Genetic risk factors are a critical component of many epidemiologic studies; however, concerns about genetic research might affect participants' willingness to enroll. The authors assessed factors associated with completion of mailed buccal-cell collection kits following telephone interviews at the Atlanta, Georgia, study site of the National Birth Defects Prevention Study. Pregnant women who were interviewed after June 30, 1999, and had an estimated delivery date of December 31, 2002, or earlier were included (n = 1,606). For this time period, overall interview participation was 71.9%. Among those interviewed, 47.6% completed the buccal-cell collection kit (61.1% of non-Hispanic Whites, 34.9% of non-Hispanic Blacks, and 39.1% of Hispanics). Non-Hispanic White race/ethnicity, an English-language (vs. Spanish) interview, receipt of a redesigned mailing packet and an additional $20 incentive, and consumption of folic acid were associated with higher buccal-cell kit participation. Among non-Hispanic White mothers, higher education, intending to become pregnant, and having a child with a birth defect were associated with increased participation. Among non-Hispanic Black mothers, receipt of the redesigned packet and$20 incentive was associated with increased participation. Among Hispanic mothers, an English-language interview, higher education, and receipt of the redesigned packet and $20 incentive were associated with increased participation. At this study site, minority groups were less likely to participate in DNA collection. Factors associated with participation varied by race/ethnicity. In recent years, epidemiologic studies have expanded to include not only demographic and environmental risk factors but also genetic risk factors. There is considerable concern among the general public about the risks of genetic testing (1). The inclusion of a genetic component in a research study might affect people's willingness to enroll in the study and, hence, the sample size available for analysis and the generalizability of the results. Analyses of participation rates in a cancer registry and in smoking studies with genetic components have shown that sociodemographic characteristics such as race/ethnicity, age, and health behavior are associated with participation (24). We were interested in determining what factors were associated with participation in the DNA collection component of the population-based National Birth Defects Prevention Study (NBDPS), an ongoing US study of risk factors for birth defects. The NBDPS requires self-collection of DNA through the completion and return of a buccal-cell collection kit. We focused our analysis on parents and children in metropolitan Atlanta, Georgia, because of the low overall rates of participation in buccal-cell collection in this population. To optimize participation in the genetic component of this study, it is important to understand the relation between participation and various demographic and potentially motivating factors. In the 1-hour NBDPS telephone interview, which precedes the buccal-cell collection, study personnel obtain extensive information on the mother's demographic characteristics, lifestyle, and health behaviors. This provides a valuable opportunity to perform detailed analyses of factors potentially associated with participation in the buccal-cell collection. ## MATERIALS AND METHODS The NBDPS is an ongoing, population-based, multisite case-control study designed to investigate environmental and genetic risk factors for major birth defects among women aged 18 years or more at the time of the interview. This analysis was limited to participants from the Atlanta study site of the NBDPS. The metropolitan Atlanta area has a diverse population and has had low participation in the buccal-cell collection portion of the study. Participants were included in the buccal-cell collection analysis if they had an estimated delivery date of December 31, 2002, or earlier and had completed an interview after June 31, 1999, when buccal-cell collection was added at the Atlanta study site of the NBDPS. The study protocol requires that the mother be contacted and the interview be completed 6 weeks to 24 months after the estimated due date. The average number of months from due date to completed interview was 9 for cases and 6 for controls. Details on the general methods of the study and DNA collection have been published elsewhere (5, 6). Children with birth defects were identified through an existing birth defects surveillance system in metropolitan Atlanta (7). Information on case infants was obtained from hospital reports and medical records and reviewed by a clinical geneticist to ensure that the infant's birth defect(s) met the NBDPS case definition for major birth defects. Case infants with recognized or strongly suspected single-gene conditions or chromosomal abnormalities were excluded from the study, because the purpose of the study is to determine the causes of birth defects of unknown origin. Over 30 major birth defects are included in the study. The most common classes of birth defects among the Atlanta case infants (n = 1,248) were heart defects (n = 529), orofacial clefts (n = 182), second- and third-degree hypospadias (n = 143), gastrointestinal defects (n = 113), and neural tube defects (n = 76). Control infants were liveborn infants with no major birth defects who were randomly selected from the same population as the case infants, using birth certificates or birth hospital records. It is the goal at the Atlanta study site to interview approximately 300 mothers of case infants and 100 mothers of control infants per year. Initially, mothers were sent an advance package that contained a description of both parts of the study (the telephone interview and buccal-cell self-collection) and a$20 money order as an incentive, which the mother was advised to keep, regardless of whether or not she chose to participate in the study. The introductory materials were followed by a telephone call asking the mothers whether they were willing to participate in the study. If consent was given, a computer-assisted telephone interview was conducted with the mother to assess environmental exposures and maternal and paternal risk factors. All components of the study interview were conducted in English or Spanish, according to the participant's preference. At the end of the interview, the mother was informed that a buccal-cell collection kit which included a second $20 money order would be sent to her address. After receiving the kit, the mother decided whether or not to participate in buccal-cell collection and was again told that the$20 could be kept regardless of whether she chose to complete the buccal-cell kit.

