Understanding the impact of conflict on health services in Iraq: information from 401 Iraqi refugee doctors in Jordan


G. Burnham, Johns Hopkins Bloomberg School of Public Health, 615 N Wolfe Street, E8132. Baltimore, MD 21205, USA. E-mail: gburnham@jhsph.edu



This study aims to assess how conflict has affected the function of Iraqi health services and its doctors.


Interviews were conducted in person or by mobile phone with 401 Iraqi doctors entering Jordan since 2003, using respondent-driven sampling methods.


Of the Iraqi doctors interviewed in 2008, 94% came from Baghdad, although 25% had moved within Iraq in the past year. They reported a steady year-by-year decline in Iraqi health services from 2003 through 2006, with perhaps some improvement in 2007. By 2006, 67% of doctors said essential drugs were present less than half of the time (95% confidence interval [CI] 54–81), and 69% (95% CI 56–84) said essential equipment was available or working half the time or less. By 2006, 95% said their facilities lacked skilled health workers, and 90% noted reduced quality of care. Violent death rates among doctors in Baghdad reached 47.6/1000/yr (95% CI 42.0–53.7) in 2006. In the same year, migration rates for Baghdad doctors moving elsewhere in Iraq were 143.8/1000/yr (95% CI 134.0–154.1), and departure from Iraq was 299.5/1000/yr (95% CI 285.3–314.3).


Deterioration of health services quality, staffing levels and violence against doctors continued from 2003 through 2006, although these may have improved slightly in 2007. In 2009 and 2010, reports suggest that assassinations of doctors and out-migration have continued. Few have returned. Copyright © 2011 John Wiley & Sons, Ltd.


Armed conflict causes major damage to health services in many ways. Both “push” and “pull” factors affect migration of doctors during conflict. Doctors and other health workers flee conflict because they or their families are direct targets of violence, or they perceive risk from generalized violence. Others may leave to seek a more stable environment in which to live and practice (Kronfol et al., 1992). Often, departure is for a combination of these reasons. In conflict situations, routine support and maintenance of health facilities may collapse, with shortages of drugs and repair of medical equipment (Black, 1993). The emphasis of care may shift away from routine services such as maternal and child health services to care for violent injuries (Ryan, 2004). This is especially so in urban warfare situations where high-intensity weapons are likely to produce a different pattern of injuries and death than has been reported from conflicts occurring in rural subsistence societies (Spiegel and Salama, 2000; Salama et al., 2004). The function of health information and surveillance systems is often an early casualty. Ongoing violence frequently prevents both patients and healthcare workers from accessing healthcare facilities (ICRC, 2008).

The 2003 US-led invasion occurred when the Iraq health system was already weakened from 23 years of dictatorship, the 1980–1988 Iran–Iraq war, the 1990–1991 Gulf War and 12 years of embargos and sanctions (Wilson, 2004). The well-developed hospital-oriented health system existing before Saddam Hussein had badly deteriorated by 1997, with a substantial decrease in services it could provide (Garfield, 1999; Amin and Khoshnaw, 2003; Forest, 2005). Following liberalization of restrictions on drugs and supplies by the United Nations Security Council, there was some improvement in services and a slow reorientation toward primary healthcare services. As the conflict intensified and stretched on from 2003, medical doctors began migrating within Iraq as well as leaving Iraq for neighboring countries. Although few official figures are available, sources have reported large numbers of doctors killed and a substantial migration of doctors (Bristol, 2006; Zarocostas, 2007). Recent efforts have been made by the Iraqi government to stem the loss of doctors and to provide encouragement to return for those who have left (Kami, 2010).

