Re: Survey says: a million dead

On Thu, 31 Jan 2008 09:51:24 -0800 (PST), Mike
<mkornecki2000@xxxxxxxxx> wrote:

On Jan 31, 12:42=A0pm, h...@xxxxxxxxxx wrote:
so please explain again how you decided you can trust the NEJM over
other sources?

Listen puppy. It's been shwon to you what was the problem with the
Lancet report. Those problems are inherent in this crapola you just
pushed.

this is not the Lancet report.


Now, It's YOUR turn. YOU take the effort to tell us what's wrong with
the New England Journal of Medicine's article.

oh, well why didn't you say so. That's an easy one

Methods

Data Sources

The IFHS is a two-stage, stratified survey of households, with an
original target sample size of 10,080 households. The originally
allocated sample sizes for the Baghdad?Karkh domain and for Anbar and
Nineveh provinces were increased by the addition of census enumeration
areas or clusters before the survey fieldwork was undertaken: 6 areas
for the Baghdad?Karkh region (a 33.3% increase), 54 areas for Anbar (a
100% increase), and 18 areas for Nineveh (a 33.3% increase). The
addition of the census enumeration areas was done to compensate for
the expected difficulties in accessing some of the selected clusters
because of security problems. The additional enumeration areas
increased the target sample size to 10,860 households located in 1086
clusters.

Because of different selection probabilities of households in the
sample, design weights were calculated on the basis of projected
population numbers according to province and stratum and were further
adjusted for nonresponses at both the cluster and household levels.
The sampling frame that was used in the southern and central provinces
was derived from the 1997 census, which had been updated for the Iraq
Living Conditions Survey 20046; the frame that was used in Kurdistan
was based on information provided by the Statistical Offices in the
region. Population estimates for Iraq for the survey period were
projected by the Central Organization for Statistics and Information
Technology. (For details on the survey design and implementation, see
the Supplementary Appendix, available with the full text of this
article at www.nejm.org.)

All deaths that occurred in the selected households from June 2001 to
the time of the survey were ascertained. The respondents were first
asked whether anyone in the household had died during this period. For
each death, data were recorded on sex, age at the time of death, the
time and place of death, whether medical attention had been sought
before the death, and the main cause of death according to the
respondent. The interviewers coded the responses with the use of a
list of 23 probable causes of death. (Additional details regarding the
IFHS report are available at www.emro.who.int/iraq/ifhs.htm.)

Rates of Death

The rates of overall and cause-specific death were calculated with the
use of information on the age and sex of household residents at the
time of the survey, as well as the age, sex, and time of death of the
deceased and the cause of death that was reported in the household
questionnaire. Exposure times to the risk of death for both living and
dead persons during the analysis period (January 2002 through June
2006) were calculated to the nearest month. The analysis period was
divided into a pre-invasion period (January 2002 through February
2003) and a post-invasion period (March 2003 through June 2006). June
was selected as the end month to allow comparison of the results with
those of a previous survey that was conducted in mid-2006.4

Assessment of Mortality Data

Of the 1086 originally selected clusters, 115 (10.6%) were not visited
because of problems with security. These clusters were located in
Anbar (61.7% of the unvisited clusters), Baghdad (26.9%), Nineveh
(10.4%), and Wasit (0.8%). Since past mortality is likely to be higher
in these clusters than in those that were visited during the IFHS, we
imputed mortality figures for the missing clusters in Anbar and
Baghdad with the use of information from the Iraq Body Count on the
distribution of deaths among provinces to estimate the ratio of rates
of death in these areas to those in other provinces with high death
rates. Data from the Iraq Body Count were used to compute ratios for
death rates in Anbar and Baghdad, as compared with the three provinces
that contributed more than 4% each to the total number of deaths
reported for the period from March 2003 through June 2006.