In October 2002, a new buccal-cell collection kit was introduced that had been redesigned for visual appeal and ease of use. Prior to October 2002, buccal-cell collection brushes, trifolded copies of the instructions, informed consent forms, a second $20 money order, and a stamped return envelope were mailed to the mother in a plain envelope as multiple loose-sheet documents. The mailing kit was redesigned to present these materials in a glossy 8- × 11-inch (20- × 28-cm) color NBDPS portfolio that conveniently stored all of the information and the brushes. Brushes for each household member were placed in color-coded envelopes in the portfolio (red for mother, blue for father, and yellow for baby). Quotes from other participants were added to the inside of the portfolio. The kit included a cover letter; illustrated instructions; brushes for mother, infant, and father; the second money order incentive; informed consent forms; and a stamped return envelope. The mailed kit contained brushes only for current household members; brushes were not included for fathers who were not current household members or for any child who was not alive at the time of the interview. The consent forms were tailored specifically for each particular household arrangement. At the same time the kit was redesigned, a third$20 incentive was added for participants who returned the buccal-cell kit, to compensate them for their time and encourage them to return the completed kit. Mothers were told of the additional incentive in the initial contact letter. Mothers received up to two reminder calls and three reminder letters asking them if they had any questions and encouraging them to return the kit.

For this analysis, a completed buccal-cell kit was defined as return of any of the buccal brushes and a signed consent form. The nonresponders included both persons who refused to participate by actively declining participation and those who simply did not return the kit.

Our analyses compared mothers who returned the buccal-cell kit with those who did not return the kit. We used multiple logistic regression to examine demographic and other factors potentially associated with participation. The following variables were assessed: maternal education (≤12 years vs. >12 years), maternal age (≤25 years vs. >25 years), household income (≤$20,000 vs. >$20,000), language of interview (English vs. Spanish), case/control status, use of folic acid supplements or multivitamins (any use during the month before pregnancy and the first month of pregnancy vs. no use during this period), pregnancy intendedness (whether the woman wanted to become pregnant at the time of conception and/or stopped using contraception to become pregnant vs. wanting to wait until later or not wanting to become pregnant at all), vital status of the baby at the time of the interview (living vs. not living), preterm delivery (<37 weeks vs. ≥37 weeks), mother's birth in the United States (yes vs. no), any smoking from 3 months before pregnancy through the end of pregnancy (yes vs. no), alcohol use from 3 months before pregnancy through the end of pregnancy (any use vs. no use during that period), employment of the mother outside the home (yes vs. no), gravidity (primigravid vs. multigravid), and receipt of the redesigned buccal-cell collection kit with the added incentive (before introduction of the new kit vs. afterward). Racial and ethnic information was self-reported in the interview as non-Hispanic White, non-Hispanic Black, Hispanic, Asian or Pacific Islander, Native American or Alaska Native, other, refusal to answer, or unknown. This racial and ethnic terminology is used in this report.

## RESULTS

From 1999 to 2002, there were 198,323 births at the metropolitan Atlanta NBDPS study site (76,889 births to non-Hispanic Whites (38.8 percent), 77,264 births to non-Hispanic Blacks (39.9 percent), 32,087 births to Hispanics (16.2 percent), 10,216 births to Asians or Pacific Islanders (5.1 percent), and 1,867 births to women of other races or for which race/ethnicity information was missing (0.9 percent)) (table 1). Among mothers eligible for the interview with births in 1999–2002, the overall participation rate in the NBDPS interview was 71.9 percent. The participation rate was highest among non-Hispanic Whites (78.3 percent), followed by Non-Hispanic Blacks (70.1 percent), Hispanics (62.9 percent), Asians or Pacific Islanders (61.0 percent), and persons of other racial/ethnic groups or with missing data (56.3 percent) (table 1). Only limited information was available on the nonparticipants in the interview.

TABLE 1.

Rates of participation in the study interview and buccal-cell collection at the Atlanta, Georgia, site of the National Birth Defects Prevention Study, by race/ethnicity, 1999–2002

Study component

Non-Hispanic White

Non-Hispanic Black

Hispanic

Asian or Pacific Islander

Other or missing information

Total

No.

%

No.

%

No.

%

No.

%

No.

%

No.

