Background: Theoretical/Conceptual Framework

Several evidence showed that patterns of mental health services utilization differ between migrants and reference populations [1, 2]. A recent systematic review found that migrants are more likely to be admitted to a psychiatric ward via Emergency Department (ED) and less likely to be referred from Community Mental Health Centre (CMHC) compared with the reference population [3]. The available literature has also consistently shown that migrant and ethnic minorities present an increased risk of involuntary psychiatric admissions compared [4, 5, 6]. These differences in the use of mental health services by migrants could derive from the presence of barriers at different levels, including the healthcare system, the service providers and the patients themselves [7]. The healthcare system could be responsible for long waiting lists, lack of interpreters and communication barriers (e.g., information material written in no other language than the dominant one) [7, 8]. Service providers may have poor consideration of cultural background and poor communication skills with patients from a different cultural and ethnic background. Finally, patients particular cultural beliefs may lead to misunderstandings with the healthcare staff and lack of language proficiency or poor knowledge of the healthcare system of the host country may obstacle the access to services [7, 8, 9]. It is also important to remind that, while Western psychiatric services are oriented towards community care, in the countries of origin of many migrants groups the hospital is the main facility where patients seek assistance [9].

The pathway to care, length of stay and type of discharge also differ significantly between migrants and reference populations. Our previous study found that a considerable proportion of migrants (39%) accessed public psychiatric services through non-medical pathways, in line with studies conducted in Italy or other countries [10, 11, 12]. A study conducted in Bologna by Rucci et al. in 2015 [13] showed that the length of stay in psychiatric wards for acute episodes was shorter for migrants. Furthermore, it also found that the reference population was sent for a greater number of days and times to Intensive Residential Treatment Facilities (IRTF) compared with migrants [13].

Today, in the wake of the COVID-19 pandemic and all the dramatic changes it has brought, we wonder if differences in access to and use of psychiatric services by migrants have been further amplified. A recent study conducted in Italy [14] found that the number of migrants patients who visited the outpatient service in March 2020 decreased dramatically, in line with the introduction of lockdown measures. The authors concluded that the lockdown-related reduction in numbers of patients accessing the mental health service makes it difficult to support vulnerable populations during a period of time in which their mental health needs are expected to increase. Moreover, the reduction seen in follow-up compliance increases the risk of treatment discontinuation and possible relapse. As a consequence, one could expect changes in the rates of psychiatric hospitalizations among migrants.

Our study had the main objective of evaluating the incidence of psychiatric hospitalizations among migrants compared with reference population in a well-defined area of the metropolitan city of Bologna (Eastern Bologna) served by a unique psychiatric ward, Servizio Psichiatrico di Diagnosi e Cura (SPDC) Malpighi (Department of Mental Health and Pathological Addiction–Bologna Local Health Authority), between 1st January 2018 and 31st December 2020. Our study also aimed to evaluate the effect of the COVID-19 pandemic on the incidence of psychiatric hospitalizations among migrants. Secondary objective of our study was to describe possible differences between reference population and migrants regarding pathway to care, compulsory hospitalization, the length of stay, clinical course, psychiatric diagnoses, and type of discharge. Finally, we evaluated whether observed differences in the characteristics of psychiatric hospitalizations between migrants and reference population were preserved during the pandemic period.

Methods

We conducted an observational and retrospective study on migrant and patients from the reference population admitted to the acute psychiatric ward “SPDC-Malpighi” between 01/01/2018 and 31/12/2020. The information collected from the patients’ medical records includes sociodemographic and clinical data. Based on the methods used by Cooper et al. [15], we conducted a leakage study to identify any subjects missed by checking the list of patients recorded into Bologna Local Health Authority computerized information systems.

Study Population

Data on the population at risk were obtained from the most accurate local sources of sociodemographic information (Città Metropolitana di Bologna–Servizio Studi e Statistica per la programmazione strategica; Comune di Bologna–Area di Programmazione, Controlli e Statistica, U.I. Ufficio Comunale di Statistica). The geographical area of the catchment area of the SPDC Malpighi includes the urban districts of Santo Stefano, San Donato, San Vitale, Savena of Bologna and the municipalities of Loiano, Monghidoro, Monterenzio, Ozzano nell'Emilia, Pianoro and San Lazzaro di Savena. Our study population, consisting of all residents in the considered territory aged between 15 and 64 years, consisted of a total of 167.955 subjects (midterm population, 2019). The total number of foreign residents amounted to 30,977, while the total number of residents from the reference population was 136,978 subjects (Table1). The most represented groups of origin in the catchment area of the SPDC Malpighi department came mainly from Europe (46.61%), followed by Asia (32.09%), Africa (15.6%), America (5.66%), and Oceania (0.04%).

