Exploring Risk and Resilient Profiles for Functional Impairment and Baseline Predictors in a 2-Year Follow-Up First-Episode Psychosis Cohort Using Latent Class Growth Analysis

Journal of Clinical Medicine Article Exploring Risk and Resilient Profiles for Functional Impairment and Baseline Predictors in a 2-Year Follow-Up First-Episode Psychosis Cohort Using Latent Class Growth Analysis Estela Salagre 1, Iria Grande 1,*, Brisa Solé1, Gisela Mezquida 2, Manuel J. Cuesta 3, Covadonga M. Díaz-Caneja 4, Silvia Amoretti 2, Antonio Lobo 5, Ana González-Pinto 6,7, Carmen Moreno 4, Laura Pina-Camacho 4, Iluminada Corripio 7,8, Immaculada Baeza 9, Daniel Bergé10, Norma Verdolini 1, AndréF. Carvalho 11,12, Eduard Vieta 1,* , Miquel Bernardo 2and PEPs Group †   Citation: Salagre, E.; Grande, I.; Solé, B.; Mezquida, G.; Cuesta, M.J.; Díaz-Caneja, C.M.; Amoretti, S.; Lobo, A.; González-Pinto, A.; Moreno, C.; et al. Exploring Risk and Resilient Profiles for Functional Impairment and Baseline Predictors in a 2-Year Follow-Up First-Episode Psychosis Cohort Using Latent Class Growth Analysis. J. Clin. Med. 2021,10, 73. https://doi.org/10.3390/jcm10010073 Received: 2 December 2020 Accepted: 22 December 2020 Published: 28 December 2020 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. Copyright: © 2020 by the authors. LicenseeMDPI, Basel, Switzerland. This articleisanopenaccessarticledistributed under the terms and conditions of the CreativeCommonsAttribution(CCBY) license(https://creativecommons.org/ licenses/by/4.0/). 1 Bipolar and Depressive Disorders Unit, Hospital Clinic, Biomedical Research Networking Center for Mental Health Network (CIBERSAM), August Pi I Sunyer Biomedical Research Institute (IDIBAPS), University of Barcelona, 08036 Barcelona, Spain; esalagr[email protected] (E.S.); [email protected] (B.S.); [email protected] (N.V.) 2Barcelona Clinic Schizophrenia Unit, Neuroscience Institute, Hospital Clinic of Barcelona, Biomedical Research Networking Center for Mental Health Network (CIBERSAM), Department of Medicine, Institut de Neurociències, August Pi I Sunyer Biomedical Research Institute (IDIBAPS), Universitat de Barcelona, 08036 Barcelona, Spain; [email protected] (G.M.); amor[email protected] (S.A.); bernar[email protected] (M.B.) 3 Department of Psychiatry, Instituto de Investigaciones Sanitarias de Navarra (IdiSNa), Complejo Hospitalario de Navarra, 31008 Pamplona, Spain; [email protected] 4 Department of Child and Adolescent Psychiatry, Institute of Psychiatry and Mental Health, Hospital General Universitario Gregorio Marañón, IiSGM, CIBERSAM, School of Medicine, Universidad Complutense, 28007 Madrid, Spain; [email protected] (C.M.D.-C.); [email protected] (C.M.); [email protected] (L.P.-C.) 5Department of Medicine and Psychiatry, Instituto de Investigación Sanitaria Aragón (IIS Aragón), Universidad de Zaragoza, 50009 Zaragoza, Spain; [email protected] 6Department of Psychiatry, Hospital Universitario de Alava, BIOARABA Health Research Institute, University of the Basque Country, 01009 Vitoria, Spain; [email protected] 7Centro de Investigación Biomédica en Red de Salud Mental (CIBERSAM), 28029 Madrid, Spain; [email protected] 8Department of Psychiatry, Biomedical Research Institute Sant Pau (IIB-SANT PAU), Hospital Sant Pau, Universitat Autònoma de Barcelona (UAB), 08041 Barcelona, Spain 9Biomedical Research Networking Center for Mental Health Network (CIBERSAM), Child and Adolescent Psychiatry and Psychology Department, August Pi I Sunyer Biomedical Research Institute (IDIBAPS), Hospital Clínic of Barcelona, SGR-881, Universitat de Barcelona, 08036 Barcelona, Spain; [email protected] 10 Hospital del Mar Medical Research Institute, CIBERSAM, Autonomous University of Barcelona, 08003 Barcelona, Spain; [email protected] 11 Centre for Addiction and Mental Health, Department of Psychiatry, University of Toronto, Toronto, ON M6J 1H4, Canada; andr[email protected] 12 The IMPACT (Innovation in Mental and Physical Health and Clinical Treatment) Strategic Research Centre, School of Medicine, Barwon Health, Deakin University, Geelong, VIC 3220, Australia *Correspondence: [email protected] (I.G.); [email protected] (E.V.); Tel.: +34-93-227-5400 (I.G.); +34-93-227-5400 (E.V.) † Membership of the PEPs Group is provided in the Acknowledgments. Abstract: Being able to predict functional outcomes after First-Episode Psychosis (FEP) is a major goal in psychiatry. Thus, we aimed to identify trajectories of psychosocial functioning in a FEP cohort followed-up for 2 years in order to find premorbid/baseline predictors for each trajectory. Additionally, we explored diagnosis distribution within the different trajectories. A total of 261 adults with FEP were included. Latent class growth analysis identified four distinct trajectories: Mild impairmentImproving trajectory (Mi-I) (38.31% of the sample), Moderate impairment-Stable trajectory (Mo-S) (18.39%), Severe impairment-Improving trajectory (Se-I) (12.26%), and Severe impairment-Stable trajectory (Se-S) (31.03%). Participants in the Mi-I trajectory were more likely to have higher parental socioeconomic status, less severe baseline depressive and negative symptoms, and better premorbid adjustment than individuals in the Se-S trajectory. Participants in the Se-I trajectory were more likely to have better baseline verbal learning and memory and better premorbid adjustment than those in J. Clin. Med. 2021,10, 73. https://doi.org/10.3390/jcm10010073 https://www.mdpi.com/journal/jcm J. Clin. Med. 2021,10, 73 2 of 20 the Se-S trajectory. Lower baseline positive symptoms predicted a Mo-S trajectory vs. Se-S trajectory. Diagnoses of Bipolar disorder and Other psychoses were more prevalent among individuals falling into Mi-I trajectory. Our findings suggest four distinct trajectories of psychosocial functioning after FEP. We also identified social, clinical, and cognitive factors associated with more resilient trajectories, thus providing insights for early interventions targeting psychosocial functioning. Keywords: first-episode psychosis; functional outcomes; risk factors; early intervention; neurocognition ; latent class analysis; precision medicine 1. Introduction Psychosocial functioning refers to the ability to perform in daily living activities such as work, studies or recreational activities, and to establish satisfying interpersonal relationships with others [ 1 ]. In the last 50 years, psychiatry has progressively moved from a deficit-based care (which focuses on symptomatic remission), to a model oriented towards functional recovery, meaning that helping the patient to meet his/her personal goals has become as critical as achieving symptomatic remission [ 2 , 3 ]. In fact, it is increasingly accepted that functional outcomes are more meaningful when measuring treatment response than are scores on various scales rating only psychiatric symptoms [ 4 ]—and more aligned with what the patient ultimately expects from treatment [ 5 ]. Therefore, full functional remission is currently a preeminent goal in psychiatry. Prior evidence suggests that achieving full functional recovery short after first-episode psychosis (FEP) is a stronger predictor of long-term full functional remission than symptomatic remission [ 6 , 7 ]. This evidence underscores the need to find early and modifiable factors associated with functional impairment already from early stages. Although multiple studies have investigated putative predictors of poor psychosocial functioning after FEP [ 8 ], most of them have approached this question using a dichotomous outcome, that is, presence vs. absence of functional impairment. The real picture seems far more complex, though, given the highly divergent outcomes in psychosocial functioning that individuals can experience after FEP, which encompass varying degrees of functional difficulties and different evolutions over time. Some patients will experience an early functional recovery, others might exhibit severe functional difficulties from illness onset and some subgroups might experience (persistent or transitory) mild to moderate functional impairment, which still have a negative impact on their daily life. Hence, the real challenge is to predict early in the course of the disease which individual will fall into each of these trajectories in order to be able to design earlier and more tailored treatments for social and personal recovery [2,9–11]. Statistical methods like latent class growth analysis (LCGA) can help to provide a more accurate picture of the heterogeneous course in psychosocial functioning that can be observed following FEP, as it allows considering different outcomes of the same characteristic simultaneously [ 12 , 13 ]. To our knowledge, only few studies so far have applied these statistical techniques to assess functional outcomes in FEP samples [ 14 – 16 ], and none of them has considered simultaneously sociodemographic variables, clinical features and an extensive set of cognitive domains, all of them previously related to poor functional outcomes [ 17 ]. Therefore, our main aim was to identify different trajectories of functional impairment in the 24-month follow-up of a FEP cohort and to assess putative predictors of these diverse trajectories, with a special focus on resilient trajectories. As a secondary objective, we aimed to explore diagnoses distribution within the different trajectories. 