Introduction
The improvement of everyday functioning of individuals with schizophrenia over the long term is the main outcome measure for determining optimal levels of quality of life in the real world and should be considered to be the real challenge of schizophrenia treatment. Despite recent developments in its clinical management, schizophrenia spectrum disorders (SDDs) are still associated with an uncertain long-term functional prognosis (McGorry, Reference McGorry2015). Patients repeatedly have difficulties in attaining their functional milestones in educational, residential, social and vocational domains (Harvey et al., Reference Harvey, Heaton, Carpenter, Green, Gold and Schoenbaum2012). This also raises the question of accurately measuring multiple domains of function as essential for assessing the effectiveness of treatment strategies on the course of the illness (Harvey and Bellack, Reference Harvey and Bellack2009).
Early intervention (EI) programs provide an extensive multidisciplinary clinical management during the early phases of the illness, with the major aim of achieving recovery (McGorry et al., Reference McGorry, Killackey and Yung2008). They have demonstrated their benefits in reducing the short-term adverse impact of SDD and improving function (De Maio et al., Reference De Maio, Graham, Vaughan, Haber and Madonick2015; Csillag et al., Reference Csillag, Nordentoft, Mizuno, McDaid, Arango, Smith, Lora, Verma, Di Fiandra and Jones2017). The superiority of EI over treatment as usual regarding all outcomes was evident at the 2-year follow-up (Correll et al., Reference Correll, Galling, Pawar, Krivko, Bonetto, Ruggeri, Craig, Nordentoft, Srihari, Guloksuz, Hui, Chen, Valencia, Juarez, Robinson, Schooler, Brunette, Mueser, Rosenheck, Marcy, Addington, Estroff, Robinson, Penn, Severe and Kane2018). However, whether these gains are maintained or vanish over the long term, after patients are discharged from specialized EI programs to routine outpatient mental healthcare services, or whether they may provide (as presently designed) a real improvement only for particular subgroups of patients, is still under debate. The success of EI programs broadly varies and has usually been assessed by taking into account group averages and accepting a one-size-fits-all scenario. Therefore, understanding this diversity in outcomes might help to sort out individuals or subgroups of individuals who will benefit from intensive EIs (Austin et al., Reference Austin, Mors, Budtz-Jorgensen, Secher, Hjorthoj, Bertelsen, Jeppesen, Petersen, Thorup and Nordentoft2015; Velthorst et al., Reference Velthorst, Fett, Reichenberg, Perlman, van Os, Bromet and Kotov2017).
The long-term outcome assessment in SDD is much more challenging. Apart from the multidimensional nature of outcomes, there are methodological concerns about how best to compare results from studies with different inclusion criteria and different durations of follow-up (e.g. 5 v. 25 years), questioning the validity of comparing between studies with different follow-up periods (McGrath, Reference McGrath2008). Even when comparing prospective EI cohorts with similar long-term follow-up (10 years) (Secher et al., Reference Secher, Hjorthoj, Austin, Thorup, Jeppesen, Mors and Nordentoft2015; Chan et al., Reference Chan, Hui, Chang, Lee and Chen2018) several issues concerning outcomes have been left unanswered.
In that context, there is a growing interest in depicting the heterogeneity and subtyping the outcomes, which makes cluster analyses a useful exploratory tool that provides the opportunity to group individuals using a data-driven approach rather than a predetermined set of grouping criteria (Lewandowski et al., Reference Lewandowski, Sperry, Cohen and Ongur2014; Laloyaux et al., Reference Laloyaux, Laroi, Nuyens and Billieux2018). Follow-up naturalistic studies of first episode of psychosis (FEP) cohorts allow cluster analyses to identify patient characteristics that merit specific intervention.
The current study aims to investigate the heterogeneity of function and outcomes over long term by examining a large epidemiological sample of individuals with non-affective psychosis, 10 years after their inclusion in an EI program. We sought to (1) describe subtypes of functional outcomes according to clinical, neurocognitive and sociodemographic characteristics and (2) examine likely associations of these clusters with other areas of function, such as living independently, occupational/vocational status and active social interactions.
