Acta Obstet Gynecol Scand 2000; 79: 107–112 Copyright C Acta Obstet Gynecol Scand 2000

Printed in Denmark ¡ All rights reservedActa Obstetricia et

Gynecologica ScandinavicaISSN 0001-6349

ORIGINAL ARTICLE

Perinatal deaths in a Norwegian county1986–96 classified by the Nordic-Balticperinatal classification: Geographicalcontrasts as a basis for quality assessmentJAN HOLT1, INGAR NIKOLAI VOLD2, JON ØYVIND ODLAND1 AND OLAV HELGE FØRDE3

From 1Nordland Central Hospital, Bodø, 2Lofoten Hospital, Gravdal, and 3the Institute for Community Medicine, University ofTromsø, Tromsø, Norway

Acta Obstet Gynecol Scand 2000; 79: 107–112. C Acta Obstet Gynecol Scand 2000

Background. Quality assessment of perinatal care can be carried out by classifying perinataldeaths. In the following we have analyzed the geographical contrasts in perinatal deaths accord-ing to the Nordic-Baltic perinatal death classification in a sparsely populated Norwegiancounty.Material and methods. All stillbirths (Ø28 weeks of gestation) and neonatal deaths (gestationalage Ø22 weeks; death Æ28 days) in 1986–96 from Nordland county (240 000 inhabitants) wereclassified. For comparison the county was geographically divided into six general local hospitalareas and one central hospital area.Results. The classification showed a well acceptable inter and intra observer variation. One hun-dred and seventy-one stillbirths and 155 neonatal deaths were analyzed. The death rate (pr 1000births) for single, non-malformed, antenatal stillbirths was higher in the central hospital areathan in the local hospital areas (3.22 vs. 2.02). The death rate for extreme preterm infants (22–27 weeks of gestation) was on the other hand higher in the local hospital areas (2.45 vs. 1.05).One of the general local hospital areas was singled out with an especially high neonatal deathrate among extreme preterm infants. This was to some extent explained by the death of extremepreterm twins and triplets.Conclusion. The Nordic-Baltic perinatal death classification system is a consistent and repro-ducible tool also for studying perinatal death in restricted geographical areas. The observedcontrasts in perinatal deaths were used as basis for programs aimed at improving perinatal care.The observation of an unexplained increased number of antenatal stillbirths in the central hos-pital area resulted in a program for prospective recording and better characterization of the pla-centa and umbilical cord. Proposals for a better antenatal program preventing extreme pretermbirth of twins for the whole county has been launched. In utero transfer to a hospital with aneonatal intensive care unit seems crucial in improving the prognosis for these infants.

Key words: classification; geographical area; perinatal care; perinatal death

Submitted 17 May, 1999Accepted 22 August, 1999

Stillbirths and neonatal deaths are associated withfetal, maternal and obstetrical factors as well as thequality of antenatal and perinatal care. Various

Abbreviations:NICU: neonatal intensive care unit; IUGR: intra uterinegrowth retardation; OR: odds ratio; CI: confidence interval;GA: gestational age.

C Acta Obstet Gynecol Scand 79 (2000)

ways of classifying perinatal deaths have been advo-cated, most of which propose a hierarchical modelbased on categories of single underlying causes (1–4). Others (5–6) have used the audit concept (7) ex-ploring avoidable factors, and recently a computerbased classification has been published (8).

The classification of stillbirths and neonatal

108 J. Holt et al.

deaths should relate to possible improvement ofperinatal care. The Nordic-Baltic perinatal deathclassification (9), which has been used mainly forinternational comparisons (9, 10), uses only sixvariables to group the stillbirths and neonataldeaths into 13 categories. From a quality assess-ment perspective, some of the categories have aprevention potential.

In a county with a scattered but homogenouspopulation and long distances there might be areaswith different perinatal health care service. As aperinatal death is a rather rare event – is it possibleto trace differences between populations from vari-ous hospital catchment areas and search for fac-tors in the care that may be amenable to preventiveefforts?

As a part of quality assessment we explored thedistribution of 326 perinatal deaths according tohospital catchment area, using the Nordic-Balticperinatal death classification. The objective of thisstudy was to explore if this classification is suitablefor use in a county with a scattered population.

