Entry Into Prenatal Care—United States, 1989-1997 FREE

MMWR. 2000;49:393-398

2 tables, 1 figure omitted

Assuring early initiation of prenatal care (PNC) is an important component of safe motherhood programs, which aim to improve maternal and infant health outcomes. Women who receive delayed (i.e., entry into PNC after the first 12 weeks of pregnancy) or no PNC do not receive timely preventive care or education and are at risk for having undetected complications of pregnancy that can result in severe maternal morbidity and sometimes death.^1- 2 Despite overall improvements, the national health objective for 2000 to increase to at least 90% the proportion of pregnant women who enter PNC during the first trimester of pregnancy (objective 14.11) was not met.³ To increase the proportion of mothers receiving early PNC, Congress authorized the Medicaid expansion program in the mid-1980s, which allowed states to expand Medicaid eligibility criteria to include formerly ineligible pregnant women.⁴ To examine trends in delayed entry into PNC during 1989-1997 and barriers to obtaining early PNC, CDC analyzed data from two sources. This report summarizes the results of that analysis, which indicated that although more women are obtaining early PNC, racial/ethnic disparities still exist and lack of money or health insurance was not the only barrier to obtaining PNC.

CDC analyzed 1989-1997 birth certificate data for all 50 states and the District of Columbia to examine the extent to which women in the United States received delayed PNC or had no PNC. Analysis was restricted to the approximately 4 million live births to U.S. residents each year and for whom data were available on the month of their PNC entry.

CDC then used 1997 Pregnancy Risk Assessment Monitoring System (PRAMS) data for 13 states* to assess reasons for delayed PNC or no PNC among women. PRAMS is an ongoing, state-based surveillance system that randomly samples birth certificates and collects information from mothers on pregnancy-related behaviors and experiences. The 20,345 women in the study represented approximately 842,000 women who gave birth to live-born infants in 1997 in the 13 states for which response rates exceeded 70%. SUDAAN was used for analysis, and data were weighted to adjust for survey design, nonresponse, and sampling frame noncoverage.

During 1989-1997, the percentage of women with delayed PNC or no PNC decreased from 25% to 18%, with improvement in both delayed PNC (from 22% to 16%) and in no PNC (from 2% to 1%). The decrease in no PNC during 1989-1997 resulted in an estimated 206,000 additional women entering PNC who may not have done so had the 1989 rate remained unchanged. Groups more likely to have delayed or no PNC during 1989-1997 included non-Hispanic blacks, Hispanics, women aged less than 20 years, women with less than 12 years of education, and multiparous women. However, those same groups generally had larger absolute improvements in delayed and no PNC during 1989-1997. For example, the prevalence of Hispanic women with delayed or no PNC decreased from 41% to 26%, and for women aged less than 20 years, from 47% to 32%.

Improvements in the percentages of women with no PNC occurred for all racial/ethnic groups during 1989-1995, but then leveled off for non-Hispanic black and non-Hispanic white women.

Data from 1997 PRAMS indicated that 56% of women with delayed or no PNC wanted to begin PNC earlier. Reasons for delayed or no PNC varied by racial/ethnic group, age, or method of payment for PNC. The most common reason for not receiving care earlier was "I didn't know that I was pregnant" (non-Hispanic black [44%], non-Hispanic white [37%], women aged less than 20 years [47%], and women whose PNC was paid for by private insurance [44%]). The second most cited barrier to earlier PNC entry was "I didn't have enough money or insurance to pay for my visits" (Hispanics [41%], non-Hispanic whites [36%], women aged 20-24 years [36%], and women whose PNC was paid for by a method other than public or private insurance [36%]). Among women whose PNC was paid for by public assistance (e.g., Medicaid and state programs), 33% cited the latter reason as a barrier to early care. The third most common reason for not receiving early PNC was inability to get an appointment (Hispanics [27%] and women aged greater than or equal to 35 years [36%]).

Statistics and Computer Resources Br, Program Svcs and Development Br, Pregnancy and Infant Health Br, Div of Reproductive Health, National Center for Chronic Disease Prevention and Public Health Promotion; Reproductive Statistics Br, Div of Vital Statistics, National Center for Health Statistics, CDC.

