Recent substance use among clients with early psychosis and the potential to graduate from New Journeys

This study examined the relationship between recent substance use prior to intake and program graduation among young adults with early psychosis enrolled in coordinated specialty care.


| INTRODUCTION
Coordinated specialty care (CSC) is an early intervention model that improves clinical and functional outcomes (Cather et al., 2018;Nossel et al., 2018;Oluwoye et al., 2020) and has become the benchmark in treatment for individuals experiencing first episode psychosis (FEP) (Heinssen et al., 2014;Lieberman et al., 2013). Yet, up to 40% of individuals enrolled in CSC do not graduate (Doyle et al., 2014). Research has demonstrated that individuals who do not graduate from CSC are more likely to experience symptom relapse and be hospitalized compared to those who do graduate from CSC (Birchwood & Macmillan, 1993;Dixon et al., 2016).
Among those enrolled in CSC, approximately 50% meet criteria for a substance use disorder (Cather et al., 2018;Wisdom et al., 2011) which is associated with poorer functioning, worse symptoms, and low medication adherence (Mazzoncini et al., 2010;Oluwoye et al., 2019;Patel et al., 2016;Schoeler et al., 2017). Research on the association between substance use and disengagement from CSC has been mixed. For instance, continuous substance use throughout treatment has been linked to disengagement. While other studies have found that substance use at intake is not associated with disengagement (Conus et al., 2010;Schimmelmann et al., 2006;Stowkowy et al., 2012;Turner et al., 2007;Turner et al., 2009). While the majority of studies focus on disengagement from CSC, there are inconsistencies in how disengagement has been defined and generally categorizes individuals as disengaged for a period of time (e.g., 2 months), although they may re-engage at a later point in time (Doyle et al., 2014;Reynolds et al., 2019).
This study aimed to address the inconsistencies in prior research by exploring the relationship between recent substance use at the initiation of CSC on the potential to graduate. This study sought to examine whether recent substance use (any), cannabis use, alcohol use, and alcohol-cannabis co-use at intake was associated with CSC graduation across 2 years.

| Study settings and participants
New Journeys is a CSC model that has been implemented in 12 community-based outpatient mental health agencies in Washington, U.S. The model is delivered by a multidisciplinary team and includes the following components: individual therapy, medication management, employment and educational support, family therapy, case management, and peer support. Eligibility criteria and client characteristics are reported in more detail elsewhere (Oluwoye et al., 2020).

| Assessments
Of the 481 clients enrolled, 248 clients completed intake assessment for substance use at time of analysis. From January 2016 until January 2020, clients completed a modified version of the 9-item Car-Relax-Alone-Forget-Family and Friends-Trouble (CRAFFT) Screening to assess recent substance use and related risks (Knight et al., 2002). For those who endorsed lifetime substance use, a subsequent question about recent substance (i.e., last 30 days) was asked. In February 2020, the assessment battery transitioned from the CRAFFT to the Phenx Toolkit for lifetime and recent (i.e., past 30 days) substance use. These measures were merged to assess recent substance use.
Any substance use included clients who indicated "yes" to using at least one of 11 assessed substances (e.g., stimulants, opioids, cannabis, alcohol) in the past 30 days. Cannabis use only included an endorsement of "yes" and "no" to all other substances. Similarly, alcohol use only included clients who endorsed "yes" to recent alcohol use and "no" to all other substances. Alcohol-cannabis co-use included clients who indicated "yes" to using cannabis and "yes" to recent alcohol use. All substance use variables were dichotomized into recent use and no use.
Program statuses were captured and routinely updated by clinicians for each client. Statuses include: 'Active' defined as engagement (attendance) in services or in contact with clinicians; 'No Show' defined as not currently attending scheduled appointments and clinicians are attempting re-engagement; 'Disengaged' defined as inability of clinician to re-engage client in services within 90 days of 'No Show' status and has been discharged from the program; 'Referred' defined as the client was referred to another more appropriate mental health program; and 'Graduated' defined as the client completed the program by reaching self-oriented goals while receiving services and symptoms of psychosis have stabilized to a manageable level. Prior to graduation, clients enter a 3-month 'Active Maintenance' phase, a period with reduced appointments to evaluate stability of symptoms and readiness to transition out of CSC. For the purposes of this study, program status was assessed up to 24 months from enrollment.

| Statistical analyses
Client characteristics were summarized using mean and standard deviation for continuous variables and counts and frequencies for categorical variables. To compare the probability of being in each of the statuses across time ('disengagement' and 'graduated') in the presence of competing risks (program completion and program disengagement), the cumulative incidence function non-parametrically was calculated via the Aalen-Johansen estimator.
To obtain statistical estimates of the effect of substance use, New Journeys progress was analysed using a competing risks model. Causespecific Cox models were fit with, 'active', 'referred', 'graduated', and 'disengaged' as competing statuses. Those who were 'active' and 'referred' to other services were censored. Separate models were fit for different substances use categories: any substance use, alcohol use only, cannabis use only, and alcohol-cannabis use, where the reference group for all models was no substance use. All models were adjusted for covariates; age, sex, and duration of untreated psychosis (DUP), defined as the period of time between the first expression of delusion and/or hallucinations and receiving services for these symptoms. Multiple imputation was used to manage missing values in the covariates.
Coefficients are presented as hazard ratios (HR) with 95% confidence intervals. The proportional hazards assumption of the Cox regression models was assessed for every model fit on all imputed data sets.
Aalen-Johansen estimates and cause-specific Cox models were fit using the survival package and multiple imputation was performed via the mice package. No violations of the proportional hazards assumptions were detected. All computations were performed in R (Version 3.6.3).