Method
Setting
Data for the current study were obtained from a large cohort of patients representative of the general population of individuals suffering from an FEP in an epidemiological catchment area, which is the autonomous community of Cantabria, located in the Northern coast of Spain (Pelayo-Teran et al., Reference Pelayo-Teran, Perez-Iglesias, Ramirez-Bonilla, Gonzalez-Blanch, Martinez-Garcia, Pardo-Garcia, Rodriguez-Sanchez, Roiz-Santianez, Tordesillas-Gutierrez, Mata, Vazquez-Barquero and Crespo-Facorro2008a). FEP patients were treated in a longitudinal intervention program (Programa de Atención a Fases Iniciales de Psicosis, PAFIP) conducted at the University Hospital Marqués de Valdecilla (Crespo-Facorro et al., Reference Crespo-Facorro, Gonzalez-Blanch and Pelayo-Teran2005; Crespo-Facorro et al., Reference Crespo-Facorro, Perez-Iglesias, Ramirez-Bonilla, Martinez-Garcia, Llorca and Luis Vazquez-Barquero2006; Pelayo-Teran et al., Reference Pelayo-Teran, Perez-Iglesias, Ramirez-Bonilla, Gonzalez-Blanch, Martinez-Garcia, Pardo-Garcia, Rodriguez-Sanchez, Roiz-Santianez, Tordesillas-Gutierrez, Mata, Vazquez-Barquero and Crespo-Facorro2008b). Referrals to the PAFIP came from the inpatient unit and emergency room at the, from outreach mental health services, and from other health-care workers throughout the region of Cantabria. After the initial contact by a qualified psychiatric nurse, an experienced psychiatrist carried out a formal interview for a full assessment of the patient and confirmed the presence of schizophrenia and other primary psychotic disorders (F20–F29). PAFIP includes inpatient and outpatient care and provides multidisciplinary (psychiatric nursing, psychology, psychiatry and social work) and specific and personalized clinical attention from the first contact with PAFIP staff up to 3 years. It has a strong track record with FEP subjects that has included a longitudinal component with multiple assessments. Considering that the PAFIP constitutes the only alternative form of mental health care for FEP, we could defend that the included sample constitutes an epidemiological representation of the population of patients with FEP occurring in the entire region of Cantabria. Accordingly, the patients included in the PAFIP reached an age-corrected (16–60) incidence rate for SDD of 1.38 per 10 000, a figure that is equivalent to the one reported in most epidemiological studies (Crespo-Facorro et al., Reference Crespo-Facorro, Gonzalez-Blanch and Pelayo-Teran2005).
PAFIP-10 is a long-term follow-up study at 10 years (range between 8 and 12 years) of individuals with an FEP who were initially included in PAFIP from February 2001 to July 2008. All patients included in the referred period were invited for a reassessment, which comprised the 10-year follow-up study group.
Ethics
The program, which is fully publicly funded by the regional Mental Health Services, was approved by the local institutional review board (ethics committee for clinical research, CEIC-Cantabria) in accordance with international standards for research ethics (clinical trial numbers NCT0235832 and NCT02534363). Patients who met criteria and provided written informed consent, along with their families, were entered into PAFIP and PAFIP-10 for reassessment.
Baseline inclusion criteria
All referrals to PAFIP were screened against the following inclusion criteria: were 15–60 years of age; lived in the catchment area; experienced their first episode of psychosis; and were antipsychotic medication naïve, or if previously treated, a total lifetime of adequate antipsychotic treatment of less than 6 weeks. Meeting the DSM-IV criteria for drug or alcohol dependence, having an intellectual disability and having a history of neurological disease or head injury were exclusion criteria. The diagnoses were confirmed through the use of the Structured Clinical Interview for DSM-IV (SCID-I) (First et al., Reference First, Spitzer, Gibbon and Williams1996) conducted by an experienced psychiatrist within 6 months of the baseline visit.
Baseline assessment and intervention at PAFIP
Sociodemographic information was recorded from interviews with patients, their relatives and from medical records on admission. Premorbid social adjustment was measured by the Premorbid Adjustment Scale (PAS) (Cannon-Spoor et al., Reference Cannon-Spoor, Potkin and Wyatt1982). Age at psychosis onset [defined as the age when the emergence of the first continuous (present most of the time) psychotic symptom occurred], duration of untreated illness [DUI, defined as the time from the first nonspecific symptom related to psychosis (i.e. with no return to previous level of functioning) to antipsychotic treatment onset], and duration of untreated psychosis [DUP, defined as the time from the first continuous (present most of the time) psychotic symptom to initiation of adequate antipsychotic drug treatment] were obtained.