%

Births occurring in 1999–2002
All births in study area 76,889 38.8 77,264 39.9 32,087 16.2 10,216 5.1 1,867 0.9 198,323 100
No. of mothers eligible for interview* 936  797  326  59  64  2,182
Interview refusals 115 12.3 96 12.0 40 12.3 14 23.7 14.1 274 12.6
Inability to locate mother for interview 63 6.7 126 15.8 62 19.0 15.2 14 21.9 274 12.6
Case ascertained >24 months after estimated due date 25 2.7 16 2.0 19 5.8 7.8 65 3.0
Mother interviewed 733 78.3 559 70.1 205 62.9 36 61.0 36 56.3 1,569 71.9
Interview conducted after June 30, 1999,and an estimated due date on or before December 31, 2002
Mother completed interview 730 45.5 559 34.8 230 14.3 40 2.5 47 2.9 1,606 100
Mother completed buccal-cell collection kit 446 58.3 195 25.5 90 11.8 16 2.1 18 2.4 765 100
Rate of participation in buccal-cell collection

61.1

34.9

39.1

40.0

38.3

47.6

Study component

Non-Hispanic White

Non-Hispanic Black

Hispanic

Asian or Pacific Islander

Other or missing information

Total

No.

%

No.

%

No.

%

No.

%

No.

%

No.

%

Births occurring in 1999–2002
All births in study area 76,889 38.8 77,264 39.9 32,087 16.2 10,216 5.1 1,867 0.9 198,323 100
No. of mothers eligible for interview* 936  797  326  59  64  2,182
Interview refusals 115 12.3 96 12.0 40 12.3 14 23.7 14.1 274 12.6
Inability to locate mother for interview 63 6.7 126 15.8 62 19.0 15.2 14 21.9 274 12.6
Case ascertained >24 months after estimated due date 25 2.7 16 2.0 19 5.8 7.8 65 3.0
Mother interviewed 733 78.3 559 70.1 205 62.9 36 61.0 36 56.3 1,569 71.9
Interview conducted after June 30, 1999,and an estimated due date on or before December 31, 2002
Mother completed interview 730 45.5 559 34.8 230 14.3 40 2.5 47 2.9 1,606 100
Mother completed buccal-cell collection kit 446 58.3 195 25.5 90 11.8 16 2.1 18 2.4 765 100
Rate of participation in buccal-cell collection

61.1

34.9

39.1

40.0

38.3

47.6

*

Mothers of all infants with a major birth defect meeting the study definition and a random sample of 100 infants per year with no major birth defect (controls) were eligible for interview.

Buccal-cell collection for the Atlanta study site began with interviews completed on July 1, 1999.

A total of 1,606 interviewed mothers were eligible to receive a buccal-cell collection kit (1,248 case mothers and 358 control mothers). The buccal-cell kit was completed by 47.6 percent of the interviewed mothers. Participation rates were highest among non-Hispanic Whites (61.1 percent), followed by Asians/Pacific Islanders (40.0 percent), Hispanics (39.1 percent), persons of other races or with missing data (38.3 percent), and non-Hispanic Blacks (34.9 percent) (table 1). Although women were not asked why they refused to participate in buccal-cell collection, comments made to the interviewer were recorded. Frequent reasons given were that the woman did not want to participate, was not interested, disliked genetics, was concerned about privacy, or was forbidden to participate by a family member.

We examined a number of factors that might have been associated with participation in buccal-cell collection among those women who received a buccal-cell collection kit. Descriptive analysis showed that many of these characteristics, such as maternal age, education, and income, varied by race/ethnicity for the three largest racial/ethnic groups (figure 1). Crude odds ratios showed that many of these factors were also associated with increased or decreased participation, although there were no consistent patterns of association across racial/ethnic groups (table 2).

FIGURE 1.

Distribution of interview participants' characteristics by race/ethnicity at the Atlanta, Georgia, site of the National Birth Defects Prevention Study, 1999–2002. Alcohol and smoking exposures were measured as any use during the period from 3 months before pregnancy through the end of pregnancy (B3–P9). Folic acid or multivitamin exposure was measured as any use of folic acid or a multivitamin containing folic acid during the period from 1 month before pregnancy through the first month of pregnancy (B1–P1).

FIGURE 1.

Distribution of interview participants' characteristics by race/ethnicity at the Atlanta, Georgia, site of the National Birth Defects Prevention Study, 1999–2002. Alcohol and smoking exposures were measured as any use during the period from 3 months before pregnancy through the end of pregnancy (B3–P9). Folic acid or multivitamin exposure was measured as any use of folic acid or a multivitamin containing folic acid during the period from 1 month before pregnancy through the first month of pregnancy (B1–P1).

TABLE 2.

Crude odd ratios for factors potentially associated with participation in buccal-cell collection at the Atlanta, Georgia, site of the National Birth Defects Prevention Study, by race/ethnicity, 1999–2002

