Abstracts

Rationale: Prior studies show that 20-30% of patients diagnosed with psychogenic nonepileptic seizures (PNES) fail to attend their first appointment following diagnosis. Clinical experience and studies of psychotherapeutic interventions with low rates of completion suggest that non-adherence worsens after the first appointment. The rate of drop-out at subsequent appointments and the factors correlating with non-adherence among patients with PNES have received very limited attention. Methods: We prospectively followed 123 consecutive patients newly diagnosed with PNES, from March 2013 through May 2016. Patients are referred to psychotherapy for a period of at least 12 weeks, and for a minimum of 4 regular visits with a supervising neuropsychiatrist. Starting after the initial diagnosis, visits with the neuropsychiatrist are scheduled at 1, 1.5, 1.5, and then 6 month intervals. Patients who miss an appointment are encouraged to reschedule for a new date. Patients are considered to become non-adherent with psychiatric treatment if and when they fail to attend an appointment within respectively a 2, 3, 3, or 9 month interval (for a total follow-up time of 17 months). Survival analysis is performed. Baseline data collected at the time of diagnosis includes demographics, PNES event frequency, delay to diagnosis from symptom onset, prior diagnosis of PNES, psychiatric comorbidities, history of abuse, healthcare resource utilization, quality of life (Quality of Life in Epilepsy-10 or QOLIE-10), and illness perception (Brief Illness Perception Questionnaire or BIPQ). Covariates are correlated with time to non-adherence using Cox proportional hazard regression analysis. Results: Baseline characteristics are shown (table 1). Survival analysis shows that 80% of patients remain adherent at psychiatric follow-up appointment 1, 42% attend appointment 2, 24% continue through appointment 3, and only 14% are still adherent at appointment 4 (figure). Univariate Cox proportional hazard regression identifies several covariates which correlate with time to non-adherence with p < 0.15: BIPQ, prior diagnosis of PNES, current substance abuse, and a history of physical abuse. When these covariates are entered into a multivariate model, controlled for gender and age, a higher BIPQ score (indicating greater concern about one's illness) correlates with lower risk of non-adherence (hazard ratio 0.81 per 10 point increment, p-value 0.008), and a prior diagnosis of PNES correlates with higher risk of non-adherence (HR 1.6, p 0.046). A history of current substance abuse or past physical abuse both trend toward non-adherence but do not reach statistical significance (table 2). Conclusions: This prospective cohort study shows that non-adherence with psychiatric treatment among patients with PNES continues and dramatically worsens after the first psychiatric appointment following diagnosis. While the drop-out rate at the first follow-up appointment in our population (19%) is consistent with the 1st appointment follow-up rate in other studies, survival analysis shows that subsequent drop-out rates increase with consecutive appointments and over the course of 1.5 years of follow-up reach 86% -- much higher than previously appreciated. Additionally, a lesser degree of concern about one's illness and a history of having previously received a diagnosis of PNES correlate with non-adherence. These results highlight the need for establishing comprehensive clinical programs that specifically focus on treatment retention. Also, as we understand what factors predict non-adherence to treatment, more targeted interventions can be deployed to specific PNES subpopulations. Funding: Benjamin Tolchin received research support from the AAN/ABF.