Legacy Scales & Items
A common concern among researchers is how to integrate PROMIS® into
existing legacy measurement systems. There are three issues: (a) integration with
existing electronic medical records (EMRs), (b) integration with research archives,
and (c) integration with how EMR measurement information is applied in daily care
by service providers.
Integration of PROMIS CAT
with existing electronic medical records (EMRs) is a straightforward process. PROMIS
is already SNOMED
and LOINC compliant, and can be
adapted for almost any EMR system with a modest amount of programming.
Integration of PROMIS data with previously collected patient measures is likewise
straightforward, and should not be problematic in the majority of cases. In most
cases, the application of appropriate mathematical data transformation should align
PROMIS and legacy measures on a common metric.
Because of the exhaustive literature search that precedes item bank binning and
winnowing, the odds are that the best items (if not all of them) from legacy measures
are in the final item bank. The PROMIS item bank development process relies heavily
on the existing infrastructure in measurement science. Sometimes referred to as
"tau equivalence," the underlying Item Response Theory of PROMIS aims
to ensure that banks measure domains as well or better than existing tools. Binning
and winnowing followed by expert panel reviews, cognitive interviews, and item calibration
analyses highlights problematic legacy wording that enables changes to enhance item
reliability and validity. Thus older measurement scales which require aggregation
of all items for scoring will often include items that did not benefit from modern
calibration analyses, and therefore reduce precision.
PROMIS banks have improved or dropped items shown empirically to reduce precision.
This enables tau equivalence with legacy measures of the same topical domain without
reducing validity. Research to date has shown that when compared to legacy measures
in common clinical research use, PROMIS precision is equal to, and usually superior
to, legacy measures.
Because PROMIS item banks represent the best precision allowed by the state of the
art, legacy items included in the PROMIS banks capture the same information as legacy
measures of the same topic or domain. PROMIS item banks just does it faster and more precisely.
Healthcare providers who have identified legacy cut scores to highlight clinical
relevance or special attention should in most cases be able to transform PROMIS
scores or vice versa to enable continued use of familiar heuristics used to inform
clinical decision making. Of course, flagging, using cut scores to highlight clinical
attention, should be regularly analyzed as data accrue to ensure optimum validity,
and utility to ensure patient safety regardless of the measurement tool in use.