Background
Ethnic minorities and immigrants are at increased risk of developing mental illness including psychotic disorders [
1‐
5]. Underlying this increased risk is a complex set of social factors that include socio-economic disadvantage, as well as discrimination and alienation [
6‐
9]. Despite the higher incidence of psychiatric disorders among minority groups, in many countries, they are less likely to seek and to have access to care [
10‐
13] and have a higher probability of involuntary hospitalization [
14‐
17]. Mental health treatment gaps exist also in Israel between the Jewish majority and the Arab minority - who constitute 21% of the country’s population (of whom 83% are Muslim, 9% are Druze and 8% are Christians). The Arab minority in Israel, while having achieved a remarkable improvement in health and life expectancy, still suffers from social disparity compared with the Jewish population. The Arab population has an increased risk of emotional distress which may be related to social stressors, psychosocial disempowerment and self-appraisal of lower social status [
18,
19]. They are also less likely to seek mental health care [
18‐
20] and the delay between first diagnosis and commencement of treatment is greater among Arab patients [
19]. The limited availability and accessibility of linguistically and culturally appropriate care, hampers treatment seeking [
21].
Lurie and colleagues [
22] found that among people hospitalized in a psychiatric ward in Israel between 2003 and 2013, the mean length of stay per admission was about half as long among Muslim-Arab patients compared with Jewish and Christian-Arab patients. The mean number of days between hospitalizations was also lower among Muslims, resulting in a slightly higher number of hospitalizations. Court-ordered involuntary hospitalizations were twice as common among Muslim-Arab patients (16.3%) as Jewish patients (7.3%). Over the years, the gap in rate of hospitalizations between Jewish and Arab patients diminished [
22].
A mental health reform was implemented in Israel in 2000 that led to the opening of community rehabilitation services for people with severe mental illness, alongside a reduction in the number of psychiatric hospital beds [
23]. A decade later, in 2010, the Israeli parliament officially recognized the need to allocate additional resources for patients with co-occurring substance use disorder and mental illness, or dual-diagnosis (DD) patients [
24]. Adoption of these policies catalyzed a gradual process of opening hospital-based and community-based services tailored for DD patients. Use of rehabilitation services is associated with fewer hospital days among people with severe mental illness [
25], particularly those with chronic psychotic disorders with or without DD [
26,
27].
Study Aim
In this study we examined, for the first time, hospitalization patterns [frequency of hospitalizations, mean length of stay (LOS) and annual hospital days per person] among Arab and Jewish patients in three time periods: 1991–2000, 2001–2009, and 2010–2016) - before and after implementation of the mental health reforms. Specifically, we tested three a-priori hypotheses: (1) the frequency of hospitalizations is greater and LOS shorter among Arab patients; (2) in both Arab and Jewish patients, LOS is shorter among women; (3) Arab-Jewish differences in hospitalization characteristics have converged over time.
Methods
Data for this study was extracted from the Israel’s National Psychiatric Case Registry (NPCR) of the Ministry of Health. The NPCR is the official registry of all psychiatric admissions and discharges countrywide since 1950 [
29]. The data included 18,684 patients aged 18–65 years hospitalized in a psychiatric ward during the period 1963–2016, with an ICD-10 diagnosis of schizophrenia (F20) or schizoaffective disorder (F25) at their last discharge, and with at least one hospitalization occurring during the years 2010–2015. Of the 18,684 patients, 15,145 self-identified as Jewish and 2,556 self-identified as Arabs. As previously described [
26‐
28], we restricted the sample to those persons hospitalized at least once since 2010 in order to ensure that the data and the findings are relevant and timely.
For each hospitalization, a Substance Use Disorder (SUD) diagnosis was made based on a recorded ICD-10 diagnosis of F10-F19 (excluding F17, nicotine dependence) in the categories of dependence and harmful use (DSM diagnoses are not used in Israel) and/or a psychiatrist-documented indication of alcohol and/or substance abuse. We adopted a conservative approach, and classified patients as DD only if they met SUD criteria on at least two hospitalizations, or at least 20% of their hospitalizations, rather than the more common ‘lifetime’ or first-hospitalization criteria. Each individual’s hospitalization history was documented from the first hospitalization until the end of 2016. Twenty-nine patients with anomalous numbers of hospitalizations (≥80) were excluded from the study. A total of 18,684 patients with 168,377 hospitalizations were included in the analysis [
26‐
28].