The damage from the widespread looting and destruction of health facilities in 2003 took a heavy toll on health facilities (Kapp, 2003). The US-led response was paralyzed by a mixture of inter-agency conflicts and political agenda and failed to effectively shore up health services in an effective way (Burkle and Noji, 2004). Since then, health services have been perceived as continuing to deteriorate in physical condition and to be short of medicines (Medact, 2008). Both the ongoing violence and the construction of barriers limiting travel in violence-affected areas have restricted patient access, especially in Baghdad. During the worst violence, doctors minimized their time in health facilities, remaining in the relative safety of their homes for much of the time (Salman, 2006). Violence has prevented patients from receiving critical drugs such as insulin (Lancet, 2007). At the same time, some people, unable to reach health facilities, have been utilizing informal private neighborhood clinics staffed by paraprofessionals in order to receive needed care. With the migration of experienced doctors from Iraq, there have been concerns about the quality of health services and the ability of training facilities to replace those migrating, especially those with advanced specialty training (Medact, 2007). Reports indicate that many Iraqi doctors have migrated to the Gulf States, Jordan and elsewhere in the Middle East (Burnham et al., 2009).

To document and understand the migration of doctors and their perceptions of the status of the Iraqi health system, we conducted a survey of 401 Iraqi doctors residing in Jordan during 2007 and 2008. We assessed the factors that contributed to and facilitated departure from Iraq and assessed the perceptions of the doctors about the health services in which they had been working.


Respondent-driven sampling (RDS) was used to locate potential respondents for interviews because Iraqi doctors in Jordan are an uncharacterized population. RDS is designed to overcome some limitations of snowball sampling and is a preferred method for populations where probability-based samples are not feasible (Broadhead et al., 1998; Broadhead et al., 2002). RDS starts with a small number of “seeds” and expands through successive waves of peer recruitment. Seeds recruit the first-wave respondents, and first-wave respondents recruit the second-wave respondents; the process continues until the desired sample size is attained. RDS incorporates direct recruitment of peers by peers and employs recruitment quotas to facilitate a sample that is less dependent on social network size and more representative of the target population as a whole. Respondents in RDS have been found to recruit as if they are sampling randomly from their personal social networks (Wang et al., 2005).

Seeds were selected based on several individual characteristics so that different types of Iraqi doctors in Jordan would be included in the sample. Three seeds were recruited, which include two men and one woman, of which two were employed and one unemployed. One of the seeds was a recently qualified doctor, and two were more senior doctor. Data from the seeds were not included in the analysis. Three men and two women interviewers were used; the interviewers were Iraqi physicians or Iraqi medical students living in Jordan. Interviewers were involved in the development and piloting of the questionnaire and received an additional 2 days of training before data collection. Questionnaires were developed in English, and data collectors were fluent in English. Information was collected from respondents about indicators such as changes in service provision, service utilization, availability of drugs and supplies, utilities, unfilled medical doctor positions and adverse events that occurred to doctors at public facilities where they were employed for each year from 2003 until the respondent left Iraq. To estimate rates, doctors were asked how many doctors were on the medical staff at their facility in Iraq for each year between 2003 and 2007. No unique identifiers were recorded, and incentives were not provided for participation. No religious or sectarian data were collected. Interviews lasted 30–45 min. The final sample consisted of three referral chains each with between 10 and 12 recruitment waves.

Data were entered in Excel, and analysis was performed using spss (SPSS Inc., Chicago, IL) and stata (Stata Corp, College Station, TX). For binary logistic regression, a forward conditional model was used with kidnapping or assassination attempt as the dependent variable; these categories were combined because they were the most serious types of violence experienced and have the potential for a fatal outcome. The study was approved by the Jordan University of Science and Technology and the Johns Hopkins Bloomberg School of Public Health Institutional Review Board. Permission for the survey was granted by the Jordan Ministry of Health.


