Differences in personal care product use by race/ethnicity among women in California: implications for chemical exposures

Community-based participatory process

Survey planning, design, and implementation was a collaborative process among the three community-based partners and two scientific partners that form the Chemical and Personal Care: Asian, Black and Latina Exposure (CAPABLE) Study. The first community group, Healthy Heritage Movement (HHM), works primarily with Black women in the Inland Empire (San Bernardino County and Riverside County) and Los Angeles County of Southern California. The second, California Healthy Nail Salon Collaborative (CHNSC), works with Vietnamese nail salon workers throughout California, but focused on Los Angeles and Orange Counties for this study. The third group, Clinica de Salud del Valle de Salinas (CSVS), works with Latinos in the Salinas Valley, California. The scientific partners were the California Department of Public Health’s (CDPH) Safe Cosmetics Program and University of California, Berkeley’s Center for Environmental Research and Children’s Health (CERCH). Survey design was informed by focus groups in each community, which involved 7–9 community members and was led by a representative from the community organizations who identified as the same race/ethnicity as the participants. Cognitive interviews (two per community) were conducted to test the understanding and feasibility of the survey prior to administration.

Study population

Eligible participants were at least 18 years old, living in California, and spoke either English, Spanish, or Vietnamese. Survey recruitment and administration was conducted in person by staff from the community partner organizations. Each community partner recruited a convenience sample of 75–100 participants. Recruitment locations were decided by each community partner, depending on what was most appropriate in their community. HHM recruited primarily Black women from churches, health fairs, a hair show, and through collaboration with community partner organizations in San Bernardino County, Riverside County, and Los Angeles County, California. CHNSC recruited Vietnamese women from nail salons, churches, temples, and colleges in Los Angeles County and Orange County, California. CSVS recruited Latina women from health clinics, schools, and Special Supplemental Nutrition Program for Women, Infants, and Children offices in Salinas, California. White women were recruited by all community partners at these same locations, as well as in Contra Costa County at schools and medical offices. Informed consent was obtained from all participants, who received a $20 incentive coupon upon completion. Surveys were conducted on paper and self-administered. Human subjects approval was received from California Health and Human Services Agency’s Committee for the Protection of Human Subjects.

Surveys on usual personal care product use were completed by 321 Black, Latina, Vietnamese, and non-Hispanic White women between April and December 2019. We limited the study population to Black, Latina, Asian, and White women, including women who identified as more than one of these race/ethnicities (Mixed Race). Women who checked more than one race/ethnicity were categorized as Mixed Race. We excluded two women who were missing race/ethnicity information and one woman who identified as American Indian/Alaska Native for a final sample size of 318.

Measures

The 40-question survey asked about frequency of personal care product use, what brands participants favor, where they shop, whether they avoid certain ingredients, if they bring or buy products bought from other countries, and demographic information. The survey asked about 81 specific product types that were grouped into six categories of personal care products: makeup, hair, feminine hygiene, skincare, nail, and deodorant/fragrance. Each question asked about the frequency of use of a particular personal care product: every day, 5–6 days per week, 2–4 days per week, 1 day per week, 2–3 times per month, 1 time per month, a few times per year, or never. In addition to product use, the survey asked “Do you ever use any personal care products that you or someone else brings from another country?” and “Do you ever buy personal care products in the U.S. with labels not in English?” with an open-ended follow up question of which country they come from. Lastly, the survey asked: “Are there ingredients that you try to avoid in your personal care products?” and “How do you choose which personal care products to buy?”. Surveys were translated by CSVS and CHNSC staff into Spanish and Vietnamese, respectively.

Women were asked to self-identify their race/ethnicity (White, African American/Black, Latina/Hispanic, Vietnamese, Other Asian, American Indian/Alaska Native, or Other) and could check all that applied. Information on education, income, and other basic demographics was also collected. Most women answered all questions, but the sample size ranged from 305 to 318 for questions on personal care product use.

Statistical analysis

All survey data were analyzed using STATA version 15.1 [40]. We compared basic demographic characteristics among the different racial/ethnic groups using Chi-square tests. For ease of interpretation, product frequency of use was converted from a categorical variable to a continuous variable of number of times per week to facilitate the comparison of means across groups. We converted “every day”, “5–6 days per week”, “2–4 days per week”, “1 day per week”, “2–3 times per month”, “1 time per month”, “a few times per year”, and “never” to 7, 5.5, 3, 1, 0.5, 0.25, 0.1, and 0 times per week, respectively. For frequently used products (>0.5 times per week on average), we compared mean use per week across racial/ethnic groups as a continuous variable. To control for sociodemographic differences between communities, we conducted linear regression models controlling for age and education to obtain adjusted means for each racial/ethnic group and used the “margins” and “pwcompare” commands in STATA for post hoc pairwise comparisons of differences in means from these models. We used education as a marker of socioeconomic status because it was missing for fewer respondents (three missing for education versus eight missing for income). We did not control for country of birth because it was strongly collinear with identifying as Latina or Vietnamese. For products or services that were used less frequently (<0.5 times per week), such as feminine hygiene products, certain hair products, and nail products, we compared the percent of women who used these products or services at least once per month using logistic regression to obtain age- and education-adjusted percents and p values for pairwise comparisons. Personal care products or services that did not differ significantly by racial/ethnic group in the linear or logistic regression models are not presented in tables but are noted in table footnotes. For the additional questions on reasons for purchasing, product use from other countries, and ingredients avoided, we calculated the number and percent of respondents and tallied open-ended responses of specific ingredients and products. An alpha level of 0.05 was set for statistical significance.