Table 1 Population at risk
Full size table

According to the current definition proposed by the European Migration Network (EMN, https://ec.europa.eu/home-affairs/networks/european-migration-network-emn_en), we used the term "migrant” in our study to define either a person born outside national borders (first-generation migrants) or born in Italy but with at least one foreign-born parent (second generation migrants).

Ethics

The study was approved by the EC on 17/04/2019, ref. EC AVEC n 257/2019/OSS/AUSLBO. All information is kept anonymously in respect of privacy. Data collection does not involve any treatment outside of normal departmental good clinical practice.

Measures

Information on admissions were retrieved from the informatic registries of the psychiatric unit of the psychiatric ward “SPDC Malpighi” (Hospital Sant’Orsola-Malpighi). Data were anonymized. For each participant we collected the following information:

  • - sociodemographic variables: age, gender, country of birth;

  • - clinical data: admission and discharge date, inpatient duration, pathway to care, discharge modality, compulsory treatment, discharge diagnosis.

Diagnoses were coded according to the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) [16] and grouped as following: organic disorders (OD), substance use disorders (SUD), schizophrenia spectrum disorders (SSD), mood disorders (MD), anxiety disorders (AD), personality disorders (PD), other disorders (OD). Detailed information on the diagnostic groups is provided in the Supplementary Materials (Table 5).

Analysis

Descriptive analyses were conducted using parametric or non-parametric test as appropriate to compare sociodemographic and clinical characteristics by migrant status. We performed analyses on the total sample and then by year of admission (2018–2020).

Then, we estimated the hospitalization rates for each diagnostic group (including admission for any psychiatric disorder) and the rates of compulsory admissions. We estimated the population at risk in our catchment area based on the local sources of the most accurate sociodemographic information on the population of 2019 (Città Metropolitana di Bologna–Servizio Studi e Statistica per la programmazione strategica; Comune di Bologna–Area Programmazione, Controlli e Statistica, U.I. Ufficio Comunale di Statistica). We only considered individuals aged between 15 and 64 years. Population data were stratified by age (15–24 years, then 20-year bands), sex, and migrant status. We first estimated the person-days at risk multiplying the resident population by the days of study duration. Then, we accounted for the time that each individual of our cohort spent in the psychiatric ward during the study period by subtracting the number of inpatient days to the total person-days at risk. Finally, we converted person-days into person-years. Rates with 95% confidence intervals were presented per 10,000 person-years. Poisson regression models were fitted to estimate the incidence rate ratios (IRR) of hospitalizations for psychiatric disorders by migrant status adjusting for age and gender. We used the log of person-time at risk as “offset” while estimating the Poisson regression models to take into account the population denominator over the study time. We controlled the false discovery rate using the Benjamini–Hochberg procedure [17], tolerating a 10% false discovery rate.

Analyses were conducted using IBM SPSS Statistics for Windows, version 25 (IBM Corp., Armonk, N.Y., USA) and STATA 17 (StataCorp. 2021. Stata Statistical Software: Release 17. College Station, TX: StataCorp LLC)).

Results

Table 1 reports the distribution of at-risk population.

Clinical and Sociodemographic Characteristics

Characteristics of the admissions and sociodemographic information of the admitted patients are shown in Tables 2 and 3. Our sample comprised 217 migrants (103 women and 114 men), and 695 individuals from the reference population (297 women and 398 men). The number of total admissions was 329 for migrants and 1,248 for the reference population.

Table 2 Sociodemographic characteristics of the sample
Full size table
Table 3 Clinical characteristics of the inpatient admissions
Full size table

The mean age at first admission was significantly lower among migrants (34.22 vs 42.05; t = 7.95, p < 0.001).

The large proportion of the hospitalized migrants were born in Eastern Europe (26.7%), Sub-saharan Africa (20.3%), North-Africa (17.5%), or Asia (15.2%).

