Table 1 

Validity of remission by approximation in major depression 

Validation Method 
Reference Criteria 

Remission by Approximation 
Clinical Histories 

N = 133 
Positive 
Negative 


Positive 
74 
5 
Negative 
6 
48 


Statistics 
Value (%) 
95% CI 


Validity of the measurement 

Sensitivity 
92.5 
88.0  96.9 
Specificity 
90.6 
85.6  95.6 
Positive predictive value 
93.7 
89.6  97.8 
Negative predictive value 
88.9 
83.6  94.2 
False positives 
7.5 
3.0  11.9 
False negatives 
9.4 
4.4  14.4 


Reliability 

Cronbach's alpha 
90.6 
85.6  95.6 
Area under the curve 
91.2 
86.5  96.1 


Concordance 

McNemar test 
58.1 
49.7  66.5 
Pearson correlation 
82.8 
73.1  92.5 
Weighted kappa (Cohen) 
82.8 
73.1  92.6 


Clinical utility 

Positive Probability Ratio 
9.8 
 
Negative Probability Ratio 
0.1 
 


Reference criteria: Review of patients' histories. Significance: p < 0.001 in all cases. CI: confidence intervals 

SicrasMainar et al. BMC Public Health 2010 10:64 doi:10.1186/147124581064 