2. Experimental Section 2.1. Participants The current study is based on data from the project ‘Phenotype–genotype and environmental interaction. Application of a predictive model in first psychotic episodes’ (PEPs J. Clin. Med. 2021,10, 73 3 of 20 study), a multicenter, longitudinal, naturalistic follow-up study [ 18 ]. A total of 16 centers throughout Spain participated in this study; fourteen of them were members of the Biomedical Research Networking Center for Mental Health (CIBERSAM) [ 19 ] and two were collaborator centers [ 18 ]. The study was conducted in accordance with the ethical principles of the Declaration of Helsinki. It was approved by the ethics committees at each participating center (project identification code: 2008/4232). All participants or their legal guardians signed an informed consent after providing them a full explanation of the study’s procedures. The detailed protocol of the PEPs study was published elsewhere [ 18 , 20 ]. Briefly, a total of 335 subjects with FEP were recruited by all the participating centers, from April 2009 to April 2012. Individuals were included in the PEPs study if they were between 7 and 35 years old, presented first lifetime psychotic symptoms for at least one week in the last 12 months, were fluent in Spanish language, and were willing to sign the informed consent. Intellectual disability according to the Diagnostic and Statistical Manual of mental disorders, 4th edition (DSM-IV) criteria [ 21 ], history of head trauma with loss of consciousness, and presence of an organic disease with mental repercussions constituted exclusion criteria. Patients had been under antipsychotic treatment for less than 12 months at study entry. Follow-up assessments were conducted at 2 months, 6 months, 12 months and 24 months following inclusion. 2.2. Assessment 2.2.1. Baseline Sociodemographic Data Sociodemographic data were collected from all participants at baseline, including sex, age, ethnicity, educational level, marital status, current living situation, occupation, and parental socioeconomic status (SES). Parental SES was determined using the Hollingshead Two-Factor Index of Social Position [ 22 ]. Personal and family history of somatic and psychiatric disorders was also compiled. History of drug misuse was evaluated using the adapted version of a Multidimensional Assessment Instrument for Drug and Alcohol Dependence scale [ 23 ]. The Family Environment Scale (FES), a self-report instrument, was used to assess the patients’ perception of the social climate within their families [24,25]. 2.2.2. Baseline Clinical and Functional Assessment For all subjects in the study, diagnosis was established by experienced mental health professionals using the Structured Clinical Interview for DSM-IV Axis I Disorders (SCIDI) [ 21 , 26 ]. Psychopathology was evaluated using the Spanish validated versions of the Positive and Negative Syndrome Scale (PANSS) [ 27 , 28 ], the Young Mania Rating Scale (YMRS) [ 29 , 30 ], and the Montgomery–Åsberg Depression Rating Scale (MADRS) [ 31 , 32 ]. Premorbid adjustment was estimated by means of the retrospective Premorbid Adjustment Scale (PAS) [ 33 ]. The Functional Assessment Short Test (FAST) [ 1 , 34 ] was used to determine psychosocial functioning. It comprises 24 items, which evaluate six specific functioning domains: autonomy, occupational functioning, cognitive functioning, financial issues, interpersonal relationships, and leisure time. This scale seeks to identify changes or difficulties in functionality attributable to the illness. The FAST scores range from 0 to 72. According to the cut-off classification as proposed by Bonnín et al. [ 35 ], FAST scores > 40 are indicative of severe functional impairment, FAST score between 21 and 40 indicate moderate functional impairment, FAST scores between 12–20 indicate mild impairment, and ≤ 11 points in the FAST reflect no functional impairment. This scale has shown to be sensitive to change and has been validated for FEP [ 36 ]. In all the aforementioned scales, higher scores are indicative of greater clinical severity or functional impairment. History of traumatic life events was assessed through the Spanish version of the Trauma Questionnaire (TQ) [ 37 , 38 ]. Duration of untreated psychosis (DUP), defined as the number of days elapsed between the onset of positive psychotic symptoms and the initiation of the first appropriate treatment for psychosis, was also registered. It was estimated using the Symptom Onset in Schizophrenia (SOS) inventory [39]. J. Clin. Med. 2021,10, 73 4 of 20 2.2.3. 2-Month Follow-Up Neuropsychological Assessment Participants were likewise evaluated using a comprehensive neuropsychological battery encompassing most of the cognitive domains proposed by the National Institute of Mental Health MATRICS consensus [ 40 ]. The evaluation was performed by trained neuropsychologists in the first two months after the inclusion of the participant in the study to avoid the interference of acute psychopathological manifestations on neurocognitive assessments. The neuropsychological assessment comprised the following cognitive domains: (1) estimated Intelligence Quotient (IQ) (calculated based on the performance on the Vocabulary subtest from the Wechsler Adult Intelligence Scale (WAIS-III) [ 41 ]); (2) executive function (Stroop Color-Word Interference Test [ 42 ], Wisconsin Card Sorting Test (WCST) [ 43 ] and Trail Making Test (TMT), form B [ 44 ]); (3) attention (Continuous Performance Test-II (CPT-II) [ 45 ]); (4) processing speed (TMT, form A [ 46 ] and categorical (Animal Naming) and phonemic (F-A-S) components of the Controlled Oral Word Association Test (COWAT) [ 47 ]); (5) verbal memory (Spanish version of the California Verbal Learning Test, the Test de Aprendizaje Verbal España-Complutense (TAVEC) [ 48 ]); (6) working memory (Digit span and Letter-Number sequencing subtests of WAIS-III [ 41 ]); and (7) social cognition (Mayer–Salovey–Caruso Emotional Intelligence Test (MSCEIT) [ 49 , 50 ]). The neuropsychological battery is described in further detail in the PEPsCog study [51]. 2.3. Statistical Analysis 2.3.1. Identification of Functional Trajectories: Latent Class Growth Analysis LCGA was used to identify distinct functioning trajectories over the 24-month followup. In the current analysis, individual class membership was assigned on the basis of FAST total scores measured at five time points over the two-year follow-up period, namely at baseline, 2-, 6-, 12-, and 24-month follow-up. We only included in the analysis individuals over 18 years old, as the FAST scale has only been validated in adult samples, and with information on the FAST scale in at least two follow-up assessments. This left a sample of 275 adult participants. Each model was rerun 100 times using different start values to avoid converging to local maxima [ 52 ]. To accommodate expected fluctuations over time, we estimated linear and quadratic terms. In order to determine the optimal number of trajectory classes, models with increasing number of latent classes (from 1to 4-class models) were fitted to the data and the best-fitting model was selected according to the following goodness-offit indices: Akaike’s Information Criterion (AIC), Bayesian Information Criterion (BIC), samples-size-adjusted BIC (aBIC), and entropy. Lower values of AIC, BIC, and aBIC suggest a more parsimonious model, while higher entropy also indicates better model fit. Entropy ranges from 0 to 1 and is a summary indicator of the accuracy with which models classify individuals into their most likely class. Entropy with values approaching 1 indicate clear delineation of classes [ 53 ]. Interpretability and parsimony of the model were also taken into consideration in the final selection of the model. LCGA analyses were performed on R version 3.6.3, using the ‘lcmm’ package ([ 54 ]; https://cran.r-project.org/web/packages/ lcmm/index.html). 