Associations Between Racial and Ethnic Groups and Foot Self-Inspection in People With Diabetes

More than 30 million Americans (∼1 in 10) have diabetes and are consequently at increased risk of lower-limb amputation (LLA), a preventable complication of diabetes that has substantial negative impacts on functioning and quality of life (1,2). The most common pathway to LLA in people with diabetes involves minor trauma, cutaneous ulceration, and wound-healing failure (3). Factors that contribute to wound-healing failure include neuropathy, ischemia, and infection (4). The severity of diabetic foot complications and likelihood of LLA can be reduced by timely self-recognition of the early signs of diabetic foot complications and self-referral to diabetes specialists (4). Thus, current guidelines recommend patient education and daily foot monitoring (5).

There is substantial geographic and racial/ethnic variation in LLAs, indicating that socioeconomic and health care–related factors may contribute to the variation (6–9). Specifically, among people with diabetes, African Americans (AA) have LLA rates that are 1.5–4.0 times greater than rates in whites (6–8). Possible reasons for this disparity proposed include biological factors (e.g., greater prevalence of undiagnosed diabetes, more advanced diabetic neuropathy, and peripheral artery disease) and sociocultural factors (such as less aggressive treatment once symptoms develop, which may be a function of poorer access to high-quality health care, greater reliance on self-care methods, and distrust of medical treatment) (7,10). LLA rates are also elevated among American Indian/Alaska Natives (AI/AN) and Hispanics and lower among Asians compared with whites (6–8), though relatively few studies have examined rates in these subgroups, and at least one study (11) observed no association.

Experts estimate that more than half of diabetes-related foot complications could be avoided by proper preventive measures, patient education, and foot care (12). Important components of foot care include inspecting feet daily, compliance with other foot care recommendations (e.g., keeping feet clean and dry, protecting feet from temperature extremes, and wearing appropriate footwear), and rapid self-referral when problems with the foot appear. Daily feet checking may permit earlier detection and treatment of a foot lesion and is associated with reduced risk of complications and amputation (13,14).

There is a dearth of studies on factors associated with foot self-inspection. Generally, studies found that those who are at high risk (e.g., based on history of conditions such as diabetic foot ulcers) (15–18), take insulin (16,19), and have had diabetes for longer (e.g., >20 vs. <10 years) (16,19,20) were more likely to check their feet daily than those classified as low risk, though studies are not entirely consistent (21). Receipt of diabetes self-management education was associated with higher rates of foot self-inspection in several studies including two conducted in Europe (16,17) but not in one conducted in the U.S. in rural North Carolina (22). Several studies (19,23,24) have examined foot self-inspection by racial and ethnic groups, including a review article published in 2016 (25). In this review article, Mayberry et al. (25) identified eight studies published between 2011 and 2016 that examined disparities in foot self-exams. Findings were mixed; half of studies reported that non-Hispanic AA were more likely to perform regular foot checks than non-Hispanic whites, and half of studies did not observe significant differences, indicating that more research is needed.

The objective of this study was to include a population-based sample that was large enough to assess differences in foot self-inspection by racial and ethnic groups, including those that are understudied, such as AI/AN, Asians, and Native Hawaiian/Pacific Islanders (NH/PI). This information can help us to understand possible reasons for disparities in LLA risk and aid in identifying priorities for future interventions that tailor education, monitoring, and treatment for different subgroups.

We used data from the Centers for Disease Control and Prevention 2015–2017 Behavioral Risk Factor Surveillance System (BRFSS) surveys. BRFSS completes >400,000 interviews each year in noninstitutionalized adults aged ≥18 years from all 50 states and territories (26). Surveys are conducted in English and Spanish. The selection and weighting methodology result in a population that is representative of the population of not only each state but also the entire U.S. A complete description of survey methodology is available online (26). Interviews include a core component, optional modules (including one for diabetes), and state-added questions. The diabetes optional module was completed by 36 states/territories in 2015, 18 states/territories in 2016, and 35 states/territories in 2017. Some states (four in 2015, four in 2016, and six in 2017) only surveyed a subset of respondents (Fig. 1). The University of Washington Institutional Review Board has determined that use of BRFSS data for research does not involve “human subjects” (as defined by federal regulations and guidance) and therefore requires neither institutional review board review nor an exempt determination.