Non-Hispanic White

Non-Hispanic Black

Hispanic

Asian or Pacific Islander

Other or missing information

Total

OR*

95% CI*

OR

95% CI

OR

95% CI

OR

95% CI

OR

95% CI

OR

95% CI

Maternal age ≥25 years 1.78 1.18, 2.70 0.94 0.65, 1.37 1.16 0.68, 2.00 1.85 1.37, 2.50 1.22 0.36, 4.35 1.61 1.28, 2.00
Household income >$20,000/year 2.00 1.20, 3.23 1.15 0.79, 1.67 1.18 0.65, 2.13 1.89 1.37, 2.59 0.83 0.19, 3.70 1.82 1.45, 2.33 Maternal education >12 years 1.96 1.37, 2.94 0.74 0.51, 1.05 2.63 1.45, 4.90 5.00 0.54, 50.00 1.05 0.30, 3.65 1.66 1.35, 2.05 Mother born in theUnited States 1.19 0.69, 2.07 1.38 0.72, 2.62 2.81 1.21, 6.51 0.73 0.06, 8.83 0.32 0.09, 1.12 1.51 1.19, 1.92 Mother reported working during pregnancy 1.28 0.91, 1.80 1.05 0.65, 1.70 1.45 0.83, 2.52 0.77 0.21, 2.80 1.17 0.29, 4.74 1.15 0.91, 1.44 Child had a birth defect 1.65 1.14, 2.38 1.00 0.66, 1.50 0.71 0.39, 1.28 0.44 0.11, 1.80 4.44 0.49, 40.34 1.23 0.97, 1.56 Pregnancy was intended 1.69 1.23, 2.33 0.97 0.68, 1.39 1.47 0.86, 2.50 1.32 0.35, 5.05 0.86 0.26, 2.79 1.70 1.39, 2.07 Use of folic acid or multivitamin supplements (B1–P11.48 1.06, 2.07 1.16 0.81, 1.65 1.87 1.00, 3.50 1.30 0.36, 4.70 0.75 0.23, 2.43 1.83 1.50, 2.23 Any use of alcohol (B3–P91.24 0.92, 1.68 0.84 0.58, 1.20 1.06 0.60, 1.86 2.33 0.45, 12.23 2.23 0.67, 7.40 1.35 1.11, 1.65 Any smoking (B3–P9) 0.76 0.53, 1.10 0.73 0.43, 1.24 2.38 0.87, 6.49 0.59 0.45, 0.76 0.53 0.12, 2.31 0.96 0.73, 1.25 Preterm birth (<37 weeks) 0.86 0.60, 1.24 1.00 0.69, 1.36 1.16 0.62, 2.20 1.62 0.28, 9.23 0.44 0.10, 1.93 0.85 0.68, 1.06 Baby living 0.96 0.52, 1.77 0.68 0.37, 1.25 1.32 0.48, 3.65 1.75 1.21, 2.33 1.64 1.3, 2.07 0.98 0.67, 1.43 Primigravid vs. multigravid 0.91 0.67, 1.25 1.24 0.83, 1.84 1.39 0.80, 2.42 1.56 0.42, 5.72 0.18 0.03, 0.92 1.11 0.89, 1.37 English-language (vs. Spanish) interview N/A* N/A 3.08 1.67, 5.66 N/A N/A N/A Receipt of redesigned buccal-cell collection kit with additional$20 incentive

1.09

0.75, 1.56

1.83

1.22, 2.75

1.51

0.83, 2.74

0.56

0.12, 2.56

1.41

0.41, 4.76

1.28

1.02, 1.64

Non-Hispanic White

Non-Hispanic Black

Hispanic

Asian or Pacific Islander

Other or missing information

Total

OR*

95% CI*

OR

95% CI

OR

95% CI

OR

95% CI

OR

95% CI

OR

95% CI

Maternal age ≥25 years 1.78 1.18, 2.70 0.94 0.65, 1.37 1.16 0.68, 2.00 1.85 1.37, 2.50 1.22 0.36, 4.35 1.61 1.28, 2.00
Household income >$20,000/year 2.00 1.20, 3.23 1.15 0.79, 1.67 1.18 0.65, 2.13 1.89 1.37, 2.59 0.83 0.19, 3.70 1.82 1.45, 2.33 Maternal education >12 years 1.96 1.37, 2.94 0.74 0.51, 1.05 2.63 1.45, 4.90 5.00 0.54, 50.00 1.05 0.30, 3.65 1.66 1.35, 2.05 Mother born in theUnited States 1.19 0.69, 2.07 1.38 0.72, 2.62 2.81 1.21, 6.51 0.73 0.06, 8.83 0.32 0.09, 1.12 1.51 1.19, 1.92 Mother reported working during pregnancy 1.28 0.91, 1.80 1.05 0.65, 1.70 1.45 0.83, 2.52 0.77 0.21, 2.80 1.17 0.29, 4.74 1.15 0.91, 1.44 Child had a birth defect 1.65 1.14, 2.38 1.00 0.66, 1.50 0.71 0.39, 1.28 0.44 0.11, 1.80 4.44 0.49, 40.34 1.23 0.97, 1.56 Pregnancy was intended 1.69 1.23, 2.33 0.97 0.68, 1.39 1.47 0.86, 2.50 1.32 0.35, 5.05 0.86 0.26, 2.79 1.70 1.39, 2.07 Use of folic acid or multivitamin supplements (B1–P11.48 1.06, 2.07 1.16 0.81, 1.65 1.87 1.00, 3.50 1.30 0.36, 4.70 0.75 0.23, 2.43 1.83 1.50, 2.23 Any use of alcohol (B3–P91.24 0.92, 1.68 0.84 0.58, 1.20 1.06 0.60, 1.86 2.33 0.45, 12.23 2.23 0.67, 7.40 1.35 1.11, 1.65 Any smoking (B3–P9) 0.76 0.53, 1.10 0.73 0.43, 1.24 2.38 0.87, 6.49 0.59 0.45, 0.76 0.53 0.12, 2.31 0.96 0.73, 1.25 Preterm birth (<37 weeks) 0.86 0.60, 1.24 1.00 0.69, 1.36 1.16 0.62, 2.20 1.62 0.28, 9.23 0.44 0.10, 1.93 0.85 0.68, 1.06 Baby living 0.96 0.52, 1.77 0.68 0.37, 1.25 1.32 0.48, 3.65 1.75 1.21, 2.33 1.64 1.3, 2.07 0.98 0.67, 1.43 Primigravid vs. multigravid 0.91 0.67, 1.25 1.24 0.83, 1.84 1.39 0.80, 2.42 1.56 0.42, 5.72 0.18 0.03, 0.92 1.11 0.89, 1.37 English-language (vs. Spanish) interview N/A* N/A 3.08 1.67, 5.66 N/A N/A N/A Receipt of redesigned buccal-cell collection kit with additional$20 incentive

