Skip to main content
. 2020 Nov 3;52(10):1875–1882. doi: 10.1017/S0033291720003700

Table 5.

Estimated threshold score for discriminating between feeling better and feeling the same for the PHQ9 BDI-II and GAD-7 scales according to baseline severity and related ROC parameters

Instrument Scale Severity Threshold score Threshold as % of baseline 95% change as % of baseline Spec(1) Sens(2) AUC(3) Mean (2.5% 97.5%)
PHQ9 ≤11 −2.0 48.2 (65.1 37.1) 0.78 0.35 0.57 (0.54 0.60)
12–19 −1.7 21.3 (27.9 16.7) 0.59 0.51 0.57 (0.53 0.62)
20+ −2.4 19.7 (24.2 16.4) 0.63 0.52 0.61 (0.58 0.64)
BDI ≤11 −5.0 51.7 (66.6 41.6) 0.80 0.36 0.59 (0.55 0.63)
12–19 −3.5 23.8 (30.9 18.6) 0.65 0.50 0.60 (0.55 0.65)
20+ −4.4 19.7 (23.4 16.9) 0.65 0.51 0.61 (0.58 0.65)
GAD-7 ≤11 −2.2 72.1 (97.3 54.8) 0.78 0.32 0.55 (0.53 0.59)
12–19 −1.5 26.7 (36.2 20.7) 0.51 0.62 0.59 (0.55 0.64)
20+ −0.8 8.9 (10.9 7.3) 0.55 0.54 0.57 (0.54 0.59)
(1)

Probability (improvements/reductions smaller than MCID when feeling the same).

(2)

Probability (improvements/reductions larger than MCID when feeling better).

(3)

Probability the improvement (reduction) in scores for a randomly chosen patient drawn from those reporting feeling the same is smaller than for a randomly chosen person drawn from those reporting feeling better.