2.3.2. Identification of Baseline Predictors of Functional Trajectory Membership To identify putative baseline predictors of trajectory membership, the estimated latent classes (i.e., the estimated trajectory group) derived from LCGA were imported to SPSS, version 23 (SPSS Inc., Chicago, IL, USA), for a three-step analysis: First, we created seven cognitive composites to be used as putative baseline predictors using data from the two-month follow-up neurocognitive assessment. To do so, patients’ raw scores on each neuropsychological task were standardized to z-scores based on the performance of the whole sample. The selection of the tasks within each cognitive domain was based on previous works from the PEPs group [51,55,56]. Afterwards, z-scores of different tests were summed and averaged to create the following seven cognitive composites: (1) the processing speed composite, based on the word–color task from the Stroop Test J. Clin. Med. 2021,10, 73 5 of 20 and the TMT-A; (2) the working memory composite, which included the Letter-Number Sequencing and the Digit-Span WAIS-III subtests; (3) the verbal learning and memory index, which was composed of the total trials 1–5 list A, short free recall, short cued recall, delayed free recall, delayed cued recall, and recognition scores of the TAVEC; (4) the executive function composite, calculated based on the number of categories and perseverative errors of the WCST, the Stroop Interference Test, and the TMT-B; (5) the attention composite score, which was based on several measures of the CPT-II, such as commission and reaction time; (6) the verbal fluency composite which was composed of the Category Fluency (Animal Naming) and the F-A-S Test of the COWAT; and (7) the social cognition composite, which included the Emotional Management of the MSCEIT. Whenever extreme scores in the performance of the aforementioned test were detected (i.e., more than four standard deviations (SD) above or below the mean), the scores were truncated to z = +/−4 . Since higher scores in CPT-II, WCST perseverative errors, and TMT-A and -B indicate poorer performance, z-scores obtained from measures of these tests were reversed before constructing the corresponding composite scores. Second, candidate predictors (i.e., baseline sociodemographic and clinical variables as well as the created cognitive composites) were compared between trajectory classes using Kruskal-Wallis and chi-square tests, as appropriate. The Kruskal–Wallis test was selected for continuous variables since they did not follow a normal distribution, as assessed visually and by the Kolmogorov–Smirnov test. When applicable, post-hoc comparison analyses with Bonferroni correction for multiple comparisons were performed to further clarify the presence of significant differences between trajectory classes. Third, those variables found to be statistically significant in the post-hoc analysis in at least two pair-wise comparisons were then entered into a multinomial regression model to determine which candidate factors independently predicted trajectory membership, adjusting for age and sex. For the PANSS scale, only the PANSS positive and negative subscales were entered as independent variables to avoid multicollinearity. Significant putative predictors for the multivariable model were identified using a stepwise backwards elimination process [ 57 ], with sex and age entered as fixed factors. The identified latent classes were used as the dependent variable. Since we were interested in exploring predictors of resilient trajectories, we selected the most impaired group as the reference category. 2.3.3. Diagnosis Distribution within the Identified Functional Trajectories Lastly, to explore whether diagnosis distribution differed within each functional trajectory and how it changed over time, we compared using chi-square tests the proportion of individuals with a diagnosis of Schizophrenia, Bipolar disorder, Schizoaffective disorder, and Other psychoses (including psychotic disorder not-otherwise specified, brief psychotic disorder, schizophreniform psychosis, delusional disorder, substance-induced psychosis) in each of the predicted functional trajectories at baseline, 1-year and 2-year follow-up. The level of statistical significance for all analyses was set at p< 0.05. 3. Results 3.1. Sample Characteristics and Attrition Analysis The final sample included 261 participants. A total of 14 individuals were not considered for the analyses since information on their FAST scores was only available at one time point. Therefore, they were treated as drop-outs. The baseline characteristics of the final sample are presented in Table 1. A comparison between drop-outs and non-drop-outs at baseline, 12-month, and 24-month follow-up can be found in the Supplementary Table S1. The median age of the final sample was 25.05 years old (Interquartile Range: 9) and 33% of the participants were female. Among those subjects that dropped out from the study, there was a lower proportion of Caucasian participants and of participants with a family history of psychiatric disorders. Subjects that dropped out from the study reported more frequently substance misuse at baseline too. J. Clin. Med. 2021,10, 73 6 of 20 Table 1. Baseline characteristics of the final sample (n= 261). Characteristics Median (IQR)/n(%) Age (Years) 25.05 (9) Sex (Female) 87 (33.3) Marital status (Single) 222 (85.1) Ethnicity (Caucasian) 228 (87.4) Parental socioeconomic status (Medium-high) 141 (54.6) Living situation (Living independently) 55 (21.1) Educational level (Higher education) 115 (44.2) Occupational status (Active *) 136 (52.1) Somatic comorbidity (Yes) 73 (28.0) Family history of psychiatric disorder (Yes) 146 (55.9) * Active includes workers and students. 3.2. Latent Classes of Functional Trajectories After examining fit indices, entropy, parsimony, and interpretability of the model, the 4-class model including the quadratic term was selected as optimal for our data (Table 2). Entropy was acceptable (0.76) for the 4-class model as well as post mean class probabilities (0.81 for Class 1, 0.92 for Class 2, 0.82 for Class 3, and 0.84 for Class 4). This suggests that with the 4-class model individuals were likely to be correctly assigned to their respective latent class. Table 2. Goodness-of-fit statistics of latent class growth analysis with one-to-four class solutions of psychosocial functioning trajectories. Fit Statistics a% of the Sample in Each Class Number of Classes Number of Parameters AIC BIC aBIC Entropy Class 1 Class 2 Class 3 Class 4 1 4 9029.81 9044.07 9031.39 - 100 - - - 2 8 8675.19 8703.71 8678.35 0.82 54.02 45.98 - - 3 12 8639.13 8681.91 8643.86 0.71 41.00 32.18 26.82 - 4 16 8574.11 8631.14 8580.41 0.76 38.31 18.39 12.26 31.03 Abbreviations: AIC: Akaike’s Information Criterion; BIC: Bayesian Information Criterion; aBIC: sample size–adjusted Bayesian Information Criterion. a Lower values (AIC, BIC, and aBIC) indicate a better model fit. Higher entropy indicates better model fit. Values of 0.4, 0.6, and 0.8 represent low, medium, and high entropy [58]. Bold is used here to indicate which model was selected. The mean FAST scores at each assessment point of individuals grouped according to their predicted trajectory are presented in Figure 1. One group showed mild impairment at baseline and no impairment by the end of the follow-up, and was referred to as Mild impairment-Improving trajectory (Class 1; n= 100 (38.31%)). Another group, denominated as Moderate impairment-Stable trajectory (Class 2; n= 48 (18.39%)) exhibited moderate functional impairment at baseline and throughout the follow-up. A third group presented with severe functional impairment that improved along the follow-up. It was referred as Severe impairment-Improving trajectory (Class 3; n= 32 (12.26%)). The last group, termed as Severe impairment-Stable trajectory, displayed severe-moderate functional impairment throughout the follow-up (Class 4; n= 81 (31.03%)). Thus, 50.57% of the sample showed a trajectory characterized by a functional improvement/recovery (“Improving trajectories”), while 49.42% exhibited persistent functional impairment during follow-up (“Stable trajectories”). 