Impact of the Red River Catastrophic Flood on Women Giving Birth in North Dakota, 1994–2000
- 342 Downloads
To document changes in birth rates, birth outcomes, and pregnancy risk factors among women giving birth after the 1997 Red River flood in North Dakota. We analyzed detailed county-level birth files pre-disaster (1994–1996) and post-disaster (1997–2000) in North Dakota. Crude birth rates and adjusted fertility rates were calculated. The demographic and pregnancy risk factors were described among women delivering singleton births. Logistic regression was conducted to examine associations between the disaster and low birth weight (<2,500 g), preterm birth (<37 weeks), and small for gestational age infants adjusting for confounders. The crude birth rate and direct-adjusted fertility rate decreased significantly after the disaster in North Dakota. The proportion of women giving birth who were older, non-white, unmarried, and had a higher education increased. Compared to pre-disaster, there were significant increases in the following maternal measures after the disaster: any medical risks (5.1–7.1%), anemia (0.7–1.1%), acute or chronic lung disease (0.4–0.5%), eclampsia (0.3–2.1%), and uterine bleeding (0.3–0.4%). In addition, there was a significant increase in births that were low birth weight (OR 1.11, 95% CI 1.03–1.21) and preterm (OR 1.09, 95% CI 1.03–1.16) after adjusting for maternal characteristics and smoking. Following the flood, there was an increase in medical risks, low birth weight, and preterm delivery among women giving birth in North Dakota. Further research that examines birth outcomes of women following a catastrophic disaster is warranted.
KeywordsPregnancy Natural disaster Birth outcomes
Disasters, both man-made and natural, can result in substantial disruption of individual and community resources, infrastructure, and health services [1, 2, 3]. As a result, disasters can have major negative impacts on the well-being and mental health of survivors [4, 5, 6]. Psychosocial stress, not specific to disasters, has been well-documented as a major risk factor for adverse pregnancy outcomes such as preterm birth and low birth weight [7, 8].
Despite numerous studies, limited information exists on the impact of disasters on pregnancy outcomes at a population level . Demographic studies have documented changes in fertility rates after man-made disasters, such as the Oklahoma City bombing or the New York City blackout, and natural disasters such as Hurricane Hugo [10, 11, 12]. Previous studies exploring the impact of disasters on pregnancy have found associations with adverse pregnancy outcomes, such as preterm delivery, low birth weight infants, small for gestational-age infants, and neural tube defects [13, 14, 15, 16, 17, 18, 19, 20, 21]. However, most of these studies examined outcomes from within a couple of days to less than 1 year after a disaster. To our knowledge, there has been little research published describing the maternal characteristics of women delivering infants before and after a disaster and the impact of the disaster on reproductive health outcomes several years after the disaster at a population level.
To this end, we explored the impact of the 1997 Red River flood on pregnant women and birth outcomes among North Dakota residents. This flood was the culmination of severe winter and spring storms followed by the spring thaw and consequent flooding of the Red River and other rivers, causing considerable damage and affected North Dakota and the surrounding states . Two Presidential Major Disaster Declarations were issued in 1997 for North Dakota, one in January for severe winter storms and blizzards, and one in April for severe storms and flooding . The entire state was affected by the resulting disaster, and statewide flood recovery efforts continued for the entire year. Furthermore in Grand Forks, located in one of the most severely affected counties, community disruption continued for 2–3 years. More than $225 million of federal and state aid and grants were provided to the disaster-affected area . Six counties in North Dakota (Cass, Grand Forks, Pembina, Richland, Traill, and Walsh) suffered the most damage from the flood; however, the entire state, including 53 counties and the Indian tribes living in North Dakota, received FEMA assistance.
The purpose of our study was to document changes in birth rates, birth outcomes, and pregnancy risk factors among North Dakota women who gave birth before and after the Red River disaster using population-based natality files (1994–2000). In addition, we explored whether there were differential impacts on women living in the six counties that suffered the most flood damage compared to women living in the rest of the state. The results of this study can help researchers and health care practitioners to better understand the impact of natural disasters on pregnant women living in a geographic area, and may inform public health action to appropriately care for these women and their infants.