During 1989-1997, the prevalence of delayed or no PNC improved each year among women delivering a live-born infant in the United States. Improvements in access to early PNC may, in part, be attributed to the Medicaid expansion program.⁵ States responded to this change in the Medicaid regulations by implementing various programs that differed in regards to eligibility criteria and breadth of assistance.^6- 7

Evaluations of Medicaid expansion programs have shown that as more low-income women become eligible for Medicaid, more of these women accessed early PNC.⁸ However, as of 1998, no states had achieved the national health objective for 2000 for early PNC. Although the goals emphasized resolving health disparities, only one state had reached these goals for black women and no state had achieved them for Hispanic women.⁹ In 1997, the percentages of non-Hispanic black women and Hispanic women with delayed or no PNC remained approximately two times that of white women, approximately the same as in 1989. Such continuing disparities in obtaining early PNC mirror the disparities in many reproductive health outcomes among non-Hispanic black and Hispanic women compared with non-Hispanic white women.

Although Medicaid expansion has contributed substantially to improving access to early PNC by removing financial barriers for women, a substantial proportion of pregnant women still did not receive PNC during the first trimester. More than half of women with delayed or no PNC would have liked to obtain earlier care, and these women cited various reasons for delayed entry, with these reasons varying by group. The most frequent reasons for delay were not knowing that they were pregnant, lack of money or insurance coverage, and inability to get an appointment earlier. These reasons suggest a need for improved health education, women's health services, and coverage of early PNC services.

The findings in this report are subject to at least three limitations. First, the findings examined only one variable at a time and do not account for overlaps between demographic and socioeconomic groups of women (e.g., non-Hispanic black mothers are more likely to be aged less than 20 years than non-Hispanic white mothers). Second, although the measure used for initiation of care addresses the timing of PNC initiation, it does not account for the frequency, quality, or satisfaction with the PNC received. Finally, despite being representative of all women delivering a live-born infant in their states, PRAMS data are only available for a limited number of states and are not representative of all U.S. births.

Early, high-quality PNC is one of the cornerstones of a safe motherhood program, which begins before conception, continues with appropriate PNC and protection from pregnancy complications, and maximizes healthy outcomes for women, infants, and families. Barriers that keep women from entering PNC must be better understood to improve maternal health and to eliminate racial/ethnic disparities in the health outcomes of mothers and infants. Systems such as PRAMS provide the opportunity to understand the reasons women find it difficult to begin PNC early and to monitor changes in access to PNC. Only when timely services are available and accessible to women in their communities can strategies to assure safe motherhood provide the best start for pregnant women and their infants.

*Alabama, Alaska, Arkansas, Colorado, Florida, Georgia, Maine, North Carolina, New York (excluding New York City), Oklahoma, South Carolina, Washington, and West Virginia.

MMWR. 2000;49:278-281

1 table omitted

Infections with the soil-transmitted intestinal helminths (i.e., Ascaris lumbricoides, Trichuris trichiura, and hookworm), estimated to affect approximately 1 billion persons, are among the most common and widespread human infections.¹ Among children aged <5 years, intestinal helminth infections cause malnutrition and anemia, two important causes of mortality. Anthelminthic treatment (deworming) improves nutritional status of school-aged children.¹ The World Health Organization and the United Nations Children's Fund (UNICEF) have developed guidelines that include interventions for anemia and malnutrition² in the integrated management of childhood illness (IMCI) for children aged <5 years seen at first-level health-care facilities in developing countries. Under the IMCI guidelines, in geographic areas where hookworm or Trichuris infections are endemic, children aged 2-4 years with palmar pallor are treated with an anthelminthic drug. This report summarizes an investigation of the use of palmar pallor as an indication for anthelminthic treatment among ill children aged 2-4 years seen at first-level health-care facilities in rural western Kenya; the investigation found that palmar pallor was associated with anemia but not with intestinal helminth infection.

Children eligible for enrollment in the investigation were aged 2-4 years and seen for the first consultation for an illness during July 13-August 12, 1998, in three rural government health-care facilities in Bungoma District, Kenya. Enrollment criteria included caretaker consent, absence of a severe illness requiring referral, and no reported anthelminthic treatment during the 6 months preceding the investigation based on an interview with the caretaker. Each child was examined using IMCI guidelines, and a standard questionnaire was used to collect demographic, socioeconomic, and clinical information.

