Underreporting of maternal deaths on death certificates and the magnitude of the problem of maternal mortality

OBJECTIVES: I studied the extent to which maternal deaths are underreported on death certificates. METHODS: We collected data on maternal deaths from death certificates, linkage of death certificates with birth and fetal death records, and review of medical examiner records. RESULTS: Thirty-eight percent of maternal deaths were unreported on death certificates. Half or more deaths were unreported for women who were undelivered at the time of death, experienced a fetal death or therapeutic abortion, died more than a week after delivery, or died as a result of a cardiovascular disorder. CONCLUSIONS: The number of maternal deaths is substantially underestimated when death certificates alone are used to identify deaths, and it is unlikely that the Healthy People 2010 objective of reducing the maternal mortality rate to no more than 3.3 deaths per 100000 live births by 2010 can be achieved. Increasing numbers of births to older women and multiple-gestation pregnancies are likely to complicate efforts to reduce maternal mortality.     

Evaluation of reproductive histories constructed by linking vital records

Summary We used 1.4 million fetal death and birth certificates filed in Georgia between 1980 and 1992 to construct 369 686 chains of two or more reproductive events occurring to the same woman. We evaluated these chains using both information on the certificates and information independently collected in interviews with 1311 women. Overall, 86.6% of the chains had the expected number of events, based on the certificate's information about previous pregnancies. Seventy-nine per cent of the chains had the expected number of events based on the maternal interviews. Consistency between the observed number of events in the chain and the number expected, based either on data from the certificates or from the maternal interviews, was greatest for chains with two or three events. Mothers born in Georgia were more likely to have complete chains than mothers born elsewhere. Among the 551 391 non-linked certificates, 48.7% were the mother's first birth, 40.2% were second or higher-order births to women whose previous pregnancy occurred before 1980, and 11.1% were second or higher-order births to women whose previous pregnancy occurred after 1980. Fetal death and livebirth certificates can be linked to construct pregnancy histories with reasonably low levels of underlinkage and overlinkage.     

Contrasts in the multiple causes of stillbirth, neonatal death and postneonatal death.

A large sample of stillbirth and infant death certificates for England and Wales from 1979-81 was analysed for the frequency of appearance of maternal and fetal conditions anywhere on the certificate, not just as the underlying cause. The results suggest there is presently no need to extend the use of the new stillbirth and neonatal death certificates, introduced in 1986, to the postneonatal period. Periodic multicause analysis of the old style death certificate should be sufficient to reveal the detail of conditions incriminated in postneonatal deaths.     

Mortality among women in reproductive age in the municipality of S?o Paulo, Brazil, 1986. II. Deaths by maternal causes

In continuation to the research project on the accuracy of the certification of the underlying causes of death in women of child-bearing age (10-49), resident in the Municipality of S. Paulo, Brazil, in 1986, "original" death certificates were compared with "revised" death certificates (including additional information). The maternal mortality rate rose from 44.5 per 100,000 live births (l.b.) to 99.6 per 100,000 l.b., a high rate when compared with that of other places. When these data were compared with those of previous, similar investigations in the same city, the maternal mortality rate rose in the period 1962/4 through 1972/4 and fell in 1986. The main causes of death were: hypertension complicating pregnancy, other conditions of the mother which complicated pregnancy and puerperal complications. The need to extend the 42-day period related to the concept of maternal death, as well as the relationship between the non-maternal conditions (cancer, violence) and the gravidic-puerperal cycle are discussed.     

Mortality of women in reproductive age in the municipality of S?o Paulo (Brazil), 1986. I. Methodology and general results

The purpose of this study was to evaluate the accuracy of the death certificates of a sample of a quarter of all deaths in women of reproductive age (10-49 years) resident in the Municipality of S. Paulo, SP, Brazil, in 1986. For each death, further data were gathered by means of household interviews and from medical records and autopsy information where available. Nine hundred and fifty-three deaths were analysed, for whom there were good quality death certificates except with regard to maternal deaths an terminal respiratory diseases, the former being greatly under-reported. The official maternal mortality rate was 44.5 per 100,000 live births but the true rate was 99.6 per 100,000 live births. The three main causes of death were cardiovascular diseases, neoplasms and external causes. A great proportion of smokers was found among the deceased women (40.4%). Eleven percent of the deceased consumed large amounts of alcoholic beverages regularly.     

The risk of childbearing re-evaluated.

To determine the completeness of reporting of maternal deaths after live born deliveries in Georgia for 1975 and 1976, we matched death certificates with corresponding birth certificates for women of reproductive age. For these two years, more intensive searching led to our finding a minimum of a 27 per cent higher number of maternal deaths than that found by routine death certificate surveillance. When the delivery-death interval was not restricted to 42 days, use of the record linkage method led to a 50 per cent increase in reporting of maternal deaths. We recommend that special efforts be made to obtain more complete reporting of all pregnancy-related deaths and that completeness of reporting be periodically evaluated for all states.     

