The FEP care program in Ferrara, Northern Italy, is conceptualized as a specialty within generalist care model, and it draws its mission on the recommendations by the Regione Emilia-Romagna (2016) which were first embedded into clinical practice and then formalized locally in “PDTAs” (Percorso diagnostico terapeutico assistenziale—diagnostic and therapeutic care pathway) in 2019 (Belvederi Murri et al., 2021a; Ferrara et al., 2022). Such program provides clinical high risk (CHR) and FEP patients aged between 15 and 35 years with a specialized coordinated care treatment via pharmacological therapy, individual cognitive-behavioral therapy, physical health promotion and social inclusion programs, involving a psychiatrist and a case manager who provide care based on a stress-vulnerability explanation model for psychosis (Regione Emilia-Romagna, 2024). Referral to the program occurs via acute psychiatric hospital units, community mental health services (CMHC), ProMeco (psychologist in school), social services, and general practitioners (Ferrara et al., 2022).

This study included patients who were consecutively admitted to the FEP program in Ferrara from 2012 to February 17th, 2025, who were aged between 18–35 years, had been treated for psychosis for no longer than 24 months and had an ICD-9 diagnosis of affective (296.14, 296.24, 296.34, 296.44, 296.54, 296.64) or non-affective psychosis (295.0–295.95, 297.0–297.9, 298.0–298.9, 299.9). Patients who presented intellectual disability (IQ <50) or organic psychosis are not eligible for the program.

Although the FEP program in Ferrara includes both patients with FEP and CHR, this study only considered patients who were diagnosed with psychosis, therefore CHR individuals were excluded.

Demographic and clinical information were collected at the time of entry to the first-episode services (FES). Demographic information include age at psychosis onset, age at program admission, marital and parental status, country of birth, education level and employment status.

Recorded clinical information drawn from the Electronic Health Records included the psychosis diagnosis (affective/non-affective), referral to FES, having received treatment by Child and Adolescent Neuropsychiatry, admission to Pathological Addiction Service, family history of psychosis, cannabis use, alcohol use, tobacco use, psychopharmacological treatment, and DUP. DUP was calculated in months from the first evidence of at least one positive psychotic symptom to the beginning of an antipsychotic drug treatment for psychosis (Perkins et al., 2005). The type of psychopharmacological treatment was further classified as either oral administration or long-acting injectable (LAI) formulation. Information about family history of psychosis and cannabis, alcohol and tobacco use were retrieved from patient history, obtained through structured anamnesis conducted by trained clinicians and tox screen when available.

Follow-up assessments were conducted only for individuals who remained in care throughout the study period. Each patient’s overall symptomatology and functioning were evaluated by administering the Italian version of the Health of the Nation Outcome Scale (HoNOS) at baseline and every 6 months for 24 months. HoNOS is a hetero-administered scale (Preti et al., 2012) that includes 12 items, assessing both clinical and psychosocial features. Thus, in this study HoNOS has been used to assess both symptoms severity (the higher the score, the more severe the symptoms and overall clinical presentation) as well as the recovery status (by dichotomizing the total HoNOS score). The HoNOS is a hetero-administered scale consisting of 12 items, each one rated on a Likert-scale from 0 (no problem) to 4 (severe or very severe problem) (Preti et al., 2012). As a result, the higher is the score, the more severe are the symptoms and the overall clinical presentation. Patients were considered in clinical/functional recovery if total HoNOS score was <8 since this score corresponds to a level of symptomatology that allows the discharge of the patient from the CMHC with a referral to the general practitioner’s care (Prowse and Coombs, 2009). This scale has been recommended as the standard assessment by the Regional guidelines for FEP treatment (Regione Emilia-Romagna, 2016). The cut-off has been adopted by several FEP program in the Emilia-Romagna Region to determine the recovery status (Ferrara et al., 2019; Belvederi Murri et al., 2021b; Leuci et al., 2022; Ferrara et al., 2024c). Besides, given that items with no information available are assigned the score of 9, HoNOS evaluation has been considered invalid if at least one 9-scored item was present.

Being NEET (Not in Education, Employment or Training) at 12 and 24 months after FEP program admission was also used to assess patients’ outcomes (Srihari et al., 2015).