We included individuals who lived in a state/territory that included the diabetes module; reported being told by a doctor, nurse, or other health care professional that they had diabetes; and answered the questions on race and ethnicity. We excluded individuals who reported being told that they had diabetes during pregnancy only, prediabetes, or borderline diabetes or refused to answer the questions on checking their feet or on race.

The exposure of interest for this study was self-reported race and ethnicity (27). We conceptualize race as a social/cultural construct that is used by members of society to explain perceived biological differences (27,28). Categories were determined based on responses to two questions—one on race and a second on Hispanic ethnicity. The first question asked, “Which one of these groups would you say best represents your race? White, black or African American, American Indian or Alaska Native, Asian [Asian Indian, Chinese, Filipino, Japanese, Korean, Vietnamese, other Asian], Pacific Islander [Native Hawaiian, Guamanian or Chamorro, Samoan, other Pacific Islander].” Though the BRFSS collects information on Asian ethnicity (showed above in square brackets), this detailed information is not available in the national public release data. The question on Hispanic ethnicity asks, “Are you Hispanic, Latino/a, or Spanish origin?” If the answer is yes, interviewers are trained to ask respondents to specify one or more categories: Mexican, Mexican American, Chicano/a; Puerto Rican; Cuban; or another Hispanic, Latino/a, or Spanish origin. More detailed information on Hispanic ethnicity is also not released. We used a BRFSS-constructed variable that was created from the race and Hispanic ethnicity questions with eight categories: white, AA, AI/AN, Asian, NH/PI, other, multiracial, and Hispanic. Note that if someone reported being Hispanic, they were classified as Hispanic regardless of their self-reported race; thus, all categories other than Hispanic excluded those who self-identified being Hispanic (e.g., non-Hispanic white, non-Hispanic AA, etc.).

Daily foot inspection was the primary outcome of this study and was determined based on responses to the question, “About how often do you check your feet for any sores or irritations? Include times when checked by a family member or friend, but do NOT include times when checked by a health professional.” We created two categories: “less than once a day” and “once a day or more,” with the latter being the recommended frequency. Individuals who respond “never” or “don’t know/not sure” were categorized as “less than once a day” under the assumption that those who checked their feet daily would know the usual frequency. As a secondary analysis, we reclassified foot inspection as ever versus never checking feet. Those who refused to answer the question on foot inspection or reported they did not know or were not sure (n = 2,814) were dropped from this sensitivity analysis.

We considered age (18–39, 40–49, 50–59, 60–69, ≥70 years), sex (male, female), annual household income (<$15,000, $15,000–$24,999, $25,000–$34,999, $35,000–$49,999, ≥$50,000), and education level (less than a high school degree, high school degree, some college but no degree, and college graduate) as potential confounders. Physical/mental limitations (poor mental health, disabilities, and BMI as a measure of obesity) may impair one’s ability to check feet daily and differ by our exposure of interest. Poor mental health was operationalized as a diagnosis of depression (no/yes) or frequent poor mental health days. Diagnosis of depression was determined based on responses to the question, “Have you ever been told you have a depressive disorder (including depression, major depression, dysthymia, or minor depression)?” Frequent poor mental health days was determined based on response to the question, “Now thinking about your mental health, which includes stress, depression, and problems with emotions, for how many days during the past 30 days was your mental health not good?” Individuals who reported having ≥14 poor mental health days in the past 30 were classified as having frequent poor mental health. To determine the presence of a disability, we used information from the following questions: 1) “Are you blind or do you have serious difficulty seeing, even when wearing glasses?” 2) “Do you have serious difficulty walking or climbing stairs?” 3) “Do you have difficulty dressing or bathing?” Individuals who reported “yes” to one or more of the three questions were classified as having a disability. BMI was categorized into four categories (<25, 25–29.9, 30–34.9, and ≥35 kg/m²) based on self-reported weight and height.

