Regional perinatal mortality differences in the Netherlands; care is the question
BMC Public Health
Regional perinatal mortality differences in the Netherlands; care is the question
Miranda Tromp 2
Martine Eskes 2
Johannes B Reitsma 1
Jan Jaap HM Erwich 0
Hens AA Brouwers 5
Greta C Rijninks-van Driel 4
Gouke J Bonsel 3
Anita CJ Ravelli 2
0 Department of Obstetrics and Gynaecology, University Medical Center Groningen , Groningen , the Netherlands
1 Department of Clinical Epidemiology , Biostatistics and Bioinformatics , Academic Medical Center, University of Amsterdam , Amsterdam , the Netherlands
2 Department of Medical Informatics, Academic Medical Center, University of Amsterdam , Amsterdam , the Netherlands
3 Department of Health Policy and Management, Erasmus Medical Center , Rotterdam , the Netherlands
4 Department of Obstetrics and Gynaecology, Academic Medical Centre , Amsterdam , the Netherlands
5 Department of Neonatology, University Medical Center Utrecht , Utrecht , the Netherlands
Background: Perinatal mortality is an important indicator of health. European comparisons of perinatal mortality show an unfavourable position for the Netherlands. Our objective was to study regional variation in perinatal mortality within the Netherlands and to identify possible explanatory factors for the found differences. Methods: Our study population comprised of all singleton births (904,003) derived from the Netherlands Perinatal Registry for the period 2000-2004. Perinatal mortality including stillbirth from 22+0 weeks gestation and early neonatal death (0-6 days) was our main outcome measure. Differences in perinatal mortality were calculated between 4 distinct geographical regions NorthEast-South-West. We tried to explain regional differences by adjustment for the demographic factors maternal age, parity and ethnicity and by socio-economic status and urbanisation degree using logistic modelling. In addition, regional differences in mode of delivery and risk selection were analysed as health care factors. Finally, perinatal mortality was analysed among five distinct clinical risk groups based on the mediating risk factors gestational age and congenital anomalies. Results: Overall perinatal mortality was 10.1 per 1,000 total births over the period 2000-2004. Perinatal mortality was elevated in the northern region (11.2 per 1,000 total births). Perinatal mortality in the eastern, western and southern region was 10.2, 10.1 and 9.6 per 1,000 total births respectively. Adjustment for demographic factors increased the perinatal mortality risk in the northern region (odds ratio 1.20, 95% CI 1.12-1.28, compared to reference western region), subsequent adjustment for socioeconomic status and urbanisation explained a small part of the elevated risk (odds ratio 1.11, 95% CI 1.03-1.20). Risk group analysis showed that regional differences were absent among very preterm births (22+0 - 25+6 weeks gestation) and most prominent among births from 32+0 gestation weeks onwards and among children with severe congenital anomalies. Among term births (≥ 37+0 weeks) regional mortality differences were largest for births in women transferred from low to high risk during delivery. Conclusion: Regional differences in perinatal mortality exist in the Netherlands. These differences could not be explained by demographic or socio-economic factors, however clinical risk group analysis showed indications for a role of health care factors.
Perinatal mortality is an important indicator of health and
the quality of health care . Countries or regions are
often compared using perinatal mortality rate. The
position of the Netherlands in international comparative
research is unfavourable. In 2003 the results of the
PERISTAT study showed that Dutch perinatal mortality for the
year 1999 was substantially higher compared to other
European countries (stillbirth rate of 7.4 per 1,000 total
births and early neonatal mortality of 3.5 per 1,000 live
The observed differences in perinatal mortality across
Europe are difficult to explain unequivocally because of
the many potential explanations like variation in
registration practices, differences in definitions, and variation in
demographic structure . On the national level, fair
comparisons can be achieved more easily. Dutch public
health policies aim to reduce national health inequalities
if existent . It is unknown whether the current Dutch
perinatal mortality is uniformly distributed across the
country; differences have been reported based on civil
data in the early eighties [5,6]. Although the Netherlands
is a small country, regional variation exists in the degree
of urbanisation, the number of immigrants and to a lesser
extent, in socio-economic status. Regional variation in
mortality from other causes like cardio-vascular diseases
and cancer has been reported before [7-9].
