First-Episode Psychosis and Disengagement From Treatment: A Systematic Review
Abstract
Objectives
This review examined rates and definitions of disengagement among services for first-episode psychosis (FEP) and identified the most relevant demographic and clinical predictors of disengagement.
Methods
A comprehensive search for and review of published studies that reported rates and predictors of disengagement within FEP services were conducted. The databases PubMed (1966–2012) and PsycINFO (1882–2012) were searched. Relevant terms were used to search the Internet and the reference sections of relevant papers for other pertinent studies. Independent searches for recent publications by leading researchers in the field were also conducted.
Results
Ten articles were included in the review. There was a lack of consensus on a clear definition of engagement and disengagement. However, despite differences in definitions and study settings, the evidence reviewed indicates that approximately 30% of individuals with FEP disengage from services. Variables that were consistently found to exert an influence on disengagement across studies were duration of untreated psychosis, symptom severity at baseline, insight, substance abuse and dependence, and involvement of a family member.
Conclusions
Given the importance of continuity of care for FEP, there is a need for a clearly defined and agreed measurement of service engagement and disengagement across FEP services. In particular, those who enter an FEP program without family involvement and support as well as those who maintain persistent substance abuse are at higher risk of disengagement. Early identification of such individuals and the development of approaches to reduce risk of service disengagement are likely to increase the effectiveness of these services.
Much research on psychosis has focused on early detection and the development of effective interventions. However, the effectiveness of any intervention depends on the willingness of the patient to engage with an intervention in a sustained manner. Disengagement from treatment by patients with serious mental illness is a major concern of mental health services. Up to one-third of individuals with serious mental illnesses who have had contact with a mental health service disengage from care (1). Studies indicate that long-term treatment can improve symptoms and functioning and reduce relapse risk among individuals experiencing a first episode of psychosis (2–4). Therefore, disengagement from or nonadherence to treatment is of particular concern (5).
This article examines engagement with intervention services for first-episode psychosis (FEP). FEP intervention services were set up internationally in response to evidence that effective intervention during the early stages of a first psychotic episode could have a significant impact on the course and outcome of the illness (6,7).
The effectiveness of an early intervention program is tied to the quality of its constituent parts, but the program’s benefits are also likely to be influenced by adherence to treatment and service disengagement (8). One of the principles of best practice in an FEP service is a specific focus on therapeutic engagement (7). Furthermore, monitoring and targeting engagement levels are important measures of the quality of services. An investigation by Addington and others (9) identified 24 performance measures considered by stakeholders to be essential in the evaluation of an early-intervention service. Disengagement was emphasized as one such measure. Despite the importance of treatment disengagement as a performance measure for early-psychosis treatment services, relatively few studies have examined rates and predictors of treatment disengagement from FEP services (10).
This article reviews the research on engagement, specifically in relation to the treatment of individuals with FEP. The objective of the review was to examine rates and definitions of engagement across FEP early-intervention services and to identify the most relevant predictors of engagement.
Methods
The databases PubMed (1966–2012) and PsycINFO (1882–2012) were searched. For the PubMed search, we used the following MeSH headings: psychotic disorders (entry terms included “psychosis,” “psychoses,” “schizoaffective disorder,” and “brief reactive psychosis”), schizophrenia (entry terms included schizophrenias; schizophrenic disorders; disorder, schizophrenic; disorders, schizophrenic; schizophrenic disorder; and dementia praecox), and schizophrenia and disorders with psychotic features (no entry terms were given because these are major topics). These were combined with the following free-text search terms: engagement, disengagement, engagement predictors, and drop-out.
PsycINFO was searched by the terms “psychotic disorders” and “schizophrenia,” both by using them as major topics and by combining them with the same free-text terms listed above. We also searched on the Internet by the following terms: engagement and psychosis; disengagement and psychosis; engagement and FEP or first episode psychosis; engagement and FES or first episode schizophrenia; psychosis and drop-out; psychosis and nonadherence; and psychosis and engagement predictors. Reference sections in relevant papers were reviewed for other pertinent studies. Finally, we conducted independent searches for recent publications by leading researchers in the field of early intervention in psychosis. [Details of the search strategies are available online as a data supplement to this article.]
These combined search strategies yielded over 750 abstracts, which were independently reviewed for eligibility by two authors (RD and FF). We included only studies that reported rates of engagement or disengagement from services for FEP or early-intervention psychosis among individuals with a diagnosis of FEP or first-episode schizophrenia. [Henceforth, the term FEP encompasses first-episode schizophrenia.] If a title or abstract appeared to meet inclusion criteria, the same two authors obtained the full article and independently assessed its relevance. The article was also reviewed by one of the senior authors (EL). In the case of disagreements about whether to include a study, a consensus was achieved by consulting the other senior author (MC). Studies that were not specifically relevant to early-intervention services were excluded from the review. That is, only studies about services that were specifically aimed at providing a service for FEP were included. Studies reporting rates of engagement or disengagement related to services for FEP or early intervention in psychosis and to patients with a specific FEP diagnosis (including first-episode schizophrenia) were included. Studies reporting rates of disengagement from general mental health services were excluded.
