The purpose of this study was to assess a simple, visual tool to determine if it could measure a child's own perception of the appearance of their teeth – and in particular their own of enamel opacities. The results clearly demonstrate (see Table 2) that there is an association between the self-perception of the presence of opacities and the choice of images from the tool, in the direction one would expect. Children who consider themselves to have white marks on their teeth are more likely to select those photographs with more severe representations of mottling. The same coherence was found among the examiners’ measurement using the same tool but at a lower level of severity than for the volunteers’.
Repeatability appeared to be good with regard to selection of images for pupils and very good for examiners. With only three sets of images from which to select and with very little difference between them it might have been expected that the volunteers would simply pick a set of images at random but the Kappa score of 0.65 for repeat examinations shows that this was not the case. Changing the order and the labelling of the grouped images for the second time of questioning ensured there could be no accusation of simply picking the images in the same position as the first occasion.
When considering individual responses there was poor agreement  between the examiner’s selection of images and those of the individual volunteers. This illustrates the concerns raised about other indices of enamel opacities which are measured by trained, clinical professionals and which may over- or under-estimate the prevalence, severity and impact of enamel opacities . Years of training and an underlying understanding of the biological basis for enamel opacities among examiners is in contrast to the subjective (but authentic) view of an individual based on their feelings about appearance, aesthetics and acceptability . An additional point is that individual variances between examiners and pupil volunteers are unlikely to be of importance within epidemiological surveys where data are reported at group level and the purpose of such studies is to gain a view on population impact. Using responses from individual volunteers in this way is further supported when one considers the good repeatability scores achieved by the subjects when using this tool .
In the national survey similar results were found to those reported here for a subset of PCOs . Some variations between PCOs were found which suggest some results should be viewed with caution. For example, in one or two PCOs an unusually high proportion of pupils were reported to not know which images to select as matching with their own appearance and in other PCOs a high proportion chose not to answer one or more of the self-perception questions. This might suggest that there has been unusual administration of the tool that has lead to volunteers not engaging with the activity. This may be an issue that can be addressed through better training to ensure improved compliance with the correct method of eliciting responses.
While this study attempted to reduce the various biases that can occur in this type of survey, such as selection bias by using a random sample, order bias by scrambling the arrangement of photos to be chosen by volunteers, there are some areas that could not be lessened. For example, an issue within this type of survey is that simply asking a question about white marks may incline a participant to consider this when they never had before. Additionally, in regards to the limitations of this study, as it is purely a descriptive observational survey it cannot be used to determine the construct validity of the tool at this stage, especially given the lack of an agreed reference standard. Qualitative work would be required to gain further understanding of the different selections made and to establish if this tool really reflects how the participant feels regarding the aesthetic appearance of their teeth in relation to enamel opacities.
Tools used in epidemiological studies should be simple, efficient, and economical. They should not require complicated or expensive equipment and should be impervious, if possible, to the range of settings within which such studies are undertaken. The tool described within the current work meets these objectives and can be utilised in multiple sites simultaneously and by a number of epidemiological teams within the same site. The tool was found to be easy to implement and understood by the volunteers and requires limited training and no calibration of examiners.