Determining the optimal clinical cutoff on the CES-D for depression in a community corrections sample

Academic Article


  • Background: Screening instruments are urgently needed to rapidly and efficiently identify those in need of mental health treatment, particularly among underserved populations. Although designed initially for use in research, the CES-D has become a widely used clinical screening tool for the presence of major depression. Despite four decades and three major revisions to the DSM since the CES-D was first introduced, the cutoff score of 16 remains the marker by which individuals are assessed. The present study aims to examine an optimal cutoff score in a low-income, high-risk sample of ethnically diverse adults involved in some phase of the criminal justice system. The utility of the CES-D to detect depression in this population is unknown as these individuals are unlikely to be included in community studies of mental health. Methods: A diverse sample of participants under criminal justice supervision (n = 500, ages 19–72) completed the CES-D at up to six time points and the MINI-D at two time-points over a year. Results: Using receiver operating characteristic (ROC) curves, a cut point of 21 on the CES-D was determined to produce the best overall screening characteristics (sens = 0.82, spec = 0.76) using an efficiency calculation when compared with the gold standard MINI-D, though these results varied by race and gender. The optimal cutoff for women in this sample was determined to be 23 as opposed to 15 for men, and 20 for nonwhites versus 23 for whites. Limitations: This study is limited in its generalizability to low-income individuals without criminal justice involvement or those without substance use. Conclusion: Results suggest that the CES-D is a useful screening measure for depressive symptoms among high-risk individuals under criminal justice supervision. However, the current cutoff score of 16 is inadequate for optimizing true positives and false negatives. Possible gender and racial/ethnic bias may limit the utility of this instrument in this population. The current study contributes to the understanding of mental health needs in underserved populations.
  • Published In

    Digital Object Identifier (doi)

    Author List

  • Henry SK; Grant MM; Cropsey KL
  • Start Page

  • 270
  • End Page

  • 275
  • Volume

  • 234