Methods: A convenience sample of elderly aged 60 years+ was recruited. Each subject responded to OHIP-14 twice, 1-2 weeks apart, in two randomly assigned groups: First face-to-face, then telephone interview (F-T) or first telephone then face-to-face interview (T-F). Two global measures of perceived oral health status (POHS) and perceived satisfaction on oral health (PSOH) were included. The construct validity and internal consistency of OHIP-14 were assessed by correlating the additive OHIP-14 scores with the two global questions using Spearman rank correlation coefficients (rs) and Cronbach’s alpha, respectively. To measure the agreement of OHIP-14 between the two MOA, only subjects with no difference in POHS or PSOH in the two MOA were included. Kappa coefficients between each item and Altman and Bland 95% limits of agreement (LOA) around the mean difference in OHIP-14 scores between the two MOA were calculated.
Results: Altogether 217 elderly participated (F-T:110; T-F:107). The correlation between POHS, PSOH and OHIP-14 scores (rs:0.54-0.59, all p<0.05) and the Cronbach’s alpha (F:0.88; T:0.87) were similar in both MOA indicating comparable validity and reliability. The correlation of OHIP-14 scores between the two MOA was strong (rs=0.744, p<0.05). For those subjects with no difference in POHS, the Kappa values for corresponding items were 0.22-0.38 and 95% LOA in the difference in OHIP-14 scores were 0.77±1.96x4.96 (-8.95, 10.49) between the two MOA. Similar results were found for those subjects with no difference in PSOH.
Conclusions: Although the psychometric properties of OHIP-14 were comparable with respect to each MOA, substantial individual variability was present when interviewed in both MOA. Further research employing the same MOA in repeated OHIP-14 measurements to assess the magnitude of individual variability is desired.