Hemoglobin (Hb) levels were measured from a capillary finger-stick blood specimen using a hemoglobin photometer. Blood smears were examined for malaria parasites. Stool samples were processed using a formal-ethyl-acetate concentration technique.³ The intensity of helminth infection was measured by eggs per gram of stool and categorized as light, moderate, or heavy.³

Of the 633 eligible children, 574 (91%) were enrolled; 34 (5%) children were excluded for receiving anthelminthic treatment during the 6 months before the investigation, 13 (2%) for the presence of a severe illness requiring referral, and 12 (2%) because the caretaker refused to participate. Excluded and enrolled children had similar demographic and socioeconomic characteristics. The participants' median age was 37 months (range: 24-59 months); 319 (56%) were boys. A total of 191 (33%) children had palmar pallor, 351 (61%) children had anemia (Hb: less than 11 gm/dL; normal: 11-16 gm/dL), 329 (57%) had malaria parasitemia, 32 (6%) were infected with Ascaris, 34 (6%) were infected with hookworm, and five (1%) were infected with Trichuris; 66 (12%) children had one or more intestinal helminths.

The prevalence of helminth infection was 10% among children aged 2 years, 11% among children aged 3 years, and 16% among children aged 4 years. All Trichuris infections, 97% of hookworm infections, and 78% of Ascaris infections were of light intensity. The sensitivity, specificity, and positive predictive value (PPV) of palmar pallor as an indicator for anemia were 50%, 93%, and 92%, respectively. Palmar pallor was associated with anemia (prevalence ratio [PR]=2.0; 95% confidence interval [CI]=1.8-2.3); however, no association was found between palmar pallor and helminth infection. The sensitivity, specificity, and PPV of palmar pallor for identifying children with helminth infections were 27%, 66%, and 9%, respectively. Although malaria parasitemia modified the association between palmar pallor and helminth infection, the sensitivity and PPV of palmar pallor as an indicator for helminth infections in this geographic area remained low in children with or without malaria parasitemia. In the IMCI guidelines, the anthelminthic treatment is specifically for anemia; however, no association was found between palmar pallor and hookworm or Trichuris infection (PR=0.9; 95% CI=0.5-1.8). The sensitivity, specificity, and PPV of palmar pallor for identifying children with hookworm or Trichuris infection were 32%, 67%, and 6%, respectively.

CN Wamae, Kenya Medical Research Institute; J Mwanza,'s Makama, Ministry of Health, Nairobi, Kenya. International Child Survival and Emerging Infections Program Support Activity and Epidemiology Br, Div of Parasitic Diseases, National Center for Infectious Diseases; and an EIS Officer, CDC.

The prevalence of intestinal helminth infection among a population of ill children aged 2-4 years who resided in Bungoma District, Kenya, was low and the infections identified were of low intensity. Findings of the few prevalence studies of intestinal helminth infection among healthy preschool-aged children in tropical areas are higher, ranging from 25% to 90%.^4- 7 The prevalence of intestinal helminth infections among healthy children aged 4-5 years in Kisumu District, western Kenya, was 60%⁷ compared with 16% among children aged 4 years seen for outpatient care in Bungoma District; therefore, wide variation may exist in the prevalence of helminth infections within proximate geographic areas. These differences may be environmental (eg, Kisumu and Bungoma districts are only 62 miles [100 km] apart; however, Kisumu District is warmer and more humid than Bungoma District) or socioeconomic (e.g., the prevalence of Ascaris and Trichuris infections among school children living in overcrowded conditions in Colombo, Sri Lanka, was seven to 10 times higher than that among children attending rural schools approximately 20 miles [30 km] away).⁸

The findings in this report indicate that palmar pallor was predictive of anemia but was not associated with helminth infections. Heavy hookworm infections consistently have been reported to be associated with anemia.^9- 10 The lack of association between palmar pallor and helminth infection in Bungoma District may be the result of the light intensity of hookworm infections; all but one hookworm infection was considered light.

The findings in this report are subject to at least two limitations. First, children who participated in the study may not be representative of all ill children in Bungoma District. Second, the findings may not be generalizable beyond areas with low prevalence and intensity of helminth infections.

Most children in Bungoma District with a helminth infection would not have received anthelminthic treatment, and few receiving anthelminthic treatment would have been infected with an intestinal helminth if palmar pallor were used to indicate anthelminthic treatment, as recommended in the IMCI guidelines. These guidelines have been introduced into approximately 60 developing countries; although implementing the guidelines provides a means for delivering the nutritional benefits of anthelminthic therapy to preschool-aged children, additional studies may help to determine under what conditions palmar pallor indicates the need for anthelminthic treatment. These studies should be conducted in areas with varying prevalences of intestinal helminth and malaria infections.