Migrants were more likely to be admitted via ED (61.7% vs 51.9%) and less likely to be referred from a CMHC (29.9% vs 22.2%) or from a non-psychiatric hospital unit (6.5% vs 4%) compared with the reference population (χ2 = 18.119, p = 0.003). More migrants were discharged at home (migrants: 33.0% vs reference population: 27.7%) or to other psychiatric wards (migrants: 2.4% vs reference population:1.4%), while those from the reference population were more frequently transferred to other non-psychiatric hospital units (reference population:3.8% vs migrants:1.8%) or chose to self-discharge (reference population:11.8% vs migrants:8.0%) (χ2 = 12.110, p = 0.033). There was no evidence of a difference in the inpatient duration between the two study groups.

With regard to diagnosis, migrants were more likely to be admitted due to a SSD (migrants: 44.7% vs reference population: 34.9%), while those from the reference population were more likely to be diagnosed with a MD (22.5% vs 18.5%) or SUD (13.4% vs 10.9%) (χ2 = 14.011. p = 0.030).

Comparisons by Year of Admission

Comparing the analyses by year of admission, we found that the proportion of migrants admitted to a psychiatric ward grew over the years from 20.6% (n = 79) in 2018 to 32.0% (n = 77) in 2020. (Table 2).

Migrants were consistently younger than those from the reference population at the time of admission over the years.

In 2020 we registered an increase in the admissions of migrants from Southern Europe (8.9% in 2018, 8.2% in 2019, 11.7% in 2020). Conversely the proportion of North-Africans decreased over the years (21.5% in 2018, 19.7% in 2019, 11.7% in 2020). (Table 6).

The proportion of migrants admitted via ED was consistently higher in relation to the reference population across the years, while the latter were more likely to be referred from a CMHC; though the observed differences in the pathways to care were statistically significant only in 2019 (χ2 = 13,908; p = 0.016) and in 2020 (χ2 = 11,791; p = 0.038). We found no significant differences when analyzing discharge modality or inpatient duration by year of admission. Individuals from the reference population were generally more likely than migrants to be re-admitted, though only in 2019 the difference between the two study groups had a trend towards statistical significance (reference population: 14.2% vs migrants: 7.7%; p = 0.094). (Table 7).

With regard to diagnosis, only in 2018 there were significant differences between the two groups, with migrants being more likely than reference population to be diagnosed with a SSD (migrants:50.9% vs reference population: 34.4%; χ2 = 14.089; p = 0.029). In the following years the proportion of migrants admitted due to SSD decreased (46.2% in 2019; 38.3% in 2020), while the proportion among reference population grew from 34.4% in 2018 to 38.1% in 2020. Finally, we observed an increase in the proportion of compulsory treatments for migrants over the years, being 12.7% in 2018 and 17.2% in 2020. Conversely, within the reference population, the proportion of compulsory treatments remained stable. (Table 7).

Hospitalization Rates

During the study period we observed a total of 503,834 person-years at risk, comprising 410,910 from the reference population and 92,924 migrants. As reported in Table 4, hospitalization rates due to any psychiatric disorder for migrants were slightly higher compared with the reference population (35.30/10,000 person-years vs 30.30/10,000 person-years; IRR = 1,16 95% CI = 1.02.1.31). This finding was no longer significant when controlling for false discovery rate.

Table 4 Crude and adjusted incidence rates of hospital admissions for psychiatric disorders during the study period
Full size table

