Preliminary validation of a novel patient-reported outcomes (PRO) questionnaire for patients with cutaneous T-cell lymphoma: The CTCL-PRO-18.

Abstract

Listen

Translate article

e23161 Background: Improving quality of life (QoL) is key to clinical decision-making for patients with cutaneous T-cell lymphoma (CTCL). Currently there is no measure of QoL that captures concerns of patients with advanced-stage CTCL. A thematic analysis of interviews with early- and late-stage patients with CTCL identified novel QoL domains. Using these data, we developed questionnaire items for a new QoL measure (CTCL-PRO-18). Methods: Consented patients completed our new survey and previously validated metrics: SF-36, MF/SS-CTCL QoL, VAS-Itch (past 24 hours), Skindex-29, CES-D, and PROMIS™ Sleep Disturbance. Principal component analysis (PCA) with varimax rotation reduced the number of items on our measure; Cronbach’s alpha evaluated the internal-consistency reliability of items that loaded on each component and the overall measure. We examined the construct and convergent validity of the CTCL-PRO-18 using Pearson correlations, and differences in QoL outcomes by patient demographics and disease status using analyses of variance (ANOVA). Results: 82 patients (47 male; 69 white, 11 Black, 2 Hispanic) completed the surveys. The PCA yielded an 18-item, 5-component solution: 6-item self-consciousness; 3-item burden of treatment; 3-item concern about sores; 3-item sleep/fatigue; and 3-item lack of available CTCL resources. The 5 subscales and overall CTCL-PRO-18 measure had acceptably high reliabilities; Cronbach’s alphas ranged from .74 to .93. Correlations between the CTCL-PRO-18 and each of the previously validated measures were statistically significant (P < .005). Correlations between CTCL-PRO-18 and SF-36 (subscales ranged r =-.320 to -.624), CESD ( r= .645), PROMIS™ Sleep Disturbance ( r= .4780), and VAS-Itch ( r= .665) were moderate. Correlations between CTCL-PRO-18 and MF/SS-CTCL QoL ( r= .720) and Skindex-29 ( r= .814) were high. Correlations between the overall CTCL-PRO-18 and each of its subscales were significant (P < .001) and ranged from .571 (Burden of Treatment) to .882 (Self-consciousness). ANOVAs showed no statistically significant gender or racial/ethnic (white vs. Black/Hispanic race/ethnicity) differences in the CTCL-PRO-18 or the 5 component scores. Burden of Treatment scores were significantly lower for patients with no active disease compared with patients with reduced symptoms and active disease (4.9 [SD 3.9], 6.4 [3.2], and 8.2 [3.8], respectively; P = .018). Conclusions: The CTCL-PRO-18 and its 5 subscales had high internal-consistency reliability. Correlations were higher between the CTCL-PRO-18 and other CTCL-relevant measures (Skindex-29, MF/SS-CTCL QoL) than they were between the CTCL-PRO-18 and less CTCL-specific measures, providing preliminary evidence of the construct and convergent validity of the CTCL-PRO-18. Examination of the discriminant validity of this metric is warranted.