Obstetric deaths in Bangladesh, 1996-1997.

OBJECTIVES: The purpose of this study was to measure and to describe obstetric deaths in Bangladesh. METHODS: We reviewed hospital records and interviewed health workers in clinic sites and field workers who cared for pregnant women. RESULTS: We obtained case reports of 28998 deaths of women aged 10-50, of which 8562 (29.5%) were maternal deaths. Most (7086, 82.8%) of these deaths were due to obstetric causes. The most common causes of direct obstetric death were eclampsia (34.3%), hemorrhage (27.9%), and obstructed and/or prolonged labor (11.3%). National direct obstetric death rate was estimated to be 16.9 per 100,000 women. CONCLUSIONS: Efforts to reduce fertility in Bangladesh have led to an estimated 49% reduction in the maternal mortality rate per 1000 women during the past 18 years. Variations in maternal mortality suggest the need to develop local strategies to improve obstetric care.

Missed opportunities for congenital syphilis prevention in an urban southeastern hospital.

BACKGROUND: Despite recent declines in syphilis rates nationally, recent outbreaks suggest that prevention of congenital syphilis remains an ongoing public health problem. GOALS: To identify missed opportunities for congenital syphilis prevention during prenatal care. STUDY DESIGN: Retrospective medical record review of 157 live birth or stillbirth deliveries that involved cases of congenital syphilis from Grady Memorial Hospital (Atlanta, GA). RESULTS: The hospital congenital syphilis prevalence was 8.2 cases per 1,000 live births. Six percent of case patients were HIV positive. Opportunities for earlier maternal screening, treatment, or diagnosis were missed in 60% of case patients who received timely prenatal care. Congenital syphilis cases attributable to preventable missed opportunities were significantly more common among case patients with fewer prenatal visits (66% versus 28%, P = 0.01). CONCLUSION: Provider efforts to reduce congenital syphilis in high-risk populations receiving prenatal care should focus on (1) screening patients at the first opportunity, at both the first prenatal visit and during the third trimester (i.e., 28 weeks); (2) performing on-site testing and same-day treatment; (3) providing appropriate treatment to infected women with penicillin allergy; (4) referring sex partners for treatment to prevent reinfection; and (5) screening all pregnant women attending emergency clinics.

Injury-related deaths among women aged 10-50 years in Bangladesh, 1996-97.

BACKGROUND: Few studies have examined injury-related deaths among women in Bangladesh. We did a case-finding study to identify causes and the impact of intentional and unintentional injury-related deaths among women aged 10-50 years in Bangladesh. METHODS: Between 1996 and 1997, health care and other service providers at 4751 health facilities throughout Bangladesh were interviewed about their knowledge of deaths among women aged 10-50 years. In addition, at all public facilities providing inpatient service, medical records of women who died during the study period were reviewed. The reported circumstances surrounding each death were carefully reviewed to attribute the most likely cause of death. FINDINGS: 28,998 deaths among women aged 10-50 years were identified in our study, and, of these, 6610 (23%) were thought to be caused by intentional or unintentional injuries. About half (3317) of the injury deaths were attributable to suicide, 352 (5%) to homicide, 1100 (17%) to accidental injuries, and the intent was unknown for 1841 (28%) deaths. The unadjusted rate of suicides were higher in the Khulna administrative division (27.0 per 100,000) than the other four administrative divisions of Bangladesh (range 3.5-11.3 per 100,000). Poisoning (n=3971) was the commonest cause of injury-related death--60% of all injury deaths (6610) and 14% of all deaths (28,998). Other common causes of injury deaths in order of frequency were hanging or suffocation, road traffic accidents, burns, drowning, physical assault, firearm or sharp instrument injury, and snake or animal bite. INTERPRETATION: Intentional and unintentional injuries are a major cause of death among women aged 10-50 years in Bangladesh. Strategies to reduce injury-related deaths among women need to be devised.

Maternal cigarette smoking and invasive meningococcal disease: a cohort study among young children in metropolitan Atlanta, 1989-1996.