| RESULTS
The mean age was 20.7 years (SD = 3.7 years), the majority of clients identified as male (74%) and as racially and ethnically diverse (55%).
The mean DUP was 173.5 days (SD = 188.2 days). Of the 248 clients, 50% reported use of any substance at intake. Among those who reported any substance use, 32% reported alcohol use only, 26% reported cannabis use only, and 15% reported using both alcohol and cannabis within the past 30 days. Figure 1 displays the cumulative incidence function for each substance use type at intake. As seen in Table 1, findings from the Cox model revealed clients who reported using 'any substance' at intake, had a 57% lower chance of graduating from CSC compared to those who did not report any substance use (HR = .43, CI = .21, .88); (p = .02). Similarly, clients who reported alcohol use only at intake had a 72% lower chance (HR = .28, CI = .1, .81); (p = .02) of graduating from CSC and those who reported cannabis use only had a 67% lower chance (HR = .33, CI = .12, 0.87); (p = .03) of graduating from CSC compared to individuals who reported no substance use. All other statuses by substance use models had confidence intervals overlapping a hazard ratio of 1 and were therefore not significant.

| DISCUSSION
Findings from the current study suggest that any substance use reported at intake was associated with a lower chance of graduating from CSC relative to individuals who did not use substances. At intake, approximately 30% of clients reported alcohol or cannabis use at intake, which is consistent with previous studies that have reported substance use (Marino et al., 2020;Oluwoye et al., 2019). Overall, the results highlight that recent alcohol and cannabis use prior to F I G U R E 1 Estimated cumulative incidence functions (Aalen-Johansen estimator) for New Journeys outcomes (graduation and disengagement). Separate panels are presented for separate baseline substance use (any substance, alcohol use, cannabis use, and both alcohol and cannabis use). The top left panel shows that those with any substance use at intake had a higher probability of disengagement and lower probability of graduation across the first 2 years of the New Journeys program. enrollment in services were independently associated with the reduced chance of graduating from CSC and further supports the need for the implementation of additional services that foster engagement at the time of enrollment while also addressing substance use (e.g., health education, harm reduction) (Wisdom et al., 2011). Interestingly, our findings on alcohol-cannabis co-use was not associated with graduation from CSC which could be due to a smaller sample (n = 37) of individuals that reported both alcohol and cannabis use or the severity and frequency of use in comparison to individuals who used only alcohol or only cannabis.
These findings are somewhat aligned with previous studies that highlight continuous substance use and cannabis use as a predictor for disengagement (Conus et al., 2010;Mascayano et al., 2020;Stowkowy et al., 2012). For instance, Mascayano and colleagues found that young adults who reported cannabis use at intake had roughly 1.4 times the hazard for early discharge, defined as leaving prior to reaching 1 year of services, relative to those who did not use cannabis at intake for OnTrackNY, but any substance use was not a significant predictor and alcohol use was not examined (Mascayano et al., 2020). Building on prior research, there is a need to focus on the impact of alcohol use among young adults experiencing psychosis and explore the unique experiences and communitylevel influences that are reinforce substance use and hinder engagement in CSC.
There are several limitations worth noting. The program graduation/completion involved a 3-month maintenance phase, defined above, which may not be generalizable to other programs criteria for graduation/completion. Several reviews have highlighted the lack of a standard definition for disengagement (e.g., no contact for a continuous period of 3 months) or retention (e.g., early discharge), which may not be consistent with how disengagement was defined. Substance use was captured based on self-report which could lead to the underreporting of use and the measure used to capture substance use did not account for frequency (i.e., the number of days and level of alcohol use [e.g., binge drinking]) and should be evaluated in future studies to determine their impact on rates of program completion.

| CONCLUSION
An emphasis on the implementation of evidence-based treatment for substance use within CSC programs is needed to better support and care for individuals with co-occurring FEP and substance use in addition to achieving better outcomes and completion rates within these specialized services.

FUNDING INFORMATION
This work was partially supported by the National Institute of Mental Health (Grant Number K01MH117457 to Oladunni Oluwoye). The funders had no role in study design, data collection and analysis

CONFLICTS OF INTEREST
The authors declare no conflicts of interest.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions. Note: Hazard ratios were derived from competing risks, Cox proportional hazards models. Separate models were fit for each substance, with no substance use as the reference level. Participants still active after 2 years in the New Journeys program were censored. Abbreviations: CI, confidence interval; HR, Hazard ratio.