Construct validity was demonstrated by comparing the PHQ-4 to the SF-20, demonstrating that increases in PHQ-4 scores were associated with increases in SF-20 scores (i.e., worsening functional status). Construct validity was further assessed by comparing the PHQ-4 to its component scales, the PHQ-8 (component depressive disorder scale) and the GAD-7 (component anxiety scale). Increases in level of severity reported on the PHQ-4, PHQ-8, and GAD-7 were all associated with worsening functional status (i.e., increases in mean disability days and mean physician visits reported). Additional information is available in Figure 1 and Table 2 of the source article.
Factorial validity was demonstrated by using principal-component analysis to assess the four items on the PHQ-4, demonstrating that the two depression items of the PHQ-2 and the two anxiety items of the GAD-2 explained 84% of the total variance. The anxiety items had the highest factor loadings for Factor 1 (anxiety) and the depression items had the highest factor loading for Factor 2 (depression). Additional information is available in Table 3 of the source article.