The Epidemiology of Suicide in Israel: International and Intranational Comparisons

Data Sources on Suicide for Israel

Until 1966, suicide was regarded as an act that necessitated legal intervention. Police records were made available to the Ministry of Health and the CBS. From 1966 on, however, suicide ceased to be considered a legal offense; the physician's duly completed death certificate became the basis for establishing that a person had died by suicide. In Israel, burial will not be authorized by the district medical officer without the physician's death certificate and, in all cases that arouse suspicion, the physician's affirmation that death was not due to a natural cause. In such cases, an investigation is requested. The CBS, which collects and processes mortality statistics, draws upon any available information from ad hoc sources in order to enhance the validity of its data. These sources include police records and reports from hospitals and from the National Forensic Institute. Whenever a cause of death is reported by the physician as “undetermined,” an effort is made to arrive at a diagnosis using the above-mentioned additional sources. This operation identifies a further percentage of suicide deaths. For example, in 1973 24% of the 378 cases of undetermined deaths were rediagnosed as suicide.

During the years 1967–1974, the system of data collection may not have been in full operation, thus compromising the validity of the data; therefore, these years are omitted in this report. A check on the validity of the current suicide reports is made possible by comparing the proportion of reported deaths due to undetermined external causes to deaths by suicide during the two periods (the period during which suicide was a legal offense and the current one)—the notion being that a system calling for police inquiry is less subject to extraneous pressure aimed at having suicide recorded as an external cause of death. Contrary to expectation, the ratio of death by suicide to death by undetermined external causes was three times larger in 1981–1982 (medical system) than in 1961–1962 (legal system).

International Sources

Several WHO publications provided the data for suicide and homicide rates in countries with acceptable standards of recording (WHO, 1976, 1985, 1986). A United Nations (UN) publication was consulted for divorce rates (UN, 1984).

Findings

How often do Israelis commit suicide, compared with citizens of other countries? Table 1 shows that Israeli (standardized) rates are the second lowest in a group of heterogeneous countries in Asia, North America, and Europe selected for valid recording. (The lowest rate is found in Kuwait.) These results, however, are somewhat misleading, since the rates for Israel include both Jews and non-Jews, groups of marked difference in suicidal behavior (see below). Indeed, the rates are diluted: Non-Jews have a negligible weight on the numerator, while, as noted earlier, they constituted 17.5% of the total Israeli population in 1985.

An identical finding is provided by Table 2, reproduced in full from a previous WHO publication (WHO, 1976), wherein rates have been standardized according to the following parameters: “a) stable population of a Western model; b) mortality at level 21; and c) rate of growth at 1% a year.” Israeli rates are among the lowest third of a group of European and North American countries. The reservation expressed in regard to Table 1, however, holds for this table as well.

A more appropriate basis for comparison is provided by Table 3, since in this case (Israeli) Jews only are contrasted with populations in other countries in Europe and in the United States in different years over a 30-year period. These countries, where recording is most valid, were selected from Table 2. Thus, five countries were among the upper third (ranks 21–30 in Table 2) in their respective frequency of suicide; one fell within the intermediate level (ranks 11–20); and four, including Israel, were among the lower third (ranks 1–10). The suicide rates for Israel are low, ranking just above those for Italy.

The rates for all of these countries have not remained stable over the years (Table 3). The ratios of rates for the year 1982 (for three countries, 1980; for Israel, the mean for 1982–1984) to those for the year 1952 (for Israel, the mean for 1950–1952; for Hungary, the mean for 1961–1963) show that there have been some changes. For some countries in the lower third—Norway and Italy—an upward trend is recorded. In contrast, for Israel and for England and Wales—also in the lower third—there is a noticeable reduction. For the other countries in the table, with the exception of Hungary, rates have remained stable.

One additional comparison is of high interest. The above-mentioned New York study (Gross & Reed, 1971) included Jews who surely were of Ashkenazi (European) origin, though this was never definitely stated. The rates for their Israeli counterparts for the year 1964 are remarkably close—15.6 per 100,000 for the population aged 25 years and over.

