Reliability of interview-based ADL ability measures in older adults obtained by occupational therapists, physical therapists, and nursing staff

Abstract

Purpose: To examine the reliability of ADL interview (ADL-I) ability measures when administered by different health professionals (HPs). Materials and methods: Older adults with stable ADL ability were invited to participate in three ADL-I interviews, administered by occupational therapists (OTs), physical therapists (PTs), and nursing staff (NS), respectively. Methods based on classic and modern test theory were applied. Results: Overall, n = 36 older adults and n = 11 HPs participated. Intraclass Correlation Coefficients were acceptable for research purposes (> 0.7), but not for clinical use. Mean differences in ADL ability measures were significant when comparing measures based on interviews by OTs to measures based on interviews by PTs and NS. Further, in 25 to 47% of the individuals, ADL-I ability measures differed significantly across HPs. Limits of Agreement revealed that measures based on interviews by OTs were systematically lower compared to measures based on interviews by PTs and NS. Four ADL-I items displayed Differential Item Functioning based on HP. Conclusion: When using the ADL-I, it is recommended to only involve one profession, to increase the reliability of measures. Results highlight the importance of evaluating reliability of measures based on instruments administered across HPs before implementation in rehabilitation practice and research.