An assessment of the incidence of maternal mortality in the United States.

Recent studies in several states have found that the incidence of maternal mortality is higher than traditional vital statistics reports indicate. Since no comprehensive national study has been done to evaluate the completeness of maternal mortality ascertainment through the national vital statistics reporting system, the Centers for Disease Control (CDC) undertook such a study with the assistance of the National Center for Health Statistics and state health departments. The state health departments provided CDC with death certificates for all pregnancy-related deaths occurring during 1974-78. We reviewed and classified these certificates using both International Classification of Diseases, Adapted, Eighth Revision (ICDA-8) and International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) maternal death codes. We found that the actual incidence of maternal mortality for 1974-78 was approximately 20 per cent and 30 per cent greater than that published in national vital statistics reports using ICDA-8 and ICD-9-CM, respectively.     

Consumption of PCB-contaminated sport fish and risk of spontaneous fetal death.

Spontaneous fetal death has been observed among various mammalian species after exposure to polychlorinated biphenyls (PCBs). Our exposure-based cohort study assessed the relationship between consumption of PCB-contaminated Lake Ontario sport fish and spontaneous fetal death using 1820 multigravid fertile women from the 1990-1991 New York State Angler Cohort Study. Fish consumption data were obtained from food frequency questionnaires and history of spontaneous fetal death from live birth certificates. Analyses were stratified by number of prior pregnancies and controlled for smoking and maternal age. No significant increases in risk for fetal death were observed across four measures of exposure: a lifetime estimate of PCB exposure based on species-specific PCB levels; the number of years of fish consumption; kilograms of sport fish consumed in 1990-1991; and a lifetime estimate of kilograms eaten. A slight risk reduction was seen for women with two prior pregnancies at the highest level of PCB exposure (odds ratio = 0.36; 95% CI, 0.14-0.92) and for women with three or more prior pregnancies with increasing years of fish consumption (odds ratio = 0.97; 95% CI, 0.94-0.99). These findings suggest that consumption of PCB-contaminated sport fish does not increase the risk of spontaneous fetal death.     

Sudden infant death syndrome and reported maternal smoking during pregnancy

We investigated the effect of maternal smoking during pregnancy on the relative risk of sudden infant death syndrome (SIDS) by linking data from Georgia birth and death certificates from 1997 to 2000. We estimated the effect of misclassifying smokers as non-smokers and the effect of being misclassified on SIDS rates, and we calculated the fraction of cases caused by exposure. Of all SIDS cases, 21% were attributable to maternal smoking; among smokers, 61% of SIDS cases were attributable to maternal smoking. Maternal smoking during pregnancy is associated with a significantly increased risk of SIDS.     

Perinatal Periods of Risk: Analysis of Fetal–Infant Mortality Rates in Kansas City,Missouri

Objectives: The Perinatal Periods of Risk (PPOR) technique was used to analyze resident fetal and infant death data from Kansas City, Missouri, for the period 1998–2002. Results offer important information that can be used to develop community-based prevention strategies related to racial/ethnic disparities in infant mortality rates (IMR). Methods: The PPOR approach for fetal and infant mortality can be mapped by birthweight at delivery and time of death into four strategic prevention areas: 1) Maternal Health/Prematurity (MHP), 2) Maternal Care (MC), 3) Newborn Care (NC), and 4) Infant Health (IH). For this analysis, all fetal and infant death certificates from the metropolitan Kansas City area were linked to their birth certificates and those associated with residents of Kansas City, Missouri, proper were used to create the dataset used in this analysis. Due to the small number of fetal and infant deaths among other ethnic groups, the analysis was restricted to a comparison of the disparity of IMR between Blacks, Whites, and a national non-Hispanic white reference group. The Kitagawa formula was used to determine contribution to excess deaths from birthweight-specific mortality and birthweight distribution rates. Logistic regression techniques were used to identify risk factors for death among Black fetuses and infants with very low birthweights and also deaths due to sudden infant death syndrome (SIDS). Results: The PPOR analysis showed that of the excess deaths among black infants, when compared to a national reference group, 47% was attributable to MHP and another 29% was attributable to IH. Differences in MC and NC only accounted for 27 and 8% of the total excess deaths. During the study period, rates of sudden infant death syndrome (SIDS) were found to be significantly higher among Blacks as compared to Whites (2.12 vs. 0.81 per 1,000). An analysis of maternal characteristics for SIDS deaths among blacks using a step-wise logistic regression model, found that maternal age less than 20 years old, previous births, inadequate prenatal care, and being a Medicaid recipient were significant—adjusted odds ratios of 23.7 (95% Cl 10.48, 53.67), 8.4 (95% Cl 3.64, 19.21), 2.9 (95% Cl 1.38, 6.05) and 2.5 (95% Cl 1.04, 5.84), respectively. Conclusions: PPOR is an easy to use approach that helps focus community initiatives for improving maternal and infant health. In Kansas City, Missouri, efforts to further lower IMR in blacks can be achieved through the reduction of risk factors affecting maternal health and through maternal education to improve infant health.     

