According to the World Health Organization, the term “scorecard” pertains to the reporting
of a “status” (1). Health scorecards inhabit a longstanding space where medicine and
mathematics meet. Complex health-related data sets can be distilled into simple, robust
comprehensible numeric summaries to support diagnostic, evaluative or prognostic decision-making
(2,3). Scorecard-based reporting has been embraced in such diverse healthcare contexts
as comparative performance evaluation of healthcare systems, monitoring public health
promotion initiatives, managing health conditions and summarising the overall health
and wellbeing of individuals (3). It is in this latter realm of realising individual
health improvements by applying scorecards and constituent score summaries to guide
patient self-care journeys that the present study by de Moraes et al. (4) considered
in this issue is vested.
It is not uncommon for an initial presentation for care by a patient or health consumer
to be precipitated by self-initiated engagement with a quiz in a magazine, newspaper
or (more recently) social media. Such quizzes constitute rudimentary scorecards (of
varying provenance and curation) in their own right, identifying a health status which
may prompt a decision to seek medical advice if for example, a summary score of more
than 10 positive responses out of 20 screening questions is achieved when compared
against a given preset threshold score or within a defined scoring range. On presentation
to a healthcare facility, more focused (and often evidence-based) screening questions
may then be posed by clinicians to elicit relevant patient history for triage purposes,
to inform diagnosis and treatment choices or to track health status with repeated
presentations over time.
Gone (or disappearing rapidly) are the days when patients sat in a clinician’s waiting
room with a clipboard balanced on their laps, diligently providing ‘tick and flick’
responses to paper-based questionnaires using a blunt pencil. For some, the waiting
rooms have also disappeared in deference to the necessity for COVID isolation or the
sheer convenience of engaging with their own smartphones or tablet devices at home
to provide this information outside the physical confines of a healthcare facility.
Constituting a “patient’s voice” or point of view, electronic patient reported outcome
(and experience) measures (e-PROMs and e-PREMs) harness digital devices to elicit
health data emanating directly from the patient, without interpretation of responses
by clinicians or anyone else (5). Digital devices offer previously unheralded access
to a “blank canvas” whereby pertinent and timely e-PROM-derived information can be
captured. The key difference between a health quiz in a magazine (or on TikTok) and
responding to one or more “formal” e-PROM assessments lies in the validation and testing
that underpins e-PROMs and how the resultant patient-sourced data is then applied
to effective clinical decision making for individual health improvement (or for health
system evaluation) (5-7).
Implemented as a new cross platform-compatible mHealth app, de Moraes et al. report
on development and preliminary evaluation of a novel health scorecard in this issue
which captures responses using several self-administered e-PROMs (representing selected
health domains) to quantitate an individual’s health and wellbeing status (4). For
each domain considered in this study, the authors applied a three-tiered numeric scoring
system to yield a poor, good or excellent rating. A single composite (Magenta) score
was derived as the mean of scores tabulated across all of the domains investigated.
Guidance was offered to study participants by means of defined decision trees suggesting
evidence-based interventions for health improvement based on categorisation of reported
scores, with subsequent follow-up assessment at between 3 to 5 months to gauge any
change in reported health and wellbeing status.
A recent cursory Google search on the topic “domains of health” identified anywhere
between 3 and 27 domains of health and wellbeing. de Moraes et al. leverage a selection
of evidence-based e-PROMs representing six health-related domains for inclusion in
this new scorecard, deemed to offer the greatest potential to realise health benefits
from early intervention and reflecting the setting and context for this research,
namely a Brazilian private healthcare organisation. In addition to internationally-recognised
e-PROMs, the authors incorporated localised measures when considering specific health
domains such as nutrition status, applying dietary recommendations published by the
Brazilian Ministry of Health (4).
de Moraes and colleagues offer this study as a preliminary investigation, positing
that further research is needed. The age of the study cohort reflects “digital natives”
in their early thirties engaging with a private healthcare service. Distinct from
this demographic, younger and older persons may face challenges in engaging with e-PROM
technologies without assistance, as might persons from different socioeconomic backgrounds
(8,9). The heterogeneous nature of the study group (i.e., presenting for case or disease
management, health and wellbeing, etc.) demands experimental designs and sufficient
sample sizes across target groups to facilitate robust longitudinal statistical analysis
of outcomes. Lessons can be learned from prior systematic reviews regarding study
designs suitable to demonstrate PROM and e-PROM efficacy (10,11). For example, a contemporary
Cochrane review of 116 randomised trials found low to moderate certainty regarding
evidence supporting the effectiveness of PROM feedback in improving health outcomes
(10). Key risks identified in the reviewed studies included performance and detection
biases.
Aggregation of existing e-PROMs in this new composite incarnation requires vigorous
re-testing to assert that validity, reliability and usability is maintained across
constituent e-PROMs. Variation in the order of question presentation, mix of question
types across e-PROMs, onscreen response methods and overall completion times across
e-PROMS may all affect clinical validity of the aggregated results and usability of
the new tool (5,7,12). The Magenta score calculated in this study is based on the
mean across all 6 domains investigated; the course of some health conditions or treatments
may result in reporting average scores which fail to reflect changes (or minimum significant
changes) in or between health domains (13,14). For example, sleep score may increase
over time (>750) due to a worsening in mental health in some cases (<500) with or
without a reduction in physical activity. Similarly, sleep may be disrupted (<500)
for a patient who changes habits and reduces smoking (>500).
Opportunities exist to leverage emergent technologies to augment the “patient’s voice”
constituted by e-PROMs. Many corporate health systems are already designed to capture,
assimilate and report on e-PROM data as part of a patient’s electronic medical record
(5,7). Assistive technologies available in modern digital devices offer support for
equity in healthcare access for people with varying physical abilities (e.g., vision,
dexterity) to engage with e-PROMs by means of spoken command interactions or alternate
means of data entry to navigate digital device screens to capture data (15). Patient
generated health data (PGHD) such as physical activity (steps) or sleep duration monitored
by smartphone or wearable sensors could further “amplify” the patient’s voice by (unobtrusively)
contributing quantitative health data; “silent” accumulation of such PGHD using the
patient’s own digital devices have been likened to “grains of sand” with potential
to accumulate as “clinical pearls” to further inform health improvement (16).
Potential also exists to harness emergent (and widely lauded) generative artificial
intelligence (AI) to improve the breadth, depth and context of data captured by e-PROMs,
refining score calculation algorithms and decision trees (17). Responses to e-PROM
questions driven by AI may suggest an additional line of questioning to elicit more
information in real-time engagements with “Chat”-style e-PROMs. Entirely new vistas
of research open up when considering the possible uses of AI in eliciting patient
reported outcomes, albeit demanding a completely new set of “sums” driven by vastly
more complex AI-aware e-PROMs.
Supplementary
The article’s supplementary files as
10.21037/mhealth-23-38