Results
Study selection
When combined with the free-text terms, the PubMed search of MeSH headings generated 255 results for the term “engagement,” 44 results for “disengagement,” zero results for “engagement predictors,” and 93 results for “drop-out.” The free-text combined search using quoted phrases generated 21 results for “first episode psychosis” and “engagement,” six results for “first episode psychosis” and “disengagement,” and eight results for “first episode psychosis” and “nonadherence.”
When combined with the free-text terms, the PsycINFO search of “psychotic disorders” or “schizophrenia” yielded 196 results for the term “engagement,” 36 results for “disengagement,” zero results for “engagement predictors,” and 52 results for “drop-out.” The free-text combined search using quoted phrases generated 30 results for “first episode psychosis” and “engagement,” eight results for “first episode psychosis” and “disengagement,” and nine results for “first episode psychosis” and “nonadherence.”
Combined with Internet searches, the search strategies described above yielded over 750 abstracts for review. Following review of these abstracts, 714 were immediately excluded because they were clearly not relevant. A total of 44 papers were analyzed for more detailed evaluation. Those that did not meet detailed eligibility criteria were excluded (N=34, including 27 that were not specifically about FEP and seven that did not report or investigate rates of engagement or disengagement). As a result, ten articles met inclusion criteria and were deemed suitable for the current review (Table 1). [A flow diagram illustrating the search results is available in the online data supplement.]
| Disengagement | ||||||
|---|---|---|---|---|---|---|
| Study | Location | Setting | Sample | N | Definition | Rate |
| Anderson et al., 2012 (12) | Montreal, Canada | Prevention and early intervention for psychosis program | Defined catchment area | 324 | No contact for a continuous period of 3 months prior to completion of the 2-year program | 28% |
| Stowkowy et al., 2012 (2) | Calgary, Canada | Specialist early-psychosis treatment service | Serves a majority of incidents in a population of approximately 1.3 million | 286 | Dropping out (not returning phone calls or not attending appointments for 3 months) before 30 months | 31% after 30 months |
| Conus et al., 2010 (8) | Australia | Early Psychosis Prevention and Intervention Centre (EPPIC) FEP treatment service | Treated epidemiological sample | 786 | Active refusal of any contact with treatment facility | 23.3% |
| Turner et al., 2009 (21) | NewZealand | Early intervention for psychosis service and outreach program | Representative cohort | 41 | Terminated treatment against clinicians’ advice | 34% |
| Álvarez-Jiménez et al., 2009 (17) | Australia | EPISODE II trial, cognitive-behavioral intervention for FEP | Participants in two youth-oriented specialist FEP programs | 236 | Prematurely discontinued treatment or did not attend appointments | 31.7% |
| Miller et al., 2009 (16) | New York | Study examining antipsychotic medications at 2 community hospitals | Patients presenting for the first time for treatment of schizophrenia (inpatient and outpatient) | 112 | Left treatment for more than 1 month | 20.5% (33% nonadherent to medication at some point) |
| Lecomte et al., 2008 (13) | Vancouver, Canada | FEP treatment service | Participants in cross-sectional study of an early-psychosis intervention program | 118 | Results of Service Engagement Scale (15) | Almost 50% experienced difficulties engaging with clinician or case worker |
| Turner et al., 2007 (10) | NewZealand | Early-intervention service | Representative longitudinal FEP cohort—all FEP patients in Christchurch catchment area | 232 | Termination of treatment despite therapeutic need within 12 months of entry | 24.6% |
| Schimmelmann et al., 2006 (4) | Australia | EPPIC FEP treatment service–adolescents | Epidemiological cohort (retrospective) | 134 | Refusal of any contact or nontraceable within 18 months | 23% |
| Garety and Rigg, 2001 (20) | South London, United Kingdom | Defined geographical area served by the South London and Maudsley National Health Service Trust | All presentations of FEP and second-episode psychosis in a defined catchment area | 21 | Contact with services at 1 year | 40% |
a
FEP services include services for early intervention in psychosis. All studies were limited to patients with diagnoses of FEP or first-episode schizophrenia.
Definitions of engagement
There was no accepted definition of engagement or disengagement across studies. Definitions ranged from “when there has not been a clinical resolution or agreed termination from treatment” (11) to termination of treatment despite therapeutic need (10).
In one of the largest studies examining rates and predictors of disengagement, the term disengagement was defined as present if “case notes suggested that patients actively refused any contact with the treatment facility or were not traceable” (8). Similarly, in a study by Stowkowy and others (2), disengagement was described as “dropping out of the program before 30 months,” with dropping out defined as not returning phone calls or not attending appointments for three months. A study examining determinants of negative pathways to care and their impact on service disengagement in FEP defined disengagement from the service as no contact for a continuous period of three months (12).
Lecomte and others (13) assessed engagement by using the Service Engagement Scale (14), a brief, 14-item measure that is used by case workers to rate client engagement with services (15). In a study investigating cannabis use as a risk factor for treatment dropout, participants were considered to have disengaged if they had dropped out of treatment for more than one month and were not receiving treatment elsewhere (16). Álvarez-Jiménez and others (17) investigated predictors of adherence to cognitive-behavioral therapy within an FEP program. Individuals who did not initiate the intervention, terminated against the therapists’ advice, or did not complete 80% of the treatment protocol were considered nonadherent.
