We explored three key aspects of the Geriatric Oral Health Assessment Index (GOHAI): consistency of recalling the presence of problems across time intervals; test–retest reliability of frequency ratings; and the extent to which problem severity, frequency, and their combination explain general health-related quality of life (GHRQoL). This cross-sectional study was conducted in Budapest between May 2023 and February 2024 in urban clinical and community settings, recruiting a mixed sample (n = 306) with diverse oral health status. GOHAI was re-administered after one week to 108 participants with stable oral health. When participants reported problems on any GOHAI item, they were asked whether these occurred in the past week, the past month, or whether they had persisted for more than three months. Recall inconsistencies were assessed across five time categories. Rating consistency was compared across three scenarios with different frequency response options. We also examined how problem severity and frequency contribute to OHRQoL: we predicted GHRQoL (EQ VAS) via frequency-based (GOHAI or OHIP-14) and severity-based (Oral Health Single-Question; OH-SQ) OHRQoL measures using multiple regression. At least one problem was reported by 88.6% of respondents, and frequency ratings showed 15.4% inconsistency in the retest sample. Recall inconsistency ranged between 7.4% and 11.0% across time intervals and was lowest when no problems were present. Simple count ratings and the “presence within 3 months” criterion were more consistent than five-level frequency ratings. Compared to demographics alone (R² = 0.112), adding ADD-GOHAI improved EQ VAS prediction (R2 = 0.207, ΔR² = 0.095, p < 0.001), and adding OH-SQ further increased explained variance (R² = 0.267, ΔR² = 0.155; p < 0.001). Reducing response options and combining frequency- and severity ratings may enhance reliability and sensitivity. A one-week recall period appears optimal. These findings support the development of streamlined, patient-centered OHRQoL instruments.
Oszlánszky et al. (Mon,) studied this question.