(PS4-3) Assessing the Psychometric Properties and Validity of an Ecological Momentary Assessment Version of the PHQ-9 with Sliding Scales

Background: The Patient Health Questionnaire-9 (PHQ-9) is a widely used self-report measure for assessing depression, employing a Likert-type scale to capture symptom presence over the previous two weeks. Ecological Momentary Assessment (EMA) offers an alternative approach to traditional questionnaire administration by using multiple assessments throughout the day to collect near-real-time data. Recently, a modified version of the PHQ-9 administered in an EMA format that replaced the traditional response options with sliding-scale response options was developed. This study compared the psychometric properties, stability, convergent validity, and construct validity of the traditional and modified EMA versions of the PHQ-9.

Method: Depressed participants (n = 275) were followed for 90 days. EMA surveys were administered three times daily. Patients rated symptoms on the modified PHQ-9 EMA surveys on a sliding scale ranging from 0 ("not at all") to 100 ("constantly") over the preceding 4 hours. Pre- and post-assessments included the traditional PHQ-9 and the Inventory of Depression and Anxiety Symptoms-II (IDAS-II). EMA data from the first and last two weeks were analyzed to correspond with the administration of the pre- and post-assessment measures.

Results: Psychometric property analyses revealed stronger item-total correlations for the EMA measure (r = 0.35-0.76) compared to the traditional PHQ-9 (r = 0.33-0.66). Inter-item correlations for the EMA measure (r = 0.15-0.76) were also stronger than those for the PHQ-9 (r = 0.13-0.67). The EMA total score demonstrated greater stability from pre- to post-assessment (r = 0.70) compared to the PHQ-9 (r = 0.36), with a smaller mean difference between assessment scores (5.32/900 vs. 1.79/27). Item-level stability was also higher for EMA items (r = 0.60-0.81) compared to PHQ-9 items (r = 0.27–0.39). Regarding convergent validity, the PHQ-9 post-assessment correlated strongly with the last two weeks of EMA data (r = 0.69), and individual EMA-PHQ item correlations ranged from 0.54-0.69. The EMA measure also demonstrated excellent construct validity, correlating strongly with the general depression subscale of the IDAS-II post-assessment (r = 0.66).

Discussion: The modified EMA version of the PHQ-9 demonstrates good psychometric properties, stability, and convergent and construct validity.