We computed frequencies and population-weighted percentages (reported with 95% confidence interval [CI]) for each of the CRC screening outcome variables and independent characteristics. The association of potential predictors with CRC screening was presented as odds of non-adherence and assessed by odds ratio (OR) and 95% CI. Independent variables significantly associated (P < 0.05) with screening non-adherence in univariable (unadjusted) models were included in the final multivariable logistic regression model. The independent variables included survey year, age (50–54, 55–59, 60–64, 65–69, 70+), gender, insurance (uninsured; Medi-Cal [California’s Medicaid program], age < 65; Medi-Cal, age 65+ years; Medicare; private), percent of life lived in the United States (0–40%, 41–60%, 61+%, unknown), language spoken at home (English; non-English, one language only; English and other language), BMI (underweight+normal weight [BMI<23 kg/m²], overweight [BMI 23–<27.5 kg/m²], obese [BMI≥27.5 kg/m²]), high blood pressure (yes, no, unknown), heart disease (yes, no), number of doctor visits in the past 12 months (0, 1–2, 3–5, 6+ times) and receipt of flu shot in past 12 months (yes, no, unknown).

To account for the complex sampling design of CHIS, all distribution parameters and regression statistics were computed using SAS survey procedures (SAS Institute, version 9.3, Cary, NC) that apply replicate weights to obtain jackknife variance estimates following the guidance of CHIS Methodology and Technical Reports.²⁰ The level of statistical significance was set at alpha 0.05 for all analyses. Tests for linear trend for ordinal variables were conducted by treating the ordinal categories as continuous values in one single variable. Institutional Review Board (IRB) approval and informed consent were not required as this study was based on an analysis of a public use, de-identified dataset.

There were 459 South Asian Americans aged 50 years or older in the analysis dataset, after pooling CHIS data across the five available survey waves. CHIS did not report more granular data for country of origin for South Asian Americans sampled. More than half the sample was under age 60, nearly two-thirds were male, and the vast majority (87%) were married (Table 1). About half lived in an urban zip code, and another 45% in a small city or suburban zip code. The sample was highly educated, with nearly 80% having a college degree or higher. About 20% were uninsured or had Medi-Cal health insurance. Nearly all (98.8%) were foreign born, and most were recent immigrants, with about 40% having spent ≤40% of their life in the United States. About 11% spoke only a non-English language at home, and 73% spoke another language at home in addition to English. For English proficiency, 5.7% reported having limited English proficiency, that is, spoke English not well or not at all. In terms of health status, 18.5% reported fair or poor health, almost 70% were overweight or obese based on the Asian BMI cutpoints, few had asthma, but nearly 20% had diabetes, nearly 40% had high blood pressure and nearly 12% had heart disease. There was variability within the sample in the healthcare access measures, with about half reporting they had two or fewer doctor visits over the past year, and 43% reported getting a flu shot in the past year.

Table 2 shows the percentages of South Asian Americans from 2001 to 2009 who had each of the specific CRC screening modalities and were considered up-to-date with the 2008 U.S. Preventive Services Task Force screening recommendations. Overall, 49% of the sample were considered non-adherent with CRC screening recommendations, ranging from 48.3% in 2001, 29.8% in 2003, 47.8% in 2005, 58.5% in 2007, and 60.7% in 2009. About 20% reported getting each of the specific screening modalities, except for sigmoidoscopy in the past 5 years, which was less than 7%.

Table 3 shows the unadjusted associations of each sociodemographic, health and healthcare access factor with non-adherence (i.e. not being up-to-date) per USPSTF screening recommendations. Non-adherence was high in earlier survey years (51.7% in 2001, 70.2% in 2003), among younger South Asian Americans (63.3% among 50–54), uninsured (74.8%), immigrants with <40% of life in the United States (60.4%), those who only spoke a language other than English at home (66.5%) and those who had not had a flu shot in the past year (62.2%). Non-adherence was lower among older South Asian Americans (32.7% among 70+), those with private or Medicare insurance (37.3 and 49.1%, respectively), those who have lived in the United States for more than 60% of their life (30.2%), those who speak English at home (37.2%), those who are obese (38.2%), those with health conditions (high blood pressure [38.5%], heart disease [23.1%]), those who received a flu shot in the past year (33.0%) and a non-significant difference for those who had more than two doctor visits in the past year. In the multivariable model (Table 4), the only statistically significant associations remained with percent life in the United States, language spoken at home, BMI and receipt of flu shot.

In contrast to national declines in CRC incidence and mortality rates, recent reports show a lack of decline or even increasing trends among South Asian Americans,^1,4 who also happen to have among the lowest CRC screening rates. Regarding lack of adherence to CRC screening guidelines among South Asian Americans, we found that several determinants including being a recent immigrant, not speaking English as the primary language at home and not getting a flu shot in the past year were more likely to be associated with non-adherence to CRC screening recommendations. However, South Asian Americans who were obese based on Asian BMI guidelines,¹⁹ had a greater number of physician visits or had more chronic health conditions were more likely to be adherent to recommendations. These associations suggest the importance of more proactive healthcare utilization, especially preventive care, which may have provided more opportunities for physicians to recommend CRC screening.