Factors related to regional differences in perinatal
outcome reported in other European countries after
adjustment for demographic factors, include population density
, access and use of health services , income level
and social inequality [12,13] and excess risk for certain
The objective of this study is to examine whether regional
differences in perinatal mortality in the Netherlands exist
for the period 2000–2004, and whether these differences
persist after taking into account various risk factors that
have been linked to regional variation in perinatal
Data from the Netherlands Perinatal Registry (PRN)
2000–2004 were used. The PRN is a database that
contains the linked data from three registries: the national
obstetric database by midwives (LVR-1 registry), the
national obstetric database by gynaecologists (LVR-2
registry) and the national neonatal/paediatric database (LNR
registry) . The PRN registry contains comprehensive
data on pregnancy, provided care (interventions,
referrals) and pregnancy outcomes . The coverage of the
PRN is about 96% of all deliveries in the Netherlands. All
variables are recorded by the caregiver during prenatal
care, delivery and neonatal, lying-in period. The data are
annually sent to the national registry office, where a
number of range and consistency checks are conducted.
Data on socio-economic status (SES) on the postal code
level was obtained from The Netherlands Institute for
Social Research (SCP).
The population for this study consisted of all singleton
births born between 2000 and 2004 from 22.0 gestational
weeks onwards. Gestational age was based on ultrasound
or last menstrual period. If gestational age was unknown,
children with a birth weight below 500 grams were
excluded in accordance with the World Health
Organization reporting criteria .
Perinatal mortality was our primary outcome measure.
Perinatal mortality is defined as the sum of stillbirth (≥
22.0 gestational weeks) and early neonatal mortality
(deaths of live born children during the first week of life).
Stillbirth rate and perinatal mortality rate were both
calculated per 1,000 total births. Early neonatal mortality was
calculated per 1,000 live births. Apart from mortality, the
following mediating outcome measures were also
analysed: preterm delivery (<32.0 weeks gestational age), low
birth weight (<1500 gram) and low APGAR score after five
minutes (APGAR score <4).
Provinces and regions in the Netherlands
Regional differences in perinatal mortality were analysed
on a province level and on a regional level. The
Netherlands is formally divided into 12 provinces (see Figure 1),
which form regional administrative units in between
municipalities and the national government. The
provinces were grouped into 4 regions based on their
geographical position: Northern region (Groningen,
Friesland, Drenthe), eastern region (Overijssel,
Gelderland, Flevoland), western region (Utrecht,
Noord-Holland, Zuid-Holland) and southern region (Zeeland,
Noord-Brabant, Limburg). The northern region is the
most rural area in the Netherlands, while the western
region is most urbanised (the 4 largest cities are displayed
in figure 1). The province of each woman was based on
her registered postal code (4 digits) in the registry.
Women with an unknown or invalid postal code (0.2%)
were removed from the analyses.
Demographic characteristics of included women were
compared across regions including maternal age, parity
and ethnicity. Maternal age was categorised into <20
years, 20–34 years and ≥35 years. Parity was categorised
into 0 (first birth), 1 (second birth) and 2+ (third or
PFrigouvirnece1s and regions in the Netherlands
Provinces and regions in the Netherlands.
higher birth). Ethnicity is ascribed by the woman's care
provider. For this study, we differentiated between
Western (native Dutch and other Westerners) and
non-Western (including different ethnic groups like African/
Surinamese Creole, Surinamese Hindustani, Moroccan
In addition, data on socio-economic status (SES) were
obtained from The Netherlands Institute for Social
Research/SCP on postal code level. Using the woman's
postal code (4 digits) these data could be linked to the
perinatal registry file. The SES score is based on mean
income level, the percentage of households with a low
income, the percentage of inhabitants without a paid job
and the percentage of households with on average a low
education in a postal code area . The continuous SES
score was for our purpose categorised into a high, middle
and low group based on percentile ranges (≤ 25th
percentile, middle, > 75th percentile). The data on
socio-economic status were available for the year 2002. The
categorised score was applied to the total population for
the period 2000–2004 as large changes in SES score for a
postal code area within two years are unlikely. By using
the same postal code we could add the degree of
urbanisation (number of addresses per square kilometre), a
number which is routinely available. The degree of
urbanisation was categorised into three groups: very rural, rural
to urban and very urbanised.