Aetiology of stillbirth: unexplored is not unexplained

OBJECTIVE: To describe the rate of and demographic factors associated with fetal postmortem investigation and to classify the cause of all fetal deaths that underwent postmortem investigation. To compare the proportion of deaths remaining unexplained after postmortem investigation with estimates derived from death certificates. METHOD: All fetal deaths in Western Australia (WA) from 1990 to 1999 were identified. These data were used to calculate postmortem rates and describe the characteristics of women consenting to postmortems. A multidisciplinary team classified the cause of all deaths that underwent postmortem investigation using the Perinatal Society of Australia and New Zealand Perinatal Death Classification System. The proportion of deaths that were unexplained was compared with estimates based on death certificates. RESULTS: Of the 1,619 fetal deaths recorded for 1990 to 1999, 49% (n=789) underwent complete postmortem investigation. Based on investigations, 22% of the 789 fetal deaths were unexplained and a further 18% were identified as having fetal growth restriction. Based on death certificates, 42% were unexplained and 65% were later explained by postmortem investigation. CONCLUSION AND IMPLICATIONS: Postmortem investigation rates are low. They reveal a cause of death for the majority of cases that are unexplained clinically. Epidemiological investigations of unexplained fetal death based on cases not subject to complete postmortem investigation may lead to inaccurate conclusions. A standardised definition for unexplained fetal deaths that distinguishes between cases with detailed investigation and those with limited or no investigation is needed.     

The effects of maternal smoking on fetal and infant mortality

Although maternal cigarette smoking has been shown to reduce the birth weight of an infant, previous findings on the relation between smoking and fetal and infant mortality have been inconsistent. This study used the largest data base ever available (360,000 birth, 2,500 fetal death, and 3,800 infant death certificates for Missouri residents during 1979-1983) to assess the impact of smoking on fetal and infant mortality. Multiple logistic regression was used to estimate the joint effects of maternal smoking, age, parity, education, marital status, and race on total mortality (infant plus fetal deaths). Compared with nonsmoking women having their first birth, women who smoked less than one pack of cigarettes per day had a 25% greater risk of mortality, and those who smoked one or more packs per day had a 56% greater risk. Among women having their second or higher birth, smokers experienced 30% greater mortality than nonsmokers, but there was no difference by amount smoked. The prevalence of smoking in this population was 30%. It was estimated that if all pregnant women stopped smoking, the number of fetal and infant deaths would be reduced by approximately 10%. The higher rate of mortality among blacks compared with whites could not be attributed to differences in smoking or the other four maternal characteristics studied. In fact, the black-white difference was greater among low-risk women (e.g., married multiparas aged 20 and over with high education) than among high-risk women (e.g., unmarried teenagers with low education).     

Accuracy of fetal death reports: comparison with data from an independent stillbirth assessment program

We evaluated the completeness and accuracy of reporting on Wisconsin fetal death report forms (FDF) through case by case comparison with data from the Wisconsin Stillbirth Service Project (WiSSP), which uses extensive protocols for etiologic investigation of stillborns. Fetal deaths are underreported: no FDF was submitted for 17.8 per cent of fetal deaths evaluated through the WiSSP. For those for whom FDF were submitted, fetal anomalies were often unrecognized or unreported: only 60 per cent of stillborns identified by the WiSSP as having fetal anomalies had any indication of the presence of such anomalies on FDF. When causes of death were classified into fetal, placental/cord, maternal/environmental, and unknown, comparison of reported underlying cause of death revealed marked inaccuracies on FDF. Placental/cord causes reported on FDF often could not be documented subsequently while, in contrast, fetal causes of death were underreported. Few accurate fetal diagnoses were present on FDF. Even among common lethal malformations misdiagnosis occurred frequently.     