The BRFSS diabetes module includes several questions related to diabetes, which we evaluated to describe the population and considered as potential modifiers of the association between racial and ethnic groups and foot self-inspection. We inferred diabetes type based on an approach used in other studies (29). Participants were considered to have type 2 diabetes if their age at diagnosis was at least 30 years or if their age at diagnosis was <30 years and they were currently not using insulin. Diabetes is a progressive disease, and complications (e.g., peripheral neuropathy) more frequently develop after decades. As endogenous insulin secretory capacity progressively declines with time, most people with type 2 diabetes will eventually require insulin therapy. Thus, as a proxy for diabetes severity for individuals with type 2 diabetes, we used information on diabetes duration and use of insulin. Diabetes duration was determined based on subtraction of age at interview from age at diagnosis and categorized as follows: <5, 5–9, 10–14, 15–19, and ≥20 years. BRFSS classifies those age ≥80 years as 80 years old (regardless of their actual age), which would underestimate the true time since diagnosis among those >80 years old. Individuals classified as at least 80 years of age with an age at diagnosis of diabetes >80 years were categorized as diagnosed <5 years ago. We created categories for visits to a doctor for diabetes (0, 1, 2, 3, and ≥4 times in the previous year) based on responses to the question, “About how many times in the past 12 months have you seen a doctor, nurse, or other health professional for your diabetes?” History of taking a diabetes self-management course was dichotomized (yes/no) from responses to the question, “Have you ever taken a course or class in how to manage your diabetes yourself?”

Among individuals with type 2 diabetes, we evaluated diabetes severity (described above), frequency of visit to a doctor for diabetes, and having taken a diabetes self-management course as potential effect modifiers. We hypothesized that group differences in foot self-inspection might be reduced when stratified by a proxy for diabetes severity in those who saw a doctor more frequently and who took a diabetes self-management course.

We first conducted descriptive analyses, assessing the frequency of daily foot inspection overall and within subgroups. In these and our analytic analyses, we accounted for the complex survey design of BRFSS as well as nonresponse by using survey commands and Centers for Disease Control and Prevention–calculated weights (30). Prevalence ratios (PRs) and associated 95% CIs were estimated using log-binomial linear regression models. We used an iterative approach to assess confounding by examining whether factors included in multivariable models changed the PR for daily foot inspection of AA, Hispanics, or AI/AN (the largest three groups other than multiracial) versus whites by ≥10%. We assessed the presence of effect modification through inclusion of interaction term(s) into multivariable models. Analyses were conducted using Stata, version 15 (College Station, TX).

We included 88,424 individuals with diabetes in our study: 62,175 whites, 12,793 AA, 2,120 AI/AN, 1,279 Asians, 396 Native Hawaiian/Pacific Islanders (NH/PI), 482 other, 1,671 multiracial, and 7,508 Hispanics (Table 1). There were numerous differences between groups in terms of age (e.g., 65% of NH/PI were 40–59 years old compared with 30% of whites) and sex (e.g., AA and NH/PI were more frequently female, while Asians were more frequently male) as well as other demographic characteristics. For example, a greater proportion of whites and Asians had greater educational attainment and higher incomes than other groups. A smaller percentage (20%) of Asians were obese (BMI ≥30 kg/m²) compared with the percentage of whites who were obese (55%) and of AI/AN and AA (58%). The prevalence of poor mental health, depression, disability, and use of insulin was particularly high among AI/AN. Compared with whites, the proportion of AA, AI/AN, and NH/PI who had not had their feet checked by a doctor or other health professional in the past year was lower, while the proportion of Asians and Hispanics who had not had their feet checked in the past year was higher.