Demographic information

A total of 401 Iraqi doctors migrating from Iraq to Jordan after the 2003 invasion were interviewed. Of these, 334 were working full-time or part-time in the public sector at the time of their departure from Iraq. Demographic information presented comes from the 401 doctors, but information presented concerning health services in the public sector is restricted to the 334 who had worked in public sector facilities during the study recall period. Data were collected only for the years in which the individual was employed at a public sector facility. This decreased the number of respondents over the course of the recall period, as the respondents did not provide information on health facilities for periods after they migrated out of Iraq. Among the 401 doctors interviewed, 94% resided in Amman, and outside Amman, the majority were in Irbid. Seventy per cent of those interviewed were men, and 25% were women. Ages ranged from 22 to 77 years, with a mean year of 36 (standard deviation = 13). Of the doctors interviewed, 66% were specialists. The mean age for specialists was 40 years and 29 years for non-specialized doctors. Specialists had been in practice for an average of 16.1 years, and doctors without specialty training for a mean of 4.5 years. There were no significant differences in ages or in years of specialty training between men and women. Of the respondents, 25% reported moving within Iraq in the year before coming to Jordan, and 94% came to Jordan from Baghdad. The year of departure to Jordan for the respondents was 2003—4%; 2004—16%; 2005—25%; 2006—43%; and 2007—13%. Most doctors (83%, n = 334) had worked in public sector health facilities as their primary job. Results presented are limited to public sector facilities where the respondents had worked in Iraq prior to their emigration to Jordan. Information was not collected about events that occurred in Iraq while the respondents were living in Jordan.

Health system indicators

Information was sought for each year between 2003 and the date of departure to examine change in health system capacity and performance over the course of the conflict. Availability of essential drugs, functional equipment and basic utilities that were considered to be necessary inputs to sustain a functioning health service were included (Figure 1). In general, the performance of these measures declined from 2003 to 2006. Some improvements were observed in 2007, although the number of doctors reporting 2007 events was small. Of public sector facilities in 2006, 67% (95% CI 54–81) of the doctors reported that essential drugs were available less than half of the time, and 69% (95% CI 56–84) said essential equipment was available or working half of the time or less. Utilities, such as water and electricity, were available more than half of the time in 73% (95% CI 60–89) of public facilities. Of these, 49% (95% CI 38–62) of the reported utilities were almost always available.

Figure 1.

Health systems measures for public facilities in Iraq, 2003–2007

Provision of services

The provision of services, patient utilization and quality indicators were reported for the facilities where the respondents were employed prior to immigrating and are summarized in Table 1. Because of the small number from non-Baghdad facilities, results are presented for both Baghdad and locations outside Baghdad. Performance of health services declined between 2003 and 2006. Decreased utilization of public sector health facilities was reported by 64% of the respondents in 2003 and by 94% in 2006. For 2003, 50% of the doctors said there was already a shortage of skilled health personnel in their facility, and by 2005, 95% of the doctors reported their facilities as having a decreased availability of skilled personnel. The perceived quality of patient care followed a similar but delayed trend. Decreases in the quality of care were reported by the doctors at 33% of facilities in 2003, rising to 80–90% of facilities between 2004 and 2007. The respondents were asked about patient deaths occurring in their facilities, which they felt might not have occurred previously in Iraqi health services. For 2003, an increase in such preventable deaths was reported by 65% of the respondents, and for 2005 and 2006, 83% reported an increase in deaths they perceived as preventable.

Table 1. Percentage of doctors reporting changes in utilization and quality of health services, 2003–2007
  1. CI, confidence interval.

Decrease in utilization of servicesN20722120915218
(95% CI)(54–76)(72–96)(76–100)(79–100)(47–100)
Decrease in availability of skilled personnelN20622221115318
(95% CI)(41–61)(65–88)(78–100)(80–100)(51–100)
Decrease in quality of patient careN20321720515017
(95% CI)(26–42)(69–93)(73–99)(75–100)(41–100)
Increase in preventable patient deathsN18920019014016
(95% CI)(54–78)(66–91)(70–97)(68–99)(34–100)

