Table 2

Variables which best discriminate between positive and negative genetic diagnoses of PWS

Classification results
Predicted group
Action step entered removedVars inWilks’ LambdaSigActual groupNo of cases12
(A) PWS +ve genet v control 50 PWS and 15 control had full data
    1 Floppy at birth10.148630.0000Group 195950
    2 Weak cry/inactive20.099960.0000100%0%
    3 Poor suck at birth30.083880.0000Group 226125
    4 Childhood obesity40.074390.00003.8%96.2%
Ungrouped19145
73.7%26.3%
Percent of “grouped” cases correctly classified: 99.2%
(B) PWS +ve genet v PWS -ve genet combined with control 50 PWS +ve, 4 PWS -ve and 15 control had full data
    1 Poor suck at birth10.289160.0000Group 189872
    2 Weak cry/inactive20.178860.000097.8%2.2%
    3 Vomiting30.167210.0000Group 238830
    4 Thick saliva40.157560.000021.1%78.90%
Ungrouped761
85.7%14.3%
Percent of “grouped” cases correctly classified: 92.1%