The initial social functioning assessment was taken at first contact and 3 months later, when the participants were no longer in the hospital. Thus, the 3-month assessment was used as the starting point.
All participants received 3 years of evidence-based phase-specific individual interventions: pharmacotherapy (the lowest effective dose of antipsychotic as maintenance dosage, monitoring for clinical response and side effects, such as prevent weight gain), psychoeducation, addictions’ treatment, vocational and education plans, and group interventions for patients and their families, as well as crisis intervention services, main components of FEP services (Addington et al., Reference Addington, McKenzie, Norman, Wang and Bond2013), tracking the process of outcome with the expectation of a full recovery. In addition to routine visits, patients were telephonically contacted to schedule face-to-face appointments at 1- and 3-year follow-ups for research purposes.
Tracing and long-term follow-up procedures
Baseline and 10-year characteristics, such as years of education, education level (‘only elementary education’ v. ‘other’), socioeconomic status derived from the parents' occupation (‘low qualification worker’ v. ‘other’), living area (‘urban’ v. ‘rural’, defined as more or less than 10 000 inhabitants, respectively), relationship status (‘married/cohabiting’ v. ‘single/divorced/separate or widowed’), living status (‘alone’ v. ‘other’), employment status (‘employed’ v. ‘unemployed’), and first degree family history of psychosis, which was based on the subject and family reports (‘yes’ v. ‘no’), as well as tobacco, alcohol and cannabis consumption as dichotomous (no/yes) measures were recorded. Information on suicidal behaviors, i.e. ‘potentially self-injurious behavior for which the person intended to kill himself/herself’ (Silverman et al., Reference Silverman, Berman, Sanddal, O'Carroll and Joiner2007a, Reference Silverman, Berman, Sanddal, O'Carroll and Joiner2007b), which encompasses suicide attempts and suicide completions, were collected.
Based on follow-up information, diagnoses stability v. change was considered at 10 years reassessment by means of SCID-I information.
Symptom assessment
Clinical data were collected at four different points throughout the 10-year period. The same senior consultant psychiatrist (BC-F) interviewed patients at the baseline, 1-year, 3-year and 10-year follow-up. Respondents completed face-to-face interviews. Clinical symptoms of psychosis were assessed by the Scale for the Assessment of Negative Symptoms (SANS) (Andreasen, Reference Andreasen1983) and the Scale for the Assessment of Positive Symptoms (SAPS) (Andreasen, Reference Andreasen1984). SANS-SAPS dimensions of positive (scores for hallucinations and delusions), disorganized (scores for formal thought disorder, bizarre behavior and inappropriate affect) and negative (scores for alogia, affective fattening, apathy and anhedonia) symptoms were calculated (Grube et al., Reference Grube, Bilder and Goldman1998). Manic symptoms were assessed with the Young Mania Rating Scale (YMRS) (Young et al., Reference Young, Biggs, Ziegler and Meyer1978), general psychopathology was assessed with the Brief Psychiatric Rating Scale (BPRS) (Flemenbaum and Zimmermann, Reference Flemenbaum and Zimmermann1973), and depressive symptom severity was measured using the Calgary Depression Scale for Schizophrenia (CDSS) (Addington et al., Reference Addington, Addington, Maticka-Tyndale and Joyce1992).
Neurocognition measures
The WAIS-III vocabulary subtest, a proven measure of crystallized intelligence, was used to estimate premorbid IQ. A measure of global cognitive functioning (GCF) was calculated in accordance with previous methodology (Reichenberg et al., Reference Reichenberg, Harvey, Bowie, Mojtabai, Rabinowitz, Heaton and Bromet2009). Scores on the seven cognitive domains fundamentally impaired in psychosis were converted to T-scores derived from a healthy comparison sample and converted to deficit scores that reflected the presence and severity of cognitive deficit in each cognitive domain. Deficit scores were then averaged to create the GCF measure at baseline and the 3-year and 10-year follow-up assessments.
Measures of 10-year functional outcomes
Ten-year functional outcomes were measured by three widely used scales: the Disability Assessment Scale (DAS) Spanish version (Mañá et al., Reference Mañá, Ivorra and Girón1998), the Global Assessment of Function (GAF) (Pedersen et al., Reference Pedersen, Hagtvet and Karterud2007) and the Quality of Life Scale (QLS) (Heinrichs et al., Reference Heinrichs, Hanlon and Carpenter1984).