Vacant positions

The Iraqi doctors were asked about unfilled positions for doctors at their public sector facility from 2003 to their departure. From 2003 to 2006, 92–95% of the responses reflected staffing in Baghdad health facilities. Because of the small proportion of non-Baghdad reports and the similarity in trends, combined data are presented that include sites outside of Baghdad. The proportions of unfilled positions for medical doctors at public sector facilities are summarized in Figure 2. Counts depicted in the figure exclude “unknown” responses that were reported by 5–10% of the respondents each year.1 Of the 21 medical doctors who left Iraq in 2007, 19 reported on health facility staffing levels; these are not included in Figure 2 because they are likely less representative than the previous years with more respondents but were as follows: 75–100% of positions vacant, 33% (n = 7); 50–74% of positions vacant, 24% (n = 5); 25–49% of positions vacant 14% (n = 3); and 0–24% of positions vacant 19% (n = 4). A rising trend in unfilled positions was observed over the four years. For 2003, 84% (95% CI 77–89) of the respondents reported that less than one quarter of positions for doctors were unfilled with 5% (95% CI 3–10) reporting 50% or more unfilled positions. By 2006, only 9% (95% CI 5–15) of the respondents reported that less than 25% of positions were unfilled, with 70% of the respondents (95% CI 61–77) reporting more than 50% of medical doctor positions unfilled and 18% (95% CI 12–25) reporting that 75% or more of positions were unfilled in their facility.

Figure 2.

Unfilled doctor positions at public sector facilities in Iraq, 2003–2006

Violent events

The respondents were asked about violent events that had occurred to doctors in their Iraqi facilities from 2003 until leaving Iraq (Table 2). In Baghdad, the violent death rate among doctors for 2003 was 10.6/1000/yr (95% CI 8.8–12.8). This increased annually, peaking at 47.6/1000/yr (95% CI 42.0–53.7) in 2006. Violent death rates for doctors outside Baghdad were greatest in the first year of the conflict, 2003, at 25.3/1000/yr (95% CI 16.7–36.9), more than double the Baghdad rate for 2003 but then remained flat thereafter. Kidnapping and injury rates for Baghdad doctors were similar with the highest rates reported in 2006. Kidnapping rates outside Baghdad exceeded those in Baghdad for 2005 and were similar to Baghdad rates in 2006.

Table 2. Violent events and migration reported by Iraqi medical doctors in Jordan
  1. CI, confidence interval.

Number of respondents158167167119116
Total number of physicians reported10 815949010 5605549665
Average physicians per respondent68.556.863.246.641.6
Violent events
Violent deaths (N)11513628326416
Violent death rate/1000/yr (95% CI)10.6 (8.8–12.8)14.3 (12.0–17.0)26.8 (23.8–30.1)47.6 (42.0–53.7)24.1 (13.8–39.1)
Kidnappings (N)4319326516413
Kidnappings/1000/yr (95% CI)4.0 (2.9–5.4)20.3 (17.6–23.4)25.1 (22.2–28.3)29.6 (25.2–34.4)19.5 (10.4–33.4)
Injuries (N)682233462003
Injury rate/1000/yr (95% CI)6.3 (4.9–8.0)23.5 (20.5–26.8)32.8 (29.4–36.4)36.0 (31.2–41.4)4.5 (0.9–13.2)
Moved within Iraq (N)20432965979841
Internal migrants/1000/yr (95% CI)18.9 (16.4–21.6)34.7 (31.0–38.6)62.4 (57.7–67.4)143.8 (134.0–154.1)61.7 (44.2–83.6)
Moved to another country (N)2876101,2561,66253
Emigrants/1000/yr (95% CI)26.5 (23.6–29.8)64.3 (59.3–69.6)118.9 (112.5–125.7)299.5 (285.3–314.3)79.7 (59.7–104.2)
Locations outside Baghdad
Number of respondents15141184
Total number of physicians reported1066766591451170
Average physicians per respondent71.154.753.756.442.5
Violent events
Violent deaths (N)279982
Violent death rate/1000/yr (95% CI)25.3 (16.7–36.9)11.7 (5.4–22.3)15.2 (7.0–28.9)17.7 (7.7–35.0)11.8 (1.4–42.5)
Kidnappings (N)121145115
Kidnappings/1000/yr (95% CI)11.3 (5.8–19.7)14.4 (7.2–25.7)76.1 (55.5–101.9)24.4 (12.2–43.6)29.4 (1.0–68.1)
Injuries (N)13171395
Injury rate/1000/yr (95% CI)12.2 (6.5–20.1)22.2 (12.9–35.5)22.0 (11.7–37.6)20.0 (9.1–37.9)29.4 (1.0–68.1)
Moved within Iraq (N)4251934312
Internal migrants/1000/yr (95% CI)39.4 (28.4–53.3)66.6 (49.6–87.5)157.3 (127.0–192.8)95.3 (69.0–128.4)70.6 (36.5–123.3)
Moved to another country (N)8388255914
Emigrants/1000/yr (95% CI)77.9 (62.0–96.5)114.9 (92.1–141.5)42.3 (27.4–62.4)130.8 (99.6–168.7)82.4 (45.0–138.2)


