Surveillance for Fetal Alcohol Syndrome Using Multiple Sources -- Atlanta, Georgia, 1981-1989

Fetal alcohol syndrome (FAS) is caused by heavy alcohol consumption during pregnancy and is characterized by specific anomalies of the face; prenatal and postnatal growth deficits; and a variety of central nervous system (CNS) abnormalities, including mental retardation (1). Children with either full or partial FAS * often incur severe and costly secondary disabilities (2). Despite the importance of surveillance for establishing the magnitude of FAS and in monitoring trends in the occurrence of this disease, population-based surveillance for FAS has been difficult because the syndrome can be diagnosed only by clinical observation and often is not recognized until after the child reaches school age. Although most FAS surveillance has been based on diagnoses among newborns (3), most (89%) cases (full FAS and partial FAS) are diagnosed after the age of 6 years (2). To develop a more accurate estimate of the prevalence of FAS in a defined population, in 1997 CDC linked data from the Metropolitan Atlanta Congenital Defects Program (MACDP) and the Metropolitan Atlanta Developmental Disabilities Surveillance Program (MADDSP) for children born in Atlanta during 1981-1989 (the most recent birth year for which data were available for 3-10-year-olds). This report presents a multiple-source method for FAS surveillance that is more complete than previous methods and that enables comparison of rates between states.

MACDP monitors infants born to residents of five metropolitan Atlanta counties. Infants reported to MACDP may have any of approximately 200 specific major anomalies, including FAS. Abstractors visit all birth hospitals, neonatal intensive-care units, and genetics clinics to identify and document possible cases. MADDSP identifies children aged 3-10 years living in the same five metropolitan Atlanta counties who have either mental retardation, cerebral palsy, hearing impairment, or vision impairment. The affected children are primarily enrolled in public special education programs or are receiving other special services. This study included any child born to a resident of metropolitan Atlanta during 1981-1989 who had been assigned a code for FAS or possible FAS by either system based on hospital or school-record abstractions performed by surveillance personnel.

Hospital records for all children with possible FAS were reviewed to determine maternal alcohol use during pregnancy; children's weight, length, or head circumference at various ages; and whether any facial anomalies were noted. Cognitive test results from the MADDSP database were used to identify functional CNS impairment. Criteria based on those of the Institute of Medicine (IOM) were used to categorize children as having full FAS, partial FAS, or not having FAS (1). In this analysis, the category "partial FAS" included the IOM categories of "partial FAS" and "alcohol-related neurodevelopmental disorder." The number of live-born infants was obtained from Georgia birth certificate files, which indicated that 285,538 children were born to residents of metropolitan Atlanta during 1981-1989.

During 1981-1989, MACDP and MADDSP combined identified 92 children with possible FAS: MACDP uniquely identified 50 (54%) of these possible cases; MADDSP uniquely identified 31 (34%); and both registries identified 11 (12%). Hospital records were available for 91 of the children and their mothers; of these children, clinical signs and symptoms in 70 (77%) met the case definition for an alcohol-related diagnosis; signs and symptoms in 29 (32%) children met the case definition for full FAS and in 41 (45%), for partial FAS. The observed prevalence of full FAS was 1.0 cases per 10,000 live-born infants, and the observed prevalence of both full and partial FAS was 2.5 cases per 10,000.

To estimate the completeness of case ascertainment, a capture-recapture analysis was used (4). This approach estimates the total number of cases (T) from overlapping samples of observed cases that are independently identified by two different sources. The analysis indicated that the observed 29 cases of full FAS accounted for approximately 64% of the estimated total number of cases (T=45) **; the estimated prevalence was 1.6 cases per 10,000 (95% confidence interval {CI}=0.9-2.2). The observed 70 cases of combined full and partial FAS accounted for approximately 48% of the estimated total number of cases (T=146). The estimated prevalence derived from capture-recapture analysis was 5.1 cases of full and partial FAS per 10,000 (95% CI=3.0-7.2). During 1981-1989, prevalence of full or full and partial FAS remained stable.

Reported by: Fetal Alcohol Syndrome Prevention Section, Developmental Disabilities Br, and Birth Defects and Genetic Diseases Br, Div of Birth Defects and Developmental Disabilities, National Center for Environmental Health, CDC.

Editorial Note

Editorial Note: The findings in this report indicate that, based on case confirmations conducted in 1997, the observed prevalence of full FAS among children born in Atlanta during 1981-1989 was 1.0 cases per 10,000 live-born infants and 2.5 per 10,000 for both full and partial FAS. These rates are similar to those recently reported by active surveillance systems in nine states participating in the National Birth Defects Prevention Network (range: 0.3-4.2 cases per 10,000) (5). Two other states have used linkage of multiple data sources to calculate prevalences of full FAS. In Alaska, the range of rates for races other than American Indian/Alaskan Native during 1977-1992 was comparable to the rate in this report (6). Unpublished results from the birth defects surveillance system in Colorado also indicated a similar prevalence for FAS among children born during 1989-1993 (L. Miller, M.D., Colorado Department of Public Health and Environment, personal communication, 1997).