1.09

0.75, 1.56

1.83

1.22, 2.75

1.51

0.83, 2.74

0.56

0.12, 2.56

1.41

0.41, 4.76

1.28

1.02, 1.64

*

OR, odds ratio; CI, confidence interval; N/A, not applicable.

B1–P1 is the period from 1 month before conception to the end of the first month of pregnancy.

B3–P9 is the period from 3 months before pregnancy to the end of the pregnancy.

We examined 15 factors using backward multiple logistic regression. Race/ethnicity, an English-language interview, receipt of the redesigned packet with the added incentive, and folic acid consumption were associated with changes in participation in buccal-cell collection (table 3). Because race/ethnicity was strongly associated with participation in the genetic component of the study, as well as with many of the characteristics studied, the remainder of the analyses were stratified by race/ethnicity.

TABLE 3.

Adjusted effect estimates for participation in buccal-cell collection at the Atlanta, Georgia, site of the National Birth Defects Prevention Study, 1999–2002

Odds ratio*

95% confidence interval

Race/ethnicity
Non-Hispanic White 1.00
Non-Hispanic Black 0.37 0.29, 0.47
Hispanic 1.09 0.65, 1.84
Asian or Pacific Islander 0.45 0.23, 0.87
Other or missing information 0.42 0.23, 0.78
English-language interview 3.27 1.85, 5.79
Receipt of redesigned buccal-cell collection kit with additional $20 incentive 1.44 1.12, 1.84 Folic acid use 1 month before pregnancy or during first month of pregnancy 1.34 1.07, 1.67 Odds ratio* 95% confidence interval Race/ethnicity Non-Hispanic White 1.00 Non-Hispanic Black 0.37 0.29, 0.47 Hispanic 1.09 0.65, 1.84 Asian or Pacific Islander 0.45 0.23, 0.87 Other or missing information 0.42 0.23, 0.78 English-language interview 3.27 1.85, 5.79 Receipt of redesigned buccal-cell collection kit with additional$20 incentive 1.44 1.12, 1.84
Folic acid use 1 month before pregnancy or during first month of pregnancy

1.34

1.07, 1.67

*

All effect estimates were obtained by logistic regression.

Reference category.

Among non-Hispanic Whites (n = 730), having more than 12 years of education, intending to become pregnant at the time of conception, and having a child with a major birth defect were associated with increased participation in the buccal-cell collection component of the study (table 4). In contrast, among non-Hispanic Black mothers (n = 559), only receiving the redesigned packet with the additional incentive was associated with increased participation in the buccal-cell collection component. Participation for this group increased from 31.7 percent before the packet was redesigned to 46.0 percent after the packet was redesigned and the additional incentive was offered (odds ratio (OR) = 1.82, 95 percent confidence interval (CI): 1.22, 2.78). When data were stratified by household income (≤$20,000 per year vs. >$20,000 per year), the association with the new packet was stronger among women with an annual household income of $20,000 or less (28.4 percent before the redesign vs. 61.3 percent after; OR = 4.00, 95 percent CI: 1.80, 8.85) than in those with an annual household income of more than$20,000 (35.3 percent before the redesign vs. 41.1 percent after; OR = 1.28, 95 percent CI: 0.78, 2.09). This association with income was not seen among the other racial/ethnic groups.

TABLE 4.

