Bulletin of the World Health Organization

Low infant mortality among Palestine refugees despite the odds

Flavia Riccardo a, Ali Khader a & Guido Sabatinelli b

a. United Nations Relief and Works Agency for Palestine Refugees in the Near East, Health Department Headquarters, Bayader Wadi Seer, PO Box 140157, Amman, Jordan.
b. World Health Organization, Geneva, Switzerland.

Correspondence to Flavia Riccardo (e-mail: flaviariccardo@hotmail.com).

(Submitted: 11 September 2010 – Revised version received: 03 February 2011 – Accepted: 06 February 2011 – Published online: 17 February 2011.)

Bulletin of the World Health Organization 2011;89:304-311. doi: 10.2471/BLT.10.082743


Sixty years have passed since about 900 000 Palestinians fled their land. This event affected all countries near Palestine and generated the oldest and largest protracted refugee population in the world.1

Throughout the decades that have transpired since that event, Palestine refugees have enjoyed different levels of access to health-care services in their host countries. However, in many places they face discrimination and are a socially and economically vulnerable population2,3 whose main primary health-care provider continues to be the United Nations Relief and Works Agency (UNRWA) for Palestine Refugees in the Near East. Today this agency serves almost 5 million beneficiaries in Jordan, Lebanon, the Syrian Arab Republic, the Gaza Strip and the West Bank, territories whose young populations have high fertility rates. UNRWA delivers comprehensive primary health care, among other services, and it focuses particularly on maternal and child health, with targeted activities from preconception4 throughout gestation and the post-delivery period.

Mortality among children within the first year of life (i.e. infant mortality) is an important indicator of the health status of a population and correlates strongly with more comprehensive measures of health, such as disability-adjusted life expectancy.5 Although infant mortality surveys are carried out periodically within countries hosting Palestine refugees, distinguishing between refugee and non-refugee populations is difficult2 and the use of different survey methods by the countries makes it hard to perform comparisons and to assess trends across countries over time.

This paper presents the results of an infant mortality survey conducted by UNRWA in 2008 in Jordan, Lebanon, the Syrian Arab Republic, the Gaza Strip and the West Bank and compares them with the results of two earlier UNRWA studies performed with identical methods.6 The trends presented in this paper in infant and early child mortality among Palestine refugees in the Near East for the 10-year period from 1995 to 2005 have never been published before.


In 2008, we conducted a cross-sectional survey to estimate infant mortality among Palestine refugees served by UNRWA. The preceding birth technique7 was adopted so that the results could be compared with those of two identical surveys conducted by UNRWA in 1997 and 2003.6 Using the infant mortality rates reported in the previous UNRWA surveys,6 we calculated the minimum sample size needed in each UNRWA field of operation to estimate infant mortality to the nearest 5 deaths per 1000 live births with a confidence level of 95%. Beginning on 15 February 2008, we interviewed women who were registering the birth of a neonate (index birth) at UNRWA primary-health-care facilities and who had given birth to at least one other child, until the required sample size was attained. The data collection period varied between countries, however, since the number of newly registered infants in a given period was different in each. Data collection lasted approximately 4.5 months in Jordan, 9 months in the Syrian Arab Republic and the West Bank and 12.5 months in Lebanon. Because data in the Gaza Strip were collected twice to validate the preliminary results, data collection there took 15 months. The compiled survey forms were transferred to the UNRWA Health Department in Amman, Jordan, where the data were entered, cleaned and analysed using EpiInfo 2000 (Centers for Disease Control and Prevention, Atlanta, United States of America).

Following the preceding birth technique, which is used to indirectly estimate infant and early child mortality (defined here as death within the first three years of life), we asked women whether the preceding child (i.e. the last one born before the index birth) was still alive or not. If they reported that the preceding child was dead, we asked them the age of the child at death and the main cause of death. We also asked the date of birth and sex of the preceding child; the date of birth, age at registration and sex of the neonate, and the mother’s age, parity, residence and level of education.

From the information provided by the mothers we calculated the mean interval between current and preceding births (mean birth interval), the proportion of preceding children who had died by the time of registration and the mean neonatal age at birth registration. We then estimated neonatal, infant and early child mortality rates. The age at death for early child mortality (age “q”) depended on the interval between the index birth and the preceding birth and the age at registration of the index child. Because this estimated age at death differed slightly between Jordan (2.5 years), Lebanon (3 years), the Syrian Arab Republic (2.7 years) and the Gaza Strip and the West Bank (2.5 years each), the main focus of the paper is on neonatal and infant mortality, although the findings for older ages are also presented. We also calculated the frequency of different causes of infant death.

