The Wisconsin Department of Health Services (DHS) conducted this survey during January 1, 2022–January 31, 2023. Participants were recruited through community-based convenience and snowball sampling. Community advisory group members, schools, DNR listservs, and community organizations distributed recruitment materials to potentially eligible persons. We also asked participants who completed the survey to recruit additional participants within their social networks.

Eligible participants must have met the following criteria: (1) residing ≥1 year in the following Wisconsin counties: Milwaukee, Waukesha, Washington, or Ozaukee; (2) female; (3) self-identified as 1 of 4 major Asian ethnicities in the Milwaukee area: Chinese, Filipino, Hmong, or Karen; (4) aged 18 to 50 years; (5) had consumed ≥1 meal of fish caught from waterbodies in Wisconsin in the last 12 months by the participant or someone the participant knows; (6) the only member of their household to participate in the telephone survey; and (7) had not participated in a previous project about fish consumption with Wisconsin DHS.

Interested participants completed a screening survey to determine eligibility. Screening surveys written in English, Chinese, and Hmong were administered through REDCap instruments hosted by DHS. Based on community feedback, Filipino participants received information in English. As many persons in the Karen community do not have reliable access to email, screening forms written in Karen were administered by mail.

Trained interviewers administered the survey in each participant’s preferred language (English, Chinese, Hmong, or Karen) by telephone. Prior to each interview, we mailed or emailed visual aids to participants that included a map of the waterbodies in the Milwaukee AOC and photographs of each fish species or variety evaluated. Survey items assessed participant demographics, fish consumption habits, health beliefs about fish, and awareness of fish advisories. Participants received a $50 gift card for completing the survey.

We assessed consumption of 27 sportfish species and 13 store-purchased fish varieties. For each sportfish species, interviewers asked, “In the past 12 months, how many times did you eat [species] from Wisconsin waterbodies?” Then, if participants reported any consumption of the species in Wisconsin waterbodies, the question was repeated for the Milwaukee AOC. Participants also reported their consumption of several store-purchased fish species.

Participants reported whether they had made fish consumption changes to avoid contaminants, including eating fewer fish meals, eating different types of fish, avoiding certain parts of fish, or avoiding fish caught in certain locations. Using a 5-point Likert scale, participants reported how often they consumed parts of fish that might increase contaminant exposure (ie, skin, head, guts, organs or other innards, or belly fat). They reported which preparation methods they used to grill or roast fish, including an EPA-recommended option to reduce contaminant exposure. Additionally, the EPA recommends avoiding several fish preparation options, including pan frying, deep frying, boiling or poaching, braising, or using fish or fish parts to make broth, stock, curry, or soup. Finally, 3 options without an associated EPA recommendation, including smoking or drying fish, pickling fish, and using fish to make fish paste were included.²³

Participants who indicated a prior pregnancy described any changes to their fish consumption during pregnancy and breastfeeding. Finally, all participants reported their awareness of sportfish advisories for Wisconsin and the Milwaukee AOC and FDA or EPA limits for store-purchased fish. Those who were aware of any fish advisory answered items assessing attitudes towards the advisories.

For Wisconsin sportfish, we used limits provided by the Wisconsin DNR for the Milwaukee AOC, and for general Wisconsin Inland Waters.⁵ For store-purchased fish, we used fish consumption limits set by the FDA and EPA.¹ For fish consumption advisories’ limits for consuming fish of the same species but different sizes (eg, ≤ 6 meals per year for walleye >22 inches and ≤1 meal per month for walleye < 22 inches), we used the more restrictive limit. For species that did not have a specific limit for the Milwaukee AOC, we used the Wisconsin Inland Waters limit.⁵

We summarized participant characteristics and self-reported beliefs, behaviors regarding fish consumption during pregnancy and while breastfeeding, and knowledge of fish consumption advisories. We calculated means and standard deviations for continuous variables and frequencies and percentages for categorical variables. We dichotomized Likert scale items for analysis (eg, “extremely effective” or “very effective” vs “somewhat effective,” “a little effective,” and “not at all effective”).

For fish consumption in the preceding year, we calculated the median and range for each sport-caught and store-purchased species. For each individual species, we compared self-reported consumption to advisories to determine the number of participants reporting consumption above the advisory. For all species, limits for weekly or monthly consumption were multiplied by 52 or 12, respectively, to calculate limits for annual consumption.

We compared the proportion of Asian WCBA aware of Milwaukee, Wisconsin, and EPA and FDA advisories with consumption over advisory limits for any Milwaukee or store-purchased species, respectively. We conducted chi-square tests to determine whether awareness of advisories was associated with consumption over advisory limits in the preceding year. Analyses were completed using R version 4.1.3 (R Core Team, 2021). This activity was reviewed by the Centers for Disease Control and Prevention (CDC), deemed not research, and conducted consistent with applicable federal law and CDC policy (45 CFR part 46.102(l)(2), 21 CFR part 56; 42 USC Sect 241(d); 5 USC Sect 552a; 44 USC Sect 3501 et seq).