Besides population characteristics, we analysed regional
variation in health care services. We geographically
compared the mode of delivery and risk selection at start of
delivery. Risk selection is an important feature of the
Dutch obstetric system . Healthy women with an
uncomplicated obstetric history and/or pregnancy remain
under the care of the primary level midwife and are
selected as at low risk at start of delivery. In that case a
woman can choose to deliver at home or at the hospital,
both under supervision of the midwife. If complications
occur the woman is selected as high risk and is referred to
an obstetrician at the secondary or tertiary level. We
analysed the risk selection status at the start of delivery.
Differences in adverse outcomes by region and province
were tested by Chi-Square test using all other regions/
provinces as the reference category. Differences in
population characteristics by region were tested by Chi-Square
test. After describing crude mortality, logistic regression
modelling was used to estimate differences in perinatal
mortality between regions after adjustment for
sociodemographic factors. All previously described factors were
added to the model in two successive steps. First we
adjusted only for demographic factors parity, maternal
age and ethnicity, parity and maternal age were included
as categorical variables with the category with the lowest
mortality risk as reference. In the second model we
additionally adjusted for the degree of urbanisation and SES.
In both models we included the year of registration to
incorporate changes in perinatal mortality over time. The
strength of the association between potential predictors
and perinatal mortality are expressed as odds ratios (OR)
with 95% confidence intervals (CI).
For further interpretation of possible regional differences
in perinatal mortality, the perinatal mortality risk for five
clinical relevant risk groups was examined by region.
These groups represent distinct clinical entities with
different patterns of care based on mediating risk factors
gestational age and severe congenital anomalies. Severe
congenital anomalies were defined as anomalies which
are either highly fatal or as anomalies potentially
detectable by ultrasound and severe enough for optional late
termination of pregnancy. The five groups are very preterm
births (22+0–25+6 weeks), severe congenital anomalies,
preterm births (26+0–31+6 weeks) without severe
congenital anomalies, preterm births (32+0–36+6 weeks) without
severe congenital anomalies and term births (≥ 37+0
weeks) without severe congenital anomalies.
All analyses were performed using SAS for Windows
(version 9.1, SAS Institute Inc., Cary, NC, USA).
During the period 2000–2004, there were 904,003
singletons births in the Netherlands. Nearly half of all births
were in the urbanised western region (46.1%) and only
9.7% in the northern more rural region (see table 1). The
overall perinatal mortality in the Netherlands in the
period 2000–2004 was 10.1 per 1,000 total births. The
northern region has the highest perinatal mortality rate
with 11.5 and 11.9 per 1,000 total births in provinces
Groningen and Friesland respectively. The southern
region had the lowest perinatal mortality rate, with lowest
rates in provinces Noord-Brabant and Limburg: 9.5 and
9.4 per 1,000 total births. The perinatal mortality rate in
the northern region was significantly higher than in the
other regions (Chi-square p-value < 0.01). Both stillbirth
and early neonatal mortality were high in the northern
provinces. The proportion of preterm births and children
with a low birth weight was also significantly higher in the
northern region (1.2% and 1.2% respectively).
The northern region had the largest proportion of women
from rural areas (41.8%) and with a low SES score
(38.0%) and the lowest proportion of non-western
women (7.2%) (Table 2). The western region had the
highest proportion of women aged above 35 years
(20.6%), with non-western ethnicity (22.7%), with a high
* Significantly different (p < 0.01) from all other regions/provinces excluding the region/province itself.
SES score (32.9%) and living in urban areas (36.4%). All
these regional differences were statistically significant.
Table 3 shows that women from the northern region had
a significantly higher perinatal mortality risk compared to
the western region (unadjusted OR 1.11, 95% CI 1.03–
1.19). The western region was set as reference area because
the perinatal mortality risk was the same as the overall
rate. After adjustment for demographic factors (maternal
age, parity and ethnicity), the women in the northern
region (OR 1.20, 95% CI 1.12–1.28) and eastern region
(OR 1.08, 95% CI 1.02–1.14) had a significantly higher
perinatal mortality risk compared to women in the
western region. Living in a very urban area and having a low
p < 0.0001
p < 0.0001
p < 0.0001
p < 0.0001
p < 0.0001
p < 0.0001
Adjusted Model I#
OR 95% CI
* Region west was set as reference area as the perinatal mortality rate of this region was equal to the overall rate.
# Model I was adjusted for maternal age, parity and ethnicity.
† Model II was an extension of model I with additional adjustment for degree of urbanization and SES.
