Measuring long-term psychiatric outcomes in post-acute autoimmune encephalitis.

PURPOSE: To compare the performance of different measures of long-term psychiatric outcomes in patients with post-acute autoimmune encephalitis (AE) who may require comprehensive psychiatric evaluation.

METHODS

The sensitivity of three self-reported measures of mood and anxiety symptoms (Patient Health Questionnaire [PHQ-9]; Profiles of Mood States-2 [POMS-2]; Generalized Anxiety Disorder 7-item [GAD7]) was compared with a structured clinician-administered tool (Mini Neuropsychiatric Inventory 7.0.2 [MINI]). New cutoff scores that optimized accuracy were then identified by Youden Index Method.

RESULTS

Thirty-five patients with post-acute AE completed testing a median of 3 years after symptomatic onset (range = 1-22 years). The median PHQ9 score was 5 (range = 0-18), median POMS2 Total Mood Disturbance T-Score was 52 (range = 37-93), and median GAD7 score was 3 (range = 0-17). Twenty-five patients (71 %) met criteria for a psychiatric diagnosis on the MINI. When compared with the MINI, the sensitivity and specificity of the self-reported psychiatric symptom tools using standard cutoffs were 25 % and 80 % for the PHQ9, 50 % and 87 % for the POMS-2, 23 % and 91 % for the GAD7. Accuracy was improved when cutoffs of ≥5 for the PHQ9, ≥50 for the POMS2, and ≥ 3 for the GAD7 were used, at the cost of lower specificity.

CONCLUSIONS

Patients with post-acute AE with psychiatric sequalae may be underrecognized if self-reported measures of psychiatric symptoms are used instead of clinician-administered structured interviews. If self-reported measures are used in AE, consideration should be given into using tools with higher validity in this patient population, such as the POMS-2.