3.3. Baseline Predictors of Trajectory Membership The comparison between the four psychosocial functioning trajectories on sociodemographic, clinical, and neuropsychological variables is presented in Table 3. The baseline variables found to be statistically different between groups in at least two pairwise comparisons were: parental SES, alcohol use, PANSS positive, PANSS negative, PANSS general, PANSS total, Young total, MADRS total, PAS total, verbal learning, and memory and working memory. As previously stated, for the PANSS scale, only the PANSS positive and negative subscales were entered as independent variables in the multinomial regression model. J. Clin. Med. 2021,10, 73 7 of 20 J. Clin. Med. 2021, 10, x FOR PEER REVIEW 7 of 23 2 8 8675.19 8703.71 8678.35 0.82 54.02 45.98 - - 3 12 8639.13 8681.91 8643.86 0.71 41.00 32.18 26.82 - 4 16 8574.11 8631.14 8580.41 0.76 38.31 18.39 12.26 31.03 Abbreviations: AIC: Akaike’s Information Criterion; BIC: Bayesian Information Criterion; aBIC: sample size–adjusted Bayesian Information Criterion. a Lower values (AIC, BIC, and aBIC) indicate a better model fit. Higher entropy indicates better model fit. Values of 0.4, 0.6, and 0.8 represent low, medium, and high entropy [58]. Bold is used here to indicate which model was selected. The mean FAST scores at each assessment point of individuals grouped according to their predicted trajectory are presented in Figure 1. One group showed mild impairment at baseline and no impairment by the end of the follow-up, and was referred to as Mild impairment-Improving trajectory (Class 1; n = 100 (38.31%)). Another group, denominated as Moderate impairment-Stable trajectory (Class 2; n = 48 (18.39%)) exhibited moderate functional impairment at baseline and throughout the follow-up. A third group presented with severe functional impairment that improved along the follow-up. It was referred as Severe impairment-Improving trajectory (Class 3; n = 32 (12.26%)). The last group, termed as Severe impairment-Stable trajectory, displayed severe-moderate functional impairment throughout the follow-up (Class 4; n = 81 (31.03%)). Thus, 50.57% of the sample showed a trajectory characterized by a functional improvement/recovery (“Improving trajectories”), while 49.42% exhibited persistent functional impairment during follow-up (“Stable trajectories”). Figure 1. Evolution of mean FAST scores within each of the functional trajectory groups derived from the latent class growth analysis. Higher scores in the FAST are indicative of greater functional impairment. 3.3. Baseline Predictors of Trajectory Membership The comparison between the four psychosocial functioning trajectories on sociodemographic, clinical, and neuropsychological variables is presented in Table 3. The baseline variables found to be statistically different between groups in at least two pairwise comparisons were: parental SES, alcohol use, PANSS positive, PANSS negative, PANSS general, PANSS total, Young total, MADRS total, PAS total, verbal learning, and memory and working memory. As previously stated, for the PANSS scale, only the PANSS positive and negative subscales were entered as independent variables in the multinomial regression model. Figure 1. Evolution of mean FAST scores within each of the functional trajectory groups derived from the latent class growth analysis. Higher scores in the FAST are indicative of greater functional impairment. Multinomial regression analysis (final model: R 2 Nagelkerke 53%, X 2 = 140.26; df = 24 ; p< 0.001) indicated that parental SES, total baseline scores in PANSS positive subscale, PANSS negative subscale, MADRS, and PAS, as well as verbal learning and memory contributed to differentiate among the four functional trajectories (Table 4). Specifically, subjects falling into the Mild impairment-Improving group were more likely to have a medium-high parental SES (OR: 4.14, 95% CI 1.65–10.42), lower severity of baseline negative symptoms (OR: 0.89, 95% CI 0.83–0.96) and of depressive symptoms (OR: 0.94, 95% CI 0.89–0.99), and better premorbid adjustment (OR: 0.96, 95% CI 0.94–0.98). On the other hand, compared to individuals in the Severe impairment-Stable trajectory, better premorbid adjustment (OR: 0.96, 95% CI 0.93–0.99) and higher scores in the verbal learning and memory domain (OR: 3.09, 95% CI 1.36–7.03) increased the probability of belonging to the Severe impairment-Improving trajectory group. Finally, individuals falling in the Moderate impairment-Stable trajectory were more likely to score lower in the PANSS positive subscale (OR: 0.93, 95% CI 0.87–0.99) at baseline than the Severe impairment-Stable group. 3.4. Exploring Diagnoses Distribution among Functional Trajectories throughout the Follow-Up The diagnoses distribution within each functional trajectory at baseline, one-year follow-up and two-year follow-up is depicted in Figure 2. Diagnosis distribution significantly differed between trajectory groups at baseline (n= 261; X 2 = 19.9; p= 0.02), 1-year follow-up (n= 202; X 2 = 42.6; p< 0.001) and at 2-year follow-up (n= 156; X 2 = 28.5; p= 0.001 ). A higher proportion of patients with a diagnosis of Schizophrenia was found among individuals falling into the Severe impairment-Stable and Moderate impairment-Stable trajectories compared to the Mild impairment-Improving trajectory. On the other hand, the diagnoses of Bipolar disorder and Other psychosis were more frequent among individuals falling into the Mild impairment-Improving trajectory compared to the Severe impairment-Stable trajectory. Abbreviations: Mi-I: Mild impairment-Improving; Mo-S: Moderate impairment-Stable; Se-I: Severe impairment-Improving; Se-S: Severe impairment-Stable. Figure 2represents, for each of the functional trajectories derived from the Latent class growth analysis, the number of individuals with a diagnosis of Bipolar disorder, Schizophrenia, Schizoaffective disorder, or Other psychoses at baseline, 12-month and 24-month follow-up. Other psychoses include psychotic disorder not-otherwise specified, brief psychotic disorder, schizophreniform psychosis, delusional disorder, and substanceinduced psychosis. (*) symbol indicates which diagnostic categories within the Severe impairment-Stable trajectory and the Moderate impairment-Stable trajectory show a significantly different proportion of individuals compared to the Mild impairment-Improving trajectory group. Abbreviations: Mi-I: Mild impairment-Improving; Mo-S: Moderate impairment-Stable; Se-I: Severe impairment-Improving; Se-S: Severe impairment-Stable J. Clin. Med. 2021,10, 73 8 of 20 Table 3. Comparison between groups derived from the identified functional trajectories. Mi-I (1) n= 100 Mo-S (2) n= 48 Se-I (3) n= 32 Se-S (4) n= 81 Kruskal– Wallis/X2p-Value Post-Hoc a 1 vs. 2 1 vs. 3 1 vs. 4 2 vs. 3 2 vs. 4 3 vs. 4 Sociodemographic characteristics Age (years) b25.8 (10) 24.9 (8) 24.4 (8) 24.8 (8) 1.44 0.70 Sex (Female) c29 (29.0) 18 (37.5) 10 (31.2) 30 (37.0) 1.78 0.62 Ethnicity (Caucasian) c91 (91.0) 41 (85.4) 27 (84.4) 69 (85.2) 1.97 0.58 Marital status (Single) c83 (83.0) 42 (87.5) 29 (90.6) 68 (83.9) 1.42 0.70 Living situation (Living independent) c30 (30.0) 4 (8.3) 7 (21.9) 14 (17.3) 10.19 0.02 <0.05 Educational level (Higher education) c55 (55.0) 20 (41.7) 16 (50.0) 24 (29.6) 12.71 <0.01 <0.05 Occupational status (Active d)c63 (63.0) 26 (54.2) 18 (56.2) 29 (35.8) 13.68 <0.01 <0.05 Socioeconomic status (Medium-high) c72 (72.0) 22 (45.8) 18 (56.2) 29 (35.8) 24.06 <0.001 <0.05 <0.05 Family history of psychiatric disorders (Yes) c53 (53.0) 33 (68.7) 17 (53.1) 43 (53.1) 3.92 0.27 Previous psychiatric diagnoses (Yes) c24 (24.0) 13 (27.1) 8 (25.0) 29 (35.8) 4.65 0.59 Substance use c Tobacco 70 (70.0) 33 (68.7) 24 (75.0) 55 (67.9) 0.57 0.90 Alcohol 60 (60.0) 25 (52.1) 24 (75.0) 32 (39.5) 14.05 <0.01 <0.05 <0.05 Cannabis 44 (44.0) 22 (45.8) 13 (40.6) 