Method: 8 mandibular arch orthodontic study models of identical form but with varying degrees of crowding were prepared using acrylic teeth that had been measured with digital calipers. The models were digitally scanned to quantify the crowding in each model. 62 orthodontists then examined the models to assess crowding and decide on the need for extractions. 11 of the orthodontists repeated the exercise 9 months later.
Result: All of the orthodontists opted for direct visualisation to estimate crowding, which was highly variable. The digital scan of the most crowded model had an arch length of 96.82mm, which was deducted from the total mesio-distal tooth widths of 104.27mm to give 7.45mm of crowding. Visualised estimates of this model ranged from 5mm to 20mm. Extraction choices were less variable than estimates of crowding and shifted from second to first premolars as crowding increased. Estimates of crowding and treatment decisions changed with time in the case of 28 out of 33 repeat measures.
Conclusion: Although clinicians were often poor at estimating crowding, their extraction decisions were less variable than their estimates of crowding. As crowding increased, so did the likelihood of prescribing extractions and the less variable was the extraction choice. Over time the estimates of crowding varied by up to 3mm for the same orthodontist examining the same case.