Migration, both within Iraq and to other countries, was higher in governorates outside of Baghdad in the first two years of conflict. Cross-border migration peaked in 2006 for both Baghdad and outside Baghdad. In non-Baghdad locations, physician migration in 2003 was estimated at 117.3/1000/yr (95% CI 97.6–139.7), nearly doubling to 226.2/1000/yr (95% CI 184.4–274.5) by 2006. In Baghdad, the doctor migration rate during 2003 was 45.4/1000/yr (95% CI 41.5–49.6); by 2006, departures from Baghdad had increased to 425.3/1000/yr (95% CI 41.5–49.6) or nearly half of all doctors in the respondent's health facility. At its peak in 2006, the cross-border rate of emigration of doctors from all sites in Iraq was 286.8/1000/yr (95% CI 273.4–300.7).


Differences in reported frequencies of violent events, changes in quality of services, utilization of services, preventable deaths and migration within Iraq and outside Iraq were examined by year of departure from Iraq. There were significant differences in recall of excess preventable deaths (p = 0.014), migration within Iraq (p = 0.004) and outside of Iraq (p = 0.004) when the accounts of doctors departing in 2003 were compared with those departing in later years. For violent events and changes in quality of services, there were no significant differences in reports for 2003 among those who left in 2003 compared with those who left in other years.


This survey of Iraqi doctors reveals a picture of major loss of human capital as a consequence of continuing conflict within Iraq following the 2003 invasion. It also suggests that extensive damage to the health services has been a result of sustained conflict. This is against a background of sanctions and two previous wars and a declining public expenditure on health (WHO/EMRO, 2004). Data from 2007 doctor migrants suggested that some early recovery signs of health services may have been present in that year.

The 401 Iraqi doctors in Jordan in this study were relatively young, with an average age of 36 years. This compares with a mean age of 48 years for Iraqi specialists departing from tertiary hospitals in Baghdad (Burnham et al., 2009). The peak flow of doctors into Jordan was in 2006, a pattern similar to other Iraqi refugees coming to Jordan (FAFO, 2007). Of the doctors in Jordan, 66% reported having completed specialty training. Although some Iraqi doctors in Jordan could have temporarily absented themselves from their practice in Iraq, less than 6% of those we interviewed envisioned returning to Iraq permanently in the next year. This suggests that the Iraqi health system is losing both its primary care doctors and its recently trained or early career specialists. The loss of this middle-level cohort of medical doctors can have long-term consequences for the health system.

Accounts of health system functionality provided by Iraqi doctors suggest that the Iraqi health system has deteriorated from 2003, although there were a small number of doctors who had left Iraq in 2007 to report circumstances for that year. The availability of functioning medical equipment has been deteriorating in each year. The same pattern appears for essential drugs. The difficulties in supplying chronic disease drugs at the beginning of the occupation seem not to have been resolved (Medact, 2008).

The absence of dependable utilities has been frequently noted in media accounts (Brookings Institution, 2007). However, most doctors reported that basic utilities for their health facilities were often or always available.

