A preliminary study of patterns of replication in Scoring linear enamel hypoplasias

In order to investigate replicability in scoring linear enamel hypoplasias 59 anterior teeth were evaluated by three observers. Hypoplasias were identified with low‐power microscopy and by the unaided eye. Location on the tooth was determined through caliper and micrometer measurement as well as assignment to tooth ‘thirds’. Hypoplasia size was determined through measurement of episode width and comparison with a cast designed to categorize types of hypoplasia severity. Intra‐observer analysis showed consistent scoring of about 65 per cent of individual hypoplasias between rounds, with high replicability in episode location, size and severity using all methods tested. Inter‐observer error was similarly low for episode size and location, but consistent scoring of individual hypoplasias between observer pairs ranged between 21 and 39 per cent. It is likely that the most important factor in the low level of replicability is the slight size of the majority of hypoplasias involved. Thus, the results suggest that further work should be directed toward developing the minimum threshold for scoring episodes. Other issues include measures to control ‘drift’ in scoring procedures over time, and procedures to score large cervical hypoplasias in canines, which can appear as multiple episodes with magnification. Among the various scoring methods tested, determination of episode location was highly replicable using all three techniques. Episode width was also highly replicable using the micrometer, but the cast association proved less successful.