The complexity of defining the phenomenon of engagement means that attendance rates are often used as measures of engagement (18). O’Brien and others (18) noted that although it is obvious that a patient who is not attending services cannot be described as engaged, most studies do not discriminate between persons who terminate treatment independently and those who make the decision collaboratively with their treating team. It may be beneficial for FEP services to divide the concept of engagement into two entities—one, engagement in the literal sense of appointment attendance or recommended intervention or treatment; and two, the more intangible concept of therapeutic engagement. O’Brien and others (18) concluded that engagement should be seen as a more “complex phenomenon” encompassing factors that include “acceptance of a need for help, the formation of a therapeutic alliance with professionals, satisfaction with the help already received, and a mutual acceptance and working towards shared goals.”
Service-level disengagement
Rates of disengagement from general mental health services as high as 46% have been noted (19), and overall there are large variations in the reported rates. The variation is due to the differences in setting and type of service provided and in how each study defined engagement. An early study in south London noted that 38% of persons with FEP had dropped out of a treatment intervention within one year (20).
FEP services reported a broadly consistent rate of disengagement of one-third or less (2,4,8,10,12,16,21), with rates varying between 20.5% and 40%. Stowkowy and others (2) reported an overall rate of treatment disengagement of 31%, with disengagement defined as not returning phone calls or attending an appointment for three months prior to completing 30 months of the program. Using a similar definition of disengagement (no contact for a continuous period of three months prior to completion of a two-year program), Anderson and others (12) found a comparable rate of disengagement (28%). A final analysis by Conus and others (8), who defined “disengagers” as persons who actively refuse any contact with a treatment facility, revealed a rate of disengagement of 23.3%. Schimmelmann and others (4), who investigated predictors of service disengagement among adolescents with FEP, reported a similar rate of disengagement of 23.4%. Adolescents who disengaged from the service did so within the first 18 months, and disengagement was defined as refusing any contact or being untraceable. Turner and others (21) reported a rate of disengagement from services of 34%, and a majority of those who disengaged did so within the first 12 months. Disengagement was defined as terminating treatment against a clinician’s advice. In an earlier study with the same definition of disengagement, the same research group reported a rate of disengagement of 24.6% (10).
Disengagement from specific interventions
Álvarez-Jiménez and others (17) reported a 31.7% rate of disengagement from cognitive-behavioral therapy, with disengagement defined as premature discontinuation or nonattendance of the program. A study by Lecomte and colleagues (13) noted that almost half of participants experienced difficulties engaging with their clinician or case worker. A limitation of the study, as noted by the authors, was that “true disengagers” (those who refused any contact with the service) were not included in the analysis. Indeed, this was a limitation shared by several studies discussed here (2,4,8,10,12,13,16,17,20,21), given that they did not account for “true nonengagers,” those who refuse any contact with a service from the start (1).
Overall, rates of disengagement from FEP services appeared to be one-third or less. A majority of studies reported rates somewhere between 23% and 34%, implying that one in every three or four people disengage from an FEP service.
Predictors of disengagement
Table 2 summarizes findings related to predictors of disengagement.
| Predictor | 12 | 2 | 8 | 17 | 21 | 16 | 13 | 10 | 4 | 5 |
|---|---|---|---|---|---|---|---|---|---|---|
| Sociodemographic | ||||||||||
| Age or age at onset | + | – | – | – | + | |||||
| Gender | ||||||||||
| Male | – | – | – | – | – | + | ||||
| Female | – | – | – | |||||||
| Family member involved | ||||||||||
| Yes | + | |||||||||
| No | + | + | ||||||||
| Living alone or without family | + | – | + | – | – | + | ||||
| Marital status | ||||||||||
| Not single | – | |||||||||
| Single | – | |||||||||
| Race-ethnicity | + | – | ||||||||
| Clinical | ||||||||||
| Insight | – | + | – | + | ||||||
| Duration of untreated psychosis | + | – | + | + | ||||||
| Duration of untreated prodromal phase | + | |||||||||
| Positive symptoms | – | + | + | |||||||
| Negative symptoms | + | – | + | – | ||||||
| GAF current scoreb | – | + | ||||||||
| CDSS total scorec | – | |||||||||
| QLS total scored | – | |||||||||
| General psychopathology | – | |||||||||
| Substance use during treatment | – | + | – | + | + | |||||
| PANSS total scoree | + | – | ||||||||
| Unemployed (includes no school) | + | + | – | + | ||||||
| Persistent substance use disorder (SUD) | + | |||||||||
| Premorbid | ||||||||||
| Functioning | – | + | ||||||||
| PAS scoref | – | – | – | |||||||
| GAF score | + | |||||||||
| Family history of psychosis | – | |||||||||
| Past psychiatric history | – | |||||||||
| Past SUD | – | + | ||||||||
| Forensic history | + | + | ||||||||
| History of sexual abuse | – | |||||||||
| History of physical abuse | – | + | ||||||||
| Suicide attempts | – | |||||||||
| Childhood trauma | + | |||||||||
| Lack of knowledge regarding consumer rights | + | |||||||||
| Baseline | ||||||||||
| GAF | + | + | ||||||||
| Substance abuse | + | |||||||||
| HoNOS scoreg | + | |||||||||
| Severity of illness | – | + | + | |||||||
| Cognitive score | – | |||||||||
| Diagnosis | ||||||||||
| Schizophrenia | – | |||||||||
| Schizophreniform | – | |||||||||
| Schizophrenia spectrum | – | |||||||||
