Deriving preliminary criterion-anchored cut-off scores for nurse–patient mutuality: A receiver operating characteristic analysis of the Nurse–Patient Mutuality in Chronic Illness Scale

Background Nurse–patient mutuality is a relational process associated with self-care behaviours in chronic illness. Although the Nurse–Patient Mutuality in Chronic Illness Scale (NPM–CI) has demonstrated sound psychometric properties, the absence of criterion-anchored reference points limits the interpretation of its scores across samples and settings. Objective To derive preliminary, sample-dependent, criterion-anchored cut-off values for the NPM–CI using a prespecified threshold of self-care management as an external behavioral criterion and to examine their discriminatory performance. Methods This observational cross-sectional study represents a secondary analysis of data from a multicentre study conducted in five healthcare centres in Italy. The analytic sample comprised 90 adults with chronic illnesses who had been recruited as the patient members of nurse–patient dyads in the source study. Only patient-reported NPM–CI and Self-Care of Chronic Illness Inventory (SC-CII) data were included; nurse-reported data were not analyzed. Self-care management was dichotomized using the established threshold of ≥70/100 to indicate adequate self-care. Receiver operating characteristic (ROC) curve analyses were performed for each NPM–CI domain. Discriminatory performance was evaluated using the area under the curve (AUC), and sample-specific thresholds were derived using the Youden Index. Pairwise comparisons of AUCs were conducted to assess differences between domains. Results All three NPM–CI domains showed statistically significant but limited discriminatory performance in distinguishing participants who met versus did not meet the prespecified SC-CII self-care management criterion (AUC range: 0.657–0.681; p < .01). Domain-specific cut-off values were derived using the Youden Index. Although sensitivity and specificity varied across domains at their respective Youden-index thresholds, discriminatory performance did not differ significantly between domains, as indicated by pairwise comparisons of the AUCs. Conclusions Anchoring mutuality scores to a behavioral self-care criterion enables the derivation of empirically informed thresholds that improve the interpretability of NPM–CI scores. These preliminary, sample-dependent thresholds may serve as exploratory reference points for interpreting mutuality scores in relation to self-care management. Validation in larger independent samples is required before considering any clinical application.