California birth certificate request

Objective. To evaluate the validity of racial/ethnic information in California birth certificate data.

Data Sources. Computerized birth certificate data and postpartum interviews with California mothers.

Study Design and Data Collection. Birth certificates were matched with face-to-face structured postpartum interviews with 7,428 mothers to compare racial/ethnic information between the two data sources. Interviews were conducted in Spanish or English during delivery stays at 16 California hospitals, 1994-1995.

Principal Findings. The sensitivity of racial/ethnic classification in birth certificate data was very high (94 percent to 99 percent) for African Americans, Asians/Pacific Islanders, Europeans/Middle Easterners, and Latinas (Hispanics). For Native Americans, however, the sensitivity was only 54 percent. The positive predictive value of birth certificate classification of race/ethnicity was high for all racial/ethnic groups (96 percent to 97 percent).

Conclusions. Despite limited training of birth clerks, the maternal racial/ethnic information in California birth certificate data appears to be a valid measure of self-identified race and Hispanic ethnicity for groups other than Native Americans.

Key Words. Race, ethnicity, birth certificates, vital statistics

Racial/ethnic information in birth certificate data is used extensively in monitoring differences among population groups in infant mortality, low birth weight, and utilization of prenatal care. For example, racial/ethnic data in birth certificates will be used to monitor progress toward the goal of The President's Initiative on Race to eliminate racial/ethnic disparities in infant mortality (U.S. Department of Health and Human Services [HHS] 1998). The published literature, however, does not include studies that validate this information, and the quality of information on many items in birth certificates has been questioned (Clark, Fu, and Burnett 1997; Emery et al. 1997; Green, Moore, Adams, et al. 1998; Piper, Mitchel, Snowden, et al. 1993; Woolbright and Harshbarger 1995).

In California birth certificates, "race" and "Hispanic ethnicity" are recorded as separate items. Although they receive no uniform training, birth clerks have access to the California Department of Health Service's birth registration handbook, which instructs them to obtain racial/ethnic information directly from the informant using a race identification worksheet (State of California, Department of Health Services 1994). The extent to which these guidelines are followed is unknown, and ongoing quality control or improvement measures do not exist. Furthermore, the ability of birth clerks to obtain accurate information from non-English-speaking persons is uncertain in that multilingual skills are not a requirement for the job. One concern is that birth clerks may at times base "race" and "Hispanic ethnicity" on their observations rather than on mothers' self-identification. Racial/ethnic information obtained by self-report may differ substantially from that obtained through observation, especially for some rac ial/ethnic subgroups. For example, one study showed that self-identified "race" corresponded well with observer-identified "race" for African Americans and whites, but that it showed significant discrepancies for Native Americans (Hahn, Truman, and Barker 1996).

This study was conducted to determine the validity of information in California birth certificate data on "race" and "Hispanic ethnicity." Because the "race" and "Hispanic ethnicity" assigned to infants in the United States since 1989 has been that of the mother, we focused on maternal "race" and "Hispanic ethnicity."

METHODS

Data Sources and Matching Methods

Birth certificates were matched with data from face-to-face structured postpartum interviews to compare reported racial/ethnic information between the two data sources. The interviews were part of a larger statewide study of prenatal care use among mothers who gave birth in California hospitals from August 1994 through July 1995; survey methods have been described previously (Braveman, Pearl, Egerter, et al. 1998; Braveman, Egerter, and Marchi 1999). A total of 10,165 mothers were interviewed during their delivery stays at 19 randomly selected hospitals in California. We used 1991 statewide birth certificate data to select hospitals by stratified random sampling according to their delivery population characteristics (proportion of African American births, geographic region of the state, and prevalence of privately insured deliveries). Military hospitals, hospitals participating in California's Centers for Disease Control and Prevention--affiliated Pregnancy Risk Assessment Monitoring System, and hospitals wit h fewer than 600 deliveries in 1991 were excluded. These exclusions eliminated approximately 19 percent of all women older than 15 years who delivered statewide during the study period.

For 17 of the 19 hospitals in the larger survey sample, birth records were retrieved from the Automated Vital Statistics System (AVSS) (Williams, Marinko, and Shields 1983) and the California Automated Registration and Entry (CARE), two electronic birth certificate systems (Starr and Starr 1995). Two hospitals had to be excluded because they did not use AVSS or CARE and hence did not have complete electronic birth certificate data that were accessible for this study. Interview data were linked with electronic birth certificate files through use of a matching algorithm based on a combination of variables common to both data sources: delivery hospital, delivery date, mother's birth date, mother's first name, mother's maiden name, and child's last name. (A full description of the linkage algorithm is available on request.) The match rate of one of the survey hospitals was much lower than that of the other 16 hospitals, raising concerns that a significant number of false matches could result; therefore this hosp ital was excluded. The remaining 16 hospitals included in this validation study were diverse in ownership type. Semi-structured interviews were conducted with birth clerks at the 16 hospitals to obtain information on the procedures that they used for completing birth certificates.

A woman present on the postpartum ward in a selected hospital was eligible to participate if she had a live birth during this hospital stay, spoke English or Spanish, was at least 15 years old, was an "emancipated minor" or financially self-sufficient if she was 15-17 years old, and was not incarcerated then or during her pregnancy. She was ineligible if nursing staff believed that the interview would interfere with her care. Of the women on the postpartum wards during staffed hours in the 16 hospitals (n = 14,332), 1,123 (7.8 percent) were ineligible: 458 (3.2 percent) because of language, 291 (2.0 percent) because of age/emancipated minor status, 280 (1.9 percent) because of nurse disapproval, and 94 (0.7 percent) for other reasons. Of the 13,209 women eligible to participate, 9,023 (68.3 percent) were approached for an interview, and 4,186 (31.7 percent) had been discharged before interviewers could approach them. A total of 7,712 women completed the interviews, representing 85.5 percent of the eligible w omen who were approached (58.4 percent of all eligible women). Of these women, 7,428 (96.8 percent) were successfully matched with corresponding birth certificates. Women with matched and non-matched records had similar sociodemographic characteristics (income, age, race/ethnicity, language, education, parity, and insurance).

As seen in Table 1, the validation study sample had a larger proportion of Latina women and fewer European/Middle Eastern women compared with all women who delivered in California in 1994. Other differences between the study sample and all women who gave birth in California that year were small (column 3 of Table 1).

Categorization of Race/Ethnicity

Table 2 shows the exact terms, in both the birth certificate and the postpartum survey, that were used to define the racial/ethnic categories for comparison in the validation study. (A technical appendix [available on request] includes specifications for the recording of racial/ethnic information in birth certificates and the postpartum survey.)