As reported in Table 4, both men and women with a migratory background had higher rates of hospitalization due to any psychiatric cause and the difference between migrants and the reference population was particularly pronounced among the youngest age group (15–24 years), as in this age-band rates were 3 times higher for migrants (70.97/10,000 person-years vs 21.61/10,000 person-years; IRR = 3.24 95% CI = 2.44–4.29). The rates then decreased in the following age-bands and among those aged between 45- and 64-years individuals from the reference population had a 2 fold higher risk of being admitted to the psychiatric ward for any cause (30.51/10,000 person-years vs 16,31/10,000 person-years; IRR = 0.55 95% CI = 0.41–0.72). Rates of compulsory treatments were generally higher for males and similar between the two groups, but migrants aged between 15- and 24-years had a 4-times greater risk of compulsory admission (10.65/10,000 person-years vs 2.77/10,000 person-years; IRR = 3.77 95% CI = 1.78–7.96). With regard to admissions due to a specific disorder, we found relevant differences in hospitalization rates for SSD and MD. Specifically, migrants had a 1.55 (95% CI = 1.28–1.87) higher rate of being admitted for SSD compared with the reference population (15.71/10,000 person-years vs 10.54/10,000 person years). Rates were increased for both males (22.99/10,000 person-years vs 13.83/10,000 person-years; IRR = 1.58 95% CI = 1.25–2.00) and females (9.66/10,000 person-years vs 7.22/10,000 person years; IRR = 1.46 95% CI = 1.05–2.03) of migrant origin and they were particularly high in the youngest age-group (28.39/10,000 person years vs 5.53/10,000 person years; IRR = 4.97 95% CI = 3.04–8.11). Regarding admissions due to a MD, rates were generally similar when considering the total admissions (6.56/10,000 person-years for migrants vs 6.84/10,000 person-years for the reference population; IRR = 1.17 95% CI = 0.94–1.44), but they were higher for female migrants compared with females from the reference population (8.28/10,000 person-years vs 5.86/10,000 person-years; IRR = 1.45 95% CI = 1.02–2.07) and for migrants of the youngest age-group in relation to the reference population counterpart (8.87/10,000 person-years vs 1.56/10,000 person-years; IRR = 5.90 95% CI = 2.40–14.52). When adjusting our results for multiple comparisons, the association between gender and increased risk of admission due to a MD was no longer significant.

Discussion

We found that both men and women with migration history had higher rates of hospitalization due to any psychiatric disorder compared with the reference population. The risk of admission to the psychiatric ward was around 3 times higher among the youngest migrants (15–24 years) in relation to non-migrants of the same age. Importantly, the risk of compulsory admission for migrants aged between 15 and 24 years old was almost 4 times greater. These findings are consistent with available evidence, which has consistently shown that rates of psychiatric admissions and in particular of compulsory psychiatric admissions are significantly increased among individuals of ethnic minorities [4, 5, 6, 18].

In line with previous studies [13, 19], we found disparities in services use between migrants and the reference population: migrants were more likely to be access via ED and to be discharged at home, while those from the reference population were more likely to be readmitted (revolving doors) and to self-discharge from the hospital treatment.

The reasons for the excess of compulsory treatment among migrants and ethnic minorities and other ethnic disparities in psychiatric services use has been correlated with multiple reasons, at the individual as well at the society and services organization level, such as experiences of structural or institutional racism [20]; difficulty in engaging culturally diverse people in mental health services [13, 21]; the migrants’ own beliefs and stigmatization of mental disease [22, 23]; difficulties related to local migration policies, including expulsions [24]; inadequate health literacy, low language proficiency, or other known barriers to health care [7, 9, 25, 8]. A recent study conducted in Rome found that migrants with less than 2 years of stay in the host country present about 5-time increased odds of compulsory admission [26], suggesting that recently arrived migrants are the most vulnerable to adverse pathways to care. This aligns with our finding that the migrants of the youngest age-group are those carrying the highest risk of involuntary admission.

Moreover, we found relevant differences in hospitalization rates for SSD and MD. Specifically, migrants had a 55% higher probability of being admitted for SSD compared with the reference population, with higher rates of admission for SSD for both men and women migrants and particularly the youngest age-group (16–24 years). Considering the age band, it is likely that the admissions coincide with the first episode of psychosis. A meta-analysis of studies conducted in England suggests that young-aged migrants and ethnic minorities are more likely to be involuntarily admitted to the psychiatric ward and less likely to be referred by the General Practitioner in correspondence of first-episode psychosis compared to the reference population [27]. A study conducted in Canada over subjects aged 16–35 years with first-episode psychosis yielded similar results [28]. Regarding admissions due to MD, there was some evidence of higher rates for women migrants compared with females from the reference population. A previous study found higher rate of depression among migrant women, but not among migrant men, compared with the reference population [29]. Interestingly, this study [29] did not find higher rates of psychiatric consultations among migrant women compared with women from the reference population and thus concluded that depression among migrants women could be mostly under-recognized and undertreated. Our results also showed that migrants of the youngest age-group had 5-times higher rate of admission for MD compared with the reference population. Overall, migrants of the youngest age-group and of female gender carried the highest risk of psychiatric admission in our catchment area. This result confirms previous findings from a study conducted in Paris [19].