The DAS is a semistructured interview that assesses several areas of functioning: self-care, social withdrawal, participation in the household, relationship with spouse or partner, occupational role and general interest. These areas are rated on a 6-point scale: 0, no disability; 1, minimum disability; 2, obvious disability; ⩾3, serious to maximum disability.
The GAF is a numeric scale (0 through 100) used by mental health professionals to subjectively rate social, occupational and psychological functioning. The highest ratings are 91–100, ‘Superior functioning in a wide range of activities’, and the lowest ratings 10–1, ‘Persistent inability to maintain minimal personal hygiene’.
The QLS is a 21-item scale rated from a semistructured interview providing information on four domains (interpersonal relations, instrumental role, intrapsychic foundations, common objects and activities) on a 7-point scale (0–1, severe impairment, to 5–6, normal or unimpaired functioning). The specific descriptors vary among items, but the high end of the scales (scores of 5 and 6) reflects normal or unimpaired functioning, and the low end of the scales (scores of 0 and 1) reflects severe impairment of the function in question.
Assessment interviews were performed by two experienced raters (psychiatrist and psychologist) per subject. An overall judgment was established with a full consensus rating reached together with the psychiatric nurse and the social worker.
Recovery, comprising symptomatic remission and adequate psychosocial function at 1, 3 and 10-year follow-ups, was determined by scores of 2 or less in the corresponding items of the SANS and the SAPS scales, as established by the remission work group (Andreasen et al., Reference Andreasen, Carpenter, Kane, Lasser, Marder and Weinberger2005), and a score of 1 or less in the DAS scale.
Data analyses
The data were analyzed using the R statistical computer program (R Development Core Team, 2010), version 3.5.1 (http://www.r-project.org). Cluster analyses were performed using R packages (script available upon request).
The possible relationships between the functional characteristics were explored using principal component analysis (PCA) and hierarchical clustering (HC) applied to the mean-centered and SD-scaled (z-transformed) data. The HC analysis was based on the Euclidean distance and Ward's linkage method. The results were visualized by means of dendrograms and a PCA biplot of the first two principal components.
The number of clusters examined was selected by visual inspection of the dendrograms and confirmed by discriminant function analysis. Clusters (using K-means results) were then compared on sociodemographic, clinical and cognitive variables using analysis of variance (ANOVA) or χ2. Post-hoc t tests with Bonferroni correction were conducted to examine pairwise relationships between clusters.
Results
Study description
Out of the 307 patients assessed at baseline, from February 2001 to July 2008, 10 patients died, including four deaths from suicide, during the follow-up period. Out of the 297 eligible participants who took part in the PAFIP study, 209 (70.4%) completed the 10-year follow-up reassessment (PAFIP-10). Attrition within the analysis sample seemed random; that is, the number of assessments was not associated with age, sex, negative symptoms, positive symptoms, employment, independent living, financial support or diagnosis (data available upon request).
Out of the 209 participants reassessed at the 10-year follow-up, 183 completed face-to-face interviews and 26 completed telephone interviews. The nonresponse rate was primarily accounted for by refusal to participate and loss to follow-up (see flowchart in Fig. 1). Among the 183 patients who had face-to-face, 10-year follow-up reassessments, 131 (71.6%) maintained the initial diagnosis. The dropouts did not differ from the analyzed sample in terms of sex, age or diagnosis (all p = 0.05).
Fig. 1. Flowchart depicting the inclusion of 10-year FEP patients in the study.
Establishing clusters of functional outcomes
A cluster analysis was conducted and included the following variables: the DAS score, the GAF score and scores in the four domains of the QLS (intrapsychic foundations, interpersonal relations, instrumental role and common objects and activities). The agglomeration schedule suggested a six-cluster solution. The first two principal components explained 85.83% of the variability. Cluster membership was then determined (see Fig. 2).
Fig. 2. Cluster membership.