| Psychosis not otherwise specified | – | |||||||||
| Brief psychotic disorder | – | |||||||||
| Delusional disorder | – | |||||||||
| Schizoaffective | – | |||||||||
| Other psychoses | + | |||||||||
| Other | – | |||||||||
| Substance use | ||||||||||
| Alcohol use | – | – | ||||||||
| Cannabis use | – | – | ||||||||
| Other drug use | – | – | ||||||||
| Comorbid SUD | ||||||||||
| Comorbidity excluding SUD | ||||||||||
| Psychological | ||||||||||
| High agreeableness | + | |||||||||
| Low neuroticism | + | |||||||||
| Difficulties in alliance | + | |||||||||
| Recovery style |
a
All studies except reference 5 included only patients with first-episode psychosis. +=significant; –=not significant
b
GAF, Global Assessment of Functioning
c
CDSS, Calgary Depression Scale for Schizophrenia
d
QLS, Quality of Life Scale
e
PANSS, Positive and Negative Syndrome Scale
f
PAS, Premorbid Adjustment Scale
g
HoNOS, Health of the Nation Outcome Scales
Sociodemographic predictors.
Although male gender is a known risk factor for disengagement (1,5,18), three large FEP studies did not demonstrate an effect for gender (2,8,12). Miller and others (16) concluded that male gender is a significant risk factor for cannabis abuse and dependence and for cannabis use and that this association explains males’ greater hazard of nonadherence to medication. Lecomte and others (13) found that males who had more trouble complying with the law were more likely not to adhere to medication and had a poorer level of service engagement.
A review by Nosé and others (5) indicated that young age was associated with poor compliance. In contrast, two large FEP studies found that age did not predict engagement (2,8). However, that finding may reflect the relatively narrow age range in those two studies. Notably, Anderson and others (12) found older age to be associated with increased risk of service disengagement within a large FEP service.
The study by Stowkowy and others (2) found no significant differences in marital status between those who disengaged from and those who completed treatment.
Clinical predictors.
Three studies reported an association between symptom severity and risk of disengagement (2,4,8). Conus and others (8) found that those with a lower severity of illness at baseline were almost twice as likely to disengage. They postulated that those with a higher severity of illness and their families may have greater motivation for treatment or that there may be “greater effort by the treating teams to prevent disengagement when symptoms are perceived as higher intensity.” Reporting similar results, Schimmelmann and colleagues (4) found that lower severity of illness at baseline was one of the strongest predictors of disengagement (4). In contrast, Stowkowy and others (2) did not find a significant relationship between lower illness severity at baseline and rate of disengagement. However, the authors noted that their finding may be related to the difference in time to discontinuation in their particular study (30 months) compared with other studies or to differences in the population or service delivery (2). Findings from the studies that examined severity of illness and engagement with services suggest there may be a “bimodal” relationship between the two variables (22). Although a common reason for disengagement may be that someone is feeling “better,” it may also be the case that someone is feeling too unwell to attend or engage, the latter having a far more serious impact on an individual’s road to recovery.
Stowkowy and others (2) found that shorter duration of the untreated psychosis significantly predicted disengagement from treatment. As a possible explanation, the authors noted that those who seek treatment earlier may be the quickest to leave should there be a remission of symptoms. Alternatively, they speculated, factors that influence a delay in treatment-seeking behavior may serve to maintain engagement.
Notably, Conus and others (8) reported that a shorter duration of the untreated prodromal phase was a significant baseline and pretreatment predictor of disengagement. Duration of untreated psychosis was not found to be a significant predictor of disengagement.
Contrary to Stowkowy and others (2), Turner and colleagues (10) reported a strong association between longer duration of untreated psychosis and treatment disengagement. Álvarez-Jiménez and others (17) also found that longer duration of untreated psychosis predicted poor engagement with cognitive-behavioral therapy. This association remained significant even after the authors controlled for confounding variables. These findings add weight to the premise that delay of treatment for FEP reduces the therapeutic potential of interventions (10,17). These conflicting findings with regard to duration of untreated psychosis may mirror the bimodal effect noted in the discussion of symptom severity, above.
Overall, the results indicate that there is an association between duration of untreated psychosis and risk of disengagement. However, how this predictor exerts its influence remains inconclusive.
Studies indicate that forensic history may be a significant risk factor associated with disengagement from mental health services (18). Conus and others (8) found that among individuals with FEP, those with a previous forensic history were almost twice as likely to disengage as those without such a history. Lecomte and colleagues (13) found that a forensic history was associated with lower scores on the Service Engagement Scale (15). However, Conus and others (8) noted that they did not assess other factors that may be linked to criminal activity, such as personality traits. Antisocial personality disorder or traits such as impulsivity, irresponsibility, and failure to comply with perceived authority may contribute to disengagement with services (18). Patients with this particular history may have difficulty establishing rapport and trusting relationships or dealing with authoritative figures (8).
