(PS7-23) Rates and Predictors of Treatment Noncompliance in Partial/Residential OCD Treatment

Abstract Text: Introduction
About 20% of patients do not adhere to exposure and response prevention (ERP) therapy for OCD (Ong et al., 2016). Some studies find certain symptoms and comorbidities predict noncompliance, while others find no predictors (e.g., Dowling et al., 2016; Leeuwerik et al., 2019). However, few studies have examined noncompliance in partial/residential treatment programs for OCD. We aimed to identify rates of noncompliance at this program and test a variety of clinical variables as predictors of noncompliance. We hypothesized that higher OCD and depression severity would predict noncompliance and investigated other variables using an exploratory approach.

Methods
Data was collected from adult patients (N = 247; M_age = 29.2, SD = 11.1; 80.2% non-Hispanic White; 43.3% men, 52.2% women, 4.5% nonbinary) receiving intensive OCD treatment (M_Y-BOCS = 24.9, SD = 5.3) from 2020-2024. The following variables were collected at admission: Yale-Brown Obsessive-Compulsive Scale (Y-BOCS), Obsessive Beliefs Questionnaire, Hoarding Rating Scale, Dimensional Obsessive Compulsive Scale, Hamilton Depression Scale, Generalized Anxiety Disorder 7-item, Distress Tolerance Scale, Anxiety Sensitivity Scale, and Treatment Credibility and Expectancy Questionnaire. Diagnoses assigned by clinicians were used to indicate the presence of certain comorbidities (e.g., depressive disorders).

Treatment noncompliance was measured as a binary variable (treatment response vs. refusal/noncompliance) using the reason for discharge indicated by clinicians. Patients who discharged for other reasons (e.g., insurance issues) were excluded from analyses due to inconclusive information about compliance. Logistic regressions and chi-squared tests were conducted to examine whether each variable individually predicted noncompliance. Significant predictors were then combined into a multiple logistic regression.

Results
In our sample, 50.6% of patients discharged due to treatment response (M_Y-BOCS = 24.9), 21.5% reaching program length limit (M_Y-BOCS = 27.3), 9.2% noncompliance (M_Y-BOCS = 24.9), 4.4% insurance ending coverage (M_Y-BOCS = 24.4), and 14% for other reasons (M_Y-BOCS = 26.3). Contamination/washing symptoms (OR = 1.04, p = 0.04) and symmetry/ordering symptoms (OR = 1.09, p = 0.02) individually predicted noncompliance over treatment response. When combined into one multiple logistic regression, only symmetry/ordering symptoms (OR = 1.01, p = 0.03) remained significant.

Conclusions
In a combined analysis of all significant predictors, only symmetry/ordering symptoms remained significant and increased the odds of noncompliance by merely 1%, following a prevailing trend in the literature of clinical variables being poor predictors of noncompliance. That symptom dimensions/severity do not predict noncompliance well suggests intensive ERP treatment can be accessible to patients of varying clinical presentations.

Interestingly, only 51% of patients discharged due to treatment response. A large portion (40%) discharged due to insurance, program length limits, or other logistical reasons before reaching their treatment goals. Future research should examine why these patients do not meet their goals within the program.