Discussion

What could account for the relatively low magnitude of suicide rates among Israeli Jews? Methodological issues related to diagnosis and reporting (see above) may not be at stake. Rather, ethnic, sociocultural, political, and psychiatric morbidity factors are touched upon here in a partial attempt to provide a possible explanation.

The Jewish population of Israel consists of two main ethnic groupings, the Occidentals (Ashkenazim), of European ancestry (47% of the Jewish population aged 15 years and over by 1984), and the Orientals, largely from Muslim countries (46% of the population of the same age) (Friedlander & Goldscheider, 1979). The difference, 7%, is accounted for by the native Israelis whose fathers were also born in Israel, the larger proportion of these being of Asknenazi ancestry. Israelis of Oriental origin commit suicide in a lower proportion than do the Ashkenazim (Levav & Aisenberg, 1987), thus diluting the rates via the same mechanism noted above regarding the non-Jews: They are fully included in the denominator, but appear in the numerator in a reduced proportion. Indeed, if the Ashkenazim alone were considered, the rates, though still low, would climb from 9.6 per 100,000 for the total Jewish population aged 15 years and over to 14.2 (see Table 3).

It is likely that religion—or, rather, the degree of religiosity—plays a protective role (Durkheim, 1897/1966). As noted, the Jewish faith condemns suicide, since it is up to the Creator to give and take away life. A considerable segment of the Jewish Israeli population (58%) is moderately to markedly observant of religion (Ben-Meir & Kedem, 1979). Thus, the prohibition of self-destruction is in effect for many. Moreover, the gregarious and all-encompassing character of Jewish religious practices may very well reinforce the social fabric, thus promoting the integration of the individual into the group. Note as well that Israeli rates are closer to those of Catholic countries (e.g., Italy) than to those of Protestant countries, much as Durkheim found in Germany.

If religion acts as noted above, divorce and homicide rates are indicators of the opposite trend. Divorce rates in Israel for the years 1980–1984 have remained stable at 1.3%; this figure is generally lower than those in the countries selected for this comparison (see above), except for Italy (see Table 3). With respect to homicide (Table 3), the situation is analogous. For Israeli Jews, the homicide rate was 1.6 per 100,000 persons in 1982–1984, falling in the middle range of rates for the 10 selected countries.

As for the contribution of politics to the suicide rate, the continuous risk to the security of the country posed by war and terrorist activity has generated a national feeling of common destiny that certainly buttresses the individual's sense of belonging (Stone, 1982). In general, personal commitment to the country is high (Stone, 1982). A clear relationship between war and decrease in suicide rates is shown by Handelsman and Cochavi (1976), who compared the proportion of suicides during the months of active fighting in the Yom Kippur War of 1973 with those that immediately followed the cessation of hostilities.

Finally, two psychiatric disorders that are highly correlated with suicide among males—drug and alcohol abuse—are noticeably lower in Israel than in the other countries that appear in Table 3. Some indications of the low rate of substance abuse are provided by a psychiatric community survey being conducted by Dohrenwend, Levav, Shrout, Link, Skodol, and Martin (in press), which has yielded a rate of 2.2% (males) for the Tel Aviv metropolitan area. This figure contrasts sharply with the rate obtained in the recently completed Epidemiologic Catchment Area study in the United States: 25.6%, almost 12 times higher (Regier, Myers, Kramer, Robins, Blazer, Hough, Eaton, & Locke, 1984).

As for the comparison between the Ashkenazi Jews in New York and Israel, notwithstanding methodological reservations, we are tempted to speculate that the closeness in the rates is akin to those findings showing that the ranking of suicide rates among immigrants in the host country is similar to the ranking of suicide among their respective countries of origin (Barraclough, 1971; Sainsbury, 1986). It would thus seem that the vulnerability to suicide (biological/cultural) is carried with the individual and becomes expressed despite new environmental conditions, at least for the first and second generations.