Daily foot checking differed substantially across groups: Asians had the lowest prevalence of daily foot inspection (35%) and NH/PI had the highest (71%); “other” (47%), Hispanics (53%), multiracial (55%), whites (57%), AI/AN (66%), and AA (67%) fell between these two extremes (Table 2). There was no evidence of confounding by any factors; PRs changed minimally (<10%) after adjustment for each of our a priori potential confounders; thus, unadjusted results are presented. Compared with whites, AA (PR 1.18 [95% CI 1.14, 1.23]), AI/AN (PR 1.15 [95% CI 1.07, 1.25]), and NH/PI (PR 1.25 [95% CI 1.03, 1.52]) had a 15–25% higher prevalence of daily foot inspection. The prevalence of daily foot inspection was lower among Asians (PR 0.62 [95% CI 0.48, 0.81]) and Hispanics (PR 0.93 [95% CI 0.88, 0.99]) compared with whites. Associations did not vary importantly by insulin use and years since diagnosis (Table 3). Generally, results were also similar among those who did and did not have diabetes self-management education, except for Asians, for whom the PR compared with whites was attenuated among those who had taken a diabetes self-management class. Lastly, when we considered ever versus never checking feet, associations were attenuated but similar to those for daily versus less than daily (Supplementary Table 1).

Results from our study indicate that groups at highest risk of LLA—AA, AI/AN, and NH/PI—were more likely to check their feet on a daily basis compared with whites. Hispanics, whose risk of LLA is greater than whites (7,8), were less likely to check their feet. Asians, who have lower rates of LLA compared with whites (31), were significantly less likely to check their feet daily.

The reasons for this paradox of more foot self-inspection among groups with higher rates of LLA are not clear, but we present the following speculative explanations. It is possible that AA, for example, are less adherent to other components of diabetes self-management, e.g., diet, medication use, smoking, exercise, and other aspects of foot care, though evidence is not consistent and suggests reverse disparities for self-monitoring of blood glucose (with AA more likely to self-monitor than whites) and no consistent differences for diet, exercise, or smoking (25). A second possible explanation for the unexpected findings may be that race/ethnicity is a marker for disease severity, family history of LLA, and/or socioeconomic status. Prior studies have found that those at high risk of complications (as determined by factors such as foot deformities, history of ulceration, or peripheral neuropathy) were more likely to check their feet (15,18), though generally people underestimated their risk (15). Several other studies have found that foot self-examinations were performed more frequently by those with longer diabetes duration (16,20), treated with insulin (16,32), and with a history of foot complications (16,17). We did not have information on history of foot complications, though it is possible that a greater proportion of AA had a personal or family history of foot complications. It may be that checking feet is a particularly salient self-care behavior among AA because of their higher risk for diabetes-related LLA. Nevertheless, when we stratified analyses based on diabetes duration and use of insulin as a marker for diabetes severity, associations were not meaningfully different across groups.

Although one might hypothesize that lower socioeconomic status would be associated with poorer self-management, studies are not consistent. For example, lower socioeconomic status at the county level has been associated with more foot self-checking. Using 2008–2010 BRFSS data from Appalachia, investigators found that patients living in “at-risk counties” (defined as having economic indicators that classified them in the 10th–25th centiles) were 41% more likely to perform daily foot checks (odds ratio 1.41 [95% CI 1.11–1.79]) than patients living in “competitive” counties (defined as those in the 75th–90th centiles) (33). The authors hypothesized that residents in the less affluent counties in Appalachia may have been practicing self-care as a compensatory behavior due to lack of access to medical care. In the current study, results were similar after adjustment for income and education, indicating that differences in foot checking by socioeconomic status cannot explain the differences observed. Furthermore, having feet checked by a health care professional in the past year was somewhat higher among AA, AI/AN, and NH/PI compared with whites.