SES score were significant risk factors for perinatal
mortality, while a high SES score lowered the risk. Subsequent
adjustment for urbanisation degree and social-economic
status explained a small part of the excess risk in the
northern region (OR 1.11, 95% CI 1.03–1.20). Living in
an very urban area was no longer a risk factor in adjusted
The health services patterns also exhibited regional
differences (Table 4). The northern region had the lowest
Mode of delivery
Elective Caesarean Section
Instrumental vaginal delivery
Emergency Caesarean Section
Care at start of delivery
low risk selection
low risk & home delivery
low risk & hospital delivery
from low to high risk during delivery
high risk selection
Number of hospitals
Number of tertiary centres
number of spontaneous deliveries (72.9% versus 75.6%
in the western region) and the lowest number of women
selected as low risk at start of delivery (42.7% versus
50.2% in the western region). The percentage of home
births was 19.7% in the northern region versus 23.0% in
the western region and in the eastern region the
percentage of home births was as high as 30.4%. There were only
small variations in the percentage of women transferred
from low risk to high risk during delivery (11.2% in the
northern region and 12.7% in the western region). In the
northern region most deliveries take place under
supervision of an obstetrician (57%). The northern region has the
lowest number of hospitals and only 1 tertiary hospital.
Among very preterm births (responsible for 28% of all
perinatal deaths), the perinatal mortality risk was about
the same in all regions (Table 5). The perinatal mortality
risk for children with severe congenital anomalies
(responsible for 12% of perinatal deaths) was higher in
the northern region (204 per 1,000 births) compared to
the western region (147 per 1,000). The mortality risk
among preterm births 26+0–31+6 weeks (responsible for
14% of perinatal deaths) in the northern region was lower
than in the western region (233 versus 237 per 1,000).
The mortality risk among preterm births 32+0–36+6 weeks
(responsible for 18% of perinatal deaths) was higher in
the northern region and lower in the southern region than
in the western region. For the term births (responsible for
28% of all perinatal deaths) the mortality risk was about
11% higher in the northern region compared to the
western region (3.4 per 1,000 versus 3.1 per 1,000 in region
west). Within the term group (≥37+0 weeks), the regional
mortality difference was the largest for the group of births
from women transferred from low risk to high risk during
delivery (4.0 per 1,000 in north versus 2.6 per 1,000 in
west). The perinatal mortality risk for term births from
women selected as high risk at start of delivery was
similar; in both northern and western regions 5.0 per 1,000.
The perinatal mortality in the Netherlands for the period
2000–2004 shows regional variation, with an increased
perinatal mortality in the rural northern region. The
regional variation was present in both stillbirth and early
neonatal mortality. The elevated risk in the northern
region could not be explained by regional variation in
demographic risk factors like maternal age, parity and
ethnicity. Socio-economic status and urbanisation grade only
explained a small part of the excess risk. Analyses focussed
on clinical relevant subgroups showed regional
differences were most prominent among births from 32+0
weeks gestation onwards and especially among term
births from women transferred from low to high risk
Data from a period of five years could be analyzed
including 904,003 pregnancies in the Netherlands. The
Netherlands Perinatal Registry contains the combined
information on pregnancy, childbirth and the neonatal
period derived from three separate registries that have
recently been linked using probabilistic record linkage
techniques. This enabled us to adjust for a combination of
demographic, care related and socio-economic factors in
relation to perinatal outcome [16,20].
Data from general practitioners providing obstetric care
were not available from the Netherlands perinatal registry.
General practitioners more often provide obstetric care in
rural areas, which are found in the northern region but
also in the eastern and southern regions . Overall the
proportion of deliveries that took place under supervision
of a general practitioner is estimated at 4%. However over
99% of hospital deliveries were included and a woman is
transferred to an obstetrician by the general practitioner in
case of high risk. This is in accordance with the finding
when the perinatal registry data were linked to civil
registry data in a pilot study, more foetal deaths were registered
in the perinatal registry, especially the very premature
foetal deaths. Medical registries suffer from entry errors by
professionals as any database. Limited entry options and
data checks by professionals combined with validated
linkage procedures [15,22] have confined errors to a
minimum. The current perinatal registry does not contain
information on smoking (only reporting on heavy
smoking with clear underestimation), food intake, folic acid
intake, maternal education and body mass index (BMI),
Clinical risk groups
Very preterm births <26+0 weeks
Severe congenital anomalies
Premature 26+0–31+6 weeks
Premature 32+0–36+6 weeks
Term ≥ 37+0 weeks
Prev = prevalence.