The advantages of using existing multiple data sources to calculate the prevalence of FAS are the feasibility of the method and more complete casefinding. Because of the use of existing data sources, the method is less costly than establishing a new surveillance program or actively ascertaining cases by examining a representative selection of the members of a community. In addition, compared with single-source methods (e.g., counting International Classification of Diseases {ICD} codes on newborn hospital discharges), linkage of data sources improves completeness of ascertainment. This completeness can then be further assessed by capture-recapture methods. Use of a uniform case definition, rather than relying on hospital ICD-coded cases of FAS, also enables comparisons of findings from different surveillance systems. To further improve the comparability of different surveillance systems, each clinical criterion of the FAS definition should be standardized.

The findings in this report are subject to at least three limitations. First, because procedures to identify children with FAS did not include a confirmatory physical examination by a dysmorphologist, the validity of this surveillance system could not be formally analyzed. Second, the observed FAS prevalence documented in Atlanta is considered to be a minimum prevalence because of underdiagnosis of FAS, insufficient documentation of the criteria, and loss to follow-up or death before FAS could be diagnosed. The children identified with FAS whose diagnoses were confirmed are most likely those who were more severely affected, a finding supported by the higher rate of completeness of capture for full FAS than for full and partial FAS. Third, the prevalence derived using the capture-recapture method probably is an underestimate because of some "positive dependence" between the two surveillance systems (7) (i.e., children identified by MACDP may have had an increased likelihood of being identified by MADDSP because a child already identified as having FAS probably is more likely to have cognitive evaluations and to enter special education).

The approach and analysis used in this report can be employed by other health jurisdictions in conducting surveillance for FAS. Surveillance is necessary to monitor patterns of FAS and the impact of prevention efforts. For many states, birth defect registries are an existing data source that can be used to develop multiple-source surveillance systems for FAS. Thirty-one states now operate birth defects registries (either active or passive), and five states have recently been funded by CDC to establish multiple-source FAS registries.

References

1. Stratton K, Howe C, Battaglia F. Fetal alcohol syndrome: diagnosis, epidemiology, prevention, and treatment. Washington, DC: National Academy Press, 1996.

2. Streissguth AP, Barr HM, Kogan J, Bookstein FL. Understanding the occurrence of secondary disabilities in clients with fetal alcohol syndrome (FAS) and fetal alcohol effects (FAE). Seattle, Washington: University of Washington Publication Services, 1996.

3. CDC. Update: trends in fetal alcohol syndrome -- United States, 1979-1993. MMWR 1995;44:249-51.

4. Stroup DF. Special analytic issues. In: SM Teutsch, RE Churchill, eds. Principles and practice of public health surveillance. New York, New York: Oxford University Press, 1994:136-49.

5. National Birth Defects Prevention Network. Birth defects surveillance data from selected states. Teratology 1997;56:115-72.

6. Egeland GM, Perham-Hester KA, Gessner BD, Ingle D, Berner J, Middaugh JP. Fetal alcohol syndrome in Alaska, 1977-1992: an administrative prevalence derived from multiple data sources. Am J Public Health 1998 (in press).

7. Brenner H. Use and limitations of the capture-recapture method in disease monitoring with two dependent sources. Epidemiology 1995;6:42-8.

* Full FAS=all four of the following criteria: 1) confirmed maternal alcohol exposure; 2) characteristic facial anomalies (e.g., short palpebral fissures, smooth philtrum, and thin upper lip); 3) growth retardation; and 4) structural central nervous system abnormalities (e.g., microcephaly). Partial FAS=1, 2, and either 3, 4, or 5) a pattern of cognitive or behavioral abnormalities. Alcohol-related neurodevelopmental disorder=1 and either 3, 4, or 5 (1). 

** T={(R + 1)(S + 1)/(C + 1)}-1. Variance (T)=(R + 1)(S + 1) N1N2 / {(C + 1)2 (C + 2)}. The 95% con-fidence interval of (T)=T plus or minus 1.96 Variance (T), where R=the number identified by system 1, S=the number identified by system 2, C=the number identified by both systems, N1=the number uniquely identified by system 1, and N2=the number uniquely identified by system 2.

Disclaimer   All MMWR HTML versions of articles are electronic conversions from ASCII text into HTML. This conversion may have resulted in character translation or format errors in the HTML version. Users should not rely on this HTML document, but are referred to the electronic PDF version and/or the original MMWR paper copy for the official text, figures, and tables. An original paper copy of this issue can be obtained from the Superintendent of Documents, U.S. Government Printing Office (GPO), Washington, DC 20402-9371; telephone: (202) 512-1800. Contact GPO for current prices.

Page converted: 09/19/98

HOME  |  ABOUT MMWR  |  MMWR SEARCH  |  DOWNLOADS  |  RSS |  CONTACT POLICY  |  DISCLAIMER  |  ACCESSIBILITY

Morbidity and Mortality Weekly Report Centers for Disease Control and Prevention 1600 Clifton Rd, MailStop E-90, Atlanta, GA 30333, U.S.A Department of Health and Human Services

This page last reviewed 5/2/01