Accuracy and validity of space analysis and irregularity index measurements using digital models

Abstract

Abstract Background Digital copies of study models may avoid the storage and retrieval issues of plaster study models, but measurements made on digital models may not be as accurate as measurements made on traditional study models. Aim To determine the reliability and validity of tooth size–arch length discrepancies (TALD), irregularity indices and arch lengths (four- and six-segment analyses) measured directly on study models with digital calipers with the same measurements measured on digital copies of the study models with proprietary software. Methods The irregularity indices and TALDs (four- and six-segments) were measured on 50 sets of pretreatment plaster models. The plaster models were measured using manual calipers with a digital readout. The models were then couriered to OrthoCAD and digital copies emailed to the authors. The digital models were measured with the proprietary software provided with the digital models. Repeat measurements of the TALDs and the irregularity indices were subjected to intraclass correlations (ICC) to assess the reliability. The least squares means of variation was used to assess validity and the impact of measuring arch length (four- and six-segments) on the digital models, and the implications on the TALDs. Results There were high correlations (ICC) ranging from 98.6–99.9 per cent for both the irregularity indices and the TALDs. The choice of manual over computer and four-segment over six-segment analysis had a significant effect when measuring lower arch lengths (p < 0.05), but they had no effect on the upper arch findings. Conclusions Reliable measurements of the irregularity index and the TALD can be made on digital models. Computer measurements of TALDs on digital models were more consistent than manual measurements of TALDs on plaster models. Six-segment analyses of lower arch lengths on digital and plaster models gave more consistent findings than the four-segment analyses.