Maternal exposures to hazardous waste sites and industrial facilities and risk of neural tube defects in offspring

PURPOSE: We examined the relationship between maternal proximity to hazardous waste sites and industrial facilities and neural tube defect (NTD) risk. METHODS: Texas Birth Defects Registry cases were linked with their birth or fetal death certificates; controls (without defects) were randomly selected from birth certificates. Distances from maternal addresses at delivery to National Priority List (NPL) and state superfund sites and Toxic Release Inventory (TRI) facilities were determined for 655 cases and 4368 controls. RESULTS: Living within 1 mile of an NPL or state superfund site was not related to NTD risk (adjusted odds ratio [OR] = 1.0; 95% confidence intervals [CI] = 0.6, 1.7). Living within 1 mile of a TRI facility carried a slight risk (adjusted OR = 1.2; 95% CI = 1.0, 1.5). The effect was highest among mothers 35 years and older (OR = 2.7; 95% CI = 1.4, 5.0) and among non-Hispanic white mothers (OR = 1.8; 95% CI = 1.1, 2.8). CONCLUSIONS: Hazardous waste sites posed little risk for NTDs in offspring. Close proximity to industrial facilities with chemical air emissions was associated with NTD risk in some subgroups. Further investigation is needed to determine if the effects are real or due to unresolved confounding or bias.     

Identifying Multiple Gestation Groups Using State-Level Birth and Fetal Death Certificate Data

Purpose: Birth and fetal death certificates classify individuals as twins or higher order multiples, but do not identify multiple gestation groups. As a result, multiple gestations are consistently excluded from maternal and child health research studies despite the surge in multiple births since the early 1980s and the health risks associated with them. A standardized methodology for states to identify multiple gestation groups is proposed to allow researchers to account for multiple gestations in analyses, improve the accuracy of the incidence of multiple gestations and further knowledge of the impact of multiple gestations on birth outcomes. Methods: Using 3 years of Massachusetts birth and fetal death certificate data from 1998 to 2000 (247,959 births and 1358 fetal deaths), we assigned matching multiple gestation group numbers to records with identical combinations of mother's first name, last name, date of birth, and month of delivery. To validate our methodology, we calculated plurality and compared it to plurality reported on the existing birth and fetal death data. Results: This method correctly identified 10,765 records out of 10,795 validated multiple gestation deliveries (99.8%). Our method identified 71 additional multiple gestation deliveries, which were not identified by the birth and fetal death files. This method resulted in only 4 false positives and 51 false negatives over 3 years. Conclusions: This algorithm provides much needed information on multiple gestation groupings, and as an additional benefit, improves the identification of multiple gestation deliveries. This method has proven easy to use, employs state-level data, and offers numerous new analytic opportunities. Presentations: Maternal and Child Health Epidemiology annual meeting, Clearwater, FL, December 2002. Annual meetings of the American Public Health Association, San Francisco, CA, November 2003.     

Causes of death in Huntington disease as reported on death certificates

Huntington disease (HD) is a late onset autosomal dominant neurological disorder. Two hundred fifty-three death certificates of HD-affected individuals from four midwestern states were examined to determine the completeness of reporting HD on the death certificates. Overall, 66% of death certificates indicated HD as a primary or contributory cause of death. There was significantly better reporting on more recent death certificates, but even since 1979, 16% did not report HD. The implications to those researching HD family histories and to the accuracy of mortality rates are discussed.     

Discrepancies between published data on racial classification and self-reported race: evidence from the 2002 North Carolina live birth records

OBJECTIVES: We compared data on race as reported by the mother on North Carolina birth certificates with data on race in officially reported statistics. We also determined to what extent differences in the classification of race affect measures of racial disparity in maternal and child health indicators. METHODS: We examined how data on race are collected, coded, and tabulated in North Carolina via live birth certificates, death certificates, the Behavioral Risk Factor Surveillance System (BRFSS) telephone survey, and the Central Cancer Registry case records. We showed how the data on race collected through North Carolina birth and death certificates are translated into 10 fixed racial categories designated by the National Center for Health Statistics (NCHS) for use in official vital statistics. We compared race as reported by the mother on birth certificates to racial tabulations used in the official published birth statistics. We also examined to what extent differences in the determination of race affect measures of racial disparity in maternal and child health indicators. RESULTS: Out of nearly 118,000 live births in North Carolina in 2002, mothers reported more than 600 different versions of race on birth certificates. These entries were collapsed into the 10 standard racial categories outlined in federal coding rules. Approximately two-thirds of mothers of Hispanic ethnicity report their race with a label that can be categorized as "Other" race, but nearly all of these births are re-coded to "white" for the official birth statistics. Measures of racial disparity vary depending on whether self-reported or officially coded race is used. CONCLUSIONS: This study shows that, given the opportunity to report their own race, North Carolinians describe their race using a wide variety of terms and concepts. In contrast, health statistics are usually reported using a few standardized racial categories defined by federal policy. The NCHS rules for coding race should be reexamined. As the ethnic and racial diversity of the United States continues to increase, these rules will become increasingly antiquated.