Related Experiment Video
Updated: Jul 4, 2026
A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
Improving the efficiency of screening for major depression in people with spinal cord injury
Daniel E Graves1, Charles H Bombardier
1Department of Physical Medicine and Rehabilitation, Baylor College of Medicine, Houston, Texas, 77030, USA. dgraves@bcm.tmc.edu
Background/Objective:
To investigate the metric properties, relative efficiency, sensitivity and specificity, and positive predictive value of a short form of the Patient Health Questionnaire-9 (PHQ-9) that may be used as a screening test for depression.
Methods:
Data from the National Spinal Cord Injury Statistical Center Database containing 3,652 records with complete data for the PHQ-9 were analyzed using Confirmatory Factor Analysis, Item Response Theory Graded Response Model analysis, and sensitivity and specificity analysis of classification.
Results:
A scale comprised of items 1, 2, and 6 from the PHQ-9 has a relative efficiency of 0.66 compared to the 9-item scale. Using this 3-item scale and a cutoff score of 3 or more provides specificity of 0.93 and sensitivity of 0.87; a cutoff of 4 provides specificity of 0.95 and sensitivity of 0.82. The shorter version of the scale reduces the effect of response bias caused by gender. The relative efficiency of the 9-item scale is 0.88 for women compared to men; the 3-item scale increases the relative efficiency to 0.93.
Conclusion:
A 3-item scale provides adequate information for clinical screening purposes. Cutoff scores of either 3 or 4 are acceptable and present options for decision making within a particular clinical setting. Additionally, the 3-item scale reduces the effect of gender of the respondent on the score obtained.
