Presenting comparative study PRO results to clinicians and researchers: beyond the eye of the beholder
Presenting comparative study PRO results to clinicians and researchers: beyond the eye of the beholder
Michael Brundage 0 1 2 4 5 6
Amanda Blackford 0 1 2 4 5 6
Elliott Tolbert 0 1 2 4 5 6
Katherine Smith 0 1 2 4 5 6
Elissa Bantug 0 1 2 4 5 6
Claire Snyder 0 1 2 4 5 6
locations) 0 1 2 4 5 6
0 Katherine Smith
1 Elliott Tolbert
2 Amanda Blackford
3 Michael Brundage
4 Department of Oncology, Queens Cancer Research Institute, Queen's University Kingston , Kingston, ON , Canada
5 Claire Snyder
6 Elissa Bantug
Purpose Patient-reported outcome (PRO) results from clinical trials can inform clinical care, but PRO interpretation is challenging. We evaluated the interpretation accuracy and perceived clarity of various strategies for displaying clinical trial PRO findings. Methods We conducted an e-survey of oncology clinicians and PRO researchers (supplemented by one-on-one clinician interviews) that randomized respondents to view one of the three line-graph formats (average scores over time for two treatments on four domains): (1) higher scores consistently indicating “better” patient status; (2) higher scores indicating “more” of what was being measured (better for function, Vol.:(011233456789)
-
Presented in part at the 2016 International Society for Quality of
Life Research Annual Meeting.
worse for symptoms); or (3) normed scores. Two formats
displayed proportions changed (pie/bar charts). Multivariate
modeling was used to analyze interpretation accuracy and
clarity ratings.
Results Two hundred and thirty-three clinicians and 248
researchers responded; ten clinicians were interviewed.
Line graphs with “better” directionality were more likely
to be interpreted accurately than “normed” line graphs (OR
1.55; 95% CI 1.01–2.38; p = 0.04). No significant differences
were found between “better” and “more” formats. “Better”
formatted graphs were also more likely to be rated “very
clear” versus “normed” formatted graphs (OR 1.91; 95% CI
1.44–2.54; p < 0.001). For proportions changed, respondents
were less likely to make an interpretation error with pie
versus bar charts (OR 0.35; 95% CI 0.2–0.6; p < 0.001); clarity
ratings did not differ between formats. Qualitative findings
informed the interpretation of the survey findings.
Conclusions Graphic formats for presenting PRO data
differ in how accurately they are interpreted and how clear they
2
3
4
Johns Hopkins School of Medicine, 550 N. Broadway, Room
1111, Baltimore, MD 21205, USA
Johns Hopkins School of Medicine, Baltimore, MD, USA
Johns Hopkins Bloomberg School of Public Health, 624
N. Broadway, Room 725, Baltimore, MD 21205, USA
Johns Hopkins Bloomberg School of Public Health, 624
N. Broadway, Room 726, Baltimore, MD 21205, USA
Johns Hopkins Bloomberg School of Public Health, 624
N. Broadway, Room 649, Baltimore, MD 21205, USA
Sidney Kimmel Comprehensive Cancer Center at Johns
Hopkins, 1650 Orleans Street, Baltimore, MD 21287, USA
Cancer Clinic of Southeastern Ontario, 25 King Street West,
Kingston, ON K7L 5P9, Canada
are perceived to be. These findings will inform the
development of best practices for optimally reporting PRO findings.
Introduction
An emphasis on patient-centered care has increased the
demand for patient-reported outcomes (PROs), data
collected directly from patients about health conditions and
impacts of treatments [
1, 2
]. PROs can inform patient care
in a variety of ways [
3, 4
], including data representing ‘the
voice of the patient’ in randomized clinical trials to inform
decision-making by patients and clinicians based on trial
results [
4–7
]. Oncologists have endorsed the use of PROs
for this purpose, and there is evidence that PRO results can
influence treatment choices [
8–11
].
The optimal integration of PRO results from clinical
trials, and other comparative research into clinical care requires
that clinicians understand and interpret PROs accurately
and be able to communicate PRO findings to their patients
where appropriate. This understanding can be challenging
because of the variety of PRO questionnaires, variation in
their scoring (e.g., higher scores indicating better or worse
outcomes), their scaling (e.g., scores ranging from 0 to 100
as worst-to-best, or scores normed to a defined population),
and how statistical and clinical significance of the findings
are addressed [
9, 12–15
].
This study was part of a larger research program designed
to examine approaches for presenting PRO data to promote
their understanding and use. Previously, we evaluated
existing approaches to presenting study PRO results [
15
]. We
then partnered with stakeholder workgroups of clinicians
and patients to develop improved graphical presentation
approaches [
14
]. In a separate research stream, we focused
on the communication of PROs to patients (e.g., in
educational materials or decision aids) [
16
]. In this study, we
focused on the reporting of PROs from clinical trials to
clinicians and PRO researchers, building on our earlier (...truncated)