For instance, we compared the ratio of the rate of death in Baghdad
relative to the rate in three high-mortality provinces reported by the
Iraq Body Count (3.08) with the rate ratio reported by the IFHS for
the same provinces (1.56). To obtain the same ratio, overall mortality
in Baghdad would need to have been 1.97 times as high as that in the
three other provinces on the basis of the visited clusters only. This
corresponds to a rate of death in the missing clusters that is 4.0
times as high as that in the visited clusters; the corresponding
numbers for Anbar were 1.43 and 1.70, respectively (Table 1 of the
Supplementary Appendix). This adjustment involves some uncertainty,
since it assumes that completeness of reporting for the Iraq Body
Count is similar for Baghdad and other high-mortality provinces. Since
the Iraq Body Count did not collect information on age and sex for all
deaths, the adjustment factors were not stratified according to these
factors.

Adjustment for Reporting Bias

In general, the underreporting of deaths is likely to be common in
household surveys. The most serious concern is household dissolution
after the death of a household member. Several demographic assessments
have suggested that there has been an underreporting of deaths in the
IFHS. The application of the growth balance method,7 with the use of
the age distribution of deaths in the population obtained from the
household roster, indicates that the level of completeness in the
reporting of death was 62%. However, this estimation needs to be
interpreted with caution, since a basic assumption of the method ? a
stable population ? is violated in Iraq. Furthermore, the comparison
is not made to a rate of death derived from two successive censuses,
as is usually done, but from the age distribution of the households in
the IFHS.

Analysis of Sibling Data

The IFHS also asked female respondents between the ages of 15 and 49
years for the sex and survival status of all siblings with the use of
the sibling history module developed by the Demographic and Health
Surveys.8,9 For surviving siblings, respondents were asked for the
current age; for siblings who were reported to have died, respondents
were asked for the age at the time of death and the time period. These
data were used to make separate estimates of rates of death that were
specific to age and sex for the pre-invasion and post-invasion periods
for comparison with the death rates estimated from the questions on
deaths reported in households. The analysis of death rates among
siblings was not used in the preparation of the final estimate of
post-invasion deaths. Methods proposed by Gakidou and King10 were
adapted to adjust for two forms of bias in the estimation of death
rates from data regarding the survival of siblings. (For more
information on the adjustment, see the Supplementary Appendix.)

For comparison, rates of death were also estimated with the use of
microdata from the Iraq Body Count5 and from the 2006 study by Burnham
and colleagues.4 These data were provided to the WHO by the principal
investigators of each study. We used projected midyear population
numbers that were adjusted for net migration to convert the rates of
death into numbers of violent deaths (Table 2 of the Supplementary
Appendix).

Response Rates

Table 1 shows the results of the household questionnaires and reasons
for nonresponse, according to geographic region. Of the 9710
households that were visited, successful interviews were conducted in
9345 households, which represented a national response rate of 96.2%
percent, including 95.8% for southern and central regions. Of the
households that did not respond, 0.7% were absent for an extended
period of time, and 1.1% of households were vacant dwellings. Only
0.4% of households declined to complete the questionnaire. Analysis of
the level of nonresponse did not show significant differences
according to the rate of reported household mortality.

View this table:
[in this window]
[in a new window]
Get Slide
Table 1. Final Disposition of Household Questionnaires,
According to Region.


Statistical Analysis

Rates of death from any cause and from violent causes were calculated.
Robust confidence intervals were estimated with the use of the
jackknife procedure,11 and all results were weighted with the use of
survey normalized weights.

To estimate the most probable rate of violent deaths after the
invasion and the range of uncertainty, we performed Monte Carlo
simulations that took into account the survey sampling errors that
were estimated with the use of the jackknife procedure and uncertainty
regarding the missing cluster-adjustment factors, the level of
underreporting, and the projected population numbers. We assumed that
the level of underreporting was 35% (95% uncertainty range, 20 to 50),
and its uncertainty was normally distributed. Uncertainty in the
projected population for the post-invasion period reflected
uncertainty in the total migration of 1.49 million persons (95%
uncertainty range, 1.00 to 2.00), with an assumption of normal
distribution. Uncertainty in the missing cluster-adjustment factors
was difficult to quantify, since we assumed that the excess risk of
mortality in missing clusters in Baghdad and Anbar was normally
distributed, with standard deviations of 0.2 and 0.1, respectively.
All analyses were performed with the use of Stata statistical
software, version 9. All P values are two-sided.