Overall, participants reported annually consuming a median of 11 (interquartile range [IQR] 5–23) Wisconsin sportfish meals, including 6 (IQR 2–18) from the Milwaukee AOC, and a median of 20 store-purchased fish meals (IQR 9–46). Among the 4 ethnic groups, Chinese participants reported the highest store-purchased fish consumption, and Filipino participants reported the highest sportfish consumption (Table 2). One-third of participants (n=50, 33%) reported consumption above advisory levels for ≥1 species from the Milwaukee AOC (Table 2). For store-purchased fish, 33 participants (22%) reported consumption above advisory levels for ≥1 store-purchased species (Table 2). However, average self-reported consumption for each individual fish species was low; only 1 sport-caught species (total Wisconsin white bass) and 2 store-purchased species (salmon and tilapia) had median consumption above zero (Supplemental Tables 1 and 2). Most participants who reported fish consumption exceeding advisory levels reported a limited number of meals for a fish species with zero limit (ie, species listed as, “do not eat” or “choices to avoid”). For example, 44 (29%) participants reported eating carp (listed “do not eat”) from the Milwaukee AOC (Supplemental Table 1). However, among this group, the median annual carp consumption was only 2 meals (Supplemental Table 1). For store-purchased fish, 23 participants ate king mackerel, 4 ate shark, 10 ate swordfish, and 3 ate tilefish – all of which were classified as fish “choices to avoid” (Supplemental Table 2).

Certain participants reported they had made behavior changes to avoid harmful contaminants from fish. Sixty (39%) participants reported eating fewer fish meals, 76 (50%) reported eating different species of fish, and 80 (52%) avoided eating certain parts of the fish. However, most participants reported keeping the skin (86%) or head (78%) of fish at least some or most of the time while preparing fish. Additionally, 146 (95%) participants reported sometimes using cooking methods that can trap fat (and consequently, contaminants) within the fish. This included pan-frying (n = 129, 84%), boiling or poaching (n = 98, 64%), or using the fish in a broth or soup (n = 90, 59%). Hmong participants (n = 27, 52%) reported eating fewer fish meals to reduce contaminant exposure, compared with approximately one-third of Chinese, Filipino, and Karen participants. Other behavior changes varied across ethnic groups. For example, 21 (57%) Chinese participants reported avoiding certain types of fish or fish caught at certain locations, whereas 27 (52%) Karen participants reported avoiding eating certain parts of the body (Table 3).

Among 113 participants who reported a previous pregnancy, 89 (79%) reported eating ≥1 fish meal during their most recent pregnancy. Of these 89 women who did consume fish, 62 (70%) reported that fish consumption remained the same or increased during pregnancy. Although 79 (89%) did not change the species of fish they consumed during pregnancy, 31 (35%) of 89 reported that they avoided certain fish species while pregnant.

Among 24 participants who did not eat fish during their most recent pregnancy, 19 (79%) reported that they avoided fish because of concerns that chemicals in fish were harmful (Table 4). Among 75 participants who reported breastfeeding after their last pregnancy, 57 (76%) consumed fish while breastfeeding. Of those, 49 (86%) reported fish consumption the same or higher while breastfeeding than before pregnancy, and most did not change the species they were eating. However, 39 (68%) reported avoiding certain species while breastfeeding (Table 5).

Only 29 (19%) respondents were aware of local fish consumption advisories, 39 (25%) were aware of Wisconsin sportfish advisories, and 34 (22%) were aware of FDA or EPA advisories. Karen participants had the lowest advisory awareness, with 2 (4%) aware of local advisories, 1 (2%) aware of state advisories, and zero aware of FDA or EPA advisories. Awareness was highest for Chinese and Filipino participants, of whom approximately 33% (16/49) were aware of local advisories, 47% (23/49) were aware of state advisories, and 33% (17/49) were aware of FDA or EPA advisories (Supplemental Figure). Among 29 respondents who were aware of local advisories, 10 (34%) reported that they knew “some,” “quite a bit,” or “a great deal” about the advisories. Of 39 respondents who were aware of Wisconsin sportfish advisories, 10 (26%) reported that they knew at least “some,” “quite a bit,” or “a great deal” about the advisories. Among 34 participants who were aware of FDA or EPA advisories, 11 (32%) reported that they knew “some,” “quite a bit,” or “a great deal” about the advisories.

Overall, sportfish consumption did not vary meaningfully by advisory awareness. Eight (28%) participants who were aware of Milwaukee advisories and 41 (34%) participants who were not aware of Milwaukee advisories consumed ≥1 meal above Milwaukee AOC advisories (P = 0.74). Few participants reported consumption above Wisconsin sportfish advisory limits, including 4 (10%) of 39 participants who were aware of Wisconsin advisories and 8 (7%) of 114 who were not aware of Wisconsin advisories (P = 0.48). For store-purchased fish, 14 (41%) of 35 participants who reported awareness of FDA or EPA fish advisories ate ≥1 “choices to avoid” fish meal, compared with 19 (16%) of 114 participants who were not aware of store-purchased fish advisories (P = 0.28).

This project has multiple limitations. First, we recruited participants by convenience and snowball sampling, and findings might not represent their overall WCBA communities. Second, food frequency questionnaires have been shown to underestimate overall consumption.^29,30 We attempted to increase the reliability and validity of our survey by providing pictures of fish species and a map of the Milwaukee AOC. However, misclassification of fish consumption data was likely. Third, participant recruitment fell short of recruitment goals for the Chinese and Filipino communities. Findings for these groups might not represent community views or behaviors. However, although the limited number of Filipino participants in this study reported higher average consumption than Hmong or Karen participants, community feedback suggests that their recruitment shortfall occurred because sportfishing is not common among Milwaukee’s Chinese and Filipino communities. Thus, the findings of this survey might not be as relevant to them as it is to Hmong and Karen communities. Despite its limitations, however, our findings provide disaggregated insight into the fish consumption behaviors of Milwaukee’s heterogeneous Asian communities. With multilingual surveys conducted over the phone and screening surveys conducted both on paper and online, we were able to conduct a representative evaluation.