Bivariate analyses were performed to assess socio-demographic and clinical characteristics between the two groups (males vs. females) at admission to FES, while HoNOS scores and other clinical characteristics were analysed both at baseline and at 12 and 24 months follow up. The chi-squared test was conducted to analyse categorical data between sexes. The Mann–Whitney test was used to compare continuous non-normally distributed variables, while t-test was used to compare normally distributed data. The normality of continuous variables was detected through the Shapiro–Wilk test. For analysis using parametric statistics, DUP scores were log10 transformed because of positive data skewness.

Symptoms severity (average total HoNOS scores) and the recovery probability between the two groups at several endpoints (baseline, 6 months, 12 months, 18 months, and 24 months) were analyzed, 95% confidence intervals (CI) and effect size were reported. In summary, the two sex groups were then compared, at 12 and 24 months follow up evaluations, with respect to (a) proportion of patients who were considered to be in recovery status (HoNOS <8), (b) NEET status. Outcomes over time were compared between groups using mixed effects models repeated measures analysis of variance (MMRM). MMRM is advantageous because it includes all existing data in the model, without imputation or substitution of missing data. All existing data comprise the model.

Statistical analyses were conducted by using the software R. Statistically significant threshold was identified as p-value ≤0.05.

During the observed time interval, a total of 219 patients were admitted to the FEP program in Ferrara, but only 174 met the study criteria and were eligible for the study. Figure 1 shows the study participants’ selection. Table 1 shows the demographic and clinical characteristics of the total sample at admission.

In the study sample most were of male sex (74.1%), and single (96.6%). Women, compared to men, were more often married (4.4% vs. 0.8%) and had children (4.4% vs. 1.5%). No substantial disparities were found in terms of age at admission, education level or employment status. Of note, fewer women were born in Italy compared to men (66.7% vs. 81.4%, p = 0.04).

Regarding clinical characteristics, mean DUP in the whole sample was 42.5 weeks and it was longer in women compared to men (44.9 vs. 41.6 weeks, p = 0.82), although not statistically significant. Most of the sample (83.9%) had been diagnosed with non-affective psychosis, but no differences were found between the two sexes in terms of psychosis type (affective vs. non-affective). Compared to men who were mostly referred by the inpatient unit (53%), women were mainly referred to the program by the community outpatient psychiatric service (51.1%). Women had previously received individual treatment by the Child and Adolescent Neuropsychiatry service more often than men (22.2% of women vs. 10.8% or men).

At admission to the program, considerably more men than women used cannabis (56.6% vs. 22.2%, p < 0.001), tobacco (62% vs. 28.9% p < 0.001) and had alcohol abuse problems (51.2% vs. 15.5%, p < 0.001). The average HoNOS total score at program entry was 18.3 (± 6.8) indicating a moderate overall illness severity, but no significant difference was observed between the sexes. The pharmacological treatment at admission to the program differed between the sexes: specifically, LAI antipsychotics were prescribed more frequently to men than to women (28.5% vs. 9.5%, respectively), whereas women were more often prescribed oral treatment only (90.5% in women vs. 71.5% in men).

A total of 139 (102 M, 79.1%) and 104 individuals (78 M, 76.5%) were eligible, respectively for the 12 and 24 month follow up (Figure 1): the attrition rate did not show any statistical difference between the sexes at both time points (p = 0.58, p = 0.52).

At both 12- and 24-months follow-up, the clinical outcomes investigated (clinical severity Figure 2, recovery status Figure 3) showed no significant differences between the sexes (Table 2). At both time points, men had a tendency to a lower overall illness severity [HoNOS total score 9.2 vs. 11.3, effect size = 0.37 (CI –0.07; 0.8); 8.3 vs. 9.6, effect size = 0.24 (CI –0.26; 0.75)], even though this difference was not significant. Of note, the only difference observed in terms of outcomes was a worst functional outcome observed in men vs. women at the 24 months time point: more men than women in fact were NEET at the 24 months follow-up (26.3% vs. 8%, p = 0.04).

Further analyses were performed to assess how the overall symptomatology (HoNOS total score) and recovery status (HoNOS total <8) progressed over time by sex (Table 3). At 6- and 12-months men had a significant lower HoNOS scores compared to women, respectively 11.9 ± 1.3 in males and 14.5 ± 2.1 in females at 6-months (p = 0.03), and 9.4 ± 1.3 in males and 11.9 ± 2.2 in females at 12-months (p = 0.05), so men seem to improve faster than women. Even though men scored lower than women at all endpoints, baseline, 18 and 24 months were not statistically significant.