The six clusters were as follows: cluster 1 (42.2% of the patients) was characterized by high scores on function (GAF and QLS), and low scores on disability (DAS), and was labeled ‘Normal overall’; cluster 2 (9.6%) demonstrated low scores on DAS, high scores on the GAF and all QLS measures except for intermediate scores in the interpersonal relationships domain and was labeled ‘Interpersonal intermediate’; cluster 3 (12.7%) demonstrated low scores on DAS, and high scores on the GAF and all QLS measures except for intermediate scores in the instrumental role domain and was labeled ‘Instrumental intermediate’; cluster 4 (12%) was characterized by high scores on QLS intrapsychic foundations, common objects and activities, and interpersonal relationships, intermediate scores on the GAF and DAS measures, but low scores in the QLS instrumental role domain, and was labeled ‘Instrumental severe’; cluster 5 (13.9%) was characterized by intermediate scores on the GAF, DAS and QLS intrapsychic foundations and common objects and activities domains, but low scores in the interpersonal relationships and instrumental role domains, and was labeled ‘Instrumental and interpersonal severe’; cluster 6 (9.6%) was characterized by low scores on GAF and QLS, and high scores on DAS and was labeled ‘Overall impaired’. The profile of each cluster is depicted in Fig. 3.
Fig. 3. Profile of each cluster.
Comparisons between clusters
The results of ANOVAs revealed significant effects of cluster membership on several variables. Descriptive statistics for each cluster and ANOVAs for sociodemographic, clinical and cognitive variables, including the results of the post-hoc analyses, are reported in Tables 1–3. Post-hoc comparisons of clusters with larger effects are summarized below:
– Cluster 1. The patients in cluster 1 had significantly better premorbid adjustment in childhood, early and late adolescence and adulthood than cluster 6, and better adjustment in early adolescence than cluster 5. They showed less severe symptomatology (lower BPRS scores) than clusters 4 and 6, particularly less severe negative symptoms (lower SANS scores) and better functioning (higher GAF scores) than clusters 4, 5 and 6, less severe positive symptoms than cluster 6 and less severe negative symptoms (negative dimension) than cluster 2 at the 10-year follow-up. This cluster showed significantly better GCF than cluster 6 at the 10-year follow-up. These patients were also more frequently in a working/studying occupational status than clusters 5 and 6 at baseline and the 3-year follow-up and clusters 3, 4, 5 and 6 at the 10-year follow-up; they were also more frequently married or in a relationship 10 years later (47.1%) and showed functional recovery (95.7%), remission (91.4%) and stability (92.9%). In this cluster, 19 patients (27.1%) were not prescribed antipsychotic treatment.
– Cluster 2. The patients in cluster 2 showed a similar functional outcome to patients in cluster 1. However, they presented measurably worse baseline global functioning (GAF at first contact) than cluster 1, with more premorbid social disabilities (premorbid DAS) than clusters 3, 4 and 5. Their premorbid IQ was significantly higher than that in cluster 6, and they had also reached significantly more years of education than the patients in clusters 5 and 6. This subgroup presented less severe positive symptoms than cluster 6 at the 10-year follow-up and less severe negative symptoms at the 3- and 10-year follow-ups.
– Cluster 3. The patients in cluster 3 showed less frequent functional recovery than patients in cluster 1, which was characterized by a significantly higher unemployment rate (42.9%), disability status (71.4%) and/or financial state support rate (81%) than cluster 1 at the 10-year follow-up. Similar to patients in cluster 2, their premorbid IQ was significantly higher than that of patients in cluster 6, and they presented less severe positive symptoms than cluster 6 at the 10-year follow-up and less severe negative symptoms than cluster 6 at the 3- and 10-year follow-ups.
– Cluster 4. The patients in cluster 4 did not achieve functional recovery relative to clusters 1 and 2 and had more frequent suicidal behaviors. These patients were more frequently unemployed and on disability and/or dependency conditions at the 10-year follow-up when compared with patients in cluster 1. All patients in this cluster were maintained on antipsychotic treatment (20 out of 20) and received financial state support; the percentage of patients using community-based rehabilitation resources (30%) was similar to that observed in cluster 6 (31.3%).
– Cluster 5. The patients in cluster 5 were characterized by the presence of more severe negative symptoms than the patients in clusters 1, 2 and 3 at the 10-year follow-up. The absence of occupational activities at baseline, 3-year and 10-year follow-ups was significant when compared to clusters 1 and 2, with larger proportions on disability (69.6%) and receiving financial state support (73.9%) for their condition.