We excluded from the study mothers whose preceding pregnancy had ended in stillbirth but compensated for this by applying a corrective factor of 1.09 when calculating mortality rates, as required by the preceding birth technique.8 If the preceding pregnancy had ended in abortion, we obtained information on the previous child.

The preceding birth technique makes it unnecessary to ask an infant’s exact date of death. Since birth registration coverage in UNRWA clinics is high – already in 2003, over 95% of infant births were being registered within 6 weeks of delivery – the preceding birth technique seemed appropriate for the study setting. The method is a valid alternative to prospective analysis when conducting household surveys is complicated or unfeasible, as is the case among Palestine refugees, who live mostly outside refugee camps and are difficult to target. The technique was modified in 1997 to categorize the time of a child’s death as follows: during the first month, neonatal death; between months 1 and 12, postneonatal death; and within 1 year, infant death.6


We sampled 14 202 neonates registered by UNRWA in 2008. They were distributed among fields as shown in Table 1. The sexes were equally represented and the mean neonatal age at birth registration was less than one month in all fields, although intrasample variations were wide. The mean interval between the birth of the neonates and of the preceding children is shown in Table 2.

Globally, 318 deaths were recorded and 291 of them occurred within the first year of life. Infant mortality among Palestine refugees during 2005–2006 ranged from 19.0 deaths per 1000 live births in Lebanon to 28.2 per 1000 in the Syrian Arab Republic. Early child mortality, as per age “q,” ranged from 20.8 per 1000 live births in Lebanon to 29.6 per 1000 in the Syrian Arab Republic (Table 3).

The three main causes of death in the first year of life among Palestine refugees were conditions related to low birth weight or prematurity (30.2%), congenital malformations (22.7%) and respiratory infections (13.4%). Some deaths were also caused by preventable conditions such as birth trauma and hypothermia (Fig. 1). Birth trauma deaths were relatively frequent; they accounted for almost 4% of all infant deaths and for approximately 5% of all neonatal deaths. One death due to hypothermia was documented during the survey. Almost 5% of infant deaths were attributed to the sudden infant death syndrome; 43% of these sudden deaths occurred in the neonatal period and 57% in the post-neonatal period.

Fig. 1. Infant deathsa among Palestine refugees by cause and field of operation of the United Nations Relief and Works Agency, 1995–2005
a Infant deaths recorded during this study totalled 291.

Between 59% and 74% of Palestine refugee infants died during the first month of life. Almost half (43.2%) of these neonates died from causes related to low birth weight or prematurity. Communicable diseases, primarily respiratory infections (10.4%), septicaemia/meningitis (3.1%) and gastroenteritis (0.5%), accounted for 15% of all neonatal deaths. Postneonatal infant deaths were mainly caused by congenital malformations (29%) and communicable diseases (30%), among which the three leading causes were respiratory infections (20%), gastroenteritis (6.2%) and septicaemia/meningitis (1.0%).

Although in the five UNRWA fields of operation we studied, infant mortality rates differed from the rates found in the 2003 survey, none of the differences was statistically significant.

After infant mortality dropped sharply between 1995 and 2001 in Jordan, Lebanon and the Gaza Strip (according to the previous UNRWA surveys), by 2005–2006, the timeframe to which the present survey results can be attributed, rates had stabilized at the levels required to meet MDG 4 (Fig. 2). Only the West Bank experienced a slight, non-significant increase in infant mortality with respect to 2003. If current infant mortality rates are maintained, the MDG 4 target for infant mortality will be met by the Palestine refugee communities in Lebanon, in the Gaza Strip and the West Bank. In Jordan this target seems well within reach, while the Syrian Arab Republic appears to be lagging behind.

Fig. 2. Infant mortality rates (per 1000 live births) among Palestine refugees according to United Nations Relief and Works Agency surveys for 1997, 2003 and 2008
Fig. 2. Infant mortality rates (per 1000 live births) among Palestine refugees according to United Nations Relief and Works Agency surveys for 1997, 2003 and 2008

In the decade from 1995 to 2005 early child mortality in the Palestine refugee community followed the same general trend observed for infant mortality:11 a slight increase was observed only in the West Bank. However, because the interval between births has increased progressively during this period, the “q” time frame does not overlap exactly. Thus, more in-depth comparisons cannot be drawn.