OBJECTIVES: This study assessed the association between maternal cigarette smoking during pregnancy and the risk of invasive meningococcal disease during early childhood. METHODS: Using a retrospective cohort study design, cases from an active surveillance project monitoring all invasive meningococcal disease in the metropolitan Atlanta area from 1989 to 1995 were merged with linked birth and death certificate data files. Children who had not died or acquired meningococcal disease were assumed to be alive and free of the illness. The Cox proportional hazards analysis was used to assess the independent association between maternal smoking and meningococcal disease. RESULTS: The crude rate of meningococcal disease was 5 times higher for children whose mothers smoked during pregnancy than for children whose mothers did not smoke (0.05% vs 0.01%). Multivariate analysis revealed that maternal smoking (risk ratio [RR] = 2.9; 95% confidence interval [CI] = 1.5, 5.7) and a mother's having fewer than 12 years of education (RR = 2.1; 95% CI = 1.0, 4.2) were independently associated with invasive meningococcal disease. CONCLUSIONS: Maternal smoking, a likely surrogate for tobacco smoke exposure following delivery, appears to be a modifiable risk factor for sporadic meningococcal disease in young children.

No fathers' names: a risk factor for infant mortality in the State of Georgia, USA.

Many studies have explored maternal and infant factors as risks for infant mortality, but little attention is given to paternal factors. In Georgia, listing a father's name on the birth certificate is optional for married couples and possible after paternal acknowledgment for unmarried couples. The authors evaluated father's name reporting as a paternity measure and risk for infant mortality. Using the linked 1989-1990 birth and death certificates of singleton Georgia infants to calculate relative risks (RRs), infant mortality rates for 38,943 infants with no father's names listed were compared to rates for 178,100 with father's names listed. Compared with the rate for married women listing names, the death rates were higher for unmarried mothers not listing fathers (relative risk, RR = 2.5; 95% CI 2.3-2.7), unmarried mothers listing fathers (RR = 1.4; 95% CI 1.3-1.6), and married women not listing fathers (RR = 2.3; 95% CI 1.6-3.1). Increased risks remained after stratifying by maternal race, age, adequacy of prenatal care and medical risks; and congenital malformations, birthweight, gestational age, and small-for-gestational age. Using logistic regression to examine for effect modification and to adjust for these factors together, the adjusted relative risks for death varied across different groups without fathers' names, regardless of marital status. For example, it remained statistically higher for infants with no father listed and without effect-modifying conditions such as low birthweight (estimated RR = 2.0; 95% CI 1.6-2.4). Although these findings suggest paternal involvement, as measured by listing fathers' names, is protective against low birthweight and infant mortality, further evaluation is needed.

Child fatality review in Georgia: a young system demonstrates its potential for identifying preventable childhood deaths.

BACKGROUND: Child fatality review (CFR) by interagency teams can contribute to the prevention of childhood deaths. We investigated the potential usefulness of Georgia's CFR, legislated in 1990 primarily to prevent death from child maltreatment, for identifying preventable deaths from injury and sudden infant death syndrome (SIDS). METHODS: Using CFR report data and death certificate data, we examined reviewed and nonreviewed childhood deaths in Georgia in 1991 and examined data by etiology, county, risk factors, and preventability. RESULTS: Injury or SIDS caused 33.2% of childhood deaths in Georgia in 1991; CFR reviewed 29.4% of these. Child fatality review was most sensitive for investigating death from intentional injury (40.5%) and SIDS (35.3%). Review teams reassigned the cause of five deaths (2.0%) to child abuse or neglect. County participation was low (31.4%). Overall, 29.0% of deaths were judged preventable. CONCLUSIONS: Georgia's CFR has potential for identifying preventable childhood deaths. Refinements in the system can increase the number and accuracy of death investigations. By participating in the system, physicians may make meaningful contributions to preventing childhood death in their own communities.

Differences in the risk of homicide and other fatal injuries between postpartum women and other women of childbearing age: implications for prevention.

OBJECTIVES: This study compared injury deaths between postpartum women and other women aged 15 to 44. METHODS: Risk ratios and 95% confidence intervals (CIs) were computed for injury fatality rates. RESULTS: Fifty percent (29/58) of postpartum injury deaths were homicides, compared with 26% (427/1648) of injury deaths among nonpregnant, nonpostpartum women. For females aged 15 to 19, the homicide rate was 2.6 times higher (95% CI = 1.17, 5.95) for postpartum females than for other females. The motor-vehicle fatality rate was lower for postpartum females than for nonpregnant, nonpostpartum females (risk ratio = 0.30, CI = 0.18, 0.48). CONCLUSIONS: Postpartum females aged 15 to 19 years were at higher risk of homicide. Postpartum women were at reduced risk of motor-vehicle fatalities.

Delayed entry into prenatal care: effect of physical violence.

