Discordance in Treatment Response Assessment Between Clinicians and Patients With Skin Chronic Graft-vs-Host Disease.

Importance: Clinician-reported and patient-reported outcomes are critical measures of therapeutic efficacy in cutaneous chronic graft-vs-host disease (cGVHD) but are not always correlated. Discordance in treatment response between clinicians and patients hinders interpretation of outcomes in clinical trials and complicates therapeutic decision-making in clinical practice. Objective: To identify factors associated with discordance in clinician-reported and patient-reported treatment response assessments and to evaluate the association of clinician-reported and patient-reported responses with survival. Design, Setting, and Participants: This multicenter longitudinal cohort study included adults 18 years and older with cutaneous cGVHD at study enrollment, assembled from 2 observational studies and 1 randomized clinical trial. Data were collected from August 2007 to March 2024, and data were analyzed from July 2024 to May 2025. Main Outcomes and Measures: A global 8-point cutaneous cGVHD treatment response assessment (with 1 indicating resolved and 8 indicating very much worse) was reported by clinicians and patients 3 to 6 months after study enrollment. Clinician-reported and patient-reported treatment responses were categorized into improved, stable, and worse from the 8-point scale, and discordance was defined as a difference in response between clinicians and patients. Positive clinician discordance indicates the clinician reported a better response than the patient, and negative clinician discordance indicates the clinician reported a worse response than the patient. The association of clinician-reported and patient-reported responses with survival was measured by nonrelapse mortality. Results: Of 489 adults with cutaneous cGVHD, 192 (39.3%) were female, 297 (60.7%) were male, and the median (IQR) age was 55 (43-62) years. A total of 321 adults (65.6%) had concordant responses and 168 (34.4%) had discordant responses between clinician-reported and patient-reported treatment responses. Patients with sclerotic cGVHD had greater odds of discordance compared with those without sclerosis, with clinicians reporting both better and worse treatment response than patients (positive clinician discordance: adjusted odds ratio, 3.14; 95% CI, 1.41-6.95; P = .005; negative clinician discordance: adjusted odds ratio, 2.33; 95% CI, 1.19-4.56; P = .01). Worsening compared with improved overall cutaneous cGVHD was associated with nonrelapse mortality when reported by clinicians (adjusted hazard ratio, 2.28; 95% CI, 1.46-3.54; P < .001) and patients (adjusted hazard ratio, 1.86; 95% CI, 1.12-3.08; P = .02), while only patient-reported worsening was significantly associated with nonrelapse mortality in patients with sclerotic disease (adjusted hazard ratio, 2.00; 95% CI, 1.02-3.90; P = .04). Conclusions and Relevance: In this cohort study, discordance in treatment response assessments between clinicians and patients was common in cutaneous cGVHD, yet clinician-reported and patient-reported treatment responses were both associated with survival. In patients with sclerosis who were more likely to experience discordance, patient-reported response was a critical treatment end point, and approaches should be developed to bridge discordance.