Familiarity, Navigation, and Comprehension: Key Dimensions of Health Literacy in Pap Test Use Among Korean American Women

Abstract

Health literacy consists of multiple dimensions such as print and oral literacy or numeracy. Different dimensions of health literacy may have more salient impact on certain health behaviors and outcomes. Yet, evidence is limited regarding which dimensions particularly affect cervical cancer screening. The objective of this study was to examine the role of different dimensions of health literacy in cervical cancer screening among Korean American women. We used baseline data obtained from 560 Korean American women in a community-based health literacy-focused intervention study. Backward stepwise logistic regression analysis revealed that familiarity (adjusted odds ratio [AOR] = 1.20, 95% confidence interval [CI] = 1.11–1.31) and navigational health literacy (AOR = 1.10, 95% CI = 1.04–1.16) were associated with lifetime Pap test use and comprehension (AOR = 1.08, 95% CI = 1.02–1.14) with triennial Pap test screening. Prior exposure to healthcare settings and knowing how to navigate the healthcare system were more important than other health literacy dimensions for lifetime Pap test use. Understanding cancer screening-related words was most relevant to triennial Pap test use. In addition to addressing system factors such as insurance and physicians’ recommendations, interventions to increase Pap test screening targeted at Korean American women are needed to address certain dimensions of health literacy such as familiarity, navigation, and comprehension.

Cervical cancer is the fourth most common cancer in women worldwide, and the second leading cause of cancer death in women aged 20 to 39 years (Siegel, Miller, & Jemal, 2018). Despite national statistics indicating the high incidence and late diagnosis of cervical cancer (Innos & Horn-Ross, 2003; Lin et al., 2002; Liu, Zhang, Deapen, Bernstein, & Ross, 2003), Korean American (KA) women in the United States (U.S.) continue to report lower cervical cancer screening rates than those of other ethnic groups (American Cancer Society, 2018; Juon, Choi, & Kim, 2000; Maxwell, Bastani, & Warda, 2000; Moskowitz, Kazinets, & Tager, 2003; Ponce, Gatchell, & Brown, 2003; Swan, Breen, Coates, Rimer, & Lee, 2003). Screening for cervical cancer is recommended for women aged 21 to 65 years with a Pap test every three years (U.S. Preventive Services Task Force, 2012). Health literacy—”the degree to which individuals have the capacity to obtain, process, and understand basic health information and services to make appropriate health decisions”—is the key to acquiring access to safe, quality healthcare (U.S. Department of Health and Human Services, 2000). Low health literacy has been identified as a major barrier to effective utilization of cancer screening tests, independent of race or socioeconomic status (Bennett et al., 1998; Lindau et al., 2002). In addition, studies have shown positive correlations between low health literacy and disproportionately higher cancer incidence and mortality (Garbers & Chiasson, 2004; Lindau, Tomori, McCarville, & Bennett, 2001; Lindau et al., 2002).

Several studies have shown significant health literacy challenges in certain groups such as recent Asian immigrant in the U.S. who often lack English proficiency (Berger, Huang, & Rubin, 2017; Jun & Nan, 2018; Sentell & Halpin, 2006). KAs, one of the largest Asian populations in the U.S., are predominantly foreign born (92.5%) and monolingual (70% do not speak English well) (Shin, 2003). Patients experiencing language discordance are more likely to have misconceptions and lack appropriate understanding of information provided; these differences negatively affect knowledge, attitude, and self-efficacy with regard to adequate utilization of preventive care (Davis et al., 1996; Lindau et al., 2002; Scott, Gazmararian, Williams, & Baker, 2002; Sentell, Braun, Davis, & Davis, 2015; Sentell & Halpin, 2006). Even highly educated individuals with high English proficiency skills find it difficult to correctly convert and interpret proportions commonly used to report cancer risks, resulting in increased misperceptions and lowered personalization of such risks (Lipkus, Samsa, & Rimer, 2001). This problem is heightened among individuals in the U.S. with low English proficiency who are also predisposed to low health literacy (Fiscella, Franks, Doescher, & Saver, 2002; Jacobson, Hund, & Soto Mas, 2016; Sentell et al., 2015; Sentell & Halpin, 2006).

Health literacy is a multidimensional concept and is often measured as such (Baker, 2006; Sorensen et al., 2012). The common dimensions of health literacy include print literacy (reading and understanding text), oral or communicative literacy (speaking and listening effectively), and numeracy (ability to use quantitative information) (Baker, 2006; Speros, 2005). Additional dimensions include critical literacy (critically analyzing information to make health decisions), interactive literacy (actively deriving meaning from different forms of communication and applying this to changing circumstance), cultural literacy (recognizing and using collective beliefs, customs, and social identity to interpret and act on health information), civic literacy (being aware of public issues and becoming involved in the decision making process), or digital literacy (using information and communication technologies to locate, evaluate, create, and communicate digital information) (American Library Association, 2013; Nutbeam, 2000; Zarcadoolas, Pleasant, & Greer, 2005). Recent research shows that different dimensions of health literacy may have more salient impact on certain health behaviors and outcomes. For example, numeracy has been shown to be a key dimension of health literacy for adequate self-management of chronic diseases such as diabetes (Bailey et al., 2014; Osborn, Cavanaugh, Wallston, & Rothman, 2010) and asthma