Results

IFHS

At the time of the survey, 61,636 persons were living in the sampled
households, with 13.1% of the respondents living in Kurdistan. The
male:female ratio in the households was 1.02, the average household
size was 6.4 persons, and two-thirds of household members were between
the ages of 15 and 59 years. (Detailed analyses are available at
www.emro.who.int/iraq/ifhs.htm.)

Table 2 shows the distribution of 1325 reported deaths according to
major cause for the periods relative to the 2003 invasion among
children under the age of 15 years, men and women between the ages of
15 and 59 years, and persons over the age of 60 years. (Cause-specific
data are available in Table 3 of the Supplementary Appendix.) Overall,
the proportion of deaths from injuries increased from 10.5% before the
invasion to 23.2% after the invasion. The increase was most dramatic
among men between the ages of 15 and 59 years, among whom deaths from
injuries increased from 31.2% before the invasion to 63.5% after the
invasion and became the leading cause of death in this age group.

View this table:
[in this window]
[in a new window]
Get Slide
Table 2. Distribution of Main Causes of Death Reported from
January 2002 through June 2006, According to Sex and Age.


Table 3 shows the crude rates of death per 1000 person-years for the
period from January 2002 through June 2006, according to region, time
period, sex, and age, after adjustment for missing clusters in Anbar
and Baghdad. The adjustment increased the post-invasion rate of death
from any cause by nearly 5%, from 5.73 to 6.01 per 1000 person-years;
it also increased the rate of violence-related death by 36%, from 0.80
to 1.09 per 1000 person-years. (For additional details, see Table 4 of
the Supplementary Appendix.)

View this table:
[in this window]
[in a new window]
Get Slide
Table 3. Rates of Death per 1000 Person-Years before and after
the Invasion, According to Age, Sex, and Region.


Death from any cause per 1000 person-years was 5.31 (95% confidence
interval [CI], 4.89 to 5.77). The rate was higher in the southern and
central regions of Iraq than in Kurdistan and higher among men than
women. In Kurdistan, a nonsignificant increase in the rate of death
was observed. In southern and central Iraq, the adjusted rate of death
per 1000 person-years increased significantly, from 3.19 (95% CI, 2.67
to 3.82) to 6.36 (95% CI, 5.78 to 7.02); the increases were seen in
all age groups but were most prominent in men between the ages of 15
and 59 years. Mortality from nonviolent causes was significantly
higher per 1000 person-years in the post-invasion period (4.92; 95%
CI, 4.49 to 5.41) than in the pre-invasion period (3.07; 95% CI, 2.61
to 3.63) (Table 4 of the Supplementary Appendix).

The independent estimate of mortality of persons between the ages of
15 and 59 years that was derived from data on sibling survival showed
reasonable consistency with the above-mentioned estimates (Table 5 of
the Supplementary Appendix). On the basis of household data, for the
period 5 to 15 years before the survey date, death rates per 1000
person-years for persons between the ages of 15 and 59 years were 2.2
for men and 1.2 for women, as compared with 2.0 and 0.8, respectively,
for the 2-year period before the invasion. In this age group, death
rates for the period from 2003 to 2006 were 3.2 for men and 1.2 for
women, as compared with 5.9 and 2.0, respectively, for the
post-invasion period on the basis of household data. The estimates
that were based on reported sibling survival may be subject to similar
or higher levels of underreporting than were estimates based on
household data. Completeness of the reporting of sibling deaths might
well be lower than that of households if siblings had been out of
recent contact.2

Comparison with Other Sources

Our comparison between the results of the study by Burnham et al. and
that of the Iraq Body Count focused on the distribution of violent
deaths according to the region of the country, time trends, and rates.
Figure 1A shows the proportional distribution of deaths for four
groups of provinces: Kurdistan, Baghdad, high-mortality provinces (in
which each province accounted for ?3.5% of total mortality reported by
the Iraq Body Count, including Anbar, Babylon, Basra, Diyala, Nineveh,
and Salahuddin), and low-mortality provinces (<3.5% of total mortality
reported by the Iraq Body Count) for the post-invasion period. All
three sources agreed on the low mortality in Kurdistan. Of all the
violent deaths occurring in Iraq, the proportion in Baghdad was 54% in
the IFHS, 60% in the Iraq Body Count, and only 26% in the study by
Burnham et al. The Iraq Body Count probably overestimated this
proportion, since press coverage is probably better in Baghdad than it
is elsewhere in the country. It should also be noted that the rate of
death in Baghdad was adjusted with the use of data from the Iraq Body
Count, as reported above. The Iraq Body Count does not include
combatant deaths among Iraqis, which would have been picked up by the
household surveys. The most striking difference in rates of death was
between those in the study by Burnham et al. and those in the two
other data sources for the six high-mortality provinces, which
accounted for 64% of all deaths in the study by Burnham et al.