OBJECTIVE: To assess whether women who experienced physical violence by their partner during the 12 months before delivery were more likely to delay entry into prenatal care than were women who had not experienced physical violence. METHODS: We analyzed data from the Pregnancy Risk Assessment Monitoring System. The sample included 27,836 women who delivered live infants during 1993-1994 in nine states and were surveyed 2-6 months after delivery. We calculated risk ratios and 95% confidence intervals (CIs) to measure the association between physical violence within the 12 months before delivery and entry into prenatal care. RESULTS: The prevalence of delayed entry into prenatal care (entering after the first trimester) was 18.1% and that of reported physical violence was 4.7%. Overall, women who experienced physical violence were 1.8 times more likely (95% CI 1.5, 2.1) to have delayed entry into prenatal care than women who had not experienced such violence. When stratifying by selected maternal characteristics, this association was found only for groups of women who were 25 years of age or older or were of higher socioeconomic status. CONCLUSION: Older women and women of higher socioeconomic status who reported physical violence were more likely to delay entry into prenatal care than younger or less affluent women.

Quality assessment of fetal death records in Georgia: a method for improvement.

OBJECTIVES: Although more fetal deaths than neonatal deaths occur, routinely collected fetal death data are seldom used for perinatal epidemiologic research because of data quality concerns. We developed a strategy for identifying and correcting errors in birthweight and gestational age in fetal death records. METHODS: Using data from Georgia for 1989 and 1990, we detected singleton fetal death records having improbable or missing birthweight or gestational age by comparing these values with referent values. To verify the questionable values, we contacted 100 reporting hospitals in 1992. RESULTS: In 817 of 2226 records, values were either improbable (60.1%) or missing (39.9%). We were able to contact the hospitals to verify data for 716 (88%) of these records. Verification resulted in corrections to 405 (57%) records, and 48% of unreported birthweights were obtained. CONCLUSIONS: Many errors in recorded gestational age and birthweight were identified by this method. Rather than deleting or inputting problem data for analyses, researchers should consider efforts to verify them. Efforts to improve this information should include improved reporting, strict quality assurance, and procedures for routine verification and correction of records.

An evaluation of the use of dried blood spots from newborn screening for monitoring the prevalence of cocaine use among childbearing women.

A collaborative March of Dimes study was designed to examine the utility of dried blood spot (DBS) materials routinely collected from newborns as a source for monitoring cocaine exposure and to assess the prevalence of cocaine use among childbearing women in Georgia. We used a modified urinary radioimmunoassay (RIA) to anonymously detect the cocaine metabolite benzoylecgonine (BE) in DBSs. Extensive efforts were undertaken to assure absolute nonlinkage of BE data to any individual. The positive results found by RIA were confirmed by a mass spectrometry (MS) method specifically developed to detect BE in DBSs. BE was measured in 23,141 DBSs collected during 2 months of routine newborn screening in Georgia. A good correlation was observed for RIA results versus MS results (r2 = 0.97). The estimated minimal statewide BE prevalence was 4.8 per 1000 childbearing women. We demonstrated that immunoassay testing for cocaine without confirmatory testing can yield falsely elevated prevalence rates. When proper confirmatory testing is done, DBSs are a valuable source for population-based monitoring of substance abuse among childbearing women.

Patterns of prenatal care initiation in Georgia, 1980-1992.

OBJECTIVE: To determine whether characteristics in a woman's first pregnancy were associated with the trimester in which she initiated prenatal care in her second pregnancy. METHODS: Data for white and black women whose first and second pregnancies resulted in singleton live births between 1980 and 1992 were obtained from Georgia birth certificates (n = 177,041). Adjusted relative risks (RRs) for early prenatal care in the second pregnancy were computed by logistic regression models that included trimester of prenatal care initiation, infant outcomes, or maternal conditions in the woman's first pregnancy as the exposure and controlled for maternal age, education, child's year of birth, interval between first and second pregnancy, presence of father's name on the birth certificate, and the interaction between prenatal care and education. Models were stratified by race. RESULTS: Women of both races who initiated prenatal care in the first trimester of their first pregnancies were more likely than those with delayed care to initiate prenatal care in the first trimester of their second pregnancies (RR = 1.25 and 1.63 for white and black women educated beyond high school, respectively). Both white and black women who delivered a baby with very low birth weight (RR = 1.06 and 1.15, respectively) or who suffered an infant death (RR = 1.09 and 1.31, respectively) in their first pregnancies were more likely than those who did not experience these events to begin prenatal care in the first trimester of their second pregnancies. CONCLUSION: Women with some potentially preventable adverse infant outcomes tend to obtain earlier care in their next pregnancy. Unfortunately, women who delayed prenatal care in their first pregnancy frequently delay prenatal care in their next.

