gms | German Medical Science

Objective: In the era of patient-centered neurooncologic therapy, the association between health-related quality of life (HRQoL), psychosocial distress, and supportive care needs should be in the focus of research. In a prospective study those needs were assessed using three established instruments in an outpatient setting. The aim of our analysis is to investigate the relationship between HRQoL, psychosocial needs and distress, to evaluate the consistency of three screening instruments with respect to psychosocial burden and to determine a cut-off value for HRQoL scales indicating pathologic distress and need for supportive care.

Method: In a prospective trial outpatients with high-grade gliomas were screened for their need for psychosocial support using Distress Thermometer (DT), EORTC QLQ C-30+BN-20 (HRQoL) and Supportive Care Needs Survey-SF34-G (SCNS). Items of the above mentioned survey tools were grouped in 7 categories corresponding to the HRQoL function scales and selected symptom scales. We estimated convergent validity of single items and scores across the instruments using Pearson's and Spearman's correlation coefficients. Cut-off values of HRQoL for pathologic DT and SCNS scores were calculated using ROC curves.

Results: 175 patients were screened at 3 centers yielding 167 complete datasets. The strongest correlations of HRQoL and DT was found for cognitive function, global health perception score (GHS), emotional and role function, future uncertainty (FU) and fatigue as well as between HRQoL and SCNS for FU, emotional and role function (range |0.4-0.7|; p<0.01). GHS, emotional and social function scores showed a fair discrimination between pathologic and normal DT scores (AUC 0.75, AUC 0.77, AUC 0.749 p<0.01) with best discrimination for emotional and social function. A HRQoL cut-off of <54.2 (GHS) and <58.5 (emotional function) accurately predicted a pathologic DT. A HRQoL cut-off of <54.2 (emotional function) and <52.8 (FU) predicted the respective need for supportive care (AUC 0.82, AUC 0.86; p<0.01) with sensitivity >0.8 and specificity >0.7.

Conclusions: Reduction in HRQoL correlates significantly with higher scores in DT and SCNS. We established cut-off values that might impact patients' referral to further treatment. However, not all aspects of DT and SCNS are detected with a sufficient accuracy using HRQoL only. Thus, combining several screening instruments seems to be inevitable to assess all needs in glioma patients.