Three factors predicting irregular versus regular dental attendance: a model fitting to empirical data.