Figure 1
View larger version (24K):
[in this window]
[in a new window]
Get Slide
Figure 1. Percent Distribution of Violent Deaths among
Provinces and the Number of Violent Deaths per Day from March 2003 to
June 2006, According to Three Data Sources.

Panel A shows the proportions of violent deaths among provinces in
Iraq according to three data sources: the Iraq Family Health Survey
(IFHS), the Iraq Body Count, and a study by Burnham et al.4 Panel B
shows the number of violent deaths per day according to the year and
data source.


The rates and time trends of violent deaths differed considerably
among the three sources (Table 4). On the basis of population
estimates shown in Table 2 of the Supplementary Appendix, the IFHS
data indicate that every day 128 persons died from violence from March
2003 through April 2004, 115 from May 2004 through May 2005, and 126
from June 2005 through June 2006. The Iraq Body Count numbers were 43,
32, and 55 civilian deaths per day for the same periods. In the study
by Burnham et al., there was a much higher rate of death from violence
and a sharp increase during the 3-year period, with 231, 491, and 925
deaths per day, respectively (Figure 1B). There was greater agreement
regarding mortality from nonviolent causes between the IFHS study (372
deaths per day) and the study by Burnham et al. (416 deaths per day)
(Table 2 of the Supplementary Appendix).

View this table:
[in this window]
[in a new window]
Get Slide
Table 4. Violence-Related Deaths after the Invasion, According
to Time Period and Data Source.


The pre-invasion rates of adult mortality from any cause per 1000
person-years were 2.0 for men and 0.8 for women in the IFHS, with a
relatively small proportion of deaths attributed to violent causes. In
a regional comparison for 2002, a study by the WHO12 estimated that in
Syria and Jordan, the rates of death for adults were 4.2 for men and
2.8 for women. In Iran, the rates were 4.7 and 2.9, respectively. If
we assume that the rate of death in Iraq would have been at similar
levels without the invasion, underreporting of adult deaths in the
IFHS would be as much as 55% for men and 70% for women for reported
deaths occurring in 2001 and 2002. The underreporting of deaths was
expected to be lower for more recent years.

A comparison of the results of the IFHS and those of the Multiple
Cluster Indicator Survey (MICS) 2006,13 conducted by the United
Nations Children's Fund (UNICEF), provides some insight into the
consistency of child-mortality estimates. Both surveys included
questions regarding all children who were ever born and those who were
still alive to estimate child-mortality trends, which could be used to
estimate the rate of death for children under the age of 5 years with
the use of model life tables.7 The estimates per 1000 live births were
54 deaths in the IFHS and 49 deaths in the MICS for the period from
2002 through 2006. On the basis of the simulation that took into
account the sampling errors and the uncertainty in factors for missing
clusters, the level of underreporting, and the projected population
numbers, we estimated that there were 151,000 violent deaths in Iraq
(95% uncertainty range, 104,000 to 223,000) during the post-invasion
period from March 2003 through June 2006.

Discussion

The IFHS 2006 was a large health survey that included the collection
of data regarding rates of death. In spite of the difficult
circumstances in Iraq, the experienced survey coordinating teams
managed to visit 89% of the 1086 selected clusters, and household
response rates were high throughout the country.

Recall of deaths in household surveys with very few exceptions suffer
from underreporting of deaths. None of the methods to assess the level
of underreporting provide a clear indication of the numbers of deaths
missed in the IFHS. All methods presented here have shortcomings and
can suggest only that as many as 50% of violent deaths may have gone
unreported. Household migration affects not only the reporting of
deaths but also the accuracy of sampling and computation of national
rates of death.