factors which may (partly) explain the regional differences
in perinatal mortality [23-25]. Additional adjustment for
BMI and smoking on a province level for women in the
reproductive age (data from Statistics Netherlands and
STIVORO) did not change the elevated perinatal mortality
risk in the northern region (data not shown). Risk factor
behaviour in pregnancy is related to both socio-economic
class and ethnicity [26,27]. After adjustment for these
factors the risk status of the northern region remained high;
therefore we believe that regional variation in
unmeasured risk factors is unlikely to explain the observed
differences in mortality in our adjusted models. One could
challenge the use of the SES score on a neighbourhood
level rather than on the individual level. However,
previous research on socio-economic inequalities have
demonstrated that this is a valid approach [28,29].
This is the first time that regional differences in perinatal
mortality were studied in the Netherlands using the
national linked perinatal registry data on 904,003
pregnancies. Previous regional analyses were based on
aggregate data on 11 provinces for the period 1979–1982 
and on 40 economic sub-regions for the period 1980–
1984 , rather than on individual data. Treffers et al.
found differences in the percentage of hospital deliveries
(versus home deliveries) per province, but could not
relate this to regional perinatal mortality rate . We also
found differences in hospital deliveries per region and
found that the regional differences were most pronounced
among term women transferred from low to high risk
during delivery. Mackenbach et al. reported perinatal
mortality rate to depend on mean income, part of the population
living in a large municipality and the presence of a
leveltwo hospital . We applied individual demographic
adjustment, and used more refined variables to account
for SES and urbanisation. We had access to urbanisation
and SES on the neighbourhood level based on postal
code, which showed large variation between regions.
Social factors have been reported as explanatory factors
for perinatal mortality differences [12,13], however
adjustment for SES and urbanisation only explained a
small part of the excess risk in the northern region in our
study. The presence of fewer hospitals in the northern
region may have played a role. The differences in regional
perinatal mortality are sizable, and consistent with
recently observed mortality differences for other causes
(cardiovascular, cancer) .
Regional variation in health outcomes can be caused by
variation in incidence of complications and/or variation
in prognosis. Health status of the women and preventive
and obstetric care if applicable can influence incidence
and prognosis. Both stillbirth and neonatal mortality were
increased in the northern region, which indicates a role
for factors common to both. Population composition
factors and environmental risk factors, undetected by the
direct and indirect adjustment factors can be present, but
their presence is less likely given the adjustments. The
analyses suggest that prevention and care factors may have
played a role. Potential candidates are the following.
The uptake of prevention (general – smoking, specific –
folic acid, screening) may be less or the intensity or
effectiveness of health services may be lower . As
differential access will have been partly covered by the adjustment
factor urbanisation, explanations at the care level are
more likely. The number of clinical facilities in the
northern region is smaller, and only 1 tertiary centre is
available. Perhaps intensity and quality of preventive and
delivery care is less in areas with low population density
[32,33]. The elevated mortality risk for children with
congenital anomalies in the northern region (while
prevalence was similar) might also point to differences in care
shortly after birth. Late neonatal mortality (7–27 days)
showed the same regional pattern, excluding mortality
differences by different care management during the first
week and subsequent delay of mortality. As an increased
mortality risk in the northern region is present in both
preterm and term births also differences in hospital
supply services (obstetrical, neonatal) has to be considered,
and delay due to the on average larger travelling distances
in case of intended home births with an emerging risk
requiring hospital admission. The group transferred from
low to high risk during delivery is at higher risk for
perinatal mortality than women who deliver under care of a
midwife completely [16,34]. For term births the regional
mortality differences were most pronounced in the group
transferred from low to high risk during delivery, possibly
indicating a role for travel distance. Further exploring the
role of care factors rests on more detailed analysis of
clinical risk groups for perinatal mortality and of stillbirth and
neonatal mortality separately, but also on audit studies
[35,36]. Audit studies could also provide information on
causes of death, which is not registered in the current PRN
registry. Against the background mortality observed in the
other regions, the observed mortality in the northern
region of 11.2 per 1,000 births and about 17,500
deliveries annually, this excess risk in the northern region
accounts for about 19 deaths a year.