For each patient, the following hospitalization measures were calculated: average length of stay (LOS), annual number of hospitalizations, and annual number of hospitalization days. For a complete case analysis, repeated-measures ANOVA (with Greenhouse-Geisser correction) was used to measure associations with population group (Arabs, Jews), time-period (Period1: 1991–2000, Period2: 2001–2009, Period3: 2010–2016) and gender, as well as interactions between these variables on hospitalization measures. By way of a sensitivity analysis, linear mixed effects models with random intercept at the participant level adjusted for gender and ethnicity were also run to compare all individuals across the periods. The data were analyzed using IBM® SPSS® Statistics, Version 24.0 and STATA 16.
The study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Review Board of the Israel Ministry of Health. Data was anonymized prior to being released to the researchers.
Discussion
Schizophrenia is estimated to develop in about 0.5% of adults, with little variation across countries [
30,
31]. The Israeli Arab minority constitutes 21% of the country’s population [
32], yet the proportion of Arab patients among persons hospitalized with a diagnosis of schizophrenia or schizoaffective disorder in Israel was 14.4%. Assuming there is no major underlying difference in prevalence between the two population groups, it would seem that a greater proportion of Arabs with psychotic disorder in need of hospital care, are not hospitalized.
As described above, and in keeping with our hypothesis, the large differences in psychiatric hospitalization patterns between Jewish and Arab patients that were noted in earlier years, narrowed over time. The most noticeable gap was in the pre-reform study period (1991–2000) when LOS and annual days of hospitalization were nearly two-fold higher among Jewish than Arab patients. A dramatic decrease in hospitalization days among Jewish patients, and a modest decrease among Arab patients, followed enactment of the Community Rehabilitation of Persons with Mental Disability Law in 2000. The law brought about a reduction of hospitalization beds and the opening of community-based rehabilitation services, primarily though, for the Hebrew-speaking (Jewish) population. Till today, there is a substantial shortage of rehabilitation frameworks for the Arab-minority. In Jerusalem, for example, the ratio of Hebrew-language rehabilitation services to Arabic-language service is over 20:1, although Arabs comprise one-third of the city’s population. In addition, only one residential detoxification facility and few community-based SUD services are available for the Arabic-speaking population [
33]. Arabs comprise nearly one-quarter of the population of Israel, yet less than 2% of psychiatrists in the country are [
34]. These are especially critical service gaps for young Arab adults who require a culturally appropriate therapeutic approach and who have difficulty adjusting to treatment settings geared for the Jewish population which pose linguistic and cultural challenges that likely impair patient-therapist communication.
Consistent with the tendency in the minority Arab population to avoid hospitalization and thus postpone initial admission [
18‐
20], Arab patients are, on average, a year and a half older than Jewish patients at the time of their first hospitalization. Several factors, beyond the limited number of treatment settings and programs, may contribute to hesitancy to seek mental health care. Knowledge and awareness about mental disorders generally lags behind that for physical diseases [
35], particularly among socially disadvantaged and marginalized groups. Furthermore, social stigma about mental disorders is dominant in the traditional Arabic society [
36‐
38], especially among women, as it is in the ultra-Orthodox Jewish community [
39,
40].
It is also possible that greater family support in the Arab population of Israel, and an extended family living arrangement helps prevent exacerbations that require hospitalization [
41,
42]. Social and familial support may also explain the shorter length of hospital stay among Arab patients, as there may be fewer cases in which social intervention for housing is required, such as searching for an alternative post-discharge place of residence, which prolongs the hospital stay. A greater tolerance for aberrant psychotic behavior, reliance on traditional healers, and a mistrust of the public hospital system that is regarded as “westernizing medicine” by some members of the Arabic community, may also play a role [
20‐
22].