Prenatal smoking in two consecutive pregnancies: Georgia, 1989-1992.

OBJECTIVE: To explore the patterns of prenatal smoking among women whose first and second pregnancies ended in live births. METHODS: We used population-based data to explore prenatal smoking among 14,732 white and 8968 black Georgia residents whose first and second pregnancies ended in live births during 1989-1992. Smoking status was obtained from birth certificates linked for individual mothers. Because of demographic differences, we analyzed white and black women separately. RESULTS: Approximately 15% (2253) of white women and 4% (318) of black women smoked during their first pregnancy. Of those smokers, 69% (1551) of white women and 58% (184) of black women also smoked during their second pregnancy. For both white and black nonsmokers during the first pregnancy, low education was the most significant predictor of smoking during the second pregnancy, after adjusting for consistency of the father's name on the birth certificate, prenatal care, birth interval, mother's county of residence, and birth outcome. CONCLUSIONS: The prevalence of smoking in this study may be low because of underreporting of prenatal smoking on birth certificates. The majority of women who smoked during their first pregnancy also smoked during their second, suggesting that these women exposed their first infant to tobacco smoke both in utero and after delivery. Practitioners should offer smoking cessation programs to women during, as well as after, pregnancy. Pediatricians should educate parents on the health risks to young children of exposure to environmental tobacco smoke and refer smoking parents to smoking cessation programs.

Evaluation of an intervention to reduce playground hazards in Atlanta child-care centers.

In 1988, we identified playground hazards at 58 child-care centers (CCCs) and intervened by showing the director the hazards and distributing safety information. In 1990, we evaluated the 58 intervention CCCs as well as 71 randomly selected control CCCs. Intervention centers had 9.4 hazards per playground; control centers had 8.0. We conclude that the intervention was ineffective.

Pregnancy-related mortality in New York City, 1980 to 1984: causes of death and associated risk factors.

To identify causes and risk factors for pregnancy-related mortality in New York City, we analyzed 224 pregnancy-related deaths that occurred from 1980 to 1984. The leading causes of death were ectopic pregnancy complications, embolism, intrapartum cardiac arrest, and hypertension. Mortality ratios were determined by comparing the characteristics of the women whose death was pregnancy-related with those of women who had survived delivery of a live infant in New York City during the same period. Black and Hispanic women had mortality ratios that were respectively 4.2 and 2.0 times higher than those for white, non-Hispanic women. In comparison with women aged 20 to 24, those older than 30 were more than twice as likely to die from pregnancy-related causes, and those older than 40 were five times as likely to do so. Other factors that were associated with an increased risk of pregnancy-related mortality included 9 to 11 years of education, lack of private medical insurance, more than five previous pregnancies, and fewer than five prenatal visits. This study suggests that changes in current maternal-health and family-planning services will be required to achieve further reductions in preventable pregnancy-related mortality.

Maternal mortality surveillance, United States, 1980-1985.

To better define the incidence, causes, and risk factors associated with maternal deaths, the Division of Reproductive Health, Center for Chronic Disease Prevention and Health Promotion, CDC, coordinated a study by the Maternal Mortality Collaborative, a Special Interest Group of the American College of Obstetricians and Gynecologists (ACOG). In 1983, this group established voluntary surveillance of maternal deaths for the years 1980-1985. The Maternal Mortality Collaborative reported 601 maternal deaths from 19 reporting areas for 1980-1985, representing a maternal mortality ratio of 14.1 per 100,000 live births. Overall, 39% more maternal deaths were reported by the Maternal Mortality Collaborative than by the National Center for Health Statistics for these reporting areas. Overall, women over 30 years of age had a higher risk of dying than did younger women. For each age group, women of black and other races who were 30 years and older having the highest risk. The leading causes of maternal deaths were embolism, hypertension in pregnancy, sequelae from ectopic pregnancy, hemorrhage, cerebrovascular accidents, and anesthesia complications. Of the 111 nonmaternal deaths, 90 (82%) were attributed to unintentional or intentional injuries. As a result of the success of this voluntary reporting system, the Division of Reproductive Health initiated National Pregnancy Mortality Surveillance in January 1988.

Maternal mortality in the United States: report from the Maternal Mortality Collaborative.

