IADR Abstract Archives
Full-scale Psychometric Assessment of the OHIP-14 in 291 Pregnant Women
Objectives: Background: The oral health impact profile 14-item version (OHIP-14) has been highly influential as a measure for oral health-related quality of life (OHQoL) since its publication in 1997. However, intensive empirical studies on its psychometric characteristics is sparse, especially among pregnant women.
Objectives: To fully investigate the psychometric characteristics of the OHIP-14 among pregnant women.
Methods: Methods: We collected the OHIP-14 data from 291 pregnant women in our two clinics. Then we assessed the OHIP-14’s psychometric characteristics at the item-, dimension-, and measure-level, including its factor structure by confirmatory factor analysis (CFA) and exploratory factor analysis (EFA).
Results: Results: At item-level, all 14 items had severely skewed distribution among the five response categories (from “0 = never” to “4 = very often”) with a median as “0.0”. Two items (#2 and #14) had spearman correlation with the “overall condition of oral health” less than 0.1 (“small”). At dimension-level, all pairs of items within each of the original seven dimensions correlated well, with Spearman correlation coefficients ranged from 0.47 to 0.62. But the Cronbach’s alphas ranged from 0.64 to 0.77, indicating suboptimal reliability. At measure-level, CFA results did not support the original conceptual 7-dimensional factor structure of the OHIP-14. The results from EFA indicated that at most four factors can be extracted from the 14 items, and that a 3-dimensional factor structure best-fitted the data. Finally, another CFA confirmed this new 3-dimensional factor structure: Root Mean Square Error of Approximation (RMSEA) = 0.044, CFI = 0.993, TLI = 0.992, Standardized Root Mean Square Residual (SRMR) = 0.039. And the three new dimensions had Cronbach’s alphas as 0.92, 0.72, and 0.71, indicating adequate reliabilities.
Conclusions: Conclusions: The findings from this study suggest that the original OHIP-14 needs modifications at the item-, dimension-, and measure-level, for better interpretation and application in pregnant women.
Objectives: To fully investigate the psychometric characteristics of the OHIP-14 among pregnant women.
Methods: Methods: We collected the OHIP-14 data from 291 pregnant women in our two clinics. Then we assessed the OHIP-14’s psychometric characteristics at the item-, dimension-, and measure-level, including its factor structure by confirmatory factor analysis (CFA) and exploratory factor analysis (EFA).
Results: Results: At item-level, all 14 items had severely skewed distribution among the five response categories (from “0 = never” to “4 = very often”) with a median as “0.0”. Two items (#2 and #14) had spearman correlation with the “overall condition of oral health” less than 0.1 (“small”). At dimension-level, all pairs of items within each of the original seven dimensions correlated well, with Spearman correlation coefficients ranged from 0.47 to 0.62. But the Cronbach’s alphas ranged from 0.64 to 0.77, indicating suboptimal reliability. At measure-level, CFA results did not support the original conceptual 7-dimensional factor structure of the OHIP-14. The results from EFA indicated that at most four factors can be extracted from the 14 items, and that a 3-dimensional factor structure best-fitted the data. Finally, another CFA confirmed this new 3-dimensional factor structure: Root Mean Square Error of Approximation (RMSEA) = 0.044, CFI = 0.993, TLI = 0.992, Standardized Root Mean Square Residual (SRMR) = 0.039. And the three new dimensions had Cronbach’s alphas as 0.92, 0.72, and 0.71, indicating adequate reliabilities.
Conclusions: Conclusions: The findings from this study suggest that the original OHIP-14 needs modifications at the item-, dimension-, and measure-level, for better interpretation and application in pregnant women.