Adjusted effect estimates for factors associated with participation in buccal-cell collection at the Atlanta, Georgia, site of the National Birth Defects Prevention Study, stratified by race/ethnicity, 1999–2002

Race/ethnicity* and factor

Odds ratio

95% confidence interval

Non-Hispanic White
Maternal education >12 years 1.80 1.23, 2.62
Pregnancy was intended 1.52 1.10, 2.12
Child had a birth defect 1.68 1.16, 2.44
Non-Hispanic Black or African-American
Received redesigned buccal-cell collection kit with additional $20 incentive 1.82 1.22, 2.78 Hispanic Interviewed in English 3.00 1.56, 5.78 Maternal education >12 years 2.14 1.13, 4.07 Received redesigned buccal-cell collection kit with additional$20 incentive

1.96

1.03, 3.72

Race/ethnicity* and factor

Odds ratio

95% confidence interval

Non-Hispanic White
Maternal education >12 years 1.80 1.23, 2.62
Pregnancy was intended 1.52 1.10, 2.12
Child had a birth defect 1.68 1.16, 2.44
Non-Hispanic Black or African-American
Received redesigned buccal-cell collection kit with additional $20 incentive 1.82 1.22, 2.78 Hispanic Interviewed in English 3.00 1.56, 5.78 Maternal education >12 years 2.14 1.13, 4.07 Received redesigned buccal-cell collection kit with additional$20 incentive

1.96

1.03, 3.72

*

No significant associations were observed for Asians/Pacific Islanders.

All variables are dichotomous, with the unexposed used as the referent group.

All effect estimates were obtained by logistic regression.

Of the women who identified themselves as non-Hispanic Black, 8.8 percent had been born outside the United States. Buccal-cell participation rates were similar between the two groups of non-Hispanic Black women; 28.4 percent of those born outside the United States completed buccal-cell kits compared with 35.5 percent of those born in the United States (OR = 0.73, 95 percent CI: 0.38, 1.39).

Among Hispanic mothers (n = 230), completing the telephone interview in English (vs. Spanish) was associated with increased participation in the buccal-cell collection component of the study, as was receiving the redesigned buccal-cell kit with the additional incentive (table 4). Having more than 12 years of education was also associated with increased participation. Seventy-two percent of the Hispanic mothers completing the interview at the Atlanta study site identified themselves as Mexican, Mexican-American, or Chicana, with the remainder identifying themselves as members of over 25 other Hispanic groups.

The logistic regression model showed no significant associations for Asians or Pacific Islanders. At the Atlanta study site, this group of participants is small (n = 40), with very diverse backgrounds. For example, 37 of the women had been born outside of the United States in one of 11 different countries.

## DISCUSSION

Our findings suggest that participation in the buccal-cell collection component of the NBDPS was strongly associated with general characteristics such as race/ethnicity, language preference, and inclusion of additional incentives as part of a redesigned buccal-cell collection kit. Of particular interest, non-Hispanic Black mothers were willing to participate in an extensive personal interview about pregnancy exposures (70.1 percent); however, only 34.8 percent of those who completed the interview were willing to contribute a DNA sample. The Asian/Pacific Islander and Hispanic populations also showed lower participation rates for both the interview (61.0 percent and 62.9 percent, respectively) and the DNA collection (40.0 percent and 39.1 percent) as compared with non-Hispanic Whites. These low participation rates will limit both the generalizability of the study findings and the ability to assess the impact of genetic polymorphisms common in mothers from diverse racial/ethnic backgrounds.

The association of race/ethnicity with participation in this study of birth defects was consistent with findings in other studies of participation in DNA collection. An analysis of women who had previously participated in a population-based case-control study in North Carolina and who were invited to join a cancer genetics registry found that participation rates in those contacted were 58 percent among Whites and 37 percent among Blacks (3). Race was the only predictor of participation in that study (3). Additionally, Kozlowski et al. (4) showed that 57 percent of respondents to a random digit dialing survey about smoking agreed to have a buccal-cell kit sent to them to consider participation in DNA collection. Of those who agreed to receive the kit, 45 percent returned it. White race and a college education were associated with increased participation (4). A study of primary-care patients aged 65 years or more who agreed to participate in an at-home interview showed that race and age were strongly associated with participation (8). The participants were asked in person to provide a buccal swab for DNA testing and were then asked whether the sample could be banked for later use of the DNA. The participation rate for buccal-cell collection among those interviewed at home was 88 percent, and consent for DNA banking was 78 percent. Increased age (≥80 years) was associated with reduced participation in buccal-cell collection. Compared with Whites, African Americans were less likely to participate in buccal-cell collection.

These studies show that, even among people willing to participate in other types of research, participation in genetic components might be lower, especially among non-Hispanic Blacks. Although the at-home study of older participants (8) was quite different from our study, the results were similar, suggesting that reluctance to participate in genetic studies among non-Hispanic Blacks is not limited to women of childbearing age.