36 (44.4) 0.22 0.97 Cocaine 9 (9.0) 8 (16.7) 6 (18.7) 12 (14.8) 3.04 0.39 Clinical measures DUP (days) b85.0 (165) 133.0 (263) 110.0 (168) 162.0 (216) 14.36 <0.01 <0.05 PANSS b PANSS positive 14.0 (14) 16.0 (10) 23.5 (10) 21.0 (9) 32.14 <0.001 <0.05 <0.05 <0.05 <0.05 PANSS negative 14.0 (11) 19.0 (13) 19.5 (11) 23.0 (9) 57.27 <0.001 <0.05 <0.05 <0.05 <0.05 PANSS general 29.5 (19) 34.5 (17) 45.5 (19) 43.0 (14) 58.89 <0.001 <0.05 <0.05 <0.05 <0.05 PANSS total 57.5 (39) 72.0 (28) 86.5 (29) 88.0 (24) 65.12 <0.001 <0.05 <0.05 <0.05 <0.05 Young total b2.0 (14) 2.0 (14) 13.5 (19) 7.0 (18) 15.26 <0.01 <0.05 <0.05 MADRS total b6.0 (10) 13.0 (12) 16.0 (17) 16.0 (12) 39.05 <0.001 <0.05 <0.05 <0.05 PAS total b30.0 (24) 43.0 (27) 36.0 (31) 57.0 (33) 51.58 <0.001 <0.05 <0.05 <0.05 TQ b1.0 (2) 0.0 (1) 0.0 (2) 0.0 (2) 2.69 0.44 FES b J. Clin. Med. 2021,10, 73 9 of 20 Table 3. Cont. Mi-I (1) n= 100 Mo-S (2) n= 48 Se-I (3) n= 32 Se-S (4) n= 81 Kruskal– Wallis/X2p-Value Post-Hoc a 1 vs. 2 1 vs. 3 1 vs. 4 2 vs. 3 2 vs. 4 3 vs. 4 Cohesion 52.0 (13) 52.0 (13) 52.0 (15) 47.0 (17) 5.30 0.15 Expressiveness 53.0 (16) 50.0 (16) 53.0 (14) 47.0 (16) 4.14 0.25 Conflict 49.0 (9) 49.0 (9) 49.0 (9) 49.0 (17) 7.19 0.07 Independence 51.0 (11) 51.0 (11) 51.0 (14) 51.0 (17) 3.50 0.32 Achievement-orientation 47.0 (10) 47.0 (10) 47.0 (15) 47.0 (16) 1.35 0.72 Intellectual-cultural orientation 51.0 (23) 47.0 (14) 47.0 (14) 42.0 (19) 5.78 0.12 Active-recreational orientation 53.0 (14) 48.0 (19) 48.0 (21) 44.0 (4) 19.69 <0.001 <0.05 Moral-religious emphasis 44.0 (10) 49.0 (15) 44.0 (10) 44.0 (15) 3.34 0.34 Organization 54.0 (10) 51.5 (19) 49.0 (20) 49.0 (15) 4.51 0.21 Control 45.0 (14) 49.0 (14) 49.0 (14) 49.0 (14) 4.06 0.25 Cognitive measures n= 93 n= 44 n= 28 n= 76 IQ b100 (20) 92.5 (24) 93.5 (19) 90.0 (20) 10.18 0.02 <0.05 n= 91 n= 43 n= 25 n= 67 Verbal Fluency b0.20 (1.2) −0.08 (1.4) 0.25 (1.1) −0.50 (0.9) 20.69 <0.001 <0.05 n= 85 n= 36 n= 23 n= 53 Attention b0.18 (0.4) 0.03 (0.7) 0.10 (0.6) −0.11 (0.5) 13.54 <0.01 <0.05 n= 94 n= 43 n= 28 n= 73 Working memory b0.14 (1.1) −0.01 (1.3) 0.15 (0.9) −0.17 (1.2) 14.10 <0.01 <0.05 <0.05 n= 92 n= 40 n= 25 n= 67 Verbal Learning and Memory b0.38 (1.2) 0.20 (1.3) 0.25 (0.9) −0.43 (1.4) 24.18 <0.001 <0.05 <0.05 n= 93 n= 41 n= 30 n= 70 Processing Speed b0.32 (0.8) 0.14 (0.9) −0.11 (1.4) −0.16 (1.0) 14.78 <0.01 <0.05 n= 92 n= 40 n= 28 n= 58 Executive function b0.29 (0.7) −0.03 (0.8) 0.14 (0.8) 0.00 (1.0) 13.26 <0.01 <0.05 n= 89 n= 43 n= 26 n= 66 Social cognition b−0.33 (1.3) −0.01 (1.4) 0.04 (1.3) −0.07 (1.4) 3.47 0.32 a Tukey or Z statistic, as appropriate. Significance values have been adjusted using the Bonferroni correction for multiple tests. Bold type indicates p< 0.05. b Values are indicated as median (Interquartile Range). c Values are indicated as n(%). d Active includes workers and students. Abbreviations: Mi-I: Mild impairment-Improving; Mo-S: Moderate impairment-Stable; Se-I: Severe impairment-Improving; Se-S: Severe impairment-Stable; DUP: Duration of Untreated Psychosis; PANSS: Positive and Negative Syndrome Scale; MADRS: Montgomery–Åsberg Depression Scale; YMRS: Young Rating Mania Scale; PAS: Premorbid Adjustment Scale; TQ: trauma questionnaire; FES: Family Environment Scale; IQ: Intelligence Quotient. J. Clin. Med. 2021,10, 73 16 of 20 Appendix A Post-hoc mediation analyses Given that previous works on FEP samples have suggested that premorbid adjustment may influence psychosocial functioning through verbal memory and negative symptoms [ 59 ], we decided to test how the identified predictors interact to impact functioning in our sample. For that, we examined mediation using a regression-based bootstrapping approach [ 60 ]. Analyses were performed with PROCESS [ 61 ]. Before beginning the analyses, two dummy variables for trajectory membership were created, one including only Mild impairment-Improving and Severe impairment-Stable trajectories and another including only Severe impairment-Improving and Severe impairment-Stable trajectories. First: we used PROCESS model 4 to test a simple mediation model with trajectory membership (Severe impairment-Improving vs. Severe impairment-Stable, with Severe impairment-Stable trajectory as the reference category) as the outcome variable (Y), baseline PAS score as the predictor variable (X) and baseline verbal learning and memory as the mediator variable (M) (Figure A1). Age and sex were included as covariates. The data are consistent with the claim that better premorbid adjustment positively impacts verbal learning and memory, which in turn increases the probability to belong to the severe and improving functional impairment trajectory (indirect effect = − 0.011; 95% CI = −0.030 to −0.001 ). The mediation partially explains the effect of premorbid adjustment on trajectory membership; in addition, premorbid adjustment influences class membership independently from the proposed mechanism (b=−0.05 , p= 0.002). Hence, we infer complementary partial mediation [62]. Second: we used a series mediation model to assess mediation between predictors of Mild impairment-Improving vs. Severe impairment-Stable trajectory. In this model, trajectory membership (Mild impairment-Improving vs. Severe impairment-Stable, with Severe impairment-Stable trajectory as the reference category) was the outcome variable (Y) and parental SES the predictor variable (X). Baseline PAS score (M1) and baseline PANSS negative subscale scores (M2) were included, in this order, as mediator variables. Total MADRS score was not considered as a mediator as no association with parental SES was found in a preliminary analysis. We could establish a serial mediation from parental SES through premorbid adjustment and through baseline negative symptoms to trajectory membership (indirect effect = − 0.249; 95% CI: − 0.551 to − 0.086). In addition, parental SES had an indirect effect on class membership only through premorbid adjustment (indirect effect = −0.525, 95% CI: −1.097 to −0.171) and only through baseline negative symptoms ( indirect effect = −0.504 , 95% CI: − 1.076 to − 0.135). Finally, there was a direct effect of parental SES on trajectory membership (b= − 0.932, p= 0.029), indicating complementary partial mediation. J. Clin. Med. 2021, 10, x FOR PEER REVIEW 19 of 23 independently from the proposed mechanism (b = −0.05, p = 0.002). Hence, we infer complementary partial mediation [62]. Second: we used a series mediation model to assess mediation between predictors of Mild impairment-Improving vs. Severe impairment-Stable trajectory. In this model, trajectory membership (Mild impairment-Improving vs. Severe impairment-Stable, with Severe impairment-Stable trajectory as the reference category) was the outcome variable (Y) and parental SES the predictor variable (X). Baseline PAS score (M1) and baseline PANSS negative subscale scores (M2) were included, in this order, as mediator variables. Total MADRS score was not considered as a mediator as no association with parental SES was found in a preliminary analysis. We could establish a serial mediation from parental SES through premorbid adjustment and through baseline negative symptoms to trajectory membership (indirect effect = −0.249; 95% CI: −0.551 to −0.086). In addition, parental SES had an indirect effect on class membership only through premorbid adjustment (indirect effect = −0.525, 95% CI: −1.097 to −0.171) and only through baseline negative symptoms (indirect effect = −0.504, 95% CI: −1.076 to −0.135). Finally, there was a direct effect of parental SES on trajectory membership (b = −0.932, p = 0.029), indicating complementary partial mediation. Figure A1. Mediation analyses. #: S-S was used as the reference category. Abbreviations: S-I: Severe impairment-Improving; S-S: Severe impairment-Stable; M-I: Mild impairment-Improving. PAS: Premorbid Adjustment Scale; SES: Socioeconomic Status, PANSS-N: Positive and Negative Syndrome Scale-Negative Subscale. References 1. Rosa, A.; Sanchez-Moreno, J.; Martinez-Aran, A.; Salamero, M.; Torrent, C.; Reinares, M.; Comes, M.; Colom, F.; Van Riel, W.; Ayuso-Mateos, J.; et al. Validity and reliability of the Functioning Assessment Short Test (FAST) in bipolar disorder. Clin. Pr. Epidemiol. Ment. Health CP EMH 2007, 3, 5, doi:10.1186/1745-0179-3-5. 