Objectives: The 14-item Oral Health Impact Profile (OHIP-14) is the most frequently used instrument to measure oral health-related quality-of-life (OHRQoL) in adults. Despite its popularity, its psychometric properties have been predominantly investigated based on the classical test theory while the fundamental principles of measurement have not been fully assessed. Therefore, our aim was to investigate to what extent the OHIP-14 meets the fundamental requirments of measurements.
Methods: We used the Rasch model to explore person-item-targeting, unidimensionality, local independence of items, invariance (differential-item-functioning, DIF), and the order of thresholds between response-options in the German version OHIP-14. We used data from osteoarthritis patients because hand disabilities and joint pain might influence oral hygiene. Furthermore, osteoarthritis in the temporomandibular-joint directly affects oral functioning.
Results: Five-hundred sixteen patients were included (mean age 66.5 years [±10.2; ranging from 34 to 89]; 71.3% [368] females). The OHIP-14 median total score was 0 (interquartile-range from 0 to 4), indicating a right-skewed distribution because many patients reported good OHRQoL. The instrument was found unidimensional. However, there was strong evidence of local dependency, disordered thresholds between response-options, and age-related DIF for item 5. A revised scoring scheme with three instead of five answer-options in all items and eliminating two items resolving local dependency, the newly adapted OHIP-12, showed better reliability and item-fit to the Rasch model than the original OHIP-14.
Conclusions: This study assesses, for the first time, the OHIP-14 in terms of fundamental principles of measurement and proposes an item-reduced OHIP-12 as a psychometrically more accurate version of the instrument.
Clinical significance: The Rasch model is essential to ensure instruments' precision and clinical meaningfulness when measuring OHRQoL in clinical practice and research. The OHIP-12, derived from the OHIP-14 by deleting two items due to local dependency, with a revised scoring scheme for all items distinguishing three answer-options instead of five, represents a psychometrically improved version of the instrument.
Keywords: Evidence-based dentistry; Measurement accuracy; Patient-reported outcome measures; Psychometric properties; Quality-of-life; Rasch model.
Copyright © 2021. Published by Elsevier Ltd.