Comparing the proportion of the different ethnic groups among the patients admitted to the psychiatric ward with the proportion of the same group in the general population, we observed that migrants and minorities from sub-Saharan Africa (20.7% of the admitted vs 6.2% in the general population) and North-Africa (17.9% of the admitted vs 8.7% in the general population) were about 2–3 times overrepresented in the psychiatric ward. Conversely, migrants and minorities from Asia (11.8% of the admitted vs 30.8% in the general population) were underrepresented. The area of origin disproportionality in the context of acute psychiatric ward is highly suggestive of disparities in mental healthcare provision among different ethnic groups. Evidence of disparities in access and use of mental health services is well documented across the globe [30, 31], and various initiatives and guidelines have been promoted [32, 33]. Migrants, especially those from certain World regions, are more exposed to violence and traumatic events before migrating and during the migration journey, with detrimental effects on their mental health [34]. Nevertheless, prior research has established that social disparities in the country of resettlement increase the risk of severe psychiatric disorders, such as psychosis, independently of trauma [35, 36]. Thus, appropriate interventions at multiple levels (primary, secondary and tertiary prevention) are required to intercept the mental health needs of these population groups and to provide adequate and accessible services.

The rates of voluntary and compulsory admissions of migrants increased over the years, particularly during the year of COVID-19 pandemic (2020), while the voluntary and the compulsory rates of psychiatric hospitalizations among the reference population remained quite stable over time. This result is in line with the finding of an increase in compulsory admissions already found by previous studies for migrants during pandemic [37]. This phenomenon might be attributed to an increase in mental health disorders in ethnic minorities [38, 39]. Increased precariousness, financial constraints, stigmatization by the non-migrant community, as well as the impossibility of implementing preventative social distancing measures due to a lack of living space have been hypothesized to contribute to the higher vulnerability of migrants to the psychological and emotional trauma of the COVID-19 pandemic [40]. We also found an increase in the proportion of psychiatric admissions for migrants from Southern-Europe and Asia, along with a reduction of the representation of North- and sub-Saharan Africans in the last year of COVID-19 pandemic (2020). We can hypothesize that some ethnic groups had to change their attitudes to seeking care in countries of origin or from traditional healers because of the limitation of travels and social contacts imposed during pandemic.

Strengths and Limitations

To our knowledge, this study is one of the first investigations on acute psychiatric admissions before and during the pandemic. This study has been conducted in acute psychiatric ward covering a well-defined catchment area. Nevertheless, we recognize that the single-center design may limit the generalizability of our findings. Leakage studies were conducted to minimize under-ascertainment [15]. No calculation of the sample size was done as, in this naturalistic study, we aimed to include all the patients consecutively admitted to the psychiatric ward over a defined timeframe.

Some patients, despite residing in our catchment area, may have been admitted in other psychiatric wards and thus may not have been included in the study. However, psychiatric regulation in Bologna establishes that patients must be treated in the ward which the area where they reside. Those who are firstly admitted to another psychiatric ward are generally transferred to the appropriate psychiatric ward. Thus, we think that very few cases could have been missed and that our estimates of hospitalization rates are reliable.

Denominators for the population at risk of psychiatric admissions were derived from the most accurate local sources of sociodemographic information. We used population of 2019 to estimate the person-years at risk. We acknowledge that the true dynamic population at risk over the survey period may have varied, but we expect this variation to be slight and thus not biasing our estimates.

Some important information, such as educational, occupational, and marital status were not fully available. This did not allow adjusting for potentially relevant confounders.

Conclusions

The higher rates of the acute psychiatric admissions for migrants that we found in Bologna suggest that the issue of equal provision of mental health intervention is far from solved and more public health attention is needed for addressing migrants’ mental health needs. The increasing of psychiatric admissions for migrants during 2020 enlighten that the pandemic increased the migrants’ vulnerability to psychiatric disorders and weakened their access to community psychiatric care. Further studies are urgently needed to better understand the barriers faced by migrants with mental disorders seeking care and to define the culturally competent interventions that should be implemented to ameliorate their pathway to care. Finally, special attention should be paid to women and youngest migrants.