In conclusion, our study revealed persistent adverse
perinatal outcome in the northern part of the Netherlands
even after adjustment for demographic and
socio-economic factors. Analysis of clinical risk groups showed
perinatal mortality differences were most pronounced
among children with severe congenital anomalies and
among term births from women transferred from low to
high risk during delivery. The results provide an incentive
to explore the role of health care factors, both at the
prenatal and delivery stage of care.
The authors declare that they have no competing interests.
AR, MT, ME, JR and GB contributed to the conception and
design of the study. JJE, HB and GR were involved in
acquisition of the data. MT and AR conducted the analyses
and MT drafted the manuscript. AR, GB, JR and ME
provided advice on the analysis and interpretation of the
data. All authors critically revised the draft versions and
approved the final version of the manuscript.
We gratefully acknowledge the investment of numerous caregivers in the
Netherlands providing the registry information. We thank The Netherlands
Perinatal Registry http://www.perinatreg.nl for her permission to use the
1. Buitendijk S , Zeitlin J , Cuttini M , Langhoff-Roos J , Bottu J : Indicators of fetal and infant health outcomes . Eur J Obstet Gynecol Reprod Biol 2003 , 111 (Suppl 1): S66 - S77 .
2. Zeitlin J , Wildman K , Breart G , Alexander S , Barros H , Blondel B , et al.: PERISTAT: indicators for monitoring and evaluating perinatal health in Europe . Eur J Public Health 2003 , 13 : 29 - 37 .
3. Graafmans WC , Richardus JH , Macfarlane A , Rebagliato M , Blondel B , Verloove-Vanhorick SP , et al.: Comparability of published perinatal mortality rates in Western Europe: the quantitative impact of differences in gestational age and birthweight criteria . BJOG 2001 , 108 : 1237 - 1245 .
4. Mackenbach JP , Stronks K : A strategy for tackling health inequalities in the Netherlands . BMJ 2002 , 325 : 1029 - 1032 .
5. Mackenbach JP , van Leengoed PL : [Regional differences in perinatal mortality: the relationship with various aspects of perinatal care] . Ned Tijdschr Geneeskd 1989 , 133 : 1839 - 1844 .
6. Treffers PE , Laan R : Regional perinatal mortality and regional hospitalization at delivery in The Netherlands . Br J Obstet Gynaecol 1986 , 93 : 690 - 693 .
7. de VE, Schouten LJ , Visser O , Eggermont AM , Coebergh JW : Rising trends in the incidence of and mortality from cutaneous melanoma in the Netherlands: a Northwest to Southeast gradient ? Eur J Cancer 2003 , 39 : 1439 - 1446 .
8. Kunst AE , Looman CW , Mackenbach JP : Determinants of regional differences in lung cancer mortality in The Netherlands . Soc Sci Med 1993 , 37 : 623 - 631 .
9. Mackenbach JP , Kunst AE , Looman CW : [Regional differences in mortality from ischemic heart disease] . Ned Tijdschr Geneeskd 1990 , 134 : 596 - 601 .
10. Daltveit AK , Vollset SE , Irgens LM : Population density and perinatal mortality in Norway and Sweden 1975-1988 . Scand J Public Health 1999 , 27 : 213 - 219 .
11. Gissler M , Keskimaki I , Teperi J , Jarvelin M , Hemminki E : Regional equity in childhood health - register-based follow-up of the Finnish 1987 birth cohort . Health Place 2000 , 6 : 329 - 336 .
12. Lauria L , De Stavola BL : A district-based analysis of stillbirth and infant mortality rates in Italy: 1989-93 . Paediatr Perinat Epidemiol 2003 , 17 : 22 - 32 .
13. Martuzzi M , Grundy C , Elliott P : Perinatal mortality in an English health region: geographical distribution and association with socio-economic factors . Paediatr Perinat Epidemiol 1998 , 12 : 263 - 276 .
14. Serenius F , Winbo I , Dahlquist G , Kallen B : Regional differences in stillbirth and neonatal death rate in Sweden with a cause-ofdeath specific analysis . Acta Paediatr 2001 , 90 : 1062 - 1067 .
15. Meray N , Reitsma JB , Ravelli AC , Bonsel GJ : Probabilistic record linkage is a valid and transparent tool to combine databases without a patient identification number . J Clin Epidemiol 2007 , 60 : 883 - 891 .
16. Ravelli AC , Eskes M , Tromp M , Huis AM , Steegers EAP , Tamminga P , et al.: [Perinatal mortality in the Netherlands 2000-2006; risk factors and risk selection] . Ned Tijdschr Geneeskd 2008 , 152 : 2728 - 2733 .