To better define the incidence, causes, and risk factors associated with maternal deaths, the Maternal Mortality Collaborative in 1983 initiated national voluntary surveillance of maternal mortality. The Maternal Mortality Collaborative reported 601 maternal deaths from 19 reporting areas for 1980-1985, representing a maternal mortality ratio of 14.1 per 100,000 live births. Overall, 37% more maternal deaths were reported by the Maternal Mortality Collaborative than by the National Center for Health Statistics for these reporting areas. Older women and women of black and other races continued to have higher mortality than younger women and white women. The five most common causes of death for all reported cases were embolism, nonobstetric injuries, hypertensive disease of pregnancy, ectopic pregnancy, and obstetric hemorrhage. Compared with national maternal mortality for 1974-1978, ratios were lower for all causes except for indirect causes, anesthesia, and cerebrovascular accidents. Fatal injuries among pregnant women are not commonly reported to maternal mortality committees. As maternal mortality from direct obstetric causes continues to decline, clinicians will need to emphasize preventing deaths from nonobstetric causes.

Sterilization regret among Puerto Rican women.

This study examined sterilization regret among Puerto Rican women with contraceptive tubal sterilizations (TS) using retrospective data from a population-based survey of women aged 15 to 49 years. Twenty-one percent of the 846 respondents expressed some regret, with 11% stating definite dissatisfaction with the decision. Factors associated with regret were young age at TS, absence of daughters, someone else making the decision to be sterilized, medical indications for TS, sterilization failure, and living with a new marital partner. There were no significant effects associated with the timing of sterilization, whether interval or postpartum, or with the type of last delivery. Other factors not independently associated with regret included years since TS, parity, education, and age at follow-up.

PIP: Reasons for sterilization regret in Puerto Rican women were examined from interview data from the 1982 Puerto Rico Fertility and Family Planning Assessment. 846 of the 3175 women surveyed had contraceptive tubal sterilization, i.e., medical sterilizations, vasectomies and hysterectomies were excluded. Sterilization took place between 1954 and 1982, in women aged 15-49 years old at the time of the interviews. 42% of the subjects lived in rural areas. Regret was expressed on the questionnaire as definitely yes, maybe yes, maybe no, or definitely no. 21% of these women expressed some regret, and 11% felt definite regret for having been sterilized. The most common reasons for regret were desire for more children then, or change of heart later, and not being consulted about the operation. Factors associated with regret were age under 25 at operation, having no daughters, husband or physician having decided on sterilization, medical indications for sterilization, sterilization failure and living with a new partner. Factors not significantly related to regret included timing of sterilization with respect to childbirth, years since sterilization, parity, education and age at survey.

Proximate determinants of fertility and policy implications in Beijing.

Data from a 1982 sample survey of 3,830 married women below age 50 in a district of Beijing City are applied to the Bongaarts model of the proximate determinants of fertility. A total fertility rate (TFR) of 0.75 births per woman is estimated from the model, compared with a TFR of 1.24 actually observed from the survey. The estimated TFR (0.75) results from the assumed total fecundity (TF) of 15.30 births per woman being inhibited by the indexes of the proximate determinants of fertility, most notably non-marriage (.280), contraception (.315), and induced abortion (.586). The effects of two other proximate determinants, lactational infecundability and spousal separation, were negligible. Comparative data from other countries confirm that the study area has very low levels of fertility and marriage, a very high prevalence of induced abortion, and a small effect of lactational infecundability. The extremely low level of fertility was, thus, achieved through a high level of induced abortion. Future research needs and policy implications associated with the study are discussed.

PIP: Data from a 1982 sample of 3,830 married women age 50 in the Xi Cheng district of Beijing City are applied to the Bongaarts model postulating 7 proximate determinants of fertility, through which social, cultural, and economic conditions affect fertility. The sample was selected from a total of 115,279 eligible women, using a 2-stage areal sampling technic. A total fertility rate (TFR) of 0.75 births/woman is estimated from the model, compared to a TFR of 1.24 actually observed from the survey. The estimated TFR (0.75) results from the assumed total fecundity (TF) of 15.30 births/woman being inhibited by the indexes of the proximate determinants of fertility, most notably non-marriage (.280), contraception (.315), and induced abortion (.586). The effects of 2 other proximate determinants, lactational infecundability and spousal separation, were negligible (even though spousal separation was especially considered, on the belief that it is strongly affected by employment patterns). Comparative data from other countries confirm that the study area has very low levels of fertility and marriage, a very high prevalence of induced abortion, and a small effect of lactational infecundability. The extremely low level of fertility was, thus, achieved through a high level of induced abortion. This situation was considered undesirable from the standpoints of both health and the burden on the health service system. Future research needs and policy implications of the study include the low prevalence of breastfeeding, and high levels of contraceptive failure.