Materials and methods

Geographic area and population

Nordland county has 240 000 inhabitants and islocated within the Arctic circle with one centralhospital, six general local hospitals and three mat-ernity homes. There are 1200–1300 deliveries an-nually at the central hospital and almost 2000 atthe other units. The central hospital has an obstet-rical department for referral from the other hospi-tals and a neonatal intensive care unit (NICU).

Although improving in the last few years, ratherfew obstetricians have been attending the generallocal hospitals, and we assumed that there mightbe some differences in antenatal and perinatal careaccording to where patients lived. For analyticalpurposes we therefore divided the county geo-graphically into a central hospital area and six lo-cal hospital areas. The county consists of 45 com-munities, 10 in the central hospital area. As ap-proximately 50% of all new-borns from threecommunities in the local hospital areas were bornat the central hospital, these three communities arein the present analysis included in the central hos-pital catchment area. Each case was grouped ac-cording to the maternal permanent address.

Since the latter part of 1985 a perinatal commit-tee has consecutively gathered information con-cerning stillbirths and neonatal deaths of residentsfrom Nordland county. To these data were addedinformation from the Medical Birth Registry ofNorway concerning deaths outside the county. Asthe last completed year of register data was 1996,we selected the study period to be 1986–96.

C Acta Obstet Gynecol Scand 79 (2000)

Study protocol

Stillbirths of gestational age more than 27 weeksand neonatal deaths (Æ28 days) were studied.Neonatal deaths before 22 completed weeks wereexcluded. Data for each case were gathered fromthe perinatal committee annals, hospital records,antenatal care recordings, autopsy reports and, ina few instances, by interviewing health personnelinvolved. To ascertain that every case had beentraced, information was checked against listsfrom the Medical Birth Registry, delivery proto-cols at the delivery units and recordings from theNICU. Information on missing cases was re-quested.

The Nordic-Baltic perinatal death classificationwas applied without using category IV (Antenataldeath; before 28 weeks of gestation) and categoryVII (Intrapartum death; after admission; before 28weeks of gestation). Stillbirths with gestational ageof less than 28 weeks were excluded because thenumber of these infants are not calculated in theperinatal death rate in Norway, and thus every casemay not be reported to the Medical Birth Registry.The Nordic-Baltic perinatal classification is basedon six variables: Time of death in relation to deliv-ery and admission, fetal malformation, single vs.multiple births, growth retardation, gestational ageand Apgar score at five minutes.

Infants with major malformations, metabolicanomalies or chromosomal disorders were classi-fied as fetal malformation.

Routinely, a second-trimester ultrasound exami-nation was used to estimate the gestational age. Incases where this examination was not performed(approximately 10%), the last menstrual periodrule was used. To decide whether or not an infantwas large for gestational age or intrauterinegrowth retarded (IUGR) (birthweight∫2s.d.), ul-trasound-estimated fetal growth curves were used(11). For single growth-retarded fetuses an ap-proximate time of death was recorded to determineif this might have altered the classification.

The deaths were classified by two obstetricians(INV, JØO) separately. For each case the obstetri-cians were given information regarding the sixvariables, place of death and whether or not fetalheart sounds could be traced on admission. Theycarried out the classification without knowing thematernal place of living and without any other in-formation about the classification system than areprint of the original paper (9). Using the distrib-uted perinatal death classification, the two obstet-ricians agreed on 304 cases (kappaΩ0.92). The co-ordinator (JH) revised the cases after the firstclassification and the guidelines for classificationwere emphasised:

Perinatal deaths in various hospital areas 109

Table I. Distribution of relevant characteristics among 171 stillbirths and 155neonatal deaths in Nordland county 1986–96

Stillbirths Neonatal deaths(nΩ171) (nΩ155)

n (%) n (%)

GA estimated by ultrasound 151 (88) 125 (81)Gender: female 75 (44) 70 (45)Congenital abnormalities 12 (7) 35 (23)Multiple births 10 (6) 25 (16)Necropsy performed 106 (62) 93 (60)

Birthplace– Central hospital 79 (46) 86 (55)– Local hospital 74 (43) 47 (30)– At home 3 (2) 3 (2)– Nursing homes 2 (1)– During transport 1 (0.6)– Outside the county 15 (9) 16 (10)

GA: gestational age.

– Intrapartum death should be considered onlyif it happened after admission.

– Antenatal death in hospital is classified as cat-egory II, III or V.