– Cluster 6. The patients in this group showed significantly more severe positive and manic symptoms than the patients in the other clusters. The severity of positive symptoms (related to not achieving remission, stability or recovery) was particularly significant in this cluster at the 10-year follow-up. For negative symptoms, the severity in this cluster was significant when compared with all clusters at the 10-year assessment. The patients were characterized by their poor premorbid adjustment in early and late adolescence and adulthood, the low level of education and the persistent absence of occupational activities (not been working or engaged in any other activity such as formal studying). Ten years later, these patients were less frequently without a partner (81.3%) and were more likely to be receiving financial state support (16 out of 16), presenting a disability (75%), being in a dependence situation (43.8%), and using community-based rehabilitation resources (31.3%).
Table 1. Cluster comparisons at baseline on sociodemographic, clinical, cognitive and functional characteristics
DUI, duration untreated illness; DUP, duration untreated psychosis; CGI, clinical global impression; YMRS, Young Mania Rating Scale; CDSS, Calgary Depression Scale for Schizophrenia; SAPS, Scale Assessment Positive Symptoms; SANS, Scale Assessment Negative Symptoms; PAS, Premorbid Adjustment Scale; DAS, Disability Assessment Scale; GAF, Global Assessment of Functioning.
Table 2. Cluster comparisons at 10-year follow-up on sociodemographic, clinical, cognitive and functional characteristics
CGI, clinical global impression; YMRS, Young Mania Rating Scale; CDSS, Calgary Depression Scale for Schizophrenia; SAPS, Scale Assessment Positive Symptoms; SANS, Scale Assessment Negative Symptoms; DAS, Disability Assessment Scale; GAF, Global Assessment of Functioning; FEP, First Episode Psychosis.
Table 3. Cluster comparisons at 3-year follow-up on sociodemographic, clinical, cognitive and functional characteristics
CGI, Clinical Global Impression; YMRS, Young Mania Rating Scale; CDSS, Calgary Depression Scale for Schizophrenia; SAPS, Scale Assessment Positive Symptoms; SANS, Scale Assessment Negative Symptoms; DAS, Disability Assessment Scale; GAF, Global Assessment of Functioning.
Discussion
Summary of findings
Mental health is defined as ‘a state of welfare in which the individual realizes his own abilities, can cope with the normal stresses of life, can work productively and fruitfully, and is able to make a contribution to his community’ (WHO, 2008). This definition expresses what constitutes good functional outcome after an FEP. It is very good news that our results showed that 42.2% of FEP patients, those considered with normal overall functioning, showed good outcomes, and only 9.6%, those overall impaired, experienced poor long-term functional outcomes. A systematic review carried out in 2012 (Clemmensen et al., Reference Clemmensen, Vernal and Steinhausen2012), including studies published between 1980 and 2012, showed that only 15.4% of early onset schizophrenia patients showed a good outcome, while 60.1% experienced poor outcomes. However, Clemmensen and colleagues remarked that the rate of poor outcomes was significantly higher in the low attrition samples compared to the high-attrition samples (rates ranging from 20% to 60%) and that these types of findings depended on the definition used for good or poor functioning in terms of occupational/vocational status, independent living and active social interactions. In our study, with a more than acceptable retention rate (70.4%), a good outcome was confirmed in 42.2% of the patients, those who 10 years later had attained their functional milestones in educational, residential, social and vocational domains, being really low the percentage of those with poor outcome, supporting that rates of good outcome may decrease in longer studies, while rates of poor outcome do not necessarily increase with time as would be expected in a progressive deteriorating illness (Menezes et al., Reference Menezes, Arenovich and Zipursky2006). Thus, using the words of Zipursky et al. (Reference Zipursky, Agid and Remington2018), 30 years of studying individuals with FEP has transformed our understanding of the outcomes of schizophrenia, once thought to be a progressive, deteriorating mental disorder and currently seen as a manageable condition for many people.
Overall, our results also highlighted the fact that it would be too naive to assume that the functional outcomes would simply group into good or poor outcomes. Menezes et al. (Reference Menezes, Arenovich and Zipursky2006) reported that good and poor outcomes (grossly defined) were reported in 42% and 27% of FEP, respectively. If we had classified patients in this manner, we would not have taken into account almost 50% of the sample, i.e. those patients that show a wide diversity in functional outcomes. Indeed, in our study, the two clusters (clusters 2 and 3) close to normal function (cluster 1) encompassed 22.3% of patients, with cluster 2 being moderately impaired in interpersonal relationships and cluster 3 being moderately impaired in instrumental roles. Clusters 4 and 5 comprised patients who were more homogeneous on the GAF and DAS measures than the QLS domains and presented more severe deficits.