Women experienced less frequent hospitalizations and fewer hospital days than men, as has been previously reported [
22]. Although various studies did not find significant differences in LOS between men and women [
43,
44], in our study women’s LOS was shorter. The average LOS in Israel is higher than in the US and other Western countries [
43‐
45], and it is therefore possible that in longer hospitalizations, as in Israel, the better prognosis of women is also expressed in a shorter LOS compared to that of in men [
31,
46]. This gender difference was more pronounced among Arab patients, perhaps due to greater pressure for early discharge to bring the woman back home. Firstly, maintaining religious practices and preserving modesty in hospital settings is a challenge among Muslim women, about which much has been written [
48‐
52]. Also, although the Israeli Arab society is transitioning toward modernization, women are still bound by traditional cultural norms that emphasize the woman’s responsibility for “internal” tasks within the home []. Among women, the onset age of schizophrenia is typically in the mid-twenties, on average 3–5 years later than in men [
31,
53,
54] (with a second peak after age 40 [
31,
55]). At these ages, Arab women are more likely than Jewish women to be married and have children. In 2010, 77% of Arab women aged 25 − 20 were married compared with 51% of Jewish women, and in 2020 these percentages were 70% and 47%, respectively [Central Bureau of Statistics of Israel, 2023]. Being married is itself associated with a better prognosis and fewer hospitalizations [
31]. Historically, Muslim women had much higher fertility rates than Jewish women, although the gap declined from a three-fold difference in the 1960’s to a two-fold difference even as recently as the early 2000’s and has all but disappeared in recent years [Central Bureau of Statistics of Israel, 2023]. Notwithstanding, Arab women still suffer from greater social disadvantage compared to men, compounded by their subordinate role in a patriarchal society, thus affecting their contact with mental health services [
18,
56,
57].
The dramatic decrease in hospitalization days seen among Jewish patients from Period
1 to Period
2 is likely due to the opening of rehabilitation services in the community and the reduction of hospitalization beds. Among Arab patients, only a modest decrease in hospitalization days was noted, perhaps related to the ongoing limited availability of suitable community-based services. It is plausible that members of the Arab population may have reservations about seeking rehabilitation, similar to their hesitancy towards hospitalization. The decrease in hospital days coincided with an increase in the number of hospitalizations in both groups. It is possible that among the Jewish patients the increase in the number of hospitalizations is related to the shortening of LOS among patients who were released before reaching full remission due to the reforms described, when the rehabilitation system was still in its infancy. Among Arab patients, the shortening of LOS is relatively minor and therefore unlikely to be the primary cause of the significant increase in hospitalizations. The increasing use of drugs (e.g., synthetic cannabinoids and cannabis) in both groups may have also contributed to the increase in the number of hospitalizations in Period
2 [
27].
Strengths and Limitations.
This study utilized data from the Israel Mental Rehabilitation Register and the National Psychiatric Case Register that captures virtually all psychiatric hospitalizations. Complete hospitalization histories were obtained for all in-patients diagnosed with schizophrenia or schizoaffective disorder in Israel in the period 2010–2015. However, the retrolective design of the study precluded the analysis of some important demographic and clinical variables unavailable in the national register, such as socioeconomic status, severity of the psychotic disorder, and level of functioning. Restricting the analyses to persons hospitalized between 2010 and 2015, rather than including all patients with the relevant diagnoses who were hospitalized during the entire study period (from 1991), limited our ability to assess whether differences in hospitalization characteristics between Arab and Jewish patients have converged over time. In addition, Twenty-nine individuals with more than 80 hospitalizations were not included in the analyses. The decision to exclude them stemmed from an impression that the excessive numbers of hospitalizations might have been due to double reporting, however they could also represent extreme cases of revolving-door patients. Regrettably, the data for these individuals is unavailable and we are unable to rerun the analyses with them included to assess the impact of the exclusion. We believe the impact is negligible given the small number excluded relative to the size of the study population.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.