Barriers to foot care include physical limitations (e.g., vision problems and physical disabilities that impaired accessing feet), knowledge, education (18), and psychological factors such as emotional distress (34). A study by Hernandez et al. (32) found that across racial/ethnic subgroups, lower diabetes-related distress was correlated with a higher frequency of foot care activities among AA and Latinos. We considered disability, depression, and poor mental health days as potential confounders and found that none confounded the association between racial and ethnic groups and checking feet. Furthermore, results were similar when we stratified on diabetes self-management education.

The literature with respect to variation in self-foot examination by racial and ethnic groups is not consistent, and relatively few studies have examined this question. Mayberry et al. (25) summarized the literature published between 2011 and 2016 and found that four of eight studies reported evidence of significant racial/ethnic disparities in foot self-examination; non-Hispanic AA performed regular foot self-checks more frequently than both whites and Hispanics, and Hispanics were less likely to perform foot self-exams than whites. Notably, population diversity (e.g., single states vs. multiple states), study sample size (from 100–200 to >1,000), and outcome of interest (daily foot inspection vs. never checking feet) of prior studies varied substantially, which may account for the heterogeneity in results. Few studies have investigated prevalence of foot self-examination in Asian Americans. Our results are consistent with prior research indicating lower screening rates among Asian Americans. Specifically, a large cross-sectional study found that despite their higher risk of diabetes, Asian Americans were the racial/ethnic group least likely to be screened for the presence of diabetes (35). Based on results of the current study, it appears that Asian Americans are less likely to check their feet, to have their feet checked by a health care professional, and to receive diabetes self-management education. Notably, our results stratified on diabetes self-management education indicated that among those who obtained self-management training, racial/ethnic disparities were reduced.

Major strengths of this study include its large sample size, national breadth, and inclusion of rarely studied groups. An important potential limitation is that foot inspection was self-reported and thus subject to social desirability bias. However, for overreporting of foot inspection to induce bias in PRs, whites would need to substantially underreport foot inspection and/or AA, AI/AN, and NH/PI would have to substantially overreport foot inspection. We have no reason to suspect that such differential reporting exists. A second limitation relates to the substantial heterogeneity within groups, including, for example, Asians. Data from the 2017 U.S. census provide some insight into the breakdown of countries of origin for Asians living in the U.S. and indicate that 22% are Chinese, 20% are Asian Indian, 18% are Filipino, 9% are Vietnamese, 8% are Korean, 7% are Japanese, and 16% include other Asian ethnicities (e.g., Pakistani, Thai, Laotian, Hmong, etc.) (36). Furthermore, since race and ethnicity are social constructs, and many people belong to more than one group, the approach employed by BRFSS impairs our ability to better understand important social and cultural factors that may impact behaviors. Future studies should collect information to allow for better classification by ethnicity by including questions on birthplace, language, religion, and family origins as well as other cultural/behavioral factors that might contribute to differences.

In conclusion, in the largest study to our knowledge of foot self-examination to date among persons with diabetes, high-risk groups including AA, AI/AN, and NH/PI were more likely to practice this preventive behavior daily, indicating that efforts to increase foot self-examination in these populations may be effective. Nevertheless, increased self-checking does not appear to be effective in eliminating the elevated risk of LLA in these groups, as recent studies indicate that AA, AI/AN, and NH/PI remain at higher risk of diabetes-related LLA (7,8,37,38). Continued efforts are needed to develop strategies and interventions to reduce disparities in LLA.

Funding. This work was unfunded. Partial support for this research came from a Eunice Kennedy Shriver National Institute of Child Health and Human Development research infrastructure grant (P2C HD042828) to the Center for Studies in Demography and Ecology at the University of Washington.

Duality of Interest. E.J.B. served on an advisory board for Bayer AG. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. A.J.L. conceived and designed the study, conducted the literature search, performed statistical analyses, interpreted data, and drafted and revised the manuscript. C.J.K. contributed to the analysis design, interpretation of data, and drafting and revision of the manuscript. E.J.B. and S.E.H. contributed to the analysis design, interpretation of data, and revision of the manuscript. All authors read and approved the final manuscript. A.J.L. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.