2. Kahn, R.S.; Sommer, I.E.; Murray, R.M.; Meyer-Lindenberg, A.; Weinberger, D.R.; Cannon, T.D.; O'Donovan, M.; Correll, C.U.; Kane, J.M.; van Os, J.; et al. Schizophrenia. Nat. Rev. Dis. Primers 2015, 1, 15067, doi:10.1038/nrdp.2015.67. 3. Weissman, M.M.; Sholomskas, D.; John, K. The assessment of social adjustment. An update. Arch. Gen. Psychiatry 1981, 38, 1250– 1258, doi:10.1001/archpsyc.1981.01780360066006. 4. Keck, P.E., Jr. Defining and improving response to treatment in patients with bipolar disorder. J. Clin. Psychiatry 2004, 65 (Suppl S15), 25–29. 5. Michalak, E.E.; Murray, G. A clinician’s guide to psychosocial functioning and quality of life in bipolar disorder. In Practical Management of Bipolar Disorder; Young, A.H., Michalak, E.E., Ferrier, I.N., Eds.; Cambridge University Press: Cambridge, UK, 2010, doi:10.1017/CBO9780511776922.016pp. 163-174. 6. Álvarez-Jiménez, M.; Gleeson, J.F.; Henry, L.P.; Harrigan, S.M.; Harris, M.G.; Killackey, E.; Bendall, S.; Amminger, G.P.; Yung, A.R.; Herrman, H.; et al. Road to full recovery: Longitudinal relationship between symptomatic remission and psychosocial recovery in first-episode psychosis over 7.5 years. Psychol. Med. 2012, 42, 595–606, doi:10.1017/s0033291711001504. 7. Birchwood, M.; Todd, P.; Jackson, C. Early intervention in psychosis. The critical period hypothesis. Br. J. Psychiatry. Suppl. 1998, 172, 53–59. Figure A1. Mediation analyses. #: S-S was used as the reference category. Abbreviations: S-I: Severe impairment-Improving; S-S: Severe impairment-Stable; M-I: Mild impairment-Improving. PAS: Premorbid Adjustment Scale; SES: Socioeconomic Status, PANSS-N: Positive and Negative Syndrome Scale-Negative Subscale. J. Clin. Med. 2021,10, 73 17 of 20 References 1. Rosa, A.; Sanchez-Moreno, J.; Martinez-Aran, A.; Salamero, M.; Torrent, C.; Reinares, M.; Comes, M.; Colom, F.; Van Riel, W.; Ayuso-Mateos, J.; et al. Validity and reliability of the Functioning Assessment Short Test (FAST) in bipolar disorder. Clin. Pr. Epidemiol. Ment. Health CP EMH 2007,3, 5. [CrossRef] [PubMed] 2. Kahn, R.S.; Sommer, I.E.; Murray, R.M.; Meyer-Lindenberg, A.; Weinberger, D.R.; Cannon, T.D.; O’Donovan, M.; Correll, C.U.; Kane, J.M.; van Os, J.; et al. Schizophrenia. Nat. Rev. Dis. Primers 2015,1, 15067. [CrossRef] [PubMed] 3. Weissman, M.M.; Sholomskas, D.; John, K. The assessment of social adjustment. An update. Arch. Gen. Psychiatry 1981 ,38, 1250–1258. [CrossRef] [PubMed] 4. Keck, P.E., Jr. Defining and improving response to treatment in patients with bipolar disorder. J. Clin. Psychiatry 2004 , 65 (Suppl. S15) , 25–29. [PubMed] 5. Michalak, E.E.; Murray, G. A clinician’s guide to psychosocial functioning and quality of life in bipolar disorder. In Practical Management of Bipolar Disorder; Young, A.H., Michalak, E.E., Ferrier, I.N., Eds.; Cambridge University Press: Cambridge, UK, 2010; pp. 163–174. [CrossRef] 6. Álvarez-Jiménez, M.; Gleeson, J.F.; Henry, L.P.; Harrigan, S.M.; Harris, M.G.; Killackey, E.; Bendall, S.; Amminger, G.P.; Yung, A.R.; Herrman, H.; et al. Road to full recovery: Longitudinal relationship between symptomatic remission and psychosocial recovery in first-episode psychosis over 7.5 years. Psychol. Med. 2012,42, 595–606. [CrossRef] [PubMed] 7. Birchwood, M.; Todd, P.; Jackson, C. Early intervention in psychosis. The critical period hypothesis. Br. J. Psychiatry. Suppl. 1998 , 172, 53–59. [CrossRef] [PubMed] 8. Santesteban-Echarri, O.; Paino, M.; Rice, S.; González-Blanch, C.; McGorry, P.; Gleeson, J.; Alvarez-Jimenez, M. Predictors of functional recovery in first-episode psychosis: A systematic review and meta-analysis of longitudinal studies. Clin. Psychol. Rev. 2017,58, 59–75. [CrossRef] 9. Vieta, E.; Salagre, E.; Grande, I.; Carvalho, A.F.; Fernandes, B.S.; Berk, M.; Birmaher, B.; Tohen, M.; Suppes, T. Early Intervention in Bipolar Disorder. Am. J. Psychiatry 2018,175, 411–426. [CrossRef] 10. Silva Ribeiro, J.; Pereira, D.; Salagre, E.; Coroa, M.; Santos Oliveira, P.; Santos, V.; Madeira, N.; Grande, I.; Vieta, E. Risk Calculators in Bipolar Disorder: A Systematic Review. Brain Sci. 2020,10, 525. [CrossRef] 11. Bernardo, M.; Cabrera, B.; Arango, C.; Bioque, M.; Castro-Fornieles, J.; Cuesta, M.J.; Lafuente, A.; Parellada, M.; Saiz-Ruiz, J.; Vieta, E. One decade of the first episodes project (PEPs): Advancing towards a precision psychiatry. Rev. De Psiquiatr. Y Salud Ment. 2019,12, 135–140. [CrossRef] 12. Miettunen, J.; Nordström, T.; Kaakinen, M.; Ahmed, A.O. Latent variable mixture modeling in psychiatric research—A review and application. Psychol. Med. 2016,46, 457–467. [CrossRef] [PubMed] 13. Van der Nest, G.; Lima Passos, V.; Candel, M.J.J.M.; Van Breukelen, G.J.P. An overview of mixture modelling for latent evolutions in longitudinal data: Modelling approaches, fit statistics and software. Adv. Life Course Res. 2020,43. [CrossRef] 14. Hodgekins, J.; Birchwood, M.; Christopher, R.; Marshall, M.; Coker, S.; Everard, L.; Lester, H.; Jones, P.; Amos, T.; Singh, S.; et al. Investigating trajectories of social recovery in individuals with first-episode psychosis: A latent class growth analysis. Br. J. Psychiatry J. Ment. Sci. 2015,207, 536–543. [CrossRef] [PubMed] 15. Hall, M.H.; Holton, K.M.; Öngür, D.; Montrose, D.; Keshavan, M.S. Longitudinal trajectory of early functional recovery in patients with first episode psychosis. Schizophr. Res. 2019,209, 234–244. [CrossRef] [PubMed] 16. Chang, W.C.; Chu, A.O.K.; Kwong, V.W.Y.; Wong, C.S.M.; Hui, C.L.M.; Chan, S.K.W.; Lee, E.H.M.; Chen, E.Y.H. Patterns and predictors of trajectories for social and occupational functioning in patients presenting with first-episode non-affective psychosis: A three-year follow-up study. Schizophr. Res. 2018,197, 131–137. [CrossRef] [PubMed] 17. Suvisaari, J.; Mantere, O.; Keinänen, J.; Mäntylä, T.; Rikandi, E.; Lindgren, M.; Kieseppä, T.; Raij, T.T. Is It Possible to Predict the Future in First-Episode Psychosis? Front. Psychiatry 2018,9, 580. [CrossRef] 18. Bernardo, M.; Bioque, M.; Parellada, M.; Saiz Ruiz, J.; Cuesta, M.J.; Llerena, A.; Sanjuan, J.; Castro-Fornieles, J.; Arango, C.; Cabrera, B. Assessing clinical and functional outcomes in a gene-environment interaction study in first episode of psychosis (PEPs). Rev. De Psiquiatr. Y Salud Ment. 2013,6, 4–16. [CrossRef] 19. Salagre, E.; Arango, C.; Artigas, F.; Ayuso-Mateos, J.L.; Bernardo, M.; Castro-Fornieles, J.; Bobes, J.; Desco, M.; Fananas, L.; Gonzalez-Pinto, A.; et al. CIBERSAM: Ten years of collaborative translational research in mental disorders. Rev. De Psiquiatr. Y Salud Ment. 2019,12, 1–8. [CrossRef] 20. Salagre, E.; Grande, I.; Vieta, E.; Mezquida, G.; Cuesta, M.J.; Moreno, C.; Bioque, M.; Lobo, A.; González-Pinto, A.; Moreno, D.M.; et al . Predictors of Bipolar Disorder Versus Schizophrenia Diagnosis in a Multicenter First Psychotic Episode Cohort: Baseline Characterization and a 12-Month Follow-Up Analysis. J. Clin. Psychiatry 2020,81, 19m12996. [CrossRef] 21. APA. DSM-IV: Diagnostic and Statistical Manual of Mental Disorders, 4th ed.; American Psychiatric Association: Washington, DC, USA, 1994. 22. Hollingshead, A.B.; Redlich, F.C. Social class and mental illness: A community study. 1958. Am. J. Public Health 2007 ,97, 1756–1757. 23. Kokkevi, A.; Hartgers, C. EuropASI: European Adaptation of a Multidimensional Assessment Instrument for Drug and Alcohol Dependence. Eur. Addict. Res. 1995,1, 208–210. [CrossRef] 24. Moos, R.H.; Moos, B.S. Family Environment Scale Manual; Consulting Psychologist Press: Palo Alto, CA, USA, 1981. 25. Fernández-Ballesteros, R.; Sierra, B. Escalas de Clima Social FES, WES, CIES y CES.; TEA: Madrid, Spain, 1989. J. Clin. Med. 2021,10, 73 18 of 20 26. First, M.S.R.; Gibbon, M.; Williams, J. Structured Clinical Interview for DSM-IV Axis I Disorders; Administration booklet; American Psychiatric Press Inc.: Washington, DC, USA, 1994. 