As a group, non-Hispanic White mothers had the highest buccal-cell collection participation rate; 61.1 percent of those who completed the interview returned the buccal-cell kit. Among non-Hispanic White mothers, higher educational level, having a child with a birth defect, and wanting to become pregnant at the time of conception were associated with greater participation in the genetic component of the study. The association between higher education and greater participation among non-Hispanic White mothers is consistent with results reported by Audrian et al. (9), who found that for non-White parents of adolescents, participation rates in genetic studies of smoking behavior remained at 76 percent with increased education, whereas for Whites, participation increased from 77 percent to 89 percent with more than a high school education. Higher education was not associated with increased buccal-cell collection participation among non-Hispanic Black mothers in our study (table 2). In fact, there was a trend towards an inverse association between education and buccal-cell collection participation among non-Hispanic Black mothers in our study.

The low rate of participation of non-Hispanic Black and Hispanic mothers in the buccal-cell collection portion of our study is of particular concern, because it could limit our ability to identify genetic risk factors and gene-environment interactions in these populations. Genetic polymorphisms often vary by race/ethnicity, and analysis of genetic risk factors among particular racial/ethnic groups will not necessarily yield information that is informative about the risks to other groups (10). For example, Bamshad (10) showed that many common genetic variants were not shared between populations. The racial/ethnic disparity in participation will result in a disparity in our understanding of the causes of birth defects.

Our data suggest that participation among non-Hispanic Black and Hispanic mothers was substantially improved by the use of the redesigned buccal-cell collection kit and the addition of a third \$20 incentive following return of the completed kit. As such strategies are shown to be effective, it is important to incorporate protocol changes into epidemiologic studies to minimize disparities in participation.

Among our non-Hispanic White mothers, as well as in the total sample, pregnancy-intendedness was associated with increased participation. Currently, there is no literature on the relation between pregnancy planning and participation in scientific research. It is possible that these mothers were more likely to participate because their pregnancies were planned, or this measure could have been a proxy for some other characteristic not ascertained by our interview. Among non-Hispanic Whites, mothers were more likely to participate in buccal-cell collection if their child was affected by a birth defect. It is possible that having a child with a birth defect could increase the perceived value of the study, despite the parent's knowledge that participation in the study will provide no personal benefit. Having a child with a birth defect might increase the parent's knowledge of birth defects and the importance of research in this field. Additional research is needed to elucidate how these factors affect participation in genetic research.

NBDPS research offers no direct benefits from genetic testing to the participants. However, predictive clinical testing of women with a family history of breast cancer for BRCA1/2 has shown reduced participation rates for African Americans, even when there is the potential for a clinical benefit to the individual. Armstrong et al. (2) found that African Americans with a family history of breast cancer were less likely to undergo BRCA1/2 genetic counseling and testing even after adjustment for the probability of carrying the mutation, socioeconomic status, cancer risk perception and worry, primary-care physician counseling, and attitudes about the benefits of BRCA1/2 testing. The reasons for lower participation in genetic testing and genetic research must be elucidated so that the potential benefits of these technologies can have an equal impact in all communities.

Thompson et al. (11) interviewed women at a hospital in Harlem, New York, about their perceptions of genetic testing. In a hierarchical linear regression model, only Spanish language preference and medical mistrust among African Americans were significant predictors of concern about abuse of genetic testing (11). This is consistent with the results from our study. Spanish language preference was strongly associated with reduced participation in the genetic component of the present study. Latino and African-American focus group participants have expressed concerns about whether or not members of their racial/ethnic group will actually benefit from genetic research or whether the benefits will go primarily to others (12). In addition, the Hispanic population comprises persons of many diverse ethnic and cultural backgrounds that may function differently from one another (13). The buccal-cell kit instructions are translated into traditional Spanish, and issues of general readability and readability across Spanish-language dialects need to be explored. Anecdotal reports from NBDPS interviewers suggest that some Spanish-speaking mothers have difficulty reading or understanding the materials.

It is possible that our low participation rates might have been due to specific difficulties with the collection of buccal-cell DNA. However, buccal-cell collection is a minimally invasive technique in which a soft brush is used to collect cells from the inside of the cheek. The kits are mailed to the participants' homes, and a postage-paid return envelope is provided. This method is convenient for mothers of small children because it requires a limited amount of time and omits the difficulty and expense of travel. Blood collection, although ideal for researchers because of the higher quality and quantity of DNA, is often not practical for infants and can require difficult scheduling and expensive travel by the family or a phlebotomist. Mailed mouthwash kits are a good option because of ease of completion and a high-quality DNA yield (14) but are impractical for infants. Previous studies have shown that mailed buccal-cell collection kits are suitable for epidemiologic studies (15) and might be the best option for this population of parents and infants (16, 17). Although we think this method of DNA collection is currently the best option for this population, there is a critical need for information on barriers to participation experienced by different groups and strategies for increasing participation. The availability of a health-care professional for in-person buccal-cell collection may increase participation for persons who are uncomfortable with self-collection or with mailing the kit. Additional research into this collection method and the associated increase in cost is needed.

One of the strengths of this study was that this was a large, diverse, population-based sample from the metropolitan Atlanta area. However, because this study was only of participants in metropolitan Atlanta, the results might not be generalizable to participants in rural areas or other regions of the country. Another strength of this study is that women were interviewed before the buccal-cell kits were mailed and thus had an opportunity to consent separately to the two different components of the study. We were therefore able to analyze data on many factors provided by the participant in the interview, while in many other studies the characteristics of nonparticipants are unknown.