Negative symptomatology and premorbid adjustment as explanatory variables for outcome variability
The majority of individuals in the PAFIP cohort experienced a reduction and stabilization of positive symptoms, whereas negative symptoms showed less variation over 10 years. There is little doubt that the negative symptom dimension of psychotic disorders adversely influences functioning and outcomes (Galderisi et al., Reference Galderisi, Mucci, Buchanan and Arango2018). As stated many times, people with sustained and persistent negative symptoms may experience significantly poorer functioning, worse psychological outcomes and lower rates of recovery compared to people who display reductions in negative symptoms over time (Austin et al., Reference Austin, Mors, Secher, Hjorthoj, Albert, Bertelsen, Jensen, Jeppesen, Petersen, Randers, Thorup and Nordentoft2013). Our results showed that all patients, except for those clustered as normal overall (cluster 1), failed to show a significant improvement in negative symptoms. However, clusters 2, 3, 4 and 5 displayed mild levels of negative symptoms, while only those clustered as overall impaired (cluster 6) displayed severe levels at the 10-year follow-up. Some studies of FEP patients have found that severe negative symptoms can be present at illness onset (Dominguez et al., Reference Dominguez, Saka, Lieb, Wittchen and van Os2010) and can be predicted by being male, the particular schizophrenia diagnosis and a lengthy DUP (Austin et al., Reference Austin, Mors, Budtz-Jorgensen, Secher, Hjorthoj, Bertelsen, Jeppesen, Petersen, Thorup and Nordentoft2015). Galderisi et al. (Reference Galderisi, Mucci, Bitter, Libiger, Bucci, Fleischhacker and Kahn2013) found that persistent negative symptoms (not confounded by depression) were present in 6.7% of their FEP patients’ sample, being blunted affect the symptom that more often persisted. These patients had a poorer psychopathological outcome, a worse global functioning and more frequent discontinuation after 1 year of treatment. It therefore comes as no surprise that these patients were less likely to achieve long-term optimal functioning. On the other hand, poor social functioning prior to and at illness onset could be an important marker for negative symptoms and eventually of poor functional outcomes (Santesteban-Echarri et al., Reference Santesteban-Echarri, Paino, Rice, Gonzalez-Blanch, McGorry, Gleeson and Alvarez-Jimenez2017). Thus, the combination of poor premorbid adjustment and enduring negative symptoms could potentially be present in those patients showing poor functional outcomes.
We found that the overall normal group (cluster 1) differed significantly from the patients who presented more severe functional impairments in their premorbid adjustment. Several studies have shown that the differences in functioning prior to disease onset may contribute to the heterogeneity of schizophrenia outcome (Larsen et al., Reference Larsen, Friis, Haahr, Johannessen, Melle, Opjordsmoen, Rund, Simonsen, Vaglum and McGlashan2004; Cole et al., Reference Cole, Apud, Weinberger and Dickinson2012; Quee et al., Reference Quee, Meijer, Islam, Aleman, Alizadeh, Meijer and van den Heuvel2014). Using a similar cluster analysis to that in our study, Quee et al. (Reference Quee, Meijer, Islam, Aleman, Alizadeh, Meijer and van den Heuvel2014) identified six clusters in premorbid adjustment profiles, labeled normal, social intermediate, academic decline, overall decline, overall intermediate and overall impaired adjustment, with each differing from the others on cognitive, clinical and functional characteristics after disease onset. Cole et al. (Reference Cole, Apud, Weinberger and Dickinson2012) identified three latent trajectory classes (poor-deteriorating, insidious-onset, good-stable) from premorbid adjustment to functioning after onset, suggesting that subtyping schizophrenia patients in terms of premorbid history may be useful for untangling the heterogeneity of schizophrenia outcomes. These studies are in line with the theories that propose different etiologies in the development of psychosis, suggesting patterns of premorbid maladjustment may reflect a neurodevelopmental pathway to psychosis (Larsen et al., Reference Larsen, Friis, Haahr, Johannessen, Melle, Opjordsmoen, Rund, Simonsen, Vaglum and McGlashan2004). It is not clear where the boundaries are, but the confluence of poor premorbid adjustment, lower premorbid IQ, fewer years of education and worse GCF at baseline in those patients who presented an overall impairment in long-term function may support more severe forms of the disease (Cuesta et al., Reference Cuesta, Sanchez-Torres, Cabrera, Bioque, Merchan-Naranjo, Corripio, Gonzalez-Pinto, Lobo, Bombin, de la Serna, Sanjuan, Parellada, Saiz-Ruiz and Bernardo2015). It will be difficult for researchers to establish whether these premorbid and neurocognitive characteristics, repeatedly found representative of the more deteriorating courses, are a cause or consequence of illness. However, there is no doubt that negative symptoms are a key, and hopefully modifiable, factor in the long-term functional outcomes.
