Development, validation, and item bias assessment of the Self-Care Inventory-Short-Form among racially and income-diverse adolescents living with type 1 diabetes and their caregivers.

OBJECTIVE

Youth with type 1 diabetes and their families engage in complex health behaviors to help manage blood glucose levels and to reduce the risk of serious health complications. Given well-documented health disparities in pediatric diabetes, rapid and equitable assessment of diabetes self-care can help identify potential areas where support is needed. This purpose of this study was to (a) develop a short-form of the Self-Care Inventory-Revised (SCI-R) and (b) evaluate validity, reliability, and differential item functioning (DIF) across several socio-illness-demographic variables of the full and short-form SCI-R.

METHODS

Participants were 181 adolescent-caregiver dyads from the Predicting Resiliency in Youth with Type 1 Diabetes (PRYDE) study who completed a 15-item version of the SCI-R. Adolescents (age: M = 14.64, SD =1.70, range 12-18 years), identifying as female (n = 92), male (n = 92), Black/AA (57%), and/or White (43%), also completed diabetes-specific measures of stress and Health-related Quality of Life (T1D-HRQoL). Youths' most recent hemoglobin A1c (HbA1c) values were extracted from their medical records.

RESULTS

Initial items selected based on the six highest factor loadings of 15 were evaluated using DIF analyses. The final 6-item youth and caregiver SCI-SF6 totals correlated with HbA1c, stress, and T1D-HRQoL. They were invariant across participant dyads at the scalar level (root mean squared error of approximation =.077, 90% CI: .056-.098; comparative fit index =.971, standardized root mean square residual = .0490). DIF analyses failed to reject the null hypothesis for item bias factor loadings or item thresholds across age, family income, HbA1c, racial category, gender, and illness duration.

CONCLUSIONS

The results support initial evidence of equitable assessment, measurement invariance, validity, and reliability for SCI-SF6 across important socio-illness-demographic variables.