This study is part of a national multicenter observational project promoted by the Italian Society of Psychiatric Epidemiology (SIEP), aimed at identifying clinical, demographic, and organizational factors associated with IPT across Italian mental health services. The current analysis focuses on data collected at the Reggio Emilia site, located in Northern Italy, under the coordination of the Department of Mental Health and Substance Use of the Azienda USL–IRCCS of Reggio Emilia. Within the department, there is a single Psychiatric Diagnosis and Treatment Unit, located in the town of Correggio and comprising 16 beds. This unit is specifically designated to admit and treat all patients undergoing IPT across the province, which has a population of approximately 530,000 inhabitants. Data collection at this site took place over a 12-month period, from March 20th, 2023, to March 19th, 2024.

All patients admitted to the Psychiatric Diagnosis and Treatment Unit under IPT during the study period were included in the analysis. Only adult patients aged 18 years or older were eligible.

Data were gathered using standardized event forms developed by the SIEP coordinating group. These forms were completed by trained clinical personnel immediately following each IPT admission. Collected information included sociodemographic data (age, gender), clinical data (diagnosis categorized as schizophrenia spectrum disorders, bipolar disorder, personality disorders, or other), toxicological screening results at admission, and whether mechanical restraint was applied during hospitalization.

In addition, data were collected on service-related factors, including whether the patient was already in contact with a Community Mental Health Center (CMHC), whether the hospitalization represented the first-ever contact with mental health services, and whether the IPT followed a voluntary admission. Legal and procedural variables were also documented, such as the source and setting of IPT validation (e.g., emergency department, inpatient unit, or territorial services), whether the IPT was renewed after the initial seven days, and whether the patient filed a legal appeal. Among the variables collected, the duration of the IPT in days was also recorded. All data were anonymized and transferred to a centralized national database for analysis.

All statistical analyses were conducted using IBM SPSS Statistics 27 (24) and JASP version 0.19.3 (25). Descriptive statistics were computed for all variables, including frequencies and percentages for categorical variables, and means and standard deviations or medians and interquartile ranges for continuous variables, depending on their distribution.

The normality of distribution for continuous variables, particularly IPT duration, was assessed using the Kolmogorov–Smirnov test, along with visual inspection of histograms and Q–Q plots. If the assumption of normality was met, comparisons between two independent groups were conducted using independent samples t-tests, while comparisons across more than two groups were analyzed using one-way ANOVA. In the case of non-normally distributed variables, Mann–Whitney U tests and Kruskal–Wallis tests were applied. Where appropriate, Dunn’s post-hoc tests with Bonferroni correction were used to explore pairwise differences.

Associations between categorical variables were analyzed using the Chi-square test of independence, or Fisher’s exact test when expected frequencies were below the recommended threshold. All tests were two-tailed, and statistical significance was set at p<0.05.

Given the observational nature of the study and the limited set of covariates available in the SIEP event forms, analyses were conducted as exploratory, hypothesis-guided comparisons rather than as predictive modelling. No single primary outcome was predefined; instead, we examined three pre-specified IPT-related outcomes: duration (days), renewal (yes/no), and mechanical restraint (yes/no).

Between March 20, 2023, and March 19, 2024, all patients who underwent involuntary psychiatric hospitalization (IPT) during the study period were enrolled, resulting in a total sample of 214 individuals admitted under IPT to the Psychiatric Diagnosis and Treatment Unit. The sample comprised 112 males (52.3%), with the most represented age group being between 41 and 60 years (39.7%). Regarding psychiatric diagnoses, the most frequent were schizophrenia spectrum disorders (53.7%), followed by bipolar disorder (14%). Other diagnoses (18.7%), including major depressive disorder and substance-induced psychotic disorders. In addition, toxicological screening at admission indicated recent substance use in 20.1% of patients. For 67.3% of the patients involved (N = 144), this was their first IPT since the beginning of the study.

At the time of admission, 65.4% of patients were already under the care of a Community Mental Health Center (CMHC), while 35 were first-time contacts with mental health services. Additionally, 8.4% of the sample had initially been voluntarily admitted, with their status subsequently converted to involuntary hospitalization.

For 37 patients, the IPT was renewed at least once beyond the initial seven-day period. Only one patient filed an appeal against the TSO during the study period. The majority of IPTs were validated by community psychiatric services (N = 95) and emergency departments (N = 70).

All these results are presented in Table 1.

The average duration of IPT was 7.4 days (SD= ± 3.8), with a minimum of 1 day and a maximum of 28 days. Mechanical restraint was used in 13 cases, corresponding to 6.1% of the sample, while it was not applied in 87.4% of cases. These findings are summarized in Table 2.

Our first set of hypotheses concerned the role of previous contact with mental health services, prior history of coercion, and the modality of admission (direct vs. conversion from voluntary). We expected that first-ever patients would experience longer IPTs, that prior IPT would predict renewal, and that conversions from voluntary admission would be shorter than direct compulsory admissions.