The IFHS results for trends and distribution of deaths according to
province are consistent with what has been reported from the scanning
of press reports for civilian casualties through the Iraq Body Count
project. The estimated number of deaths in the IFHS is about three
times as high as that reported by the Iraq Body Count. Both sources
indicate that the 2006 study by Burnham et al. considerably
overestimated the number of violent deaths. For instance, to reach the
925 violent deaths per day reported by Burnham et al. for June 2005
through June 2006, as many as 87% of violent deaths would have been
missed in the IFHS and more than 90% in the Iraq Body Count. This
level of underreporting is highly improbable, given the internal and
external consistency of the data and the much larger sample size and
quality-control measures taken in the implementation of the IFHS.

At present, there are no better methods available to provide more
accurate estimates of the death toll due to the humanitarian conflict
in Iraq in the wake of the 2003 invasion. Rapid small-scale surveys of
households are likely to yield unreliable estimates. Surveys of a
large number of respondents with carefully prepared household
interviews and multiple methods for collecting data on mortality still
run into reporting problems because of the insecurity, instability,
and migration associated with the conflict situation. The clustering
of violent deaths may further affect uncertainty related to
sampling,14 even though more than 1000 clusters were selected for the
IFHS. It is unlikely that more accurate estimates of the death toll
during the post-invasion period can be obtained by conducting more
household surveys with recall questions on mortality. On the basis of
press reports, the Iraq Body Count is also affected by considerable
underreporting but is likely to be a valuable way to monitor trends
over time. Further investment in such mechanisms is justified,
especially if ways can be found to assess the level of underreporting
and the consistency of the reporting mechanisms over time. Other
methods, such as systematic reporting by mortuaries and hospitals and
the strengthening of vital registrations with the use of sentinel
sites, will also need to be explored.

On the basis of direct reporting of deaths by households respondents
that were adjusted for missing clusters, it was estimated that the
violence-related rate of death from March 2003 through June 2006 was
1.09 per 1000 person-years (95% CI, 0.81 to 1.50). About half of the
violent deaths were estimated to have occurred in Baghdad, and
virtually all such deaths occurred in the southern and central
regions.

Overall mortality from nonviolent causes was about 60% higher in the
post-invasion period than in the pre-invasion period. Although recall
bias may contribute to the increase, since deaths before 2003 were
less likely to be reported than more recent deaths, this finding
warrants further analysis.

During the same period, the Iraq Body Count registered 47,668 civilian
deaths from violence. A much smaller mortality survey (1849 households
in 47 clusters) by Burnham et al. came up with a best estimate of
601,027 violent deaths. The best estimate on the number of deaths from
March 2003 through June 2006 based on the IFHS data analysis and
comparisons with other sources is three times as high as that reported
by the Iraq Body Count but only one fourth of that reported by Burnham
et al. On the basis of simulations that took into account survey
sampling errors and estimated probable uncertainty in the adjustment
factors for missing clusters, in the level of underreporting, and in
projected population figures, we estimated that there were 151,000
violent deaths in Iraq during this period (95% uncertainty range,
104,000 to 223,000). Although this number is substantially lower than
that estimated by Burnham et al., it nonetheless points to a massive
death toll in the wake of the 2003 invasion ? and represents only one
of the many health and human consequences of an ongoing humanitarian
crisis.

Supported by United Nations Development Group Iraq Trust Fund,
European Commission and the WHO.

No potential conflict of interest relevant to this article was
reported.

We thank Dr. Hafia Madi for her valuable support and Suhad Adnan and
Dana Mohammed for their assistance with data management. This article
is dedicated to our colleague and coauthor Louay Hakki Rasheed, deputy
director of the Iraqi Central Organization for Statistics and
Information Technology, who was killed on August 2, 2007, on his way
to work in Baghdad.

* Affiliations of the Iraq Family Health Survey study group are listed
in the Appendix.