MMWR. 2000;49;398-400

In December 1998, the Chester County Health Department (CCHD) in Pennsylvania received reports of four cases of scombroid fish poisoning among patrons at a local restaurant. This report summarizes the investigation of these cases by CCHD, the Pennsylvania Department of Agriculture (PDAg), and the Pennsylvania Department of Health (PDOH). Findings from this investigation suggest that initial processes that are not regulated by the Food and Drug Administration (FDA) (i.e., from hooking the fish to unloading the fish on the dock) may permit scombrotoxin formation.

On December 3, 1998, four adults became ill after eating tuna-spinach salad at the restaurant. Symptoms of illness included a burning sensation in the mouth, a metallic taste, facial flushing, nausea, diarrhea, sweating, and headache; symptoms occurred approximately 5 minutes to 2 hours after eating the salad. One patient was taken to the local emergency department and treated with diphenhydramine, cimetidine, and epinephrine. The other three patients were not examined by physicians and their symptoms resolved within a few hours. A presumptive diagnosis of scombroid fish poisoning was made based on clinical and epidemiologic features of the illness.

A sample of the remaining fish obtained from the restaurant was sent to PDOH for testing. The fish was positive for coliform and Escherichia coli, and tests were positive for histamine levels greater than 50 ppm (fresh fish normally contain histamine levels of less than 10 ppm¹) using an enzyme-linked immunoabsorbent assay.

CCHD and PDAg conducted a traceback investigation of the source of the tuna. The wholesale-to-retail chain of events involved transporting the fish across national, state, and municipal borders and involved five transporters and four processors. The tuna was from a 40-60 lb yellow-fin tuna caught by a commercial fishing boat in the Gulf of Mexico during late November 1998. The fish was caught using the long-line method, which uses a mainline up to 60 miles long with a series of suspended hook lines. The water temperature where the fish was caught was 78.5 F (25.8 C). The catch of tuna was shipped from the fishing boat in iced vats by truck to a processor on November 24. The average temperature of the fish was 32 F-33 F (0 C-1 C). Of this catch, 785 lbs of tuna were shipped the same day to the wholesaler in Pennsylvania. The wholesaler received the shipment on November 27, and the average temperature of the fish was recorded as 36 F (2 C). Three of these fish were delivered to the retail supplier; two large fillets, weighing 11.1 lbs each and noted to be in good physical appearance, were delivered to the restaurant on November 27. The fish was divided into 30 portions, kept in the freezer, and removed for thawing as needed for use. During November 28-December 4, 17 portions of the fish were served. The only four persons reporting illness ate the tuna-spinach salad on December 3.

CCHD and PDAg reviewed the records of each distributor involved in the wholesale-to-retail process of the tuna. All of the fish plants involved were inspected regularly by the FDA and/or PDAg and have Hazard Analysis and Critical Control Point (HACCP) procedures. No deviations in HACCP procedures in the wholesale-to-retail distribution of the tuna could be identified. However, the long-line method of fishing is not covered as part of the FDA Seafood HACCP regulations.

JP Maher, MD, JA Worth, J Arvay, K Raum, Chester County Health Dept, West Chester; L Iampetro, Pennsylvania Dept of Health Bur of Laboratories; JR Welte, Food Safety and Laboratory Svcs, Region 7, Pennsylvania Dept of Agriculture. Food and Drug Administration. Div of Bacterial and Mycotic Diseases, National Center for Infectious Diseases; and an EIS Officer, CDC.

Scombroid fish poisoning has been associated primarily with the consumption of tuna, mahi-mahi, and bluefish. It is caused by histamine and other products produced by certain bacteria on some types of fish; these bacteria grow in warm temperatures and produce the enzyme histidine decarboxylase that converts free histidine in fish flesh to histamine and other products.^2- 6

National surveillance data on scombroid fish poisoning is based on outbreaks of acute foodborne disease reported by state health departments to CDC.^7- 8 During 1988-1997, scombroid fish poisoning was reported in 145 outbreaks involving 811 persons from at least 20 states^7- 8; however, many cases probably are not reported.