It is also the case that offenses may be committed while someone is acutely unwell. Offenses committed during an acute phase of psychosis may signify a link between severity of illness and insight into their need for treatment, which may have a subsequent impact on service disengagement. Past forensic issues may also be linked with other issues of a practical nature, such as difficulties with housing and employment, which in turn “may lead to a more disorganized lifestyle” (8).
Poor insight has been proposed as one of the most pertinent factors impeding service engagement among individuals with psychosis. However, the evidence for a direct association between insight and engagement remains inconclusive (23). Three large studies did not find level of insight to be a significant predictor of disengagement (2,4,8).
In contrast, Álvarez-Jiménez and colleagues (17) reported that a lower level of insight was one of the strongest predictors of poorer engagement with cognitive-behavioral therapy; the association remained strong and significant even after adjustment for confounding variables. Turner and others (10) reported that those who disengaged from an early-intervention program had significantly lower level of insight than those who did not.
A novel investigation by Tait and others (14) examined whether insight, symptoms, and recovery style predicted service engagement during an acute episode of psychosis and at three- and six-month follow-ups. Insight measured at three months had no significant correlation with engagement as measured by the Service Engagement Scale at six months. Thus insight scores at three months did not predict the Service Engagement Scale scores at six months. There was also no correlation between insight scores and the Service Engagement Scale at six months. Notably, the authors found that recovery style contributed more to engagement than did insight.
There is no doubt that insight or lack of insight is of relevance when considering engagement or disengagement (18); however, the role of insight is likely to involve other factors.
Most of the research studies reported that substance abuse and dependence play a significant role in determining rates of disengagement in this population. A review by Kreyenbuhl and coauthors (1) reported that individuals with schizophrenia who disengage from services typically have a co-occurring substance use disorder.
Miller and others (16) found that 44% of individuals who had experienced a first episode of schizophrenia also had a diagnosis of cannabis abuse and dependence, which increased the risk of nonadherence to medication and of treatment dropout, independent of age, race, socioeconomic status, and gender. This finding is in line with several other studies (24,25). The authors concluded that psychiatric symptoms, addictive disorders and environmental factors combine to jeopardize treatment adherence, similar to what they have observed clinically (16). The study also indicated that many first-episode schizophrenia patients return to cannabis use early in treatment. Such behavior can alter the chances of poor adherence and subsequent relapse. A limitation of the study was that cannabis was the primary substance used by participants, who were enrolled in a program of medication research, and, therefore, the results may not generalize (16).
Stowkowy and coauthors (2) found that multivariate analysis of baseline substance abuse did not predict disengagement from treatment. However, univariate analysis of the initial group comparisons indicated that use of cannabis or other substances at baseline was significantly associated with treatment dropout within the first six months. The authors noted that the results should be interpreted with caution, given that the study investigated current rather than persistent substance use. Similarly, in a two-year outcome study of individuals attending an FEP service, individuals who disengaged were marginally more likely than those who completed treatment to report substance abuse at baseline (21).
Conus and others (8) found that comorbid substance abuse and dependence were also linked to a higher probability of disengagement from treatment. Notably, they concluded that substance abuse and dependence were linked to disengagement only if the misuse persisted throughout the treatment period.
Consistent with the findings described above, Turner and colleagues (10) reported that current substance abuse and dependence were significantly associated with disengagement. In another study, persistent substance use was associated with a 2.6-fold increased risk of disengagement from treatment services among adolescents with FEP (4).
However, Lecomte and colleagues (13) did not find substance abuse to be a predictor of service engagement. The authors noted that this finding was not due to a ceiling or a floor effect. A ceiling effect is the level at which an independent variable—in this case, substance abuse—no longer has an effect on the dependent variable—in this case, service engagement. A floor effect occurs when low scores on the dependent variable mask a potential effect on the independent variable. Similar to the study by Miller and others (16), this study represented persons who were in contact with an FEP service or outpatient service and, therefore, excluded persons who had disengaged initially.
Historically, research on the role of the family has examined the potential negative impact of high expressed emotion (2). Recently, it has become evident that family support is a critical element of “success in the treatment of FEP” (8). Stowkowy and others (2) concluded that a lack of family involvement significantly predicted disengagement from treatment. Living without a family member was also a significant predictor of disengagement, according to Conus and others (8).
Schimmelmann and others (4) found that adolescents living without family at baseline were more likely to disengage from the service. In fact, the authors concluded, those without family support during treatment were at a 4.8-fold increased risk of disengagement.
In contrast, Anderson and others (12) reported that living away from family was associated with a reduced risk of service disengagement. A possible explanation for this finding suggested by the authors is that clinicians may use more assertive follow-up for those without a support system and those who have strong social support may perceive that they are less in need of services.
Nonetheless, Turner and colleagues (10) found no significant differences related to living with parents between those who disengaged from and those who completed treatment. Álvarez-Jiménez and others (17) found that family measures had little predictive power in relation to adherence to cognitive-behavioral therapy for FEP.
Psychological predictors.