Also, only at the 12-month follow up men had a statistically significant higher recovery probability than women [p = 0.04, effect size = 0.16 (CI 0.01; 0.33)]. Overall, the probability of recovery increased more rapidly in men than in women at the 6- and 12-month follow-ups, but the trend became similar between sexes at 18 and 24 months [effect size at 24 months = 0.14 (0.01; 0.3)] (Figure 3). The proportion of men who reached recovery status was greater than that of women at all endpoints (Figure 4).

However, the sex by time interaction effect over the 24 months of treatment was not significant for both illness severity (p = 0.52) as well as probability of being in recovery (p = 0.40).

Our study has several strengths. First, to our knowledge this is the first study to analyse sex differences in outcomes in FEP patients in Italy thus pioneering the research on how sex and gender could influence clinical trajectories in psychosis. Second, clinical and socio-demographic characteristics were examined in real-life clinical practice, offering a realistic depiction of clinical outcomes. The analysis highlighted some differences both at clinical admission as well as in outcomes that could help tailoring the extant FEP program to sex characteristics and needs. Men could benefit from gender-sensitive approaches that address stigma and acknowledge male-specific help-seeking behaviors, like improving engagement and retention in care in Pathological Addiction Services or facilitating access to vocational rehabilitation programs. On the other hand, women could benefit from more intensive interventions in the first 12 months after a first-episode psychosis, and quality improvement initiatives that take in consideration the potential gender biases in pharmacological choices, expecially regarding the prescription of LAI.

Some limitations need to be taken into account when interpreting our results. First, individuals who discontinued care—either due to significant clinical improvement or for other reasons such as relocation or disengagement—were not included in the follow-up. Considering that a longer follow-up period is important when considering recovery as an outcome, this may be responsible for many non-statistically significant findings (Ferrara et al., 2024c). Second, in our sample, sex prevalence was uneven, with 74.1% men and 25.9% female, a proportion within this age range (18–35) that was expected (Ferrara et al., 2019; Belvederi Murri et al., 2023; Jongsma et al., 2018) but the attrition rate was similar for both the sexes. Moreover, the underrepresentation of women in our sample may also be due to their scarce presence in FEP programs due to the restricted age criteria (Ferrara and Srihari, 2021; Ferrara et al., 2024a; Ferrara et al., 2024b), which excludes women who have their onset at a more mature age (Ferrara et al., 2024a; Ferrara et al., 2024b; Salvadé et al., 2024), thus leading to broaden sex iniquity (Lappin et al., 2016; Ferrara and Srihari, 2021; Seeman, 2021). This selection bias could limit the generalizability of the results to Services such as those in the UK which admit FEP above the 35 years old thereshold; however these findings apply to most FEP programs both at a national as well as international contest that admit only young adults up to age 35. Third, due to this study being observational, no strict cause-effect relation could be established, thus permitting to reveal only mere association between the variables. Moreover, the study design could explain the presence of missing data on older charts that could have influenced the results. Fourth, individuals were selected if their DUP was within 2 years, according to the regional recommendations, but differently from other international program; this might limit the generalizability of the results. Last, given that many patients were born in Italy and only around 20% are foreigner, results may not be generalizable to ethnically diverse samples that could be found in other geographic areas.

To our knowledge this is the first study that showed a gender difference in timing of recovery trends, and this finding could help tailoring the pace and the quality of interventions to males and females needs. Poor vocational outcomes (NEET rates) were significantly higher in males than females, suggesting that symptomatic recovery does not necessarily equate to functional recovery, particularly in male patients (Ferrara et al., 2024c) as others have reported (Chang et al., 2011; Thorup et al., 2014). This highlights the need for psychosocial and vocational interventions tailored to support vocational attainment in men. This suggests that FEP programs should personalize identification strategies and treatment based on sex in order to detect high risk patients before or right after their psychosis onset, since this could lead to DUP reduction and improvement of life quality and global functioning, shaping a better disease trajectory. As a recent qualitative investigation has outlined, mental health providers are aware of possible sex differences in the epidemiology and presentation of FEP, however they also acknowledge the lack of specific training including the influence of the menstrual cycle on symptoms relapse and response to pharmacological treatment (McGinty et al., 2025).

Last, this study focused only on biological sex and did not consider gender and gender roles. Given that gender might differ from biological sex, it should be carefully assessed at admission as it could influence pathways to care, clinical presentation and outcomes (Ferrari et al., 2018).