Table 3.

Accuracies for predicting PCSI/PedsQL using Models 1–5 and AUC values

Prediction	Models	Correlation		MSE		AUC for classifying PPCS versus recovered
		Train	Test	Train	Test	Train	Test
PCSI prediction	Model 1	$r$=0.47, p = 1.20 × 10–4	$r$= 0.57, p = 2.07 × 10–3	293.96	304.22	0.8	0.71
	Model 2	N. S	N. S	N/A	N/A	N/A	N/A
	Model 4	$r$= 0.41, p = 8.00 × 10–4	N. S	340.28	458.34	N/A	N/A
	Model 5	N. S	N. S	N/A	N/A	N/A	N/A
PedsQL prediction	Model 1	$r$= 0.71, p = 1.76 × 10–10	$r$= 0.59, p = 1.20 × 10–3	120.78	132.67	0.70	0.56
	Model 2	$r$= 0.42, p = 6.53 × 10–4	$r$= 0.50, p = 6.80 × 10–3	137.15	161.60	0.59	0.68
	Model 3	$r$= 0.74, p = 1.72 × 10–11	$r$= 0.71, p = 3.89 × 10–5	106.24	106.75	0.70	0.63
	Model 4	$r$= 0.36, p = 3.90 × 10–3	N. S	167.96	194.41	N/A	N/A
	Model 5	N. S	N. S	N/A	N/A	N/A	N/A

N. S. denotes not significant. Accuracies were reflected by the correlation and MSE value between predicted PCSI/PedsQL and true PCSI/PedsQL values. AUC values were for classifying PPCS versus recovered using the predicted PCSI/PedsQL values