Since December 18, 1997, all processors of fish are required by FDA to conduct a hazard analysis of their operation and to implement a HACCP plan to control each identified hazard.⁹ The HACCP plan must be specific for each location where fish and fish products are processed and for each species processed.¹⁰ The fish implicated in these scombroid fish poisonings was caught by the long-line method of fishing, which consists of suspending a monofilament line, up to 60 miles long, with up to 3000 baited hooks in the water. The retrieval process may take up to 12-14 hours, and the fish may be retained on the lines up to 20 hours. Although no deviations in HACCP procedures were documented in this outbreak, the time from hooking the fish to unloading the fish on the dock is not covered by HACCP. Conditions permitting histamine production could have occurred while the fish were in warm water suspended on the long line.

Scombrotoxin formation also could have resulted from fish handling practices anywhere along the distribution chain after the fish was caught to serving at the restaurant. The reportedly good color and appearance of the fish at the retailer and the lack of other reported illnesses may indicate that scombrotoxin formation occurred at the restaurant during processing and handling of the fish.

This outbreak suggests interventions that could reduce the risk for scombroid poisoning. First, consideration should be given to limiting the amount of time that fish can remain on the line during the long-line method of fishing. Second, efforts should focus on maintaining adequate refrigeration of fish during distribution and in restaurants to prevent conditions favorable for scombrotoxin production. The key to prevention of scombroid fish poisoning is continuous icing or refrigeration at less than or equal to 32 F (≤0 C) of all potential scombrotoxin-producing fish from the time they are caught until they are cooked.

MMWR. 2000;49:1-4

1 table omitted

Neural tube defects (NTDs) are common and serious malformations that originate early in pregnancy. In the United States, approximately 4000 pregnancies each year are affected by the two most common NTDs (spina bifida and anencephaly). In 1992, the Texas Department of Health (TDH), with support from a CDC cooperative agreement, implemented the Texas Neural Tube Defect Project (TNTDP), a program of NTD surveillance and risk-reduction activities in the 14 counties that border Mexico. The project was initiated in response to an anencephaly cluster identified during 1990-1991 in Brownsville (Cameron County), Texas.¹ Whether the high anencephaly rate (19.7 per 10,000 live births) was unique to Cameron County or was characteristic of the entire border was unknown. This report summarizes NTD surveillance rates for the 14 Texas-Mexico border counties for 1993-1998 and presents preliminary results of TNTDP efforts to prevent the recurrence of NTDs by providing folic acid to high-risk women. Findings indicate that the baseline rate along the border is high (13.4 per 10,000 live births) and largely reflects the rate among Hispanics (13.8). Although a longer period is needed to obtain definitive results, folic acid appears to be effective for reducing the risk for NTD recurrence in Hispanics.

The TNTDP surveillance system involved prospective case finding (International Classification of Diseases, Ninth Revision [ICD-9], codes 740, 741, and 742.0, for all gestational ages) using the following data sources: hospitals; birthing centers; ultrasound centers; abortion centers; prenatal clinics; genetics clinics; and birth attendants including lay midwives, certified nurse midwives, and nonhospital physicians. Data on NTD cases were collected by three field teams (El Paso, Harlingen, and Laredo), abstracted onto standardized forms, and sent to TDH with confirmatory medical records. Denominator data (live birth, death, and fetal death records) were derived from the Bureau of Vital Statistics at TDH; 91% of the resident live births in the border counties were to Hispanic women of Mexican ancestry.

For 1993-1998, NTD surveillance rates include cases at all gestational ages for the 14 Texas-Mexico border counties. The surveillance system identified 360 resident NTD-affected births/terminations (cases) not otherwise accompanied by a known trisomy, triploidy, or syndrome (e.g., Turner, Meckel, or amniotic band). Of these cases, 324 (90%) occurred in the four most populous border counties—Cameron, El Paso, Hidalgo, and Webb. The overall NTD rate in the border counties for 1993-1998 was 13.4 per 10,000 live births (6.1 for anencephaly, 6.3 for spina bifida, and 1.0 for encephalocele). The craniorachischisis (contiguous opening of brain and spinal column; included in anencephaly) rate in the border counties was 0.5.

