Table 3. Sensitivity, specificity, positive and negative predictive values, and regression statistics of week two diagnoses to predict week nine depression and anxiety.

Week two predictor	Regression statistics			Positive predictive value	Negative predictive value	Sensitivity	Specificity	% correctly identified
	χ2	p	Odds ratio
Outcome: Depression at week nine
ASD (DSM-5)	14.18	<.0001	5.04	0.54	0.81	0.31	0.92	77.6
- 4+ symptoms	20.64	<.0001	5.01	0.47	0.85	0.52	0.82	75.1
ASD (DSM-IV)	14.21	<.0001	4.36	0.49	0.82	0.38	0.88	76.1
‘Two-week PTSD’ (DSM-5)a	18.40	<.0001	5.40	0.53	0.83	0.40	0.89	77.6
Preschool PTSD	24.33	<.0001	6.53	0.55	0.84	0.46	0.89	78.5
Outcome: Anxiety at week nine
ASD (DSM-5)	25.60	<.0001	11.92	0.46	0.93	0.52	0.92	86.8
- 4+ symptoms	22.84	<.0001	8.50	0.32	0.95	0.68	0.80	78.5
ASD (DSM-IV)	21.71	<.0001	8.69	0.38	0.93	0.56	0.87	83.4
‘Early PTSD’ (DSM-5 PTSD)	22.54	<.0001	9.14	0.39	0.93	0.56	0.88	83.9
Preschool PTSD	20.84	<.0001	5.46	0.40	0.95	0.64	0.87	83.9

Note. Positive predictive value = likelihood that someone with diagnosis at two would have the relevant diagnosis at week nine. Negative predictive value = likelihood that someone without a given diagnosis at week two would not go on to have the relevant diagnosis at week nine. Sensitivity = likelihood that someone with diagnosis at week nine would have previously met criteria for relevant diagnosis at week two. Specificity = likelihood that someone without diagnosis at week nine would not have met criteria for the relevant diagnosis at week two.

^a‘Two-week PTSD’ this refers to an ‘early PTSD’ algorithm, i.e. ignores the requirement that symptoms be present for at least four weeks.