We analyzed data from the 1994–2000 birth files of the state of North Dakota with county-level identifiers for women giving birth and residing in the state. We utilized an ecological study design in the recognition that environmental determinants may operate at a geographic level on the etiology of disease risks and poor pregnancy outcomes . We compared births from the six counties most severely affected (Cass, Grand Forks, Pembina, Richland, Traill, and Walsh) with those in the rest of the state. We used the Federal Information Processing Standards county codes used by the US Census to identify the natality records of the six counties. In defining pre- and post-disaster time periods, we considered that the first disaster declaration occurred in January 1997 and the second occurred in April 1997. Thus, we defined pre-disaster from 1994 to 1996 and post-disaster from 1997 to 2000; sensitivity analyses were conducted to test our categorizations.
Birth rate measures included in the analysis were crude birth rates (births per 1,000 population) and adjusted fertility rates (AFR) (births per 1,000 women aged 15–44 years). Fertility rates were direct-age adjusted using the 1990 Census population of North Dakota and compared to the AFR of the US; however, county level population estimates by age grouping were not publicly available for adjustment.
Birth outcomes included low birth weight, preterm delivery, and infants that were small for gestational age. Low birth weight was defined as infant weight of less than 2,500 g at birth. Preterm delivery was defined as gestation of less than 37 weeks based on the last menstrual period or, if menstrual information was missing, the clinical estimate of gestation. Small for gestational age (SGA) was defined as the lowest 10th percentile of birth weight by sex and race .
Pregnancy risk factors included in the analysis were maternal age, maternal race/ethnicity, marital status, education, parity, trimester of initiation of prenatal care, the Kotelchuck Index for adequacy of prenatal care , tobacco use, and alcohol use. Medical risks explored were anemia, acute or chronic lung disease, pregnancy-associated hypertension, eclampsia, and uterine bleeding; also, the number of medical risk factors per woman was examined.
Analyses were conducted using SAS version 9.1.3 (SAS, Cary, NC). We analyzed all measures by the entire state, by the six counties most affected, and by the rest of the state. Crude birth rates, adjusted fertility rates, and percent of births resulting in a poor birth outcome were calculated by year and by pre- or post-disaster occurrence. We restricted analysis of birth outcomes to women who delivered a singleton infant. Women delivering singleton infants pre- and post-disaster were described by maternal characteristics and pregnancy risk factors. Chi-square testing with significance level at P < 0.05 was used to assess differences in proportions. Logistic regression was conducted to examine associations of low birth weight, preterm delivery, and SGA status with the disaster, adjusting for maternal age, race/ethnicity, education, marital status, smoking, and trimester of entry into prenatal care. We ran three separate models for each of the three birth outcomes, one for the entire state, one for the six counties most affected, and one for the rest of the state.
Birth rate and adjusted fertility rate in North Dakota, 1994–2000
Adjusted fertility ratee
Rest of state
Percent birth outcomes among women who delivered a singleton infant by year and pre/post disaster in North Dakota, 1994–2000
Total singleton births
Low birth weight (%)
Total singleton births
Low birth weight (%)
Rest of state
Total singleton births
Low birth weight (%)
Maternal characteristics and risks of women who delivered a singleton infant pre- (1994–1996) and post- (1997–2000) Red River Disaster, North Dakota
Rest of state
n = 24,683
n = 30,661
n = 9,561
n = 11,959
n = 15,122
n = 18,702
Second or later birth
Initiation of prenatal care (PNC)
3rd Trimester or no care
Adequacy of PNCb
Acute or chronic lung disease
Number of medical risk factors
0 medical risk
1 or more medical risks
Patterns of maternal characteristics and risks were similar in the six counties and the rest of the state. Some notable differences, however, were that in the six counties there was a significant increase in the percentage of women who used tobacco during pregnancy and who had pregnancy-associated hypertension.
