Reliability and Predictive Validity of Dental Arch Relationships Using the 5-Year-Olds' Index and the GOSLON Yardstick to Determine Facial Growth.

Aims: To determine reliability and predictive validity of the 5-year-olds'(5YO) Index and GOSLON Yardstick in 119 patients born with unilateral cleft lip and palate at 5, 7/8, 10, 15/16, and 19 years.
Methods: Five hundred thirty-four dental study models were appraised by 2 teams in 2 centers, twice in each center. Intrateam and interteam reliability in scoring the models was calculated using κ. Dental arch prediction rates were calculated as the proportion of models remaining in the same category (good-scores 1 and 2; fair-score 3; poor-scores 4 and 5) over time.
Results: Intrateam and interteam κ statistics ranged from 0.74 to 0.89 and from 0.74 to 0.81, respectively. The 5YO Index and GOSLON Yardstick at 5 years produced almost identical results. The prediction rate of 19-year-old (n = 106) outcome was >80% for those in groups 1 and 2 at 5 years, while for those in groups 4 and 5 prediction was poor (<40%). Prediction of groups 4 and 5 remained poor until 10 years when it increased to 77%. At 15/16 years prediction rate was 93% for those in groups 4 and 5. Prediction of cases in group 3 was very poor at all ages.
Conclusions: These results question the predictive value of "poor" dental arch relationships before 10 years of age. However, the predictive value of "good" dental arch relationship scores over time is good in all age groups. This has implications for audit policies to predict facial growth outcomes.