Of the 360 women identified as having had an NTD-affected pregnancy, 340 (94.4%) were Hispanic. Of the 20 non-Hispanic women, 16 (4.4%) were white, three (0.8%) were black, and one (0.3%) was Asian/Pacific Islander. The rate among Hispanics was 13.8 per 10,000 live births and the rate among non-Hispanic whites was 8.8 (p=0.08). El Paso County (the northwesternmost county) had a significantly lower NTD rate (9.0) than the rest of the border counties combined (15.6; p<0.001). The rate among Hispanics also was significantly lower for El Paso County (8.8) than that for the rest of the border counties (16.1) (p<0.001).

Of the NTD-affected pregnancies, 68 (19%) were induced or spontaneously aborted at <20 weeks' gestation, 94 (26%) were delivered or induced at 20 through 33 weeks' gestation, and 198 (55%) were delivered at ≥34 weeks' gestation. Excluding fetuses that failed to reach 20 weeks' gestation would have lowered the overall rate to 10.8 per 10,000 live births (p=0.01).

The primary objective of TNTDP is preventing recurrence of NTDs by providing folic acid to women who have had an NTD-affected pregnancy. For the folic acid intervention program, all women identified through the surveillance protocol were contacted by telephone, letter, and/or in person. Women whose index pregnancy was delivered or terminated in 1993 or later and who resided in the study area were asked to enroll in the program. The enrolled women were interviewed and provided preconception, pregnancy, and NTD risk-reduction education and counseling. If the women used contraception, they were given a multivitamin with 0.4 mg folic acid; if the women did not use contraception, they were given daily doses consisting of 4.0 mg folic acid—one multivitamin containing 1.0 mg of folic acid and three 1.0 mg tablets of folic acid. Women were followed, counseled, and provided folic acid supplements at 1- to 3-month intervals.

As of December 31, 1998, 264 (73%) of the 360 women were eligible for enrollment in the folic acid intervention program; 96 (27%) women were not eligible for enrollment (moved out of area or had tubal ligations/hysterectomies). Of the 264 eligible women, 95 (36%) refused enrollment, quit, or were lost to follow-up; 17 (6%) consented but were pending enrollment; and 152 (58%) were taking folic acid. Of 65 (34%) eligible women with induced abortions, 22 (34%) refused participation in the folic acid intervention compared with 19 (15%) of 128 (p=0.004) who had had natural outcomes (i.e., live-born infants, stillbirths, or spontaneous abortions).

Pregnancy outcomes following the index NTD-affected pregnancy were assessed by telephone, letter, and home visits for 1993-1998. Overall, 89% of the women who had a subsequent pregnancy had taken folic acid before conception; of these, 64% had taken the daily 4.0 mg dose; 28%, the 0.4 mg dose; and 8%, a physician-prescribed prenatal vitamin. A pregnancy outcome was documented for 148 pregnancies; 117 (79%) of the pregnancies resulted in non-NTD-affected live births, 24 (16%) in miscarriages or incomplete spontaneous abortions, six (4%) in elective abortions, and one (1%) in a confirmed recurrent NTD. Five women known to be pregnant were lost to follow-up. None of the six elective abortions was NTD-affected. Excluding the 24 miscarriages and five pregnancies lost to follow-up, one of the remaining 124 pregnancies resulted in a recurrent NTD.

K Hendricks, MD, R Larsen, PhD, L Suarez, PhD, Texas Neural Tube Defect Project, Texas Dept of Health. Birth Defects and Pediatric Genetics Br, Div of Birth Defects, Child Development, Disability, and Health (proposed), National Center for Environmental Health, CDC.

The preliminary results of the folic acid intervention suggest that high-risk women can reduce their risk for subsequent NTD-affected pregnancies. Each woman identified through the TNTDP surveillance protocol was at risk for recurrence and could not have been enrolled in the folic acid intervention program without being identified through surveillance. One fifth of the high-risk women in the program would have been missed if only fetuses at >20 weeks' gestation were included in the surveillance. Why women with induced abortions are less likely to take folic acid than women with natural outcomes is unclear and warrants further study. The woman who had a recurrent NTD-affected baby refused to meet with field staff and never received NTD risk-reduction education, counseling, or folic acid. The one NTD recurrence was less than the three to five that would have been expected based on a 3% to 4% recurrence rate (p=0.18, 0.10 respectively).