Hospital occupancy was consistently reported around 80% in spite of loss of medical staff and difficulties patients had with access to facilities. During 2003, nearly two-thirds of doctors felt that utilization of outpatient services had decreased, and by 2006, this number was 94%. This decrease might seem inconsistent with the continued high bed-occupancy levels. However, the proportion of hospital admissions due to violence increased from 2003 to peak in 2006. By 2004, most doctors perceived that over half of hospital admissions were related to violence, with many doctors believing that it was more than three-quarters. It is possible that trauma admissions were displacing more routine or elective hospital admissions.

By 2004, 76% of the doctors reported a decrease in the numbers of skilled health workers available, and by the next year, this was 90%. Along with this shortage of other health workers, the doctors were asked about unfilled positions for doctors in the health facilities where they were working in various years. In 2003, 80% of the doctors reported that in their facility, less than a quarter of positions for doctors were unfilled. By 2006, half of the doctors reported that 50% or greater of positions for doctors in their facility were unfilled. Although there are many news reports of doctors leaving their posts because of violence, we believe this current report is only the second such attempt to measure this in a systematic manner (UN OCHA, 2007; Kaplow, 2008; Burnham et al., 2009). The loss of nurses and pharmacists from the Iraqi health system has been noted (Mason, 2007; UN OCHA, 2007; Medact, 2008). Half of the doctors surveyed in Jordan felt that in 2003, there was already a drop in the availability of skilled health workers in the health system compared with the period immediately before the invasion. This concern has been voiced by Iraqi doctors elsewhere (Voelker, 2004).

For complex medical or surgical procedures, there was a consistent pattern in reports that less-experienced doctors were taking responsibilities that formerly would have been assumed by senior doctors. This finding is consistent with the departure of specialists documented elsewhere (Burnham et al., 2009). At the same time, there was a perception by 65% of the doctors that preventable deaths had been occurring among patients of the health services in 2003. By 2005, 83% of the doctors expressed this view. This is consistent with concerns recorded elsewhere about the medical knowledge and clinical competence of Iraqi doctors (Webster, 2009). There have been reports that a large percentage of persons admitted with violent injuries subsequently die, a reversal of the pattern often reported in conflict situations where non-injuries considerably outnumber fatalities (Aziz, 2003; Medact, 2007). Not only the declining quality of hospital care and the shortage of equipment and drugs may be part of the explanation but also the nature of the Iraqi conflict and the type of weapons.

The study revealed the extent to which violent events in Iraq had affected doctors. These violent events, such as murder, kidnapping, and injuries, affected doctors in all parts of the country. In 2003, the violent events were most common outside of Baghdad, consistent with earlier data on mortality (Roberts et al., 2004; Burnham et al., 2006). By 2004, the rates reversed, with doctors in Baghdad experiencing more violence than those outside of Baghdad, with the exception of kidnappings occurring in 2005. The escalation in these violent events affecting doctors peaked in 2006, which coincides with the peak number of Iraqis fleeing into Jordan (Doocy and Sirois, 2008). Other reports noted 2006 as the most violent year in Baghdad post invasion (UN OCHA, 2007).

In the first two years following invasion, flight abroad by doctors from outside Baghdad exceeded the number of those fleeing Baghdad itself. Internal displacement also was greater for doctors residing outside Baghdad for the first three years of the war. This is consistent with violence patterns reported in the Iraq mortality survey by Roberts et al. (2004). The overall migration trend reported by the Iraqi doctors runs parallel to that among specialists at tertiary hospitals in Baghdad, which reached 29% in 2006 (Burnham et al., 2009). However, that study found a lower level of overall migration among tertiary-hospital specialists, and the migration of those based outside of Baghdad was considerably less than that reported by the Iraqi doctors interviewed in Jordan.

A study of this nature has many potential limitations. Limitations inherent to all sampling methods apply to RDS; additionally, there are several limitations that are unique to RDS in this context that are worth noting. Network ties must be sufficient to sustain the referral process; in numerous instances in this study, one individual was referred several times. Although the population of Iraqi physicians in Jordan is relatively small and cases of repeat referrals could have been anticipated, it is also possible that individuals with broader networks may have been more likely referred into the study, which could be a potential source of bias. It is also possible that information obtained was representative of specific networks only and not the experiences of the larger population of Iraqi doctors in Jordan. However, the study's three chains with 10–12 waves of referrals in each indicate some independence of networks. The respondents may also have referred others, who shared similar opinions or were known to have experienced a violent event, for interview, whereas avoided those with differing opinions or experiences.