– Antenatal death of a stillborn growth-retardedtwin or triplet should be classified as category V(Antenatal death; multiple pregnancy).

After the classification system had been re-em-phasised the two obstetricians agreed in 325 outof 326 cases (kappaΩ0.99). For one case the finalcategory was decided after a common discussion.Fifty cases were classified twice with six monthsinterval to calculate the intraobserver variability.When classifying 50 deaths twice, the two obstetri-cians reproduced their earlier classification in 47and 49 cases (kappaΩ0.92 and 0.97).

The trial protocol was approved by the RegionalEthic Committee.

Table II. Stillbirths and neonatal deaths in Nordland county 1986–96 by the Nordic-Baltic perinatal death classification according to hospital catchment area

Local Centralhospital hospital Odds ratio

Category area area (95% CI)

I Fetal malformation 35 12 1.53 (0.79–2.94)II Antenatal death; single growth retarded fetus Ø28 weeks of gestation 35 15 1.22 (0.67–2.24)III Antenatal death; single fetus Ø28 weeks of gestation 48 40 0.63 (0.4–0.95)*V Antenatal death; multiple pregnancy 6 2 1.57 (0.32–7.78)VI Intrapartum death; after admission; Ø28 weeks of gestation 7 5 0.61 (0.21–1.81)VIII Neonatal death. 28–33 weeks of gestation. Apgar score ±6 after 5 min 12 4 1.57 (0.51–4.87)IX Neonatal death. 28–33 weeks of gestation. Apgar score ∞7 after 5 min 6 5 0.43 (0.13–1.43)X Neonatal death. Ø34 weeks of gestation. Apgar score ±6 after 5 min 5 1 2.62 (0.31–22.4)XI Neonatal death. Ø34 weeks of gestation. Apgar score ∞7 after 5 min 9 6 0.78 (0.28–2.21)XII Neonatal death; before 28 weeks of gestation 58 13 2.34 (1.3–4.3)**XIII Unclassified 1 1

* pΩ0.037, ** pΩ0.00064.

C Acta Obstet Gynecol Scand 79 (2000)

Statistical analysis

Medians with interquartiles in addition to therange are given for demographic data. To compareproportions, the chi-square test or Fisher‘s exacttest was used and comparisons between groupswere made using the Mann-Whitney U-test or theKruskal-Wallis analysis. The kappa ratio test wasused to estimate the inter- and intraobserver vari-ability. A p value of 0.05 or less was consideredsignificant. The statistical analyses were performedusing the statistical packages GraphPad Prism 2.0(GraphPad Software Inc., San Diego, CA, USA)and MedCalc (MedCalc Software, Mariakerke,Belgium).

Results

In total, 326 stillbirths and neonatal deaths oc-curred among 36072 births. Two hundred and sev-enty-three cases (83.7%) were noticed both by theperinatal committee and by the Medical BirthRegistry, 46 cases (14.1%) recorded by the MedicalBirth Registry alone and seven (2.1%) merely bythe perinatal committee. Twenty-eight (61%) of thecases not reported to the perinatal committee wereborn outside the county.

Table I gives some characteristics of thestillbirths and the neonatal death groups. The me-dian birthweight of the stillbirths was 2225 g (min240, 25th centile 1500, 75th centile 2975, max 5100)and of the neonatal deaths 1060 g (min 340, 25th

centile 718, 75th centile 2410, max 4645). The me-dian gestational age was 36 weeks (min 28, 25th

centile 32, 75th centile 39, max 43) and 28 weeksrespectively (min 22, 25th centile 25, 75th centile 36,max 42). Twenty-three infants died in the neonatalperiod after/at the age of eight days or more

110 J. Holt et al.

(major anomalies: eight; discharged: four; still inNICU (without major anomalies): eleven).

The number of deaths in each of the 13 cate-gories according to hospital catchment area (localhospital areas vs. central hospital area) are shownin Table II.

Major malformations were recorded in 47cases – 14.4% of the total number of stillbirths andneonatal deaths (heart defects 14; chromosomaldisorders seven; neural tube defects four; kidneyfour; intestinal four; multiple four; skeletal three;diaphragm two; other five).