17. World Health Organization: Definitions and Indicators in family planning maternal&child health and reproductive health . WHO press. Geneva; 2001 .
18. Sociaal en Cultureel Planbureau: Van hoog naar laag; van laag naar hoog . Rijswijk, Sociaal en Cultureel Planbureau ; 1998 .
19. Bleker OP , Hulst LAM van der, Eskes M , Bonsel GJ : Place of birth: evidence for best practice . In Recent advantage in Obstetrics and Gynaecology 23 Edited by: Bonnar J, Dunlop W. London: Royal Society of Medicine Press ; 2005 : 77 - 100 .
20. Ravelli ACJ , Tromp M , van Huis M , Steegers EAP , Tamminga P , Eskes M , et al.: Decreasing perinatal mortality in the Netherlands , 2000 - 2006 : the impact of risk factors . Journal of Epidemiology and Community Health 2009 in press.
21. Wiegers T , Hingstman L : Inventarisatie 'verloskundig actieve huisartsen ' Utrecht: Nivel; 1999 .
22. Tromp M , Ravelli AC , Meray N , Reitsma JB , Bonsel GJ : An efficient validation method of probabilistic record linkage including readmissions and twins . Methods Inf Med 2008 , 47 : 356 - 363 .
23. Bhattacharya S , Campbell DM , Liston WA , Bhattacharya S : Effect of Body Mass Index on pregnancy outcomes in nulliparous women delivering singleton babies . BMC Public Health 2007 , 7 : 168 .
24. Devlieger H , Martens G , Bekaert A : Social inequalities in perinatal and infant mortality in the northern region of Belgium (the Flanders) . Eur J Public Health 2005 , 15 : 15 - 19 .
25. Kallen K : The impact of maternal smoking during pregnancy on delivery outcome . Eur J Public Health 2001 , 11 : 329 - 333 .
26. Goedhart G , van EM , Wal MF van der, Bonsel GJ : Ethnic differences in preterm birth and its subtypes: the effect of a cumulative risk profile . BJOG 2008 , 115 : 710 - 719 .
27. Mohsin M , Bauman AE : Socio-demographic factors associated with smoking and smoking cessation among 426,344 pregnant women in New South Wales, Australia . BMC Public Health 2005 , 5 : 138 .
28. Bos V , Kunst AE , Garssen J , Mackenbach JP : Socioeconomic inequalities in mortality within ethnic groups in the Netherlands , 1995 - 2000 . J Epidemiol Community Health 2005 , 59 : 329 - 335 .
29. Mackenbach JP , Stronks K , Kunst AE : The contribution of medical care to inequalities in health: differences between socio-economic groups in decline of mortality from conditions amenable to medical intervention . Soc Sci Med 1989 , 29 : 369 - 376 .
30. Zwakhals SLN , Giesberg H : Kanker totaal . In Volksgezondheid Toekomst Verkenning, Nationale Atlas Volksgezondheid Bilthoven: RIVM ; 2006 .
31. Raatikainen K , Heiskanen N , Heinonen S : Under-attending free antenatal care is associated with adverse pregnancy outcomes . BMC Public Health 2007 , 7 : 268 .
32. Moster D , Lie RT , Markestad T : Neonatal mortality rates in communities with small maternity units compared with those having larger maternity units . BJOG 2001 , 108 : 904 - 909 .
33. Phibbs CS , Bronstein JM , Buxton E , Phibbs RH : The effects of patient volume and level of care at the hospital of birth on neonatal mortality . JAMA 1996 , 276 : 1054 - 1059 .
34. Amelink-Verburg MP , Verloove-Vanhorick SP , Hakkenberg RM , Veldhuijzen IM , Bennebroek GJ , Buitendijk SE : Evaluation of 280,000 cases in Dutch midwifery practices: a descriptive study . BJOG 2008 , 115 : 570 - 578 .
35. Merkus JM : [Perinatal mortality in The Netherlands: an audit is now more necessary than ever] . Ned Tijdschr Geneeskd 2008 , 152 : 603 - 605 .
36. Richardus JH , Graafmans WC , Verloove-Vanhorick SP , Mackenbach JP : Differences in perinatal mortality and suboptimal care between 10 European regions: results of an international audit . BJOG 2003 , 110 : 97 - 105 .