27. Kay, S.R.; Fiszbein, A.; Opler, L.A. The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophr. Bull. 1987 ,13, 261–276. [CrossRef] [PubMed] 28. Peralta, V.; Cuesta, M.J. Psychometric properties of the positive and negative syndrome scale (PANSS) in schizophrenia. Psychiatry Res. 1994,53, 31–40. [CrossRef] 29. Colom, F.; Vieta, E.; Martinez-Aran, A.; Garcia-Garcia, M.; Reinares, M.; Torrent, C.; Goikolea, J.M.; Banus, S.; Salamero, M. Spanish version of a scale for the assessment of mania: Validity and reliability of the Young Mania Rating Scale. Med. Clin. 2002 , 119, 366–371. [CrossRef] 30. Young, R.C.; Biggs, J.T.; Ziegler, V.E.; Meyer, D.A. A rating scale for mania: Reliability, validity and sensitivity. Br. J. Psychiatry J. Ment. Sci. 1978,133, 429–435. [CrossRef] [PubMed] 31. Lobo, A.; Chamorro, L.; Luque, A.; Dal-Re, R.; Badia, X.; Baro, E. Validation of the Spanish versions of the Montgomery-Asberg depression and Hamilton anxiety rating scales. Med. Clin. 2002,118, 493–499. [CrossRef] 32. Montgomery, S.A.; Asberg, M. A new depression scale designed to be sensitive to change. Br. J. Psychiatry J. Ment. Sci. 1979 ,134, 382–389. [CrossRef] 33. Cannon, M.; Jones, P.; Gilvarry, C.; Rifkin, L.; McKenzie, K.; Foerster, A.; Murray, R.M. Premorbid social functioning in schizophrenia and bipolar disorder: Similarities and differences. Am. J. Psychiatry 1997,154, 1544–1550. [CrossRef] 34. Rosa, A.; Reinares, M.; Amann, B.; Popovic, D.; Franco, C.; Comes, M.; Torrent, C.; Bonnin, C.; Sole, B.; Valenti, M.; et al. Six-month functional outcome of a bipolar disorder cohort in the context of a specialized-care program. Bipolar. Disord. 2011 ,13, 679–686. [CrossRef] 35. Bonnín, C.M.; Martínez-Arán, A.; Reinares, M.; Valentí, M.; Solé, B.; Jiménez, E.; Montejo, L.; Vieta, E.; Rosa, A.R. Thresholds for severity, remission and recovery using the functioning assessment short test (FAST) in bipolar disorder. J. Affect. Disord. 2018 ,240, 57–62. [CrossRef] 36. González-Ortega, I.; Rosa, A.; Alberich, S.; Barbeito, S.; Vega, P.; Echeburúa, E.; Vieta, E.; González-Pinto, A. Validation and use of the functioning assessment short test in first psychotic episodes. J. Nerv. Ment. Dis. 2010,198, 836–840. [CrossRef] [PubMed] 37. Bobes, J.; Calcedo-Barba, A.; Garcia, M.; Francois, M.; Rico-Villademoros, F.; Gonzalez, M.P.; Bascaran, M.T.; Bousono, M. Evaluation of the psychometric properties of the Spanish version of 5 questionnaires for the evaluation of post-traumatic stress syndrome. Actas Esp. De Psiquiatr. 2000,28, 207–218. 38. Davidson, J.; Smith, R. Traumatic experiences in psychiatric outpatients. J. Trauma. Stress 1990,3, 459–475. [CrossRef] 39. Perkins, D.O.; Leserman, J.; Jarskog, L.F.; Graham, K.; Kazmer, J.; Lieberman, J.A. Characterizing and dating the onset of symptoms in psychotic illness: The Symptom Onset in Schizophrenia (SOS) inventory. Schizophr. Res. 2000 ,44, 1–10. [CrossRef] 40. Nuechterlein, K.H.; Green, M.; Kern, R.S.; Kern, R.; Baade, L.E.; Baade, L.; Barch, D.M.; Barch, D.; Cohen, J.D.; Cohen, J.; et al. The MATRICS Consensus Cognitive Battery, part 1: Test selection, reliability, and validity. Am. J. Psychiatry 2008 ,165, 203–213. [CrossRef] [PubMed] 41. Wechsler, D. Wechsler Adult Intelligence Scale—III (WAIS-III); Psychological Corporation: San Antonio, TX, USA, 1997. 42. Golden, C.J. Stroop Color and Word Test: A Manual for Clinical and Experimental Uses; Stoelting Co.: Chicago, IL, USA, 1978. 43. Heaton, R.K. Wisconsin Card Sorting Test Manual; Psychological Assessment Resources: Odessa, FL, USA, 1981. 44. Reitan, R.M. Validity of the Trail Making Test as an indicator of organic brain damage. Percept. Mot. Ski. 1958 ,8, 271–276. [CrossRef] 45. Conners, C.K. Conners’ Continuous Performance Test; Multi-Health System: Toronto, ON, Canada, 2002. 46. Reitan, R.M.; Wolfson, D.W. The Halstead–ReitanNeuropsychological Test Battery: Theory and Clinical Interpretation; Neuropsychology Press: Tucson, AZ, USA, 1993. 47. Benton, A.L.; Hamsher, K. Multilingual Aphasia Examination Manual; University of Iowa: Iowa City, IA, USA, 1976. 48. Benedet, M. Test de Aprendizaje Verbal España-Complutense (TAVEC); Tea Ediciones: Madrid, Spain, 1998. 49. Brackett, M.A.; Salovey, P. Measuring emotional intelligence with the Mayer-Salovery-Caruso Emotional Intelligence Test (MSCEIT). Psicothema 2006,18, 34–41. [PubMed] 50. Extremera, N.; Fernandez-Berrocal, P.; Salovey, P. Spanish version of the Mayer-Salovey-Caruso Emotional Intelligence Test (MSCEIT). Version 2.0: Reliabilities, age and gender differences. Psicothema 2006,18, 42–48. 51. Cuesta, M.J.; Sanchez-Torres, A.M.; Cabrera, B.; Bioque, M.; Merchan-Naranjo, J.; Corripio, I.; Gonzalez-Pinto, A.; Lobo, A.; Bombin, I.; de la Serna, E.; et al. Premorbid adjustment and clinical correlates of cognitive impairment in first-episode psychosis. The PEPsCog Study. Schizophr. Res. 2015,164, 65–73. [CrossRef] 52. Van de Schoot, R.; Sijbrandij, M.; Winter, S.D.; Depaoli, S.; Vermunt, J.K. The GRoLTS-Checklist: Guidelines for Reporting on Latent Trajectory Studies. Struct. Equ. Modeling A Multidiscip. J. 2017,24, 451–467. [CrossRef] 53. Celeux, G.; Soromenho, G. An entropy criterion for assessing the number of clusters in a mixture model. J. Classif. 1996 ,13, 195–212. [CrossRef] 54. Proust-Lima, C.; Philipps, V.; Liquet, B. Estimation of Extended Mixed Models Using Latent Classes and Latent Processes: The R Package lcmm. J. Stat. Softw. 2017,78, 1–56. [CrossRef] J. Clin. Med. 2021,10, 73 19 of 20 55. Torrent, C.; Reinares, M.; Martinez-Arán, A.; Cabrera, B.; Amoretti, S.; Corripio, I.; Contreras, F.; Sarró, S.; González-Pinto, A.; Lobo, A.; et al. Affective versus non-affective first episode psychoses: A longitudinal study. J. Affect. Disord. 2018 ,238, 297–304. [CrossRef] [PubMed] 56. Amoretti, S.; Cabrera, B.; Torrent, C.; Mezquida, G.; Lobo, A.; Gonzalez-Pinto, A.; Parellada, M.; Corripio, I.; Vieta, E.; de la Serna, E.; et al. Cognitive reserve as an outcome predictor: First-episode affective versus non-affective psychosis. Acta Psychiatr. Scand. 2018,138, 441–455. [CrossRef] [PubMed] 57. Andersen, S.B.; Karstoft, K.I.; Bertelsen, M.; Madsen, T. Latent trajectories of trauma symptoms and resilience: The 3-year longitudinal prospective USPER study of Danish veterans deployed in Afghanistan. J. Clin. Psychiatry 2014 ,75, 1001–1008. [CrossRef] 58. Shaunna, L.; Clark, B.M. Relating Latent Class Analysis Results to Variables not Included in the Analysis. 2009. Available online: https://www.statmodel.com/download/relatinglca.pdf (accessed on 6 August 2020). 59. Jordan, G.; Veru, F.; Lepage, M.; Joober, R.; Malla, A.; Iyer, S.N. Pathways to functional outcomes following a first episode of psychosis: The roles of premorbid adjustment, verbal memory and symptom remission. Aust. N. Zealand J. Psychiatry 2018 ,52, 793–803. [CrossRef] 60. Preacher, K.J.; Hayes, A.F. Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behav. Res. Methods 2008,40, 879–891. [CrossRef] 61. Hayes, A.F. Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-Based Approach; Guilford Press: New York, NY, USA, 2013. 62. Zhao, X.; Lynch, J.G., Jr.; Chen, Q. Reconsidering Baron and Kenny: Myths and truths about mediation analysis. J. Consum. Res. 2010,37, 197–206. [CrossRef] 63. Hower, H.; Lee, E.J.; Jones, R.N.; Birmaher, B.; Strober, M.; Goldstein, B.I.; Merranko, J.; Keller, M.B.; Goldstein, T.R.; Weinstock, L.M.; et al. Predictors of longitudinal psychosocial functioning in bipolar youth transitioning to adults. J. Affect. Disord. 2019 ,246, 578–585. [CrossRef] [PubMed] 64. Milev, P.; Ho, B.C.; Arndt, S.; Andreasen, N.C. Predictive values of neurocognition and negative symptoms on functional outcome in schizophrenia: A longitudinal first-episode study with 7-year follow-up. Am. J. Psychiatry 2005,162, 495–506. [CrossRef] 65. Chang, W.C.; Ho, R.W.H.; Tang, J.Y.M.; Wong, C.S.M.; Hui, C.L.M.; Chan, S.K.W.; Lee, E.M.H.; Suen, Y.N.; Chen, E.Y.H. Early-Stage Negative Symptom Trajectories and Relationships With 13-Year Outcomes in First-Episode Nonaffective Psychosis. Schizophr. Bull. 2019,45, 610–619. [CrossRef] 66. Stouten, L.H.; Veling, W.; Laan, W.; van der Helm, M.; van der Gaag, M. Psychotic symptoms, cognition and affect as predictors of psychosocial problems and functional change in first-episode psychosis. Schizophr. Res. 2014,158, 113–119. [CrossRef] 67. Bonnín, C.M.; Jiménez, E.; Solé, B.; Torrent, C.; Radua, J.; Reinares, M.; Grande, I.; Ruíz, V.; Sánchez-Moreno, J.; Martínez-Arán, A.; et al. Lifetime Psychotic Symptoms, Subthreshold Depression and Cognitive Impairment as Barriers to Functional Recovery in Patients with Bipolar Disorder. J. Clin. Med. 2019,8, 1046. [CrossRef] [PubMed] 68. Addington, J.; Addington, D. Patterns of premorbid functioning in first episode psychosis: Relationship to 2-year outcome. Acta Psychiatr. Scand. 