The number of single growth-retarded stillbirths(II) was somewhat higher in the local hospitalsgroup compared with the central hospital group,but the difference was not significant (pΩ0.61). For47 (94%) of these fetuses it was possible to estimatean approximate time of death. The median time ofdeath before delivery was 3 days, and using thetime of death as a criterion for IUGR, three fetus-es (two from the local hospital areas and one fromthe central hospital area) were not consideredIUGR and thus might have been classified in cat-egory III. This correction did not alter the signifi-cant difference between the two groups of ante-natal stillbirths in category III (pΩ0.04). In 17cases the gestational age was 28–31 weeks, in 22cases 32–37 weeks and 11 cases had a gestationalage of 38 weeks or more. In 38 cases growth re-striction was not recorded before delivery.

Thirteen infants were characterized as large forgestational age (stillbirths six; neonatal deathsseven).

The proportion of antenatal stillbirths withoutmalformations or IUGR (III), was significantlylarger in the central hospital area compared withthe local hospital areas as a whole. When com-paring the stillbirths in the two areas, no signifi-cant differences could be traced concerning parity,smoking habits, gestational age, birthweight orgender. The major underlying causes for thesedeaths (placental abruption, impaired circulationin placenta/umbilical cord, unexplained) wereevenly distributed in the two areas. Ten stillbirths

Table III. Category XII neonatal deaths ∞28 weeks of gestational age, accord-ing to hospital catchment area (Nordland county study 1986–96)

Local Local Centralhospital hospital hospital

No. 1 area No. 2–6 areas area

Total number of births 4756 18924 12392Neonatal deaths; 22–27

weeks of gestation 21 37 13Relative risk of neonatal

death (95% CI) 4.2 (2.2–8.0) 1.9 (1.0–3.4) 1

C Acta Obstet Gynecol Scand 79 (2000)

in this category died after arrival in hospital, eightfrom the local hospital areas and two from the cen-tral hospital area. Thus, 38 women from the cen-tral hospital area and 40 from the local hospitalareas arrived in hospital without traceable fetalheart sounds making the difference between thegroups even more substantial; OR 1.82; 95% CI1.17, 2.84.

Considering infants of gestational age morethan 27 weeks who died during the neonatalperiod (VIII – IX), there was no difference in thenumber of deaths from the local hospital areascompared with the central hospital area (32 vs. 16);OR 1.04; 95% CI 0.57, 1.91. For infants Ø34 weeksof gestation dying in the neonatal period (X andXI) the underlying causes of death were: Asphyxia15 (placental abruption four); sudden infant deathsyndrome four; septicemia one; peritonitis one.

The number of infants of gestational age 22–27weeks (XII), constituted 71 cases. When assessingeach local hospital area, one population (Hospital1) contributed significantly more infants to thiscategory than from the other local hospital areas(Hospital 2–6) or from the central hospital area(Table III). Considering total number of multipleconfinements in each area, place of birth (hospitalwith NICU vs. other), birthweight, gestational ageor gender there were no significant differences be-tween the groups. In the hospital 1 population,however, one set of triplets and eight dead twinswere observed as compared with four twins fromhospital 2–6 and two twins from the central hospi-tal area.

Discussion

By using the Nordic-Baltic perinatal death classi-fication to categorize stillbirths and neonataldeaths, we have observed contrasts in the perinataldeaths from various hospital catchment areas.

In doing this kind of retrospective epidemio-logical study comparing perinatal deaths from dif-ferent areas in a county, it is mandatory to makesure that all cases are included.

By using recordings from the delivery units, theNICU and the perinatal committee, and com-paring this information to the Medical BirthRegistry data, we consider the probability of miss-ing any case to be very low. For two cases, however,the data recorded were insufficient to make a re-liable classification. Although the perinatal com-mittee has access to a steady flow of informationconcerning new perinatal deaths, 14.1% of thecases were unknown to the committee. A bettersystem for reporting perinatal deaths to the com-mittee from NICUs and delivery units outside the

Perinatal deaths in various hospital areas 111

county might reduce this deficiency by approxi-mately 60%.