We found less adherence to recommendations among recent immigrants and those who primarily spoke another language at home. This finding is consistent with the literature showing these factors to be associated with low cancer screening, specifically CRC screening among various Asian American groups.^5–10 More recent data from the National Health Interview Survey for all Asian Americans combined reported a CRC screening rate of 52.1%; data for South Asian Americans were not available. Supported by qualitative research,^9,21,22 associations with immigration factors likely relate to language barriers, lack of healthcare access and/or adherence to cultural beliefs that tend to not value cancer prevention or early detection. However, very little research to date has focused exclusively on population-level CRC screening among South Asian Americans. Data from 2008 to 2014 for South Asian Americans from the Behavioral Risk Factor Surveillance Study (BRFSS) in California documented a CRC screening rate of 40.0%, but there were only 47 South Asian Americans 50 years and older sampled across this entire time period. With respect to findings found in the larger Asian American population, we failed to identify a positive association between limited English proficiency and CRC screening, even in unadjusted models, indicating that language barriers are likely not an important driver of non-adherence to screening, at least in this California population. Although CHIS was not conducted in any South Asian languages (and therefore, data may be considered to be skewed toward more acculturated South Asian Americans), we believe that these results are likely reasonably representative of South Asians in the United States, as the vast majority (80%) of South Asian Americans are English proficient.¹⁷ Moreover, that recency of immigration and language used at home (as a proxy indicator of acculturation) remained statistically significant in our multivariable model after adjusting for health insurance, socioeconomic status and healthcare utilization suggests that cultural and health beliefs may be more relevant than language barriers, financial resources and healthcare access in this population. Qualitative research specifically among South Asian Americans, with particular attention to the diverse ethnic groups captured under that broader umbrella term, is especially needed to help identify the specific barriers to CRC screening so that clinical and community-based interventions can be targeted to these unique populations.

The findings of better screening adherence among South Asian Americans who are obese and among those who reported having had a flu shot in the previous year likely reflect more intense healthcare utilization and greater endorsement of health preventive measures. With respect to a lower likelihood of non-adherence among obese respondents, this finding contrasts with some studies²³ but is consistent with others.²⁴ In addition to increased monitoring by physicians due to cardiometabolic risks, a recent study has shown that for those who are obese and experiencing the subsequent health consequences may influence them to actively seek health education and engage in health promotion activities from other sources of information, which may include preventive screenings.²⁵ Also seen in a study of urban African Americans,²⁶ obese patients may have more comorbidities or be otherwise monitored more closely by their providers. Indeed, in unadjusted models, having high blood pressure and heart disease were significantly associated with adherence to screening recommendations; however, these associations became statistically non-significant in multivariable models while obesity remained significant. We explored whether the association with receiving a flu shot may be due to the implementation of programs across some California healthcare systems to administer FOBT among patients who come in for a flu shot. We examined this association among South Asian Americans who were members of a large integrated healthcare system that was known to have implemented such a program but did not find this association in this sub-sample. Thus, we surmise that the flu shot finding likely reflects general health preventive attitudes, which is underscored by the success of pairing preventive health efforts such as the FLU-FIT program²⁷; flu shots have been used as a proxy for preventive health in other studies.^28,29 We were not able to evaluate in these data the relevance of provider recommendations, but the obesity and flu shot associations indicate that, as with many Asian cultures, endorsement from medical authorities may be a critical leverage point in conveying to South Asian Americans the importance of CRC screening.¹⁴ In addition, it has been postulated that obtaining a flu shot may reflect a broader construct of health literacy, which may be consistent with timely uptake of CRC screening.³⁰ As more emphasis is placed on timely flu shots, coupled with the development and accessibility of a COVID-19 vaccine to address the 2020 pandemic, this may serve as an opportunity to couple CRC screening with such prioritized preventive health behaviors among South Asian Americans and other non-adherent population groups.

Several limitations in this study should be noted, including the relatively small sample despite pooling across years, the lack of certain screening determinants such as culturally embedded health beliefs, provider recommendations, health literacy, as well as the relative datedness and the long timespan of the sample, from 2001 to 2009, limited by the availability of data on CRC screening in CHIS, and limitations inherent to large population-based and self-reported surveys including low response rates, recall bias and social desirability bias. Another limitation is that CHIS does not differentiate between receipt of these CRC screening tests for routine screening or diagnostic purposes; in an evaluation of data from 2007 and 2009, the majority (71.6%) of South Asian Americans who had CRC screening reported the reason as being ‘part of a routine test’ versus ‘because of a problem’ or ‘other’ reason. Despite the limitations in these data, they are arguably the only population-based data powered in sample size for assessing determinants of CRC screening currently representation of this population group. Although 2009 was the last survey year in which CRC screening data were collected in CHIS, this is still to-date the largest, most representative data for assessing CRC screening among South Asian Americans. Moreover, it is unlikely that the determinants associated with screening found in this study are different in more current California populations.

We leveraged a population-based dataset, perhaps among the largest available for South Asian Americans, to examine determinants of CRC screening. Although some findings such as associations with immigration and language variables are similar to those for other Asian American ethnic groups, some other patterns such as prior receipt of flu shots, are unique. Thus, our findings provide support for design and implementation of interventions that are culturally and linguistically tailored to this rapidly growing South Asian American population. These results also have implications for clinical practice in that provider-driven preventive measures, such as coupling FIT with flu (and future COVID-19) vaccinations could be an effective means to encourage CRC screening in the South Asian American population.