Social factors that contribute to the variability in cluster outcome
Patients within the more dysfunctional clusters, cluster 4 (instrumental severe) and cluster 6 (overall impaired), used more community-based rehabilitation resources, designed to interface with the existing outreach mental health teams, and focused on psychosocial interventions. Dependency was particularly significant in cluster 6 (43.8%) compared to the rest of the clusters, as well as in cluster 4 (15%) when compared with cluster 1 (0%). Disability was more significant in clusters 3–6 (71.4, 90, 69.6 and 75%, respectively) than in clusters 1 and 2 (34.3% and 31.3%), and it was particularly remarkable in cluster 4 (90%). Similar results were observed for financial state support, but in this case, the percentage of patients in cluster 3 (instrumental intermediate) who received some kind of public government financial state support (81%) showed that the patients in this cluster spent more time studying than patients in cluster 6. Worldwide, the most common method of financing mental health care is taxation (60%), followed by social insurance (19%), out-of-pocket payments (16%), external grants (3%) and voluntary insurance (2%) (Saxena et al., Reference Saxena, Thornicroft, Knapp and Whiteford2007). Research evidence suggests that disability applicants with a valid diagnosis of schizophrenia have significant impairment across multiple dimensions of functioning and will typically remain impaired for the duration of the normal working ages or until new interventions are developed (Harvey et al., Reference Harvey, Heaton, Carpenter, Green, Gold and Schoenbaum2012). It was observed that only the patients in clusters 1 and 2 were more frequently employed during the 10 years of follow-up (61.4% and 56.3%, respectively). This suggests the importance of early employment for long-term functional outcomes and supports the effort of implementing specific EI services to address vocational and educational challenges after an FEP as part of the treatment.
Strengths and limitations
The major contribution of this study was differentiating the heterogeneity of functional outcomes in FEP patients using an objective evaluation in the determination of clusters. This kind of analysis enhanced the speed and sensitivity of identifying similarities and improved the consistency of annotation within clusters. Moreover, very few studies have followed a cohort of FEP patients over such a long period of time and examined functional outcomes accurately with a very low attrition rate. Some limitations should also be considered. First, the assessment points during the 10 years were not equally spaced over this period. The follow-up in the PAFIP cohort was scheduled for 3 years, at which point, the patients were discharged from EI to traditional community mental health services. For the purpose of the current study, they were contacted at 10 years, and we cannot rule out other information that was not recorded but could have affected outcomes. Second, the analyses were not focused on trajectories of changes over time, which would have been made possible only with more elaborate profiles. However, those analyses exceeded the aims of the current study. Third, the current study did not address the question of whether outcomes following treatment as usual might have been similar to the outcomes in these patients at the 10-year follow-up period.
Future research and suggestions
Our findings indicated that psychotic disorders are still mental illnesses with a rather unfavorable prognosis but with favorable rates clearly increasing. Studies using longitudinal data can contribute to a better understanding of the course of the illness and help clinicians identify critical periods that merit specific interventions and adjust the timing, length and intensity of the interventions to optimize primary outcomes (symptoms and functioning) for people who have experienced an FEP. Finally, cluster analysis methods can be leveraged to develop tailored care management interventions designed to improve specific functional outcomes.
Acknowledgements
The authors wish to thank all PAFIP research teams and all patients and family members who participated in the study.
Financial support
This work was supported by the Instituto de Salud Carlos III (PI14/00639 and PI14/00918). No pharmaceutical company has financially supported the study.
Conflict of interest
The authors have no conflict of interest to declare.
