Determinants of the access to remote specialised services provided by national sarcoma reference centres

Table 4 Adequacy and discrimination parameters of the different logistic regression models (source: IGéAS cohort, RRePS – ResOs – NETSARC databases)

Models	Details	AIC	Well-ranked %	AUC
Optimal access to diagnosis	All variables	23,258	65.3	0.65
Optimal access to diagnosis	Clinical variables	24,149	59.8	0.59
Optimal access to diagnosis	Geographical variables	24,116	59.6	0.60
Optimal access to MTB	All variables	18,852	76.2	0.76
Optimal access to MTB	Clinical variables	18,910	75.8	0.75
Optimal access to MTB	Geographical variables	22,050	53.9	0.55

AIC (Akaike Information Criterion): The model to choose has the smallest AIC
Well-ranked %: The model to choose has the highest %
AUC (Area Under the Curve, from 0 to 1): The model to choose has the highest value