Tait and others (14) explored the premise that lack of insight or denial of the need for treatment may be a way of dealing with stigma associated with psychosis. The authors assessed 50 individuals during an acute episode of psychosis and at three-month and six-month follow-ups. The measures included in the study were insight, service engagement, and recovery style. The authors investigated two recovery styles—“sealing over” and “integration.” Sealing over refers to the minimization of the significance of symptoms and a lack of curiosity about the experience. Integration refers to an acknowledgment of and curiosity about the significance of psychosis and active attempts to cope in managing the illness. The results indicated that sealing over at three months predicted poor engagement at six months. The authors concluded that recovery style contributed more to engagement than did insight.
Lecomte and others (13) found that high agreeableness among individuals with early psychosis was linked to poor medication adherence and poor service engagement. The authors measured agreeableness by using the NEO Five Factor Inventory-3 by Costa and McCrae (26). While often considered a positive personality trait, agreeableness can also be associated with a higher likelihood of being influenced by peers and with eagerness to be accepted in a group (13). In the context of service disengagement, the authors noted, high agreeableness could prompt persons with FEP to avoid stigmatizing settings and behaviors in order to ensure acceptance by peers and avoid social rejection linked to psychosis.
The study by Lecomte and others (13) also indicated that having experienced childhood physical abuse predicted poor service engagement. Early trauma could be an etiological cause of psychotic symptoms (27,28). Lecomte and others (13) noted that their findings suggest that individuals with childhood trauma may experience more difficulties in trusting others, especially authority figures. In contrast, Conus and others (8) did not find that childhood physical or sexual abuse was a predictor of service engagement.
Discussion
Robust findings
The most robust predictors of disengagement were comorbid substance abuse and dependence and the involvement or support of family. All the studies that investigated substance abuse and dependence and disengagement, with the exception of the study by Lecomte and others (13), indicated that substance use increases the risk of service disengagement (2,4,8,10,16,21). These studies indicated that use of cannabis is an increased risk factor for treatment disengagement. These findings suggest a need to address comorbid substance use in early-intervention programs for FEP. Early-intervention services should also be aware that often cannabis abuse and dependence are “long-term problem[s] or [a] lifestyle and that relapse of cannabis use in this population can significantly interfere with treatment” (16).
Family support and involvement appear to be significant factors in treatment engagement with an FEP service. Three studies, including two of the largest (N=786 [8] and N=286 [2]), concluded that lack of involvement by a family member significantly predicted treatment disengagement (2,4,8). However, as noted by Anderson and others (12), the complex mechanisms by which family involvement influence engagement warrant further examination.
Conflicting findings
Conflicting findings were reported for several predictors of disengagement from FEP services, including symptom severity, duration of untreated psychosis, and insight. The mechanisms by which these factors exert their influence are complex and often act in opposite directions across the studies.
Two studies found that a lower severity of illness at baseline significantly increased the risk of disengagement (4,8), with another study finding no significant relationship between the two (2). One possible explanation for this discrepancy could be the differences between studies in terms of service delivery or populations served (2). Alternatively, the contradictory findings could reflect a bimodal relationship between illness severity and service engagement. Qualitative investigations may help elaborate this question.
There were also conflicting findings in relation to duration of untreated psychosis, with one study reporting that shorter duration of untreated psychosis significantly predicted disengagement from treatment (2) and others reporting a strong association between longer duration of untreated psychosis and risk of disengagement (10,17). These conflicting findings, again, may mirror the bimodal relationship between symptom severity and service engagement, but key differences between the studies in service delivery and intervention could also play a significant role.
There is no doubt that insight is of relevance to disengagement, but the studies reviewed reported contrary findings, with three large studies reporting that lack of insight did not influence disengagement (2,4,8) and others reporting that low level of insight was a strong predictor of disengagement (13,19). Insight is considered a multidimensional concept underpinned by many theories; this may account for the numerous insight assessment tools used within clinical practice (29). The complexities surrounding the dimension or concept of insight being measured make comparisons across studies difficult (29). The studies in this review used different methods of assessing insight, which may explain the conflicting findings (2,4,8,10,13,17). Furthermore, as noted by Lecomte and others (13), the conflicting results may suggest that service engagement could be influenced by different variables depending on whether the individual is experiencing early psychosis or a more enduring course of illness.
The inconsistent findings of these studies indicate a clear need for qualitative investigations. Such qualitative information would help services distinguish between patients who disengage because of a perception that no more help is needed or that a level of independent coping with the illness has been achieved or because of high severity of symptoms. Clarifying the reasons for disengagement as early as possible may identify “patient-cited” barriers to service engagement (22). If barriers are identified early, attempts to address the issues and provide alternatives can be made.
There is also an important difference between those who refuse any contact at all with a service from the initial stages and those who disengage after having attended or engaged with a service for a while. Individuals may disengage immediately because of preexisting beliefs surrounding mental health or mental health services (for example, stigma) or because of severity of symptoms at baseline. Disengagement at later stages may be related to therapeutic alliance, the patient’s overall satisfaction with the service, and the perceived need for further help (22). Alternatively, it may reflect a change in symptoms over time. As noted by Stowkowy and others (2), persons in remission may stop attending, and “future research could investigate how a change in symptoms or other characteristics might also contribute to disengagement.” When addressing therapeutic alliance in FEP services, simple measures can be administered to improve the quality of care.