The data rely on recall of the doctors about the events and the circumstances happened several years previously while they were still living in Iraq. The majority of the Iraqi doctors in Jordan came from Baghdad, so this can weigh the findings toward Baghdad and not necessarily be representative of other parts of the country. The situation where conflict has been less is likely to be different. Further, it is likely that refugee medical doctors may have heightened exposure or perception of violent events than doctors who did not move or who migrated within Iraq. The analysis showed that the length of time outside Iraq had only a limited effect on recall of events. Although there is a risk that several doctors could recall the same violent event happening to a colleague, it is also likely that both would report similar numbers of doctors present at the same facility. We did not collect the names of the individual health facilities where those interviewed worked, so violent events could not be verified. This limits assessing the types of facility as a factor for different outcomes. Lastly, data reported on violent events only reflect those events that occurred while the doctor was in Iraq; in the year of migration, information on violent events after departure from Iraq was not reported. This may have resulted in death and kidnapping rates that were artificially skewed downward, especially in 2005 and in 2006 when the largest proportion of doctors departed Iraq.

Because this study was concluded in 2008, the flow of doctors from Iraq has continued, although anecdotal reports suggest that the numbers are less than in earlier years. Accounts from Iraq suggest that most of the well-established doctors (often called the “famous doctors”) who had considered flight have now left Iraq. Those who remained in Iraq have extensive family and professional ties in Iraq. These anecdotal reports indicate that assassinations of doctors continue, some for political reasons and some in revenge by families who believe a relative had died from incorrect medical treatment. Doctors who remained in Iraq complain that the health services are increasingly polarized among political factions, and unqualified personnel are making major decisions. A number of Iraqi doctors are traditionally from the Christian community, and these continue to depart with migration of Christians leaving because of religious persecution (Schemm, 2009). However, most doctors now leaving Iraq are recent graduates who are seeking training and careers outside Iraq, often in Sweden, in the UK or in the USA. They tend to be pessimistic about their own future in Iraq and discouraged about the country's health services. The Iraqi government has encouraged doctors to return, with various inducements including a law authorizing doctors to carry concealed weapons (Karadsheh, 2008; Kami, 2010; Sarhan, 2010). Although some doctors may have returned to certain areas such as Kurdistan, reports suggest that overall, few doctors are returning to Iraq. Reports suggest that returning doctors, like other refugees, are disillusioned (UNHCR, 2010).


The interviews with the Iraqi doctors migrating to Jordan suggest a serious deterioration in the quality of services in the Iraqi health system, which comes in addition to damage from war and sanctions. Of the losses sustained, the depletion of human capital is hardest to replace. In Iraq, this loss appears to have had an impact on the ability to train future medical practitioners, with less-qualified doctors reportedly more likely to undertake complex procedures than in earlier years. A substantial drop in violence could decrease continuing losses and may induce doctors still in Iraq to remain and others to return. However, present evidence does not suggest this is happening. However, based on experiences in other conflicts, it is unlikely that settled circumstances will induce large numbers of doctors to return to Iraq. The triple challenges to Iraq are to restore a heavily damaged infrastructure, rebuild administrative and support systems as well as replace lost human capital. For a health system, the last remains perhaps the greatest challenge.


We are indebted to the refugee Iraqi doctors in Jordan who generously gave of their time for these interviews. We would like to express our gratitude to Dr. Riyadh Lafta for critical review of the questionnaire and Dennis Brophy for contributions in data entry and the data quality review process. Funding was provided by the Johns Hopkins Center for Refugee and Disaster Response.

  • 1

    Because of the selectivity and small sample size for 2007, where only 19 respondents reported information on staffing, data are excluded from Figure 1.