Results only partially supported these hypotheses. Contrary to expectations, there was no significant difference in IPT duration between first-ever patients and those already in care. This finding diverges from prior evidence suggesting that lack of service engagement predicts both higher risk and longer duration of compulsory hospitalization (26). In our context, organizational factors or legal safeguards may mitigate this effect. It is also plausible that the prompt involvement of community mental health services prevents delays in care transitions, thereby avoiding unnecessary prolongation of hospitalization. This would suggest that, once IPT is initiated, acute clinical severity and risk profiles, rather than service history or admission pathway, primarily determine duration (27).

By contrast, prior history of IPT emerged as a robust predictor of renewal. Patients with previous coercive experiences had a markedly higher probability of IPT prolongation compared to first−time cases. This recurrence effect is well documented: international studies consistently show that patients who have been coerced once are at higher risk of future coercion. For example, Lay et al. (15) found that a history of previous compulsory admissions increased the hazard of re−admission by approximately 1.8 times (and even more when the prior admission involved endangerment of others). Similarly, Zervakis et al. (28) showed that prior involuntary commitment was significantly associated with increased perceived coercion during subsequent voluntary admissions (rate ratio=1.6; p=0.05). Moreover, Martinez et al. (29) reported that past coercive experiences predict higher levels of perceived pressure and fear, and lower satisfaction in later hospitalizations, even when these are voluntary. Beyond the psychological and relational aftermath of prior coercion, Müller et al. (30) demonstrated that chronicity and prior involuntary treatment significantly predict earlier and more frequent coercive measures in psychiatric inpatient care. These findings suggest that, beyond clinical relapses, relational and trust barriers generated by previous involuntary interventions reinforce a cycle of exclusion and coercion.

Finally, the hypothesis that conversions from voluntary admission would be associated with shorter IPT was not supported. The absence of differences between converted and direct IPTs suggests that the clinical severity that leads to coercion outweighs admission pathway. Similar findings have been reported in German and Swiss samples, where conversions did not significantly alter hospitalization outcomes (6, 31). Taken together, these results highlight the complex relationship between service engagement and coercion: while engagement alone may not shorten duration once IPT is initiated, a history of coercion strongly increases the risk of renewed involuntariness, underlining the need for preventive, community-based approaches for high-risk patients.

The second domain concerned diagnostic correlates. As hypothesized, diagnosis significantly influenced IPT duration, with schizophrenia spectrum and delusional disorders associated with the longest stays. These findings confirm a large body of evidence showing that psychotic disorders are the strongest predictor of both compulsory admission and prolonged coercion. Indeed, a meta-analysis by Walker et al. (13) reported that diagnosis of a psychotic disorder was associated with one of the highest risks of involuntary psychiatric admission (OR = 2.18, 95% CI 1.95–2.44). Additionally, Flammer and Steinert (32) found that, among non-organic diagnoses, psychotic disorders were the most important predictor of detention during hospitalization. These observations are consistent with findings from Iozzino et al. (16), who, in a systematic review and meta-analysis of violence in acute psychiatric wards, reported an increased risk of aggressive behavior among patients with psychotic disorders. While aggressive behavior often represents a trigger for coercive interventions, this association should be interpreted with caution, as it may reflect not only clinical features but also contextual, relational, and organizational factors influencing the use of compulsory measures.

The higher IPT duration among psychotic patients may reflect several converging factors: reduced insight and treatment adherence, risk of aggression, and complexity of stabilization. Moreover, previous studies show that clinicians tend to perceive psychosis as less amenable to consensual negotiation, thus lowering the threshold for coercion (6).

Interestingly, bipolar disorder and other diagnostic categories were associated with shorter IPT durations, and personality disorders showed the shortest IPT durations, although this difference reached statistical significance only in comparison with schizophrenia spectrum disorders. While our findings cannot be directly compared with duration-related data in other settings, they are in line with international evidence showing that affective disorders are less frequently managed under prolonged coercion (30). At the same time, some studies suggest that patients with borderline personality disorder may face high rates of short-term coercion (e.g., seclusion or restraint), often driven by crisis behavior. For example, a recent study in child and adolescent psychiatry reported that manic episodes/bipolar disorder (OR = 36.4) and personality disorders (OR = 15.2) were strongly associated with increased likelihood of seclusion or restraint (33). Our data, showing no significant difference in restraint across diagnostic groups in Reggio Emilia, suggest that in this context diagnosis mainly influences duration rather than the type of coercion.

The high proportion of schizophrenia spectrum diagnoses in our IPT cohort may differ from patterns reported in some countries, where affective disorders (including major depression) contribute substantially to involuntary admissions. This discrepancy likely reflects differences in service pathways, legal and clinical thresholds for IPT, and local crisis-management practices. In our setting, IPT appears to be more commonly activated during acute psychotic decompensations requiring urgent containment and stabilization, whereas severe depressive episodes may more often be managed through voluntary admissions or alternative pathways, unless accompanied by suicidality or psychotic features.