2005,112, 40–46. [CrossRef] [PubMed] 69. Velthorst, E.; Fett, A.J.; Reichenberg, A.; Perlman, G.; van Os, J.; Bromet, E.J.; Kotov, R. The 20-Year Longitudinal Trajectories of Social Functioning in Individuals With Psychotic Disorders. Am. J. Psychiatry 2017,174, 1075–1085. [CrossRef] [PubMed] 70. Wells, R.; Jacomb, I.; Swaminathan, V.; Sundram, S.; Weinberg, D.; Bruggemann, J.; Cropley, V.; Lenroot, R.K.; Pereira, A.M.; Zalesky, A.; et al. The Impact of Childhood Adversity on Cognitive Development in Schizophrenia. Schizophr. Bull. 2020 ,46, 140–153. [CrossRef] [PubMed] 71. Grande, I.; Sanchez-Moreno, J.; Sole, B.; Jimenez, E.; Torrent, C.; Bonnin, C.M.; Varo, C.; Tabares-Seisdedos, R.; Balanzá-Martínez, V.; Valls, E.; et al. High cognitive reserve in bipolar disorders as a moderator of neurocognitive impairment. J. Affect. Disord. 2017 , 208, 621–627. [CrossRef] 72. Amoretti, S.; Rosa, A.R.; Mezquida, G.; Cabrera, B.; Ribeiro, M.; Molina, M.; Bioque, M.; Lobo, A.; González-Pinto, A.; Fraguas, D.; et al. The impact of cognitive reserve, cognition and clinical symptoms on psychosocial functioning in first-episode psychoses. Psychol. Med. 2020, 1–12. [CrossRef] 73. Omer, S.; Finnegan, M.; Pringle, D.G.; Kinsella, A.; Fearon, P.; Russell, V.; O’Callaghan, E.; Waddington, J.L. Socioeconomic status at birth and risk for first episode psychosis in rural Ireland: Eliminating the features of urbanicity in the Cavan-Monaghan First Episode Psychosis Study (CAMFEPS). Schizophr. Res. 2016,173, 84–89. [CrossRef] 74. Xiang, L.; Su, Z.; Liu, Y.; Huang, Y.; Zhang, X.; Li, S.; Zhang, H. Impact of Family Socioeconomic Status on Health-Related Quality of Life in Children With Critical Congenital Heart Disease. J. Am. Heart Assoc. 2019,8, e010616. [CrossRef] 75. Arango, C.; Díaz-Caneja, C.M.; McGorry, P.D.; Rapoport, J.; Sommer, I.E.; Vorstman, J.A.; McDaid, D.; Marín, O.; SerranoDrozdowskyj, E.; Freedman, R.; et al. Preventive strategies for mental health. Lancet. Psychiatry 2018,5, 591–604. [CrossRef] 76. Fusar-Poli, P.; McGorry, P.D.; Kane, J.M. Improving outcomes of first-episode psychosis: An overview. World Psychiatry Off. J. World Psychiatr. Assoc. (WPA) 2017,16, 251–265. [CrossRef] 77. Fu, S.; Czajkowski, N.; Rund, B.R.; Torgalsbøen, A.K. The relationship between level of cognitive impairments and functional outcome trajectories in first-episode schizophrenia. Schizophr. Res. 2017,190, 144–149. [CrossRef] J. Clin. Med. 2021,10, 73 20 of 20 78. Tabarés-Seisdedos, R.; Balanzá-Martínez, V.; Sánchez-Moreno, J.; Martinez-Aran, A.; Salazar-Fraile, J.; Selva-Vera, G.; Rubio, C.; Mata, I.; Gómez-Beneyto, M.; Vieta, E. Neurocognitive and clinical predictors of functional outcome in patients with schizophrenia and bipolar I disorder at one-year follow-up. J. Affect. Disord. 2008,109, 286–299. [CrossRef] 79. Sanchez-Moreno, J.; Bonnin, C.M.; González-Pinto, A.; Amann, B.L.; Solé, B.; Balanzá-Martinez, V.; Arango, C.; Jiménez, E.; Tabarés-Seisdedos, R.; Garcia-Portilla, M.P.; et al. Factors associated with poor functional outcome in bipolar disorder: Sociodemographic, clinical, and neurocognitive variables. Acta Psychiatr. Scand. 2018,138, 145–154. [CrossRef] 80. Solé, B.; Bonnín, C.M.; Radua, J.; Montejo, L.; Hogg, B.; Jimenez, E.; Reinares, M.; Valls, E.; Varo, C.; Pacchiarotti, I.; et al. Long-term outcome predictors after functional remediation in patients with bipolar disorder. Psychol. Med. 2020 , 1–9. [CrossRef] 81. Albert, N.; Bertelsen, M.; Thorup, A.; Petersen, L.; Jeppesen, P.; Le Quack, P.; Krarup, G.; Jørgensen, P.; Nordentoft, M. Predictors of recovery from psychosis Analyses of clinical and social factors associated with recovery among patients with first-episode psychosis after 5 years. Schizophr. Res. 2011,125, 257–266. [CrossRef] 82. Gee, B.; Hodgekins, J.; Fowler, D.; Marshall, M.; Everard, L.; Lester, H.; Jones, P.B.; Amos, T.; Singh, P.S.; Sharma, V.; et al. The course of negative symptom in first episode psychosis and the relationship with social recovery. Schizophr. Res. 2016 ,174, 165–171. [CrossRef] 83. Bucci, P.; Mucci, A.; van Rossum, I.W.; Aiello, C.; Arango, C.; Baandrup, L.; Buchanan, R.W.; Dazzan, P.; Demjaha, A.; Díaz-Caneja, C.M.; et al. Persistent negative symptoms in recent-onset psychosis: Relationship to treatment response and psychosocial functioning. Eur. Neuropsychopharmacol. 2020,34, 76–86. [CrossRef] 84. Dickinson, D.; Coursey, R.D. Independence and overlap among neurocognitive correlates of community functioning in schizophrenia. Schizophr. Res. 2002,56, 161–170. [CrossRef] 85. Buck, G.; Lavigne, K.M.; Makowski, C.; Joober, R.; Malla, A.; Lepage, M. Sex Differences in Verbal Memory Predict Functioning Through Negative Symptoms in Early Psychosis. Schizophr. Bull. 2020,46, 1587–1595. [CrossRef] [PubMed] 86. Simons, C.J.; Bartels-Velthuis, A.A.; Pijnenborg, G.H. Cognitive Performance and Long-Term Social Functioning in Psychotic Disorder: A Three-Year Follow-Up Study. PLoS ONE 2016,11, e0151299. [CrossRef] [PubMed] 87. Treen Calvo, D.; Giménez-Donoso, S.; Setién-Suero, E.; Toll Privat, A.; Crespo-Facorro, B.; Ayesa Arriola, R. Targeting recovery in first episode psychosis: The importance of neurocognition and premorbid adjustment in a 3-year longitudinal study. Schizophr. Res. 2018,195, 320–326. [CrossRef] 88. Seidman, L.J.; Nordentoft, M. New Targets for Prevention of Schizophrenia: Is It Time for Interventions in the Premorbid Phase? Schizophr. Bull. 2015,41, 795–800. [CrossRef] 89. Salagre, E.; Dodd, S.; Aedo, A.; Rosa, A.; Amoretti, S.; Pinzon, J.; Reinares, M.; Berk, M.; Kapczinski, F.P.; Vieta, E.; et al. Toward Precision Psychiatry in Bipolar Disorder: Staging 2.0. Front. Psychiatry 2018,9, 641. [CrossRef] [PubMed] 90. Bonnin, C.M.; Reinares, M.; Martínez-Arán, A.; Balanzá-Martínez, V.; Sole, B.; Torrent, C.; Tabarés-Seisdedos, R.; García-Portilla, M.P.; Ibáñez, A.; Amann, B.L.; et al. Effects of functional remediation on neurocognitively impaired bipolar patients: Enhancement of verbal memory. Psychol. Med. 2016,46, 291–301. [CrossRef] [PubMed] 91. Bowie, C.R.; McGurk, S.R.; Mausbach, B.; Patterson, T.L.; Harvey, P.D. Combined cognitive remediation and functional skills training for schizophrenia: Effects on cognition, functional competence, and real-world behavior. Am. J. Psychiatry 2012 ,169, 710–718. [CrossRef] 92. Glenthøj, L.B.; Hjorthøj, C.; Kristensen, T.D.; Davidson, C.A.; Nordentoft, M. The effect of cognitive remediation in individuals at ultra-high risk for psychosis: A systematic review. Npj Schizophr. 2017,3, 20. [CrossRef] 93. Lyngstad, S.H.; Gardsjord, E.S.; Simonsen, C.; Engen, M.J.; Romm, K.L.; Melle, I.; Færden, A. Consequences of persistent depression and apathy in first-episode psychosis—A one-year follow-up study. Compr. Psychiatry 2018,86, 60–66. [CrossRef] 94. González-Ortega, I.; Alberich, S.; Echeburúa, E.; Aizpuru, F.; Millán, E.; Vieta, E.; Matute, C.; González-Pinto, A. Subclinical depressive symptoms and continued cannabis use: Predictors of negative outcomes in first episode psychosis. PLoS ONE 2015 , 10, e0123707. [CrossRef] 95. Edwards, C.J.; Garety, P.; Hardy, A. The relationship between depressive symptoms and negative symptoms in people with non-affective psychosis: A meta-analysis. Psychol. Med. 2019,49, 2486–2498. [CrossRef] [PubMed] 96. Clementz, B.A.; Trotti, R.L.; Pearlson, G.D.; Keshavan, M.S.; Gershon, E.S.; Keedy, S.K.; Ivleva, E.I.; McDowell, J.E.; Tamminga, C.A. Testing Psychosis Phenotypes From Bipolar-Schizophrenia Network for Intermediate Phenotypes for Clinical Application: Biotype Characteristics and Targets. Biol. Psychiatry Cogn. Neurosci. Neuroimaging 2020,5, 808–818. [CrossRef] [PubMed] 97. Kirkpatrick, B.; Strauss, G.P.; Nguyen, L.; Fischer, B.A.; Daniel, D.G.; Cienfuegos, A.; Marder, S.R. The brief negative symptom scale: Psychometric properties. Schizophr. Bull. 2011,37, 300–305. [CrossRef] 98. Kring, A.M.; Gur, R.E.; Blanchard, J.J.; Horan, W.P.; Reise, S.P. The Clinical Assessment Interview for Negative Symptoms (CAINS): Final development and validation. Am. J. Psychiatry 2013,170, 165–172. [CrossRef] [PubMed] 99. Amoretti, S.; Cabrera, B.; Torrent, C.; Bonnín, C.D.M.; Mezquida, G.; Garriga, M.; Jiménez, E.; Martínez-Arán, A.; Solé, B.; Reinares, M.; et al. Cognitive Reserve Assessment Scale in Health (CRASH): Its Validity and Reliability. J. Clin. Med. 2019 ,8, 586. [CrossRef] [PubMed]