Page 87 - Let them speak Towards a patient reported outcome measure set for

PROs used in clinical trials and core outcome sets
COS development studies, but were not specifically labeled as a COS. We
excluded COS studies including no PROs.
Search strategy, study selection
We conducted a comprehensive search in November 2022 with the support
of a research librarian (JD) in MEDLINE (Ovid), PsycINFO, Embase, and
manually in the COMET database. We compiled a list of GID using the
human phenotype ontology (HPO) database (https://hpo.jax.org/app/). We
included all terms describing a genetic condition categorized under the
subontology ID. Additionally, we used a search strategy for ID of unknown
cause combined with terms for COS. See Additional file 1 for the full
search. We identified additional studies by reference list checking. We used
the application Rayyan for screening records (26). Two members of the
team (NvS, AvE) screened all titles and abstracts for potential relevance
with a subsample of 10% screened for inter-rater reliability. We resolved
discrepancies through discussion until consensus was achieved. Two
members of the team (NvS, MvM) screened all full texts for eligibility, and
discussed discrepancies until consensus was achieved.
Data extraction
We extracted the same data as in the first scoping review (PROs used in
clinical trials). One team member (NvS) extracted the following additional
data: definition of the PRO(s).
Integrating and classifying PROs
To establish a clear set of unique PROs, we performed the following steps: all
extracted PROs used in clinical trials and included in condition-specific COS
were grouped together and assigned identical terms. We eliminated unclear
PROs (i.e., PROs that lack clear terminology or definition) or outcomes that
were not PROs. We subdivided multidimensional PROs into unidimensional
PROs when feasible (i.e., multidimensional PROs consisting of distinct
unidimensional PROs), and merged duplicate PROs. Subsequently, we
integrated the remaining PROs identified by the two scoping reviews,
and classified the PROs into a conceptual framework. This conceptual
framework has been previously described in our protocol (22), and is based
on the model from Valderas and Alonso (27) (combination of the model by
4
85