Maternal mortality in Addis Ababa, Ethiopia.

Between July and September 1983, a two-stage probability survey was conducted in Addis Ababa, Ethiopia to obtain data on pregnancy outcomes for all women aged 13-49 in 32,215 houses. The survey covered a two-year period, from 11 September 1981 to 10 September 1983. Of the 9,315 women who were pregnant during those two years, 45 died from complications of pregnancy, delivery, and the puerperium. The maternal mortality rate for 1982-83 was estimated to be 566 per 100,000 live births. Mortality was highest for nullipara, the unmarried, women employed as maids/janitresses, and students. The most common cause of death was abortion. It appears that reliable data on maternal mortality can be obtained retrospectively through a probability survey.

PIP: Between July and September 1983, a 2-stage probability survey was conducted in Addis Ababa, Ethiopia, to obtain data on pregnancy outcomes for all women aged 13-49 in 32,215 houses. Previously, information on maternal mortality (MM) had been scare: only 2 hospitals reported figures of 840 and 780 deaths/100,000 live births respectively, in 1970 and 1980. This survey covered a 2-year period from 11 September 1981 to 10 September 1983. All 25 city districts were represented, and a sample of 47 subdistricts was drawn, with the aim of identifying the outcomes of about 10,000 pregnancies, requiring visits to 26,136 houses. A pilot survey was conducted to improve the technic. The survey universe for the actual study was 182,836 persons, reporting 9,954 pregnancies for 9,315 women over 2 years, with 9,155 live births. 45 women died from complications of pregnancy, delivery, and puerperium. Data on cause of death are quite detailed: for 22 of the 24 women dying in a hospital, hospital records were available; for the 15 dying at home, 2 had post mortem documents; 6 women died in transit or outpatient facilities; and many of those in the last 2 categories were diagnosable from description of symptoms. Among direct obstetric causes of death were abortion (13); hemmorage (3); hypertensive disease of pregnancy (3); ruptured uterus (2); anesthetic death (2); puerperal sepsis (1); total 24. There were 21 deaths by indirect obstetric causes, 5 of which were unknown and 7 of which were due to infectious hepatitis, the commonest cause. The MM rate for 1982-83 was estimated at 566/100,000 live births; highest among nullipara, the unmarried, women employed as maids/janitresses, and students. Reliable data on MM can be obtained retrospectively through a probability survey, as demonstrated by this project.

Mortality associated with sterilization: preliminary results of an international collaborative observational study.

Sterilization is the contraceptive method most widely used worldwide, yet the case-fatality rate of deaths attributable to sterilization is not known. We used data collected from 1971-1979 from 28 countries by Family Health International to estimate case-fatality rates. We adjusted these rates for individuals lost to follow-up. Of 41,834 sterilizations, 23 resulted in deaths temporally associated with the procedure used. The adjusted attributable case-fatality rates were 13.4 per 100,000 for interval procedures, 53.3 per 100,000 for postabortion procedures, and 43.4 per 100,000 sterilizations after vaginal delivery. Multiple factors contributed to the deaths, including pre-existing health problems, infection and anesthesia. Prevention of deaths resulting from sterilization depends on complete ascertainment of deaths associated with sterilization and careful investigation to determine preventable risk factors. We conclude that, overall, sterilization in these programs was conducted with very low attributable mortality.

PIP: Sterilization is the contraceptive method most widely used worldwide, yet the case-fatality rate of deaths attributable to sterilization is not known. In this study data was collected from 1971-79 from 28 countries by Family Health International to estimate case-fatality rates. Rates were adjusted for individuals lost to follow-up. Of 41,834 sterilizations, 23 resulted in deaths temporally associated with the procedure used. The adjusted attributable case-fatality rates were 13.4/100,000 for interval procedures, 53.3/100,000 for postabortion procedures, and 43.4/100,000 sterilizations after vaginal delivery. Multiple factors contributed to the deaths, including pre-existing health problems, infection and anesthesia. Prevention of deaths resulting from sterilization depends on complete ascertainment of deaths associated with sterilization and careful investigation to determine preventable risk factors. Overall, sterilization in these programs was conducted with very low attributable mortality.