Findings

Completed suicides by Jews and non-Jewish national-religious groups (Muslim Arabs, Christians [largely Arabs], and Druzes) were compared over 10 years (1976–1985). The data were grouped together and averaged on a yearly basis, since the numbers of suicides among the latter groups are much too small to generate stable rates.

Table 4 shows that Israeli Jews, men and women alike, commit suicide more often than do the other groups. The rate ratio for Jews to Muslim Arabs (the largest majority; see above) is 4.2 for men, 2.4 for women, and 3.3 for both sexes combined. The ratios are based on crude rates for the population aged 15 years and over. The denominator is the midpoint of the 10-year period. The rate for the male Druzes is unexpectedly high, but it is based on small numbers (15 suicides over 10 years); thus this statistic should be interpreted cautiously. The clear differential risk for suicide is somewhat reduced once possible artifacts are considered. Indeed, it is conceivable that a certain proportion of self-inflicted deaths remain unrecorded. Thus, whereas the ratio between deaths by suicide and deaths by undetermined external causes among Jews was 2.9 (average over the 10-year period), the respective ratio among non-Jews was 0.8 for the same period. In other words, among non-Jews, as many externally caused deaths (defined as undetermined) as suicides were recorded. Hence, a more conservative comparison is achieved by grouping together both types of deaths. Using such a procedure reduces the difference between the risk for all Jews and that for Muslim Arabs by almost half: It drops from 3.3 for suicide alone to 1.9 for both causes (see Table 4).

The rate ratios for Jews to Muslim Arabs, though somewhat reduced, remains evident when age-standardized rates are computed (see Table 5). (Age standardization is needed in view of the groups' different age structure.) Among the Arab minorities, Christians have slightly higher rates than do Muslims, for both men and women, but the numbers are too small for valid comparisons.

The comparison above could be refined further by contrasting the rates for non-Jews, taken together, with those of Israeli Jews of Ashkenazi and Oriental origins. Table 4 also shows two distinct patterns: There is a marked contrast (a rate ratio of 3.4) between rates of Jews of European origin (15.8 per 100,000 population) and those of non-Jews (4.6 per 100,000 population), while only a tenuous difference (a rate ratio of 1.3) is found between Oriental Jews (of Asian and African origin; rate: 6.0 per 100,000) and those of non-Jews. The rate ratios for Jews to Muslim Arabs have been shown to “favor” the latter group. But are patterns of additional interest hidden in the overall comparison?

First, Table 6 shows the ratios for each gender and for age-specific groups. These were calculated based on a yearly average during the 10-year period 1976–1985. Perusal of this table yields two patterns: The ratios are of lower magnitude for the younger age groups (15–29), and are consistently lower for women than for men.

Second, Table 7 records the rates for the two national-religious groups at three points in time: 1962–1966, 1968–1971, and 1976–1984. (It should be noted that these are crude rates for the population aged 15 years and over.) While the rates for the Jewish males have remained stable, there has been a drop among Jewish and Arab females, and especially among Arab males.

Discussion

What follows is a partial attempt to explain the differences noted between Jews and non-Jews. At all times, however, it should be borne in mind that the current investigation is based on aggregate data “posing the methodological problems involving the ecological fallacy” (Stack, 1983). Further analysis is needed, though Gove and Hughes (1980) hold a different view since, using other techniques, they were able to confirm ecological associations in suicide research. Additional methodological issues related to problems in death ascertainment may affect the comparison. It seems likely that extraneous pressure, openly or covertly exerted, may cause death by suicide to be recorded as undetermined. Note that suicide is highly stigmatizing for the surviving relatives in a small Arab community. Yet the differences between the two main groups (Jews and non-Jews) still remain when both types of death are combined, although they are somewhat lessened.

Islam regards suicide as homicide, according to one Arab Israeli religious authority (Al-Najjar, 1978). Dublin comments that Islam “has always condemned suicide with the utmost severity in every conceivable circumstance” (1963). In contrast, the relatively more tolerant view of Judaism has been noted earlier. (Curiously enough, crude homicide rates among Arabs for the population 15 years and over are 8.5 for men and 2.2 for women per 100,000; among Jews, the respective rates are 2.3 and 1.4. See CBS, 1986.)

