABCL Anxious/depressed subscale is less
unidimensional than PROMIS CAT Anxiety, as it also captures depressive
symptoms such as loneliness, feeling unloved, and worthlessness. Another
explanation may be the high floor effects of the ABCL Anxious/depressed
subscale, suggesting that it does not adequately capture the experiences of
individuals with minimal anxiety.
PROMIS Social Roles and Activities could discriminate between adults with
and without ID. However, PROMIS Anger and Sleep Disturbance showed
less discrimination than we expected. Because these outcomes were self-
reported, only adults with mild or moderate ID were included. The inclusion
of adults with severe or profound ID in this know-group analysis, might have
resulted in greater group differences between ID and no ID for Anger and
Sleep Disturbance.
GenID outcomes
Caregivers of children with GenID (proxy-report) and adults with GenID
(self-report) reported higher scores or lower scores (in the case of mobility/
physical function, peer relationships, and cognitive function) than the Dutch
and US general population across several health domains, with deviations
ranging from mild to severe. Caregivers of children consistently rated their
child’s health and functioning worse on all PROMIS domains. These lower
levels of health and functioning are not surprising and consistent with
findings from previous studies (69-72). For caregivers of adults, self-report
versions for adults were used, so reference data was not available and
comparisons with the general population could not be made.
192