Fax +41 61 306 12 34
E-Mail karger@karger.ch
www.karger.com
Letter to the Editor
One hundred and six (35.3%) of the 300 patients completed a
face-to-face follow-up interview at 6 months. All patients met
the stressor criterion A1. Six patients (5.7%, 95% confidence in-
terval = 1.3–10.1) met all other criteria for accident-related PTSD
including A2, and the mean ( 8SD) CAPS total score of these
6 patients was 57.8 8 16.1. However, in the present report, we
adopted the manner of omitting item 8 (psychogenic amnesia in
criterion C) when making the diagnosis of PTSD [15], because we
were often unable to differentiate organic from psychogenic am-
nesia. For reference purposes, 8 (7.5%) patients met criteria for
PTSD when item 8 was included in making the diagnosis of
PTSD.
We examined the cross-country relationship between infant
mortality rate and prevalence of PTSD. Methodologies which in-
cluded the consecutive recruitment of accidentally injured pa-
tients in emergency departments or intensive care units, prospec-
tive design, and assessment according to internationally accepted
diagnostic criteria suggest the validity and comparability of these
data. Studies were excluded if they included help-seeking patients,
patients recruited from police records, or all patients with trau-
matic brain injury. The data from seven studies undertaken in six
developed countries (UK, US, Israel, Australia, Switzerland and
Japan) [2–8] as well as our current data were used. We used data
for the prevalence of chronic PTSD (4–12 months after the acci-
dent) because spontaneous remission is relatively common within
3 months of a traumatic event. Although self-reported question-
naires are likely to result in elevated PTSD estimates, we used
large-scale data from the US and the UK [2, 3] for comparison.
The infant mortality rates in each country for the year when each
study was conducted or submitted to the journal were used [16].
The relation between infant mortality rate and prevalence of
accident-related PTSD was nonlinear ( fig. 1). On the basis of
model fit, the best fitting was obtained with the quadratic model
(R
2
= 0.82, p = 0.01), though a linear model was acceptable (R
2
=
0.60, p = 0.02). Infant mortality rate is well known to be associ-
ated with levels of basic health care, well-developed technology,
and medical advances. These rates are also commonly included
as part of standard of living evaluations in economics. There are
many cultural differences among the six countries such as popu-
lation density, ethnic background, founding history, dietary hab-
it, and residential setting. The present study showed a plausible
explanation for the observed discrepancy in the prevalence of
PTSD following injury. These observations may provide clues re-
garding the estimated prevalence of accident-related PTSD and
ways to reduce the number of patients that do develop PTSD.
The present Japanese study has methodological strengths, in-
cluding consecutive sampling and standardized assessment.
However, we also consider important limitations related to repre-
sentativeness, such as sampling from only one hospital and a rel-
atively high dropout rate. Furthermore, a higher prevalence of
Not only has accidental injury been shown to account for a
significant health burden on all populations, regardless of age, sex
and geographic region [1], it is also an important risk factor for
posttraumatic stress disorder (PTSD). Recent epidemiological
studies using consecutive patients have reported inconsistent
findings in the prevalence of accident-related PTSD in developed
countries. For instance, PTSD prevalence at 4–12 months after
accident was in the range of 17–32% in the UK [2], US [3] and Is-
rael [4], but at considerably lower rates of 10.4% in Australia [5],
8.5% in Japan [6], and 1.9–3.1% in Switzerland [7, 8]. In a recent
study by Schnyder et al. [8] published in Psychotherapy and Psy-
chosomatics, the authors discussed that intercultural differences
play an important role in the development of PTSD. We agree with
their suggestion that local environmental factors, such as socio-
economic and cultural components, and health care systems are
also important. In this context, Babones [9] showed an unam-
biguously positive correlation between income inequality and
population health including infant mortality (an indicator of so-
cial circumstances and basic population health) at country level.
The aim of the present study was to reveal the prevalence of PTSD
at 6 months’ follow-up in our prospective study [10] and to exam-
ine the relation between infant mortality rate and prevalence of
PTSD in the reliable cross-country data available.
A total of 300 patients consecutively admitted to the intensive
care unit (ICU) of a teaching hospital in Tokyo due to accident-
related injury were enrolled in the study and were assessed short-
ly after admission and 6 months after their accident. The main
outcome measure was the Clinician-Administered PTSD Scale
[11, 12]. The method and sociodemographic and clinical informa-
tion have been presented in an earlier publication [10]. Briefly, the
majority of participants were men (77.7%), the average ( 8SD) age
was 36.5 8 15.0 years, the average ( 8SD) Injury Severity Score
[13] was 9.1 8 7.9, and the median Glasgow Coma Scale score [14]
was 15.0 (range 3–15).
Published online: November 18, 2009
© 2009 S. Karger AG, Basel
0033–3190/10/0791–0056$26.00/0
Accessible online at:
www.karger.com/pps
Psychother Psychosom 2010;79:56–57
DOI: 10.1159/000259418
Towards an Explanation of Inconsistent Rates of
Posttraumatic Stress Disorder across Different
Countries: Infant Mortality Rate as a Marker of
Social Circumstances and Basic Population Health
Yutaka Matsuoka
a, b, c
, Daisuke Nishi
a, b, c
, Naohiro Yonemoto
a, c
,
Satomi Nakajima
a, c
, Yoshiharu Kim
a, c
a
National Institute of Mental Health, National Center of
Neurology and Psychiatry,
b
Department of Psychiatry, National
Disaster Medical Center, and
c
CREST, Japan Science and
Technology Agency, Tokyo, Japan