The NTD surveillance data indicate that baseline rates along the border are high and largely reflect the rate among Hispanics. Some of the variability in the rates may be partially explained by the unique cultural and environmental factors along the border. For example, compared with the rest of the border, El Paso County residents have a higher standard of living and are less likely to be employed as migrant farm workers.¹ In addition, the overall Texas-Mexico border rate for craniorachischisis was 0.5, a rate significantly higher (p=0.048) than the rate for this defect in the metropolitan Atlanta area (0.1).² This suggests that an unknown risk factor may exist, especially in Hidalgo County where six (46%) of these rare defects occurred. Findings from the 1993-1998 recurrence period showed that only 9% of El Paso County women who delivered normal live-born infants reported taking periconceptional folic acid (TNTDP, unpublished data, 1999). Although the 9% usage reported for El Paso County is low compared with national reported usage (25%),³ usage for Cameron County is even lower (3%).

The findings in this report are subject to at least two limitations. First, nonresident women who migrated for birth into the United States and either returned to Mexico or to another county were not eligible for the intervention program; further, resident women who moved, were lost to followup, or had tubal ligations/hysterectomies decreased the potential intervention sample size by 40%. Second, some underestimate of cases occurred because of pregnancy outcomes that occurred outside the area.

Although a sufficient number of pregnancy outcomes have yet to occur among high-risk women to achieve statistical significance, folic acid appears to reduce the risk for NTD recurrence in Hispanic women. Unlike other U.S. surveillance systems,^4- 5 since its inception the TNTDP has included cases at <20 weeks' gestational age. These data underscore the importance of a timely and active NTD surveillance system that includes fetuses at <20 weeks' gestational age for population-based and individual NTD prevention. They also highlight the need for physicians to educate their high- and low-risk patients about the benefits of folic acid.^6- 7

MMWR. 2000;49:401

CDC and Emory University's Rollins School of Public Health will cosponsor a course, "International Course in Applied Epidemiology" on October 2-27, 2000, in Atlanta, Georgia. This basic course in epidemiology is directed at public health professionals from countries other than the United States.

The course's content includes presentations and discussions of epidemiologic principles, basic statistical analysis, public health surveillance, field investigations, surveys and sampling, and discussions of the epidemiologic aspects of current major public health problems in international health. Included are small group discussions of epidemiologic case exercises based on field investigations. Participants are encouraged to give a short presentation reviewing some epidemiologic data from their own country. Computer training using Epi-Info, a software program developed at CDC and the World Health Organization for epidemiologists is included. Prerequisites are familiarity with the vocabulary and principles of basic epidemiology or completion of CDC's "Principles of Epidemiology" home-study course (SS3030) or equivalent. Preference will be given to applicants whose work involves priority public health problems in international health. There is a tuition charge.

Additional information and applications are available from Emory University, Rollins School of Public Health, International Health Dept. (PIA), 1518 Clifton Road N.E., Room 746, Atlanta, GA 30322; telephone (404) 727-3485; fax (404) 727-4590; or email pvaleri@sph.emory.edu, or on the World-Wide Web at http://www.sph.emory.edu/EPICOURSES.*

*References to sites of non-CDC organizations on the World-Wide Web are provided as a service to MMWR readers and do not constitute or imply endorsement of these organizations or their programs by CDC or the U.S. Department of Health and Human Services. CDC is not responsible for the content of pages found at these sites.

MMWR. 2000;49:382

Statisticians, epidemiologists, and others with an interest in the application of statistical methods to public health are invited to participate in the eighth biennial Symposium on Statistical Methods. The symposium is sponsored by CDC and the Agency for Toxic Substances and Disease Registry (ATSDR) and will be held January 23-24, 2001, in Atlanta, Georgia. The theme for the symposium is "Issues Associated With Complicated Designs and Data Structures." A short course on a related topic will be offered on January 22, 2001, in conjunction with the symposium.

The symposium will include invited speakers and contributed papers. Authors can submit abstracts for contributed papers related to one or more of the session content areas listed below:

Abstracts will be considered for either oral or poster presentation and must be postmarked no later than August 1, 2000. Authors of papers accepted for either oral or poster presentation will be notified by September 30, 2000. All accepted papers will be considered for publication in a dedicated issue of Statistics in Medicine. Registration, abstract information, forms, and additional information regarding the scientific content of the symposium are available on the World-Wide Web at http://www.cdc.gov/od/ads/sag; by mail to 2001 CDC and ATSDR Symposium on Statistical Methods, 4770 Buford Highway N. E., Mailstop K-21, Atlanta, GA 30341; telephone (770) 488-5185; fax (770) 488-5967; or e-mail to CJohnson3@cdc.gov.