TY - JOUR
T1 - What do these scores mean? Presenting patient-reported outcomes data to patients and clinicians to improve interpretability
AU - the PRO Data Presentation Stakeholder Advisory Board
AU - Snyder, Claire F.
AU - Smith, Katherine C.
AU - Bantug, Elissa T.
AU - Tolbert, Elliott E.
AU - Blackford, Amanda L.
AU - Brundage, Michael D.
AU - Aaronson, Neil K.
AU - Ganz, Patricia A.
AU - Garg, Ravin
AU - Michael, Fisch
AU - Hoffman, Vanessa
AU - Reeve, Bryce B.
AU - Stotsky-Himelfarb, Eden
AU - Stovall, Ellen
AU - Zachary, Matthew
N1 - Funding Information:
This work was supported by the Patient-Centered Outcomes Research Institute (PCORI) (R-1410-24,904). Drs. Snyder and Smith are members of the Johns Hopkins Sidney Kimmel Comprehensive Cancer Center (P30CA006973).
Publisher Copyright:
© 2017 The Authors. Cancer published by Wiley Periodicals, Inc. on behalf of American Cancer Society.
PY - 2017/5/15
Y1 - 2017/5/15
N2 - BACKGROUND: Patient-reported outcomes (PROs) (eg, symptoms, functioning) can inform patient management. However, patients and clinicians often have difficulty interpreting score meaning. The authors tested approaches for presenting PRO data to improve interpretability. METHODS: This mixed-methods study included an Internet survey of cancer patients/survivors, oncology clinicians, and PRO researchers circulated via snowball sampling, plus individual in-person interviews. Clinical importance was conveyed using 3 approaches (presented in random order): normal score range shaded green, concerning scores circled in red, and red threshold lines indicating normal versus concerning scores. Versions also tested 2 approaches to score directionality: higher = more (better for function, worse for symptoms) and higher = better for both function and symptoms. Qualitative data from online comments and in-person interviews supplemented quantitative results on interpretation accuracy, clarity, and the “most useful” format. RESULTS: The survey included 1113 respondents: 627 survivors, 236 clinicians, and 250 researchers, plus 10 patients and 10 clinicians who were purposively sampled interviewees. Interpretation accuracy ranged from 53% to 100%. The formats in which higher = better were interpreted more accurately versus those in which higher = more (odds ratio [OR], 1.30; 95% confidence interval [CI], 1.07-1.58) and were more likely to be rated “very”/“somewhat” clear (OR, 1.39; 95% CI, 1.13-1.70) and “very” clear (OR, 1.36; 95% CI, 1.18-1.58). Red circle formats were interpreted more accurately than green-shaded formats when the first format presented (OR, 1.29; 95% CI, 1.00-1.65). Threshold-line formats were more likely to be rated “very” clear than green-shaded (OR, 1.43; 95% CI, 1.19-1.71) and red-circled (OR, 1.22, 95% CI, 1.02-1.46) formats. Threshold lines were most often selected as “most useful.”. CONCLUSIONS: The current results support presenting PRO data with higher = better directionality and threshold lines indicating normal versus concerning scores. Cancer 2017;123:1848–1859.
AB - BACKGROUND: Patient-reported outcomes (PROs) (eg, symptoms, functioning) can inform patient management. However, patients and clinicians often have difficulty interpreting score meaning. The authors tested approaches for presenting PRO data to improve interpretability. METHODS: This mixed-methods study included an Internet survey of cancer patients/survivors, oncology clinicians, and PRO researchers circulated via snowball sampling, plus individual in-person interviews. Clinical importance was conveyed using 3 approaches (presented in random order): normal score range shaded green, concerning scores circled in red, and red threshold lines indicating normal versus concerning scores. Versions also tested 2 approaches to score directionality: higher = more (better for function, worse for symptoms) and higher = better for both function and symptoms. Qualitative data from online comments and in-person interviews supplemented quantitative results on interpretation accuracy, clarity, and the “most useful” format. RESULTS: The survey included 1113 respondents: 627 survivors, 236 clinicians, and 250 researchers, plus 10 patients and 10 clinicians who were purposively sampled interviewees. Interpretation accuracy ranged from 53% to 100%. The formats in which higher = better were interpreted more accurately versus those in which higher = more (odds ratio [OR], 1.30; 95% confidence interval [CI], 1.07-1.58) and were more likely to be rated “very”/“somewhat” clear (OR, 1.39; 95% CI, 1.13-1.70) and “very” clear (OR, 1.36; 95% CI, 1.18-1.58). Red circle formats were interpreted more accurately than green-shaded formats when the first format presented (OR, 1.29; 95% CI, 1.00-1.65). Threshold-line formats were more likely to be rated “very” clear than green-shaded (OR, 1.43; 95% CI, 1.19-1.71) and red-circled (OR, 1.22, 95% CI, 1.02-1.46) formats. Threshold lines were most often selected as “most useful.”. CONCLUSIONS: The current results support presenting PRO data with higher = better directionality and threshold lines indicating normal versus concerning scores. Cancer 2017;123:1848–1859.
KW - cancer
KW - clinical practice
KW - interpretation
KW - patient management
KW - patient-reported outcomes
UR - http://www.scopus.com/inward/record.url?scp=85009830166&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85009830166&partnerID=8YFLogxK
U2 - 10.1002/cncr.30530
DO - 10.1002/cncr.30530
M3 - Article
C2 - 28085201
AN - SCOPUS:85009830166
SN - 0008-543X
VL - 123
SP - 1848
EP - 1859
JO - Cancer
JF - Cancer
IS - 10
ER -