Adjusted odds ratio of low birth weight, preterm, and small for gestational age (SGA) births post-disaster compared to pre-disaster among women with singleton births, North Dakota (1994–1996; 1997–2000)
Entire state (n = 55,344)
6 countiesa (n = 21,520)
Rest of state (n = 33,824)
Crude (95% CI)
Adjustedb (95% CI)
Crude (95% CI)
Adjustedb (95% CI)
Crude (95% CI)
Adjustedb (95% CI)
Low birth weight
In our study, we found that the 1997 Red River flood had significant negative associations on North Dakota women giving birth in the time period following the disaster compared to women giving birth before the disaster. There were increases in the proportion of women having more than one medical risk factor during pregnancy. Although the data show an increase in eclampsia among women giving birth statewide, it is likely that this upsurge reflects increases in women exhibiting preeclampsia. Additionally, we observed significant increases in the proportion of women reporting anemia, acute or chronic lung disease, and uterine bleeding.
The large increase in eclampsia among pregnant women after the Red River flood very likely is misclassification that includes both mild and severe preeclampsia. Coding of these conditions varies greatly in accuracy . However, a substantial rise in preeclampsia is likely to occur among pregnant women after a catastrophic disaster. Among women not affected by disaster, preeclampsia has been associated with depression and anxiety among Peruvian and Finnish women [28, 29], but in other studies this association was not supported . However, Israeli Jewish women who experienced prolonged high stress related to terrorist attacks had significantly higher systolic and diastolic blood pressure when compared to women with medium or low stress . Other disaster-related factors such as high physical demands during recovery could contribute to preeclampsia; for example, stress related to physical work has been associated with preeclampsia among non-disaster-affected women in the United States and Nigeria [30, 32]. Preeclampsia can lead to maternal complications during pregnancy, preterm delivery and growth restriction among infants, and long-term cardiovascular morbidity among mothers . Because a small proportion of women will normally develop preeclampsia, this effect needs to be studied further to determine whether it occurs consistently and among different groups after a catastrophic disaster.
Previous studies have found an association between natural disasters and poor pregnancy outcomes, specifically low birth weight and preterm delivery, though these studies did not look at population impact several years afterwards. After a flood that occurred in Poland in 1997, women who were injured during the flood were more likely to have a preterm delivery . Several studies of earthquakes showed an increase in low birth weight infants  and preterm delivery , with one study indicating that women may be more sensitive to stress in the early stages of pregnancy . A study among Hurricane Katrina survivors indicated that women who experienced specific severe events, e.g., having a loved one die, had increased risk of low birth weight, but not of preterm delivery .
A recent study of birth certificates 12 months before and after Hurricane Katrina found that total births declined in the 14 hardest hit counties . Large decreases were observed in preterm and very low birth weight rates for counties in Louisiana and large increases of preterm birth were observed in selected counties in Alabama. The authors concluded that their findings may reflect population shifts in the disaster-affected areas. For the Red River flood, there was a minor increase in total population in two counties (Cass and Richland) and minor reduction in the rest of the state from 1997–1998, which are similar to previous years . Overall, there is no evidence of significant migration in the state that could explain the differences observed in our study.
Stress and depression during pregnancy have been associated with poor pregnancy outcomes [7, 8]; therefore, the additional stress and accompanying depression of a disaster may exacerbate the effects on birth outcomes. These effects on birth weight disappeared when controlled for smoking , but our models showed an increase in the odds of low birth weight even after controlling for smoking. We cannot explain the differential impact of the disaster upon increased odds of low birth weight in the six counties most directly affected and increased odds of preterm delivery in the rest of the state. We postulate that these two outcomes may be affected differently by psychosocial stress , but our data do not describe or quantify stress and depression among women affected by the Red River flood.
Our findings of an increase in smoking among women giving birth in the six counties directly affected by the disaster mirror those of other studies where smoking and substance use increase after a disaster [36, 37, 38]. However, we found a significant decline in alcohol use after the Red River flood, which needs to be explored further. One implication is that public health practitioners should be prepared to provide disaster counseling and offer effective prenatal smoking cessation to pregnant women affected by disaster.