A parsimonious explanation for the conflicting findings in relation to predictors of disengagement may be related to differences between studies in terms of service delivery and populations served. In terms of populations served, two samples were defined as epidemiological and were drawn from patients at a specialist early-intervention service; another two samples were drawn from specialist early-intervention services that captured a majority of patients with FEP in a defined catchment area (2,12); three samples were from specialist early-intervention services with a representative FEP cohort (10,17,21); one was a cross-sectional sample from an early-intervention service (13); and one was a sample of patients presenting for the first time for treatment of schizophrenia at two community hospitals that were not a specialist early-intervention service (16).
Future directions
In a review investigating disengagement from mental health services in general, O’Brien and coauthors (18) noted that “there is a need for a reliable, validated, agreed [upon] scale to measure engagement with services.” This is also the case with specialized FEP intervention programs. The development of a clearly defined and approved measurement of service engagement and disengagement across FEP programs would help clinicians to identify predictors that influence engagement in treatment. Recently, Meaden and colleagues (30) developed the Residential Rehabilitation Engagement Scale for Psychosis, a reliable, observer-rated, multidisciplinary measure of engagement for inpatient populations. Within the scale, engagement is divided into three subcategories (active participation and openness, agreement with treatment and basic relationships, and medication compliance). The authors suggested that this division allows clinicians to identify specific aspects of engagement and to employ suitable interventions at critical time periods. Undoubtedly, the development of a similar scale that captures all relevant aspects of the complex process and multidimensional nature of engagement during treatment of FEP would be of great benefit to early-intervention services.
Limitations
This review should be considered in the light of some limitations and difficulties in the comparison of the studies due to methodological differences. Although all the studies examined engagement among persons with FEP, they had a number of shortcomings. Some studies were based on a retrospective audit of files, and others had small sample sizes. In addition, none of the studies provided data on those who disengaged from but subsequently reengaged with services. Comparison across studies was further hindered by different lengths of follow-up and definitions of engagement. None of the studies provided qualitative information about the reasons for leaving the service.
Conclusions
The evidence reviewed indicates that approximately 30% of individuals with FEP disengage from services. This suggests that individuals with psychosis are at high risk of disengaging from services, even from programs that provide specialized treatment for FEP or early-intervention services. Variables that were consistently found to exert an influence on disengagement were symptom severity at baseline, duration of untreated psychosis, insight, substance misuse or dependence, and involvement of a family member.
Continuity of care is of particular importance with FEP, given evidence suggesting that long-term care can improve symptoms and functioning and reduces relapse risk. Engagement is considered a key performance measure of an FEP service and should be monitored. Strategies to improve engagement should be a core feature of an FEP service. As part of quality improvement initiatives, services should aim to promptly detect and intervene with those who are at higher risk of service disengagement.
Clinicians should particularly focus on family involvement with and support for the individual, given that individuals without family support are at higher risk of disengagement. In addition, persons with substance abuse should be considered at risk of disengagement, and specific strategies to maintain engagement should be implemented as early as possible. Furthermore, reports of contrary findings for several predictors of disengagement suggest that qualitative studies are needed to explore why individuals disengage from early-intervention services.
Acknowledgments and disclosures
Ms. Doyle’s position is funded by grant 497 from St. John of God Hospitaller Services Research Foundation.
The authors report no competing interests.
Supplementary Material
Supplementary Material (603_ds001.pdf)
- View/Download
- 91.11 KB
References
1.
Kreyenbuhl J, Nossel IR, Dixon LB: Disengagement from mental health treatment among individuals with schizophrenia and strategies for facilitating connections to care: a review of the literature. Schizophrenia Bulletin 35:696–703, 2009
2.
Stowkowy J, Addington D, Liu L, et al.: Predictors of disengagement from treatment in an early psychosis program. Schizophrenia Research 136:7–12, 2012
3.
Addington J, Leriger E, Addington D: Symptom outcome one year after admission to an early psychosis program. Canadian Journal of Psychiatry 48:204–207, 2003
4.
Schimmelmann BG, Conus P, Schacht M, et al.: Predictors of service disengagement in first-admitted adolescents with psychosis. Journal of the American Academy of Child and Adolescent Psychiatry 45:990–999, 2006
5.
Nosé M, Barbui C, Tansella M: How often do patients with psychosis fail to adhere to treatment programmes? A systematic review. Psychological Medicine 33:1149–1160, 2003
6.
Birchwood M, Todd P, Jackson C: Early intervention in psychosis: the critical period hypothesis. British Journal of Psychiatry 172(suppl):53–59, 1998
7.
Spencer M, Birchwood E, McGovern D: Management of first-episode psychosis. Advances in Psychiatric Treatment 7:133–142, 2001
8.
Conus P, Lambert M, Cotton S, et al.: Rate and predictors of service disengagement in an epidemiological first-episode psychosis cohort. Schizophrenia Research 118:256–263, 2010
9.
Addington D, McKenzie E, Addington J, et al.: Performance measures for early psychosis treatment services. Psychiatric Services 56:1570–1582, 2005
10.
