Factor Structure and Predictive Validity of Somatic and Nonsomatic Symptoms From the Patient Health Questionnaire-9: A Longitudinal Study After Spinal Cord Injury

Objective: To investigate the factor structure and predictive validity of somatic and nonsomatic depressive symptoms over the first 2.5 years after spinal cord injury (SCI) using the Patient Health Questionnaire-9 (PHQ-9). Design: Somatic and nonsomatic symptoms were assessed at baseline during inpatient hospitalization (average of 50 days after onset) and during 2 follow-ups (average of 498 and 874 days after onset). Setting: Data were collected at a specialty hospital in the Southeastern United States and analyzed at a medical university. We performed time-lag regression between inpatient baseline and follow-up somatic and nonsomatic latent factors of the PHQ-9. Participants: Adults with traumatic SCI (N=584) entered the study during inpatient rehabilitation. Interventions: Not applicable. Main Outcome Measure: PHQ-9, a 9-item measure of depressive symptoms. Results: The inpatient baseline nonsomatic latent factor was significantly predictive of the nonsomatic (r=.40; P=.000) and somatic latent factors at the second follow-up (r=.29; P=.006), whereas the somatic factor at inpatient baseline did not significantly predict either factor. In contrast, when regressing latent factors between the 2 follow-ups, the nonsomatic factor predicted only the nonsomatic factor (r=.66; P=.002), and the somatic factor predicted only future somatic symptoms (r=.66; P=.000). In addition, the factor structure was not stable over time. Item analysis verified the instability of somatic items between inpatient baseline and follow-up and also indicated that self-harm at inpatient baseline was highly predictive of future self-harm. Conclusions: Nonsomatic symptoms are better predictors of future depressive symptoms when first assessed during inpatient rehabilitation, whereas somatic symptoms become stable predictors only after inpatient rehabilitation. Self-harm (suicidal ideation) is the most stable symptom over time. Clinicians should routinely assess for suicidal ideation and use nonsomatic symptoms when performing assessments during inpatient rehabilitation.