The death rate caused by major malformationsin this study (1.30/1000 births) is low comparedwith the rate from Denmark/Sweden in 1991 (1.62/1000 births) and from Lithuania in 1993–94 (3.28/1000 births) (10). The difference betweenDenmark/Sweden and Lithuania might be causedby a higher incidence of neural tube defects in Li-thuania. As in the study from Denmark/Swedenwe recorded less neural tube defects than cardiacmalformations and chromosomal disorders. Therate from major malformations decreased in Ice-land from 3.2/1000 births in 1976–80 to 2.2/1000births in 1981–85, mainly brought about by a re-duction in central nervous malformations (4).Antenatal ultrasound screening was in our studyperformed in 85% of the cases (missing infor-mation 5%) and might have led to termination ofpregnancies and thus a lower major malformationrate.

To reveal IUGR is a major task for the antenatalcare team, and we hypothesized that scarcity ofgynecologists in the local hospital areas might leadto more single birth growth retarded stillbirthsfrom these areas. In classifying antenatal stillbirthswe used ultrasound-produced growth curves asthese might uncover more growth retarded fetuses(12). It was, however, not possible to trace signifi-cantly more growth restricted stillbirths from thelocal hospital areas. Smoking may cause growthretardation. When comparing category II caseswith the sum of the other classes the proportion ofcategory-II-women reporting smoking in preg-nancy was 34% vs. 36% in the other classes (miss-ing information in 24% in both groups). The habitof smoking is not supposed to differ between thehospital areas. Another explanation might be thefact that inspite of a thorough antenatal care pro-gram it has been possible to trace only 7–14% ofthe growth retarded infants before delivery (13,14). Only 24% of class II stillbirths were, in ourstudy, recognized as IUGR before delivery. Asmany of these infants have a rather low gestationalage, it will perhaps never be easy to recognize theseinfants before death unless a more sensitive screen-ing test for early diagnosis of IUGR is developed.

There were more non-malformed and non-growth retarded antenatal deaths (III) living in thecentral hospital area compared with the groups ofstillbirths from the other hospital areas. Inspite ofa different antenatal care program in Lithuaniaand Denmark/Sweden (10) there was no differencein the rate of class-III stillbirths between thesecountries. The quality of the antenatal care withinour county is not likely to cause the unexpecteddifference recorded, and the statistical differences

C Acta Obstet Gynecol Scand 79 (2000)

between the groups might be explained by randomvariation (Type I error). The increased number ofwomen from the central hospital area arriving inthe delivery unit without traceable fetal heart beatshave, however, made a program for prospective re-cording and better characterization of the placentaand umbilical cord to uncover suboptimal care orsome unknown factors essential.

The intrapartum death rate may be a quality in-dicator for the management of labor, and intrapar-tum stillbirths might be prevented given an opti-mal care. The number of intrapartum stillbirths(VI) in the present study is small, but the rate(0.33/1000 births) is in the same magnitude as re-corded for Denmark/Sweden (0.27/1000 births),but lower than the recorded rate in Lithuania(1.02/1000 births) (10). Some of the infants dyingin the neonatal period might suffer from subopti-mal care during childbirth. Maybe more could begained for infants Ø34 weeks of gestation (XπXI)if fetal asphyxia before and during delivery couldbe prevented.

The group of extreme preterm infants dying inthe neonatal period (XII) constituted 21.8% of thestudy group. Early neonatal death rate in this cat-egory (1.8/1000 births) might be compared withthe rate from Lithuania (1.91/1000 births), but ishigher than the rate in Denmark/Sweden (0.93/1000 births) (10). Comparisons of these groups be-tween different populations may be difficult be-cause the definition of life at birth may be differentfrom population to population and the number ofliveborne extreme preterm infants may increasewith time as the rate of stillborn decreases ( 15).In this study the number of category XII infantswas significantly higher from the local hospitalareas than from the central hospital area. Afurther analysis of the numbers disclosed that al-most 30% of the cases stem from one hospitalpopulation with only 13% of the total childbirths.The death rate among extreme preterm infants inthis hospital area was 4.2 times higher comparedwith the rate from the central hospital area. Thedeath rate was generally higher in the local hospi-tal areas as a whole, and altogether we estimatedan excess of 33 neonatal deaths from the local hos-pital areas. An excess of twins/triplets from thehospital 1 area may to some extent explain thisdifference. A better antenatal care for multiplepregnancies especially in hospital 1 area has pre-viously been recommended. Transfer to the centralhospital of threatening preterm delivery at 23weeks of gestation or more has been advised (16).Early recognition of high-risk cases and timely ad-mission to a maternity hospital with a neonatalintensive care unit is mandatory to improve peri-natal results.