Turner M, Smith-Hamel C, Mulder R: Prediction of twelve-month service disengagement from an early intervention in psychosis service. Early Intervention in Psychiatry 1:276–281, 2007
11.
Rossi A, Amaddeo F, Bisoffi G: Dropping out of care: inappropriate terminations of contact with community-based psychiatric services. British Journal of Psychiatry 181:331–338, 2002
12.
Anderson KK, Fuhrer R, Schmitz N: Determinants of negative pathways to care and their impact on service disengagement in first episode-psychosis. Social Psychiatry and Psychiatric Epidemiology 48:125–136, 2012
13.
Lecomte T, Spidel A, Leclerc C, et al.: Predictors and profiles of treatment non-adherence and treatment dropout in early psychosis. Schizophrenia Research 102:295–302, 2008
14.
Tait L, Birchwood M, Trower P: Predicting engagement with services for psychosis: insight, symptoms and recovery style. British Journal of Psychiatry 182:123–128, 2003
15.
Tait L, Birchwood M, Trower P: A new scale (SES) to measure engagement with community mental health services. Journal of Mental Health 11:191–198, 2002
16.
Miller R, Ream G, McCormack J, et al.: A prospective study of cannabis use as a risk factor for non-adherence and treatment dropout in first-episode schizophrenia. Schizophrenia Research 113:138–144, 2009
17.
Alvarez-Jiménez M, Gleeson JF, Cotton S, et al.: Predictors of adherence to cognitive-behavioural therapy in first-episode psychosis. Canadian Journal of Psychiatry 54:710–718, 2009
18.
O’Brien A, Fahmy R, Singh SP: Disengagement from mental health services: a literature review. Social Psychiatry and Psychiatric Epidemiology 44:558–568, 2009
19.
Percudani M, Belloni G, Contini A, et al.: Monitoring community psychiatric services in Italy: differences between patients who leave care and those who stay in treatment. British Journal of Psychiatry 180:254–259, 2002
20.
Garety PA, Rigg A: Early psychosis in the inner city: a survey to inform service planning. Social Psychiatry and Psychiatric Epidemiology 36:537–544, 2001
21.
Turner MA, Boden JM, Smith-Hamel C, et al.: Outcomes for 236 patients from a 2-year early intervention in psychosis service. Acta Psychiatrica Scandinavica 120:129–137, 2009
22.
Mitchell AJ, Selmes T: Why don’t patients attend their appointments? Maintaining engagement with psychiatric services. Advances in Psychiatric Treatment 13:423–434, 2007
23.
Trauer T, Sacks T: The relationship between insight and medication adherence in severely mentally ill clients treated in the community. Acta Psychiatrica Scandinavia 102:211–-216, 2000
24.
Wade D, Harrigan S, Edwards J, et al.: Course of substance misuse and daily tobacco use in first-episode psychosis. Schizophrenia Research 81:145–150, 2006
25.
Lambert M, Conus P, Lubman DI, et al.: The impact of substance use disorders on clinical outcome in 643 patients with first-episode psychosis. Acta Psychiatrica Scandinavica 112:141–148, 2005
26.
Costa P, McCrae R: Normal personality assessment in clinical practice: the NEO Personality Inventory (NEO-FFI). Psychological Assessment 4:5–13, 1992
27.
Read J, van Os J, Morrison AP, et al.: Childhood trauma, psychosis and schizophrenia: a literature review with theoretical and clinical implications. Acta Psychiatrica Scandanavica 112:330–350, 2005
28.
Shevlin M, Dorahy MJ, Adamson G: Trauma and psychosis: an analysis of the National Comorbidity Survey. American Journal of Psychiatry 164:166–169, 2007
29.
McCormack M, Tierney K, Brennan D, et al: Lack of insight in psychosis: theoretical concepts and clinical aspects. Behavioural and Cognitive Psychotherapy (Epub ahead of print, Apr 12, 2013)
30.
Meaden A, Hacker D, de Villiers A, et al.: Developing a measurement of engagement: the Residential Rehabilitation Engagement Scale for psychosis. Journal of Mental Health 21:182–191, 2012
Information & Authors
Information
Published In
Cover: Gisele, by Elizabeth Shippen Green Elliott, published in Harper's magazine, 1908. Watercolor and charcoal drawing. Library of Congress Prints and Photographs Division, Washington, D.C.
History
Published in print: May 2014
Published online: 30 July 2014
Authors
Metrics & Citations
Metrics
Citations
Export Citations
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.
For more information or tips please see 'Downloading to a citation manager' in the Help menu.
View Options
View options
PDF/EPUB
View PDF/EPUBLogin options
Already a subscriber? Access your subscription through your login credentials or your institution for full access to this article.
Personal login Institutional Login Open Athens loginNot a subscriber?
PsychiatryOnline subscription options offer access to the DSM-5-TR® library, books, journals, CME, and patient resources. This all-in-one virtual library provides psychiatrists and mental health professionals with key resources for diagnosis, treatment, research, and professional development.
Need more help? PsychiatryOnline Customer Service may be reached by emailing [email protected] or by calling 800-368-5777 (in the U.S.) or 703-907-7322 (outside the U.S.).