Is the Give Youth a Voice questionnaire an appropriate measure of teen-centred care in paediatric oncology: a Rasch measurement theory analysis

Abstract

Background: Adolescents have their own views about the cancer care they receive and how they feel they are treated, but their opinions are rarely solicited.

Objective: To determine whether the 56-item Give Youth a Voice (GYV-56), its subscales and its 20-item short-form, are clinically meaningful and psychometrically sound instruments that can be used to measure teen-centred care (TCC) in paediatric oncology.

Design: Qualitative interviews and a questionnaire survey.

Setting and participants: Qualitative interviews with 38 childhood cancer survivors. GYV-56 data collected from 200 paediatric cancer patients and survivors.

Main outcome measure: The GYV-56, which measures the following four aspects of service delivery: Supportive and respectful relationships; Information sharing and communication; Supporting independence; and Teen-centred services.

Results: Qualitative data provided broad support for the TCC conceptual framework and GYV-56 items. After post-hoc reduction of the response options from 7 to 3 (to correct for disordered thresholds), fit to the Rasch model was good, most items showed acceptable fit residuals and chi-square P-values, scale reliability were supported and item locations defined a continuum for TCC that was well-targeted to the sample. By calibrating the items for each subscale and the short-form to the full scale, the scores obtained on each measure are directly comparable.

Conclusion: Our study found initial support for use of the GYV with a reduced response option format for examining TCC in the adolescent oncology patients. in this paediatric population. Further research using the GYV is needed to elaborate upon our findings.

Keywords: Rasch measurement theory; adolescents; family-centered care; neoplasm; processes of care; quality improvement.

Figure 1

Category probability curves for item 20 ‘…understand your feelings’. The x‐axis represents the construct of TCC with higher scores for TCC increasing to the right. The y‐axis shows the probability of endorsing the response categories, reading left to right: 0 (first curve) ‘Never’, 1 (second curve) ‘+’, 2 (third curve) ‘++’, 3 (fourth curve) ‘Sometimes’, 4 (fifth curve) ‘+++’, 5 (sixth curve) ‘++++’ and 6 (seventh curve) ‘A Lot’. Locn, location; FitRes, Fit residual; Pr, probability.

Figure 2

Item characteristic curve for item 3 ‘… treat you as an individual’. The x‐axis represents the construct (TCC), with higher scores (more TCC) increasing to the right. The y‐axis shows the expected value as predicted by the Rasch model. The black dots, which represent class intervals, are very close to the line indicating a close association between observed and expected scores. Locn, location; FitRes, Fit residual; Pr, probability.

Figure 3

Person‐Item thresholds Distribution. The x‐axis represents the construct (TCC), with higher scores (more TCC) increasing to the right. The y‐axis shows the frequency of person measure locations (top histogram) and item locations (bottom histogram).