112 J. Holt et al.

The Nordic-Baltic perinatal death classificationwas easy to use and the inter- and intraobservervariability was negligible. When used in a countywith few perinatal deaths per year the classificationsystem might give groups (II, VI, X, XI, XII) ofcases with a potential for prevention. If groups VI,X and XI are added to 33 infants from group XII,the total number of deaths with a potential for pre-vention is 66 (20%). This constitutes six cases an-nually in our county or even more if the pro-portion of antenatally recognized IUGR fetuses isincreased.

Studying a county population has given somepossible clues for improving the perinatal care: Anincreased unexplained number of women from thecentral hospital area arriving in the obstetrical de-partment without fetal heart sounds and too manyextreme preterm infants are observed in the localhospital areas, especially from one general localhospital population. To explore these differences,further studies are, however, needed.

Acknowledgment

We want to thank the Medical Birth Registry of Norway forkindly providing us with the perinatal data.

References

1. Wigglesworth JS. Monitoring perinatal mortality: a patho-physiological approach. Lancet 1980: 2: 684–6.

2. Cole SK, Hey EN, Thompson AM. Classifying perinataldeaths: an obstetric approach. Br J Obstet Gynaecol 1986;93: 1204–12.

3. Hey EN, Lloyd DJ, Wigglesworth JS. Classifying perinataldeaths: fetal and neonatal factors. Br J Obstet Gynaecol1986; 93: 1213–23.

4. Georgsdottir I, Geirsson RT, Johansson JH, Biering G,Snædal G. Classification of perinatal and late neonatal

C Acta Obstet Gynecol Scand 79 (2000)

deaths in Iceland. Acta Obstet Gynecol Scand 1989; 68:101–8.

5. Larssen KE, Bakketeig LS, Bergsjø P et al. Perinatal auditin Norway, Norwegian Institute for Hospital Research.Trondheim, 1980.

6. Pattinson RC, Makin JD, Shaw A, Delport SD. The valueof incorporating avoidable factors into perinatal audits. SAfr Med J 1995; 85: 145–7.

7. Dunn PM, McIlwaine G. Perinatal audit. A report pro-duced for the European Association of Perinatal Medicine.Prenat Neonat Med 1996; 1: 160–94.

8. Winbo IGB, Serenius FH, Dahlquist GG, Kallen BAJ.NICE, a new cause of death classification for stillbirths andneonatal deaths. Int J Epidemiol 1998; 27: 499–504.

9. Langhoff-Roos J, Borch-Christensen H, Larsen S, Lind-berg B, Wennergren M. Potentially avoidable perinataldeaths in Denmark and Sweden 1991. Acta Obstet GynecolScand 1996; 75: 820–5.

10. Langhoff-Roos J, Larsen S, Basys V, Lindmark G, Badoky-note M. Potentially avoidable perinatal deaths in Denmark,Sweden and Lithuania as classified by the Nordic-Balticclassification. Br J Obstet Gynaecol 1998; 105: 1189–94.

11. Marsal K, Persson P-H, Larsen T, Selbing A, Sultan B.Intrauterine growth curves based on ultrasonically esti-mated fetal weights. Acta Paediatr 1996; 85: 843–8.

12. Gardosi J, Mul T, Mongelli M, Fagan D. Analysis of birth-weight and gestational age in antepartum stillbirths. Br JObstet Gynaecol 1998; 105: 524–30.

13. Langhoff-Roos J, Lindmark G. Obstetric interventions andperinatal asphyxia in growth retarded term infants. ActaObstet Gynecol Scand 1997; 76 (Suppl 165): 39–43.

14. Backe B, Nakling J. Effectiveness of antenatal care: a popu-lation based study. Br J Obstet Gynaecol 1993; 100: 727–32.

15. Holt J, Weidle B, Kaaresen PI, Fundingsrud HP, Dahl LB.Very low birthweight infants: outcome in a sub-Arcticpopulation. Acta Pædiatr 1998; 87: 446–51.

16. Holt J, Fagerli I. Air transport of the sick newborn infant:audit from a sparsely populated county in Norway. ActaPædiatr 1999; 88: 66–71.

Address for correspondence:

Jan Holt, M.D.Department of Pediatrics,Nordland Central Hospital8017 BodøNorway