Births decreased significantly from 13 births per 1,000 before the flood to 12 births per 1,000 after the flood. This finding is in contrast to two studies of birth rates following natural disasters. Births increased among survivors of the 1997 flood in Poland . During the year following Hurricane Hugo in 1989, birth rates in South Carolina significantly increased, and the effects were pronounced in all counties directly affected . Further analysis could explore whether changes in birth rates differed by subpopulations (e.g., race/ethnicity, age, or socioeconomic status).
The study is subject to several limitations. Our data, based on public health records, can not ascertain disaster-affected population, or pinpoint whether experiencing the event had negative pregnancy outcomes. Because North Dakota has a low number of births per year, it may be difficult to discern the impact of medical risk on poor birth outcomes. In addition, limitations exist in the use of birth certificate data for identifying medical risks , and other data sources such as medical charts could be explored for future work on this topic, such as validating reporting of preeclampsia and eclampsia. It is unclear whether health care providers were more likely to document problems during pregnancy because of the disaster, which could reflect a reporting bias. Finally, the combined disaster spanned most of 1997 but the effect of low birth weight and preterm delivery still existed when we excluded 1997 births from the analysis. Without a comparison group, it is not possible to know whether the rise of low birth weight and preterm births resulted from the direct effect of the disaster or is part of the secular trend of these outcomes rising in the United States [41, 42]. However, our findings are similar to other studies discussed earlier.
Our study results are suggestive that there may be effects of a natural disaster on the population of women giving birth in an area several years after a natural disaster, particularly one that is catastrophic. In particular, health care providers need to be diligent in monitoring preeclampsia. However, it should be cautioned that healthcare systems and providers also suffer from the effects of disaster. Following, Hurricane Katrina, healthcare facilities suffered damages or were destroyed, and providers experienced major financial and systematic barriers in trying to re-establish their practices . As the result, many providers left the Gulf Coast region to resettle elsewhere [44, 45].
This study may have implications for areas that have recently experienced massive flooding, such as the Midwest flooding in several states in 2008 and the 2009 Red River flood in North Dakota . Additionally, more assessments are needed to document the impact of disasters on women’s reproductive health needs and to ensure that their short- and long-term needs are being addressed after a natural disaster.
- 22.Federal Emergency Management Agency. Grand Forks 1997 flood recovery: Facts and statistics. http://www.fema.gov/hazard/archive/grandforks/statistics.shtm.
- 23.Federal Emergency Management Agency. Grand Forks 1997 flood recovery, facts and statistics: Chronology. http://www.fema.gov/pdf/hazard/archive/grandforks/chronology.pdf.
- 34.Hamilton, B. E., Sutton, P. D., Mathews, T. J., Martin, J. A., & Ventura, S. J. (2009). The effect of Hurrican Katrina: Births in the U.S. Gulf Coast region, before, and after the storm. National vital statistics reports; vol 58 no 2. Hyattsville, MD: National Center for Health Statistics.Google Scholar
- 35.McMurry, M. (1999). Working paper: Demographic aftermath of the 1997 Red River Valley Floods. State Demographic Center. http://www.demography.state.mn.us/DownloadFiles/pdf/Demographicaftermathof1997redrivervalleyfloods1999.pdf.
- 37.Substance Abuse and Mental Health Services Administration. The NSDUH Report: Impact of Hurricanes Katrina and Rita on substance use and mental health. http://www.oas.samhsa.gov/2k8/katrina/katrina.htm.
- 40.Lydon-Rochelle, M. T., Holt, V. L., Cardenas, V., Nelson, J. C., Easterling, T. R., Gardella, C., et al. (2005). The reporting of pre-existing maternal medical conditions and complications of pregnancy on birth certificates and in hospital discharge data. American Journal of Obstetrics and Gynecology, 193(1), 125–134.PubMedCrossRefGoogle Scholar
- 41.Centers for Disease Control, Prevention. (2002). Infant mortality and low birth weight among black and white infants—United States, 1980–2000. MMWR Morbidity and Mortality Weekly Report, 51(27), 589–592.Google Scholar
- 44.Adams, D. (2005). Katrina displaces 6, 000 doctors. AMNews, 48, 1.Google Scholar
- 46.Federal Emergency Management Agency. North Dakota severe storms and flooding. http://www.fema.gov/news/event.fema?id=11268.