Some limitations of this study were that nonresponses included both active refusals (provided an explicit refusal statement) and passive refusals (did not return the buccal-cell collection kit, with no reason given) and that participants were not asked why they would not participate in the genetic component of the study.

One of the most interesting findings of our study was that the three racial/ethnic groups had different patterns of factors associated with buccal-cell collection participation. Currently, we are planning to conduct focus group sessions among specific minority groups to directly address reasons for hesitance in participation in the genetic component of the NBDPS. This information could help us redesign our study materials or approach to directly address the specific concerns of these populations.

Even among women who were willing to participate in an extensive interview, there was a relatively low rate of participation in the genetic component of the study. Unfortunately, lower participation in DNA collection among minority groups could limit the benefits of this new research, because the generalizability of the findings would be lessened. Strategies to increase participation in research involving DNA collection are needed so that the potential benefits of the findings can have an equal impact in all communities.

The authors thank the many staff and scientists from all study sites who contribute to the National Birth Defects Prevention Study.

Conflict of interest: none declared.

## References

1.
Bates BR, Lynch JA, Bevan JL, et al. Warranted concerns, warranted outlooks: a focus group study of public understandings of genetic research.
Soc Sci Med

2005
;
60
:
331
–44.
2.
Armstrong K, Micco E, Carney A, et al. Racial differences in the use of BRCA1/2 testing among women with a family history of breast or ovarian cancer.
JAMA

2005
;
293
:
1729
–36.
3.
Moorman PG, Skinner CS, Evans JP, et al. Racial differences in enrolment in a cancer genetics registry.
Cancer Epidemiol Biomarkers Prev

2004
;
13
:
1349
–54.
4.
Kozlowski LT, Vogler GP, Vandenbergh DJ, et al. Using a telephone survey to acquire genetic and behavioral data related to cigarette smoking in “made-anonymous” and “registry” samples.
Am J Epidemiol

2002
;
156
:
68
–77.
5.
Rasmussen SA, Lammer EJ, Shaw GM, et al. Integration of DNA sample collection into a multi-site birth defects case-control study.
Teratology

2002
;
66
:
177
–84.
6.
Yoon PW, Rasmussen SA, Lynberg MC, et al. The National Birth Defects Prevention Study.
Public Health Rep

2001
;
116
:
32
–40.
7.
Correa-Villasenor A, Cragan J, Kucik J, et al. The Metropolitan Atlanta Congenital Defects Program: 35 years of birth defects surveillance at the Centers for Disease Control and Prevention.
Birth Defects Res A Clin Mol Teratol

2003
;
67
:
617
–24.
8.
Bogner HR, Wittink MN, Merz JF, et al. Personal characteristics of older primary care patients who provide a buccal swab for apolipoprotein E testing and banking of genetic material: The Spectrum Study.
Community Genet

2004
;
7
:
202
–10.
9.
Audrain J, Tercyak KP, Goldman P, et al. Recruiting adolescents into genetic studies of smoking behavior.
Cancer Epidemiol Biomarkers Prev

2002
;
11
:
249
–52.
10.
Bamshad M. Genetic influences on health: does race matter?
JAMA

2005
;
294
:
937
–46.
11.
Thompson HS, Valdimarsdottir HB, Jandorf L, et al. Perceived disadvantages and concerns about abuses of genetic testing for cancer risk: differences across African American, Latina and Caucasian women.
Patient Educ Couns

2003
;
51
:
217
–27.
12.
Schulz A, Caldwell C, Foster S. “What are they going to do with the information?” Latino/Latina and African-American perspectives on the Human Genome Project.
Health Educ Behav

2003
;
30
:
151
–69.
13.
Hispanics: a people in motion. In: Doherty C, ed. Trends
2005
. Washington, DC: Pew Research Center, 2005:71–89. (http://pewresearch.org/assets/files/trends2005-hispanic.pdf).
14.
Cozier YC, Palmer JR, Rosenberg L. Comparison of methods for collection of DNA samples by mail in the Black Women's Health Study.
Ann Epidemiol

2004
;
14
:
117
–22.
15.
Le Marchand L, Lum-Jones A, Saltzman B, et al. Feasibility of collecting buccal cell DNA by mail in a cohort study.
Cancer Epidemiol Biomarkers Prev

2001
;
10
:
701
–3.
16.
Mulot C, Stucker I, Clavel J, et al. Collection of human genomic DNA from buccal cells for genetics studies: comparison between cytobrush, mouthwash, and treated card.
J Biomed Biotechnol

2005
;
2005
:
291
–6.
17.
Witso E, Stene LC, Paltiel L, et al. DNA extraction and HLA genotyping using mailed mouth brushes from children.
Pediatr Diabetes

2002
;
3
:
89
–94.