When IPT occurs in the context of major depressive disorder, it is most plausibly driven by acute suicidal risk, psychotic depression, or other high-risk clinical scenarios rather than by impaired decision-making per se. In the absence of structured data on suicidality, psychotic features, or imminent danger at admission, this interpretation remains inferential. Nevertheless, international literature consistently indicates that severe affective episodes accompanied by suicidality or psychotic symptoms represent common pathways to compulsory treatment (16).

Finally, these results underscore the importance of implementing early intervention services for psychosis, which have been shown to improve outcomes and, in some contexts, reduce the risk of involuntary treatment. For instance, randomized and review evidence indicates that specialized early intervention services enhance functioning, adherence, and patient satisfaction (34) and that structured crisis interventions such as joint crisis plans can reduce involuntary admissions (35). Developing such structured alternatives may help manage acute psychotic decompensation without resorting to prolonged IPT.

The third domain examined gender and care pathways, focusing on restraint use, duration, and setting of validation.

As hypothesized, male patients were significantly more likely to be mechanically restrained than females. This aligns with multiple studies demonstrating that male gender is a risk factor for coercive measures: in a recent multicenter study in Paris, male sex was independently associated with a higher likelihood of experiencing mechanical restraint (36). Similarly, Nawka et al. (37) found that physical restraint was used more frequently with male patients, while forced medication and seclusion were relatively more common among females. These findings may stem from gendered perceptions of aggression and risk—staff may perceive men as more dangerous, thus lowering the threshold for restraint. Ethically, this disparity is concerning; it indicates that gender stereotypes can color clinical decision-making and underscores the urgent need for trauma-informed, bias-aware training for healthcare professionals.

In contrast, the hypothesis that women would experience shorter IPT durations was not confirmed. Although the average duration was slightly lower for women, the difference was not statistically significant. This suggests that while gender impacts the likelihood of restraint, it does not strongly affect the length of compulsory hospitalization (38).

While international studies suggest that emergency settings may be associated with faster crisis resolution, our local findings indicate that, after correction for multiple comparisons, IPT duration is broadly comparable across emergency departments, inpatient units, and community psychiatric services. The absence of large differences across standard care pathways does not negate the role of organization but rather suggests a relatively homogeneous service culture within the province, likely shaped by shared clinical practices, legal safeguards, and integrated service governance. The only setting associated with significantly shorter IPT durations was a residual category of “other services”, suggesting that organizational context per se does not systematically determine duration, except in atypical validation pathways.

Similar patterns have been reported in other European emergency psychiatry settings, where male gender and psychotic spectrum diagnoses are consistently associated with higher use of coercive measures, while differences across care settings tend to diminish once organizational characteristics are taken into account (39).

This study has several strengths. It is based on exhaustive data collection covering all IPT cases over a full year in a defined catchment area, ensuring completeness and minimizing bias. Integration into a national multicenter initiative (SIEP) allowed standardized procedures and comparability across sites. Unlike many studies relying solely on administrative data, we included clinical, demographic, and organizational variables, offering a multidimensional perspective. The study also followed rigorous ethical standards and used validated instruments with trained staff.

Nonetheless, important limitations restrict generalizability. The single-site design is the main weakness: findings from one unit cannot represent national practice given regional variability in IPT. Although the sample included all local cases, subgroup analyses suffered from small cell counts, reducing power. Missing contextual variables—such as socioeconomic status, family support, employment, staff ratios, and ward climate—limit explanatory depth. Patient and family perspectives were absent, despite coercion being primarily experienced as traumatic (5). Future studies should adopt participatory approaches to integrate these voices.

Outcome measures were restricted: while we assessed duration, renewal, and restraint, other forms of coercion (e.g., forced medication, seclusion) and long-term outcomes (e.g., relapse, functional recovery, trust) were not captured. This narrow focus risks portraying IPT as isolated rather than embedded in broader care trajectories (6, 30).

A further limitation is the absence of structured clinical indicators that commonly trigger involuntary treatment, such as imminent danger to self or others, suicidal attempts, intentions or plans, severity of psychotic symptoms, severe self-neglect, or aggressive behavior at admission. These factors are key drivers of compulsory pathways and may cut across diagnostic categories. Their unavailability prevents us from disentangling whether observed associations (e.g., diagnosis or validation setting) reflect underlying acute risk/severity profiles rather than service-level or organizational mechanisms.

Information on the presence or involvement of legal representatives or guardians was not collected, preventing any assessment of their potential impact on consent procedures, appeals, or coercive pathways. This represents an important legal and ethical dimension that should be addressed in future studies.

We were unable to compare the diagnostic distribution of involuntary psychiatric admissions with the diagnostic distribution of all psychiatric hospitalizations in the catchment area during the same period, due to the unavailability of comprehensive local denominator data. Such comparisons would be essential to determine whether specific diagnoses are over-represented among IPT cases relative to overall admissions and should be addressed in future multicenter and pooled analyses within the SIEP project.

Finally, some effect sizes were small or negligible, warranting cautious interpretation of statistically significant results.