Source Information

The members of the writing committee (Amir H. Alkhuzai, M.D., Ihsan J.
Ahmad, M.D., Mohammed J. Hweel, M.D., Thakir W. Ismail, M.D., Hanan H.
Hasan, M.D., Abdul Rahman Younis, M.D., Osman Shawani, M.B., Ch.B.,
Vian M. Al-Jaf, M.D., Mahdi M. Al-Alak, Ph.D., Louay H. Rasheed, M.Sc.
(deceased), Suham M. Hamid, M.B., Ch.B., Naeema Al-Gasseer, Ph.D.,
Fazia A. Majeed, M.D., Naira A. Al Awqati, M.D., Mohamed M. Ali,
Ph.D., J. Ties Boerma, Ph.D., and Colin Mathers, Ph.D.) assume
responsibility for the overall content and integrity of the article.

An interview with Dr. Ali can be heard at www.nejm.org.

This article (10.1056/NEJMsa0707782) was published at www.nejm.org on
January 9, 2008.

Address reprint requests to Dr. Ali at the Department of Measurement
and Health Information Systems, World Health Organization, Geneva,
Switzerland, or at alim@xxxxxxxx

References

1. Seltzer W. Some results from Asian population growth studies.
Pop Stud 1969;23:395-406.
2. Gakidou E, Hogan M, Lopez AD. Adult mortality: time for a
reappraisal. Int J Epidemiol 2004;33:710-717. [Free Full Text]
3. Working Group for Mortality Estimation in Emergencies. Wanted:
studies on mortality estimation methods for humanitarian emergencies,
suggestions for future research. Emerg Themes Epidemiol 2007;4:9-9.
[CrossRef][Medline]
4. Burnham G, Lafta R, Doocy S, Roberts L. Mortality after the 2003
invasion of Iraq: a cross-sectional cluster sample survey. Lancet
2006;368:1421-1428. [CrossRef][ISI][Medline]
5. The Iraq Body Count project. (Accessed January 4, 2008, at
http://www.iraqbodycount.org.)
6. Ministry of Planning and Development Corporation. Iraq Living
Conditions Survey 2004. Baghdad, Iraq: United Nations Development
Programme, 2005.
7. Department of Economic and Social Affairs. Indirect techniques
for demographic estimation: manual X. New York: United Nations, 1983.
8. Bicego G. Estimating adult mortality rates in the context of the
AIDS epidemic in sub-Saharan Africa: analysis of DHS sibling
histories. Health Transit Rev 1997;7:Suppl 2:7-22. [Medline]
9. Timaeus I, Zaba B, Ali M. Estimation of adult mortality from
data on adult siblings. In: Zaba B, Blacker J, eds. Brass tacks:
essays in medical demography. London: Athlone Press, 2001:43-66.
10. Gakidou E, King G. Death by survey: estimating adult mortality
without selection bias from sibling survival data. Demography
2006;43:569-585. [CrossRef][ISI][Medline]
11. Effron B, Tibshirani RJ. An introduction to the bootstrap. New
York: Chapmann & Hall, 1993.
12. World health statistics 2005. Geneva: World Health Organization,
2005.
13. Central Organization for Statistics and Information Technology
(COSIT), Kurdistan Regional Statistics Office, Suleimaniyah
Statistical Directorate, Ministry of Health and UNICEF. Findings from
the Iraq Multiple Indicator Cluster Survey. Preliminary report. March
2007. (Accessed January 4, 2008, at http://www.childinfo.org.)
14. Working Group for Mortality Estimation in Emergencies. Wanted:
studies on mortality estimation methods for humanitarian emergencies,
suggestions for future research. Emerg Themes Epidemiol 2007;4:9-9.
[CrossRef][Medline]

.

Relevant Pages

  • Fathers: Big picture reflects balance
    ... Family, Feminism, Mating, Marriage & Divorce ... Robert J. Nebel wrote that "survey after survey say that women are ... "masculine" home chores such as lawn mowing or household and car ...
    (soc.men)
  • Fathers: Big picture reflects balance
    ... Family, Feminism, Mating, Marriage & Divorce ... Robert J. Nebel wrote that "survey after survey say that women are ... "masculine" home chores such as lawn mowing or household and car ...
    (soc.women)
  • Majority of Aussies reject drinking a mix of recycled and conventional water.
    ... The university's research fellow Dr June Marks said the survey of 2,504 ... people revealed 96.5 per cent were willing to recycle reclaimed water ... for household use such as toilet flushing. ...
    (soc.culture.singapore)