These two groups differ in several ways on selected sociodemogrpahic factors that have been found to be correlated with suicide risk. Thus, compared with Jews, (1) divorce rates are lower among non-Jews; (2) there is a lower proportion of single males; (3) there are four times fewer persons living in single households among the population aged 15 years and over; and (4) there is a higher average number of people living in a household among persons aged 65 years and over (see Table 8). In brief, among non-Jews, there is a clear pattern of living in groups rather than in isolation. Gove and Hughes (1980) have shown isolation to be highly correlated with the risk of suicide and alcoholism.

It is to be noted also that the two groups differ in their respective level of urbanization, Jews being more urbanized than non-Jews. Thus, while 46% of non-Jews live in communities with fewer than 10,000 residents, the proportion among Jews is considerably less—13% (CBS, 1986). (Recall here that suicide risk is positively correlated with degree of urbanization; Sainsbury, 1986.)

Educational level, which is also directly correlated with suicide risk (Sainsbury, 1986) differs in the two groups: The median number of years of education is 8.6 among non-Jews, whereas it is consistently higher in all Jewish subgroups (see Table 6).

Of major interest is the respective contrast of suicidal behavior between the Europe-born Jews on the one hand, and the North Africa- and Asia-born Jews on the other, with the Arab minority. The differential risk is considerably higher for the former subpopulation of Jews than for the latter. There is a certain degree of similarity between the Oriental Jews and the non-Jews on selected sociodemographic variables, such as educational level and household density (see Table 8). Additional factors could be imputed for the difference as well, such as secularization; religiosity is higher among the Oriental Jews and the Arabs than among Europe-born Jews.

But, of course, cultural factors may account for the difference in suicidal behavior. Europe-born Jews, who partook of the European culture for centuries, may have brought along to Israel the potential proneness to suicide (see Tables 1–3 for the relatively higher rates of suicide in Europe). In turn, Oriental Jews may have assimilated behaviors of the Islamic society—namely, the lower risk of suicide in Muslim countries (see the rates for Israeli Arabs in Table 4 and Kuwait in Table 1; see also a study conducted in Afghanistan— Gobar, 1970).

Cross-cultural behavioral scientists have reported that guilt (a feeling positively associated with suicidal behavior) is more common in the European than in the Oriental cultures, and accordingly is more often found in depression in the West than in the East (Marsella, 1980). Indeed, Racy (1970) and Pfeiffer (1968) pointed out that guilt and feelings of self-depreciation are naturally absent among Arabs. Thus, despite common heritage, Jewish subgroups may differ as a result of their prolonged past interaction with their respective host cultures.

Changes in the rate ratios by age may be more apparent than real—a reflection, in part, of the different ethnic composition by age of the Jewish subpopulation. There is a higher proportion of Oriental Jews (both foreign-born and Israeli-born of Oriental ancestry) among the younger than among the older groups, and they evidence a lower suicide risk than do the Ashkenazim. Yet, in part, this phenomenon could also reflect social changes in the Arab minority, especially the young—access to higher education, a less traditional orientation, and the like. The different rate ratios by gender, especially among the young, may be a sensitive reflection of that process: The Israeli Arab women's strong allegiance to their all-encompassing milieu has been challenged by their exposure to the modern Israeli state.

Comparison of rates over time during the three periods (1962–1966, 1968–1971, and 1976–1984) recorded in Table 7 is problematic because of the different sources used for data collection (see above). Such a comparison, however, does provide some additional information that reinforces the trends found among the Arab women. Note that the drop in their rate, although slightly higher than among the Jewish women, is considerably lower than that for the Arab men. Overall, however, it is interesting that the marked changes that have occurred in Israel since the Six-Day War of 1967, with the ensuing administrative incorporation of the West Bank and Gaza, do not seem to have affected the suicide rate of either the Jews or the Arab minority.