In recent years, the factors associated with a drop in mortality have been better defined through research. It is now widely acknowledged that some of the most important factors for lowering mortality are public policies targeting disease prevention and control and a viable health sector. Organized political efforts seem to have a direct effect on health, one that is largely independent of income or living standard.12 Although Palestine refugee communities face socioeconomic hardship and have high fertility rates, their infant and childhood mortality rates are among the lowest in the Arab world.6 Remarkably, however, this has been achieved despite the lack of a strong state in the Gaza Strip and the West Bank and despite limited access to public services in host countries such as Lebanon.

In the countries and reference period we have examined, infant mortality rates among Palestine refugees and non-refugee populations were very similar9 (Table 3). The sole exception is the Syrian Arab Republic, whose official infant mortality rates were lower than the rate in the refugee population. These findings support those of previous studies that recognized the impact of UNRWA’s long-standing organized efforts to decrease infant mortality.12

While there is essentially no difference between the infant mortality rates seen among refugee and non-refugee communities in Jordan, the Gaza Strip and the West Bank, in Lebanon the infant mortality rate appears to be lower among refugees than among non-refugees (19 versus 27 deaths per 1000 live births). Given that the number of deaths in these populations is small and that survey methods are different, even though the infant mortality rate among non-refugees in Lebanon exceeds the upper confidence interval found for refugees, this observation should be made with caution.

Assessing the health status of the Palestine refugee population in territories of the Near East is made difficult by the lack of appropriate comparators, even when a single, clear indicator such as infant mortality is employed. On one hand, the current survey showed that the causes of neonatal mortality among Palestine refugees are proportionally similar to those found in the most developed regions of the world.13 Non-communicable diseases are the leading causes of infant, and particularly neonatal deaths, among Palestine refugees, as they are among industrialized countries in Europe and North America. Thus, the epidemiologic profile of the Palestine refugees is not typical of the Near East.14 On the other hand, the refugee population is hard to reach and socioeconomically disadvantaged, facts that must be kept in mind to avoid misconceptions. Whereas in highly industrialized countries reductions in child mortality reflect effort and investment on the part of national health systems, Palestine refugees receive health care in a fragmented and unequal fashion across host countries, with the main common health-care provider being UNRWA. Keeping child mortality at low levels will require addressing some of its root causes in the Near East, namely, socially determined patterns of marriage and childbearing, the social and economic determinants of health and limited access to secondary- and tertiary-level hospital care. This is a daunting endeavour, particularly in those States or territories where governments are less directly involved in providing health care for Palestine refugees.

Some of the leading causes of infant deaths found in the survey, such as low birth weight or prematurity and congenital malformations, overlap somewhat and are associated with certain known determinants, including early pregnancy15 and consanguineous marriage.16,17 Most Palestinian women marry young and begin bearing children right away.2 Teenage pregnancies are high risk. Fortunately, they are declining, but in the Near East they continue to contribute to maternal morbidity and mortality.18. Moreover, consanguineous marriages are extremely common and account for one-fourth of all marriages in Lebanon and almost half of all marriages in Jordan, the Syrian Arab Republic and the Gaza Strip and the West Bank.2 But despite their importance, social patterns of marriage and childbearing are not the only factor to be considered when devising ways to avert infant deaths among Palestine refugee communities. UNRWA’s targeted preventive and curative interventions do not encompass birth delivery itself, which takes place in hospitals in the host countries. The quality of maternal and neonatal care, while beyond UNRWA’s control, is also a determinant of neonatal survival and of infant mortality rates. Nonetheless, information on the quality assessment of maternity and neonatal wards in the Near East is scanty, and even when country data are made available, they focus largely on coverage rather than quality.19 Maternity and neonatal care services have probably been most extensively documented in the Gaza Strip and the West Bank. WHO20 and Rahim et al.2 have recently raised concerns regarding the delivery of intrapartum and postpartum care in these areas.

Some may argue that the Gaza Strip and the West Bank represent a worst case scenario on account of the political instability that exists there. However, this is not borne out by perinatal mortality (stillbirths and neonatal deaths occurring in the first week of life) and maternal mortality figures, both of which are indicative of the safety of delivery.13 Perinatal and maternal mortality ratios among the population residing in the Gaza Strip and the West Bank are no worse than in UNRWA’s other fields of operations. As shown in Fig. 3 and Fig. 4, perinatal mortality is highest in Lebanon, and maternal mortality has been consistently higher in Lebanon and in the Syrian Arab Republic than in the Gaza Strip and the West Bank for the past 20 years.

Fig. 3. Perinatal mortality rate (per 1000 live births) in countries of the Near East, 1995–2007
Fig. 3. Perinatal mortality rate (per 1000 live births) in countries of the Near East, 1995–2007
Sources: Perinatal mortality estimates were compiled from World Health Organization (WHO).13,21,22 As data was not available for the occupied Palestinian territories, the data presented was compiled from: Kalter et al.23 and the WHO Regional Office for the Eastern European.24 Data for 2007 was accessed from WHO Regional Office for the Eastern European,24 Jordan Department of Statistics, USAID, UNICEF, UNFPA DOS25 and for Israeli data WHO Regional Office for Europe.26 Preference where available was given to WHO data for greater comparability. Where necessary for graphical clarification, a mean of values over a time period were provided for a country.
Fig. 4. Maternal mortality ratio (per 100 000 live births) in countries of the Near East, 1990–2009
Fig. 4. Maternal mortality ratio (per 100 000 live births) in countries of the Near East, 1990–2009
Sources: Estimates for 1990–94 were compiled from Mahaini27 and WHO/UNICEF.28 Estimates for 1995–99 from Hill et al.29 Estimates for 1995 from WHO/UNICEF/UNFPA30 and Rahim et al.31 Estimates for 2000–04 from WHO32,33 and Mahaini.27 Estimates for 2005–2009 from WHO Regional Office for the Eastern European24,34 and Hill et al.29 Preference where available was given to WHO data for greater comparability. Where necessary for graphical clarification, a mean of values over a time period were provided for a country.

Although poor access to health services seems to be the main contributor to neonatal deaths in the Gaza Strip and the West Bank,2 in Lebanon, with its highly privatized medical sector,35 the medicalization of normal pregnancies is cause for greater concern. The proportion of children delivered by Caesarean section in the country, which was 23% in 2004,36 far exceeds WHO’s recommended maximum of 15%. Lebanon also reports the highest rates of stillbirth (Fig. 5) and perinatal and maternal death in the entire study area. All of this points to the need to monitor the country’s current practices in birth attendance more closely.

Fig. 5. Stillbirth rate (per 1000 live births) in countries of the Near East, 2000–2004
Fig. 5. Stillbirth rate (per 1000 live births) in countries of the Near East, 2000–2004
Source: Stillbirth rates were compiled from reference.13

Study limitations

This study’s main limitation is the potential underestimation of child mortality due to women’s reluctance to acknowledge the death of a child. It is difficult to correct for this. However, we have presented contextual data in the form of trends over time calculated from previous surveys and have compared our estimates with those made by host countries using alternative methods, and yet we found no major discrepancies to suggest a large underestimation of mortality.

Since with the preceding birth technique infant mortality is estimated from the death rate among the children whose birth preceded that of the neonate, a woman could conceivably have lost both the preceding child and the neonate before she could register the most recent birth. This birth would then go unaccounted for in the survey. However, the low mean age at registration (between 8 and 21 days) found in the present study makes this bias unlikely. Another limitation is that the causes of death were reported by the mothers, who may have misunderstood them or not recalled them accurately. Finally, the confidence intervals reported in this paper are quite wide because there were very few infant deaths. This has been discussed in previous publications whose authors applied similar methods.12


The UNRWA and other health-care providers in the Near East have succeeded in delivering to Palestine refugees quality comprehensive primary health care with a strong focus on prevention and maternal and child health. But although strides have been made in the fight against the preventable causes of infant death, such as communicable diseases and undernutrition, the current stagnation in the decline of infant mortality among Palestine refugees in the Near East is worrisome. Comprehensive primary health care is essential in maintaining infant mortality at low levels and in preventing the genetic and pregnancy-related determinants of infant death, but it is not enough. Unless access to high-quality maternity and perinatal services is improved, infant mortality rates are unlikely to decrease substantially in the near future and the tenuous results obtained so far could be readily lost if enduring geopolitical instability jeopardizes continuous access to primary-health-care services in the region.


The 2008 survey was a joint effort of the entire UNRWA health programme. We thank field family health officers, field nursing officers, medical officers, staff nurses, as well as the administrative staff in UNRWA headquarters in Amman, Jordan, for their support during the present study.

Competing interests:

None declared.