PREDICTORS OF BEHAVIORAL INTERVENTION ATTENDANCE AND CLINICAL APPOINTMENT SHOW RATE FOR INDIVIDUALS WITH SERIOUS MENTAL ILLNESS AND DIABETES

Purpose:
   To determine factors associated with behavioral intervention attendance (A_RCT) and clinic appointment show rate (A_CLINIC).

Methods:
   This exploratory study focused on a population in which delivering effective personalized care is difficult. Specifically, we evaluated individuals with serious mental illness (SMI) and a co-morbid physical health condition, diabetes mellitus (DM). Baseline covariate data from 197 individuals with SMI-DM enrolled in an NIMH-funded Randomized Controlled Trial (RCT) was used to examine (A) predictors of attendance at a self-management intervention (N=99 for the intervention arm) and (B) predictors of presentation for any outpatient clinic appointments (2011-2013) as recorded in electronic medical records prior to the RCT. A series of generalized linear mixed effects and logistic regression models were used to identify covariates predicting individual level attendance at study sessions and outpatient appointments. Covariates included measures of demographics, social support, physical health, mental health, self-efficacy, anticipated attendance and transportation.

Results:
   The sample consisted of 48% with depression, 27% with bipolar disorder and 25% with schizophrenia. Mean age was 54.0 years (SD 9.4), with 64% women and 53% African-Americans, and median (IQR) number of outpatient clinic appointments over the two-year period was 37.0 (40.0).

   There was no observed association between any of the covariates examined and A_RCT across all twelve weeks. Age (odds ratio [95% CI] of 1.97 [1.11, 3.71] per decade, p=0.017) and mental health self-efficacy (1.54 [1.00, 2.39] per 5 units, p=0.057) were the strongest predictors in multivariate analyses for attendance to the first session only. Attendance to the first session was a strong predictor of attendance over the subsequent 11 sessions (12.5 [9.62,16.30], p < 0.001); See Table). In the multivariate models predicting A_CLINIC, significant factors (p<0.05) included Causian race (vs. African-American), older age, depression SMI diagnosis (vs. bipolar/schizophrenia), increased diabetes knowledge, greater number of co-morbid conditions, less psychotic symptoms, and higher SF36 mental component score.

Conclusions:
   The strongest predictor of future attendance was first session attendance in the RCT. Physical health and transportation measures were not good predictors of A_RCT or A_CLINIC in the SMI-DM individuals.

	Total # of sessions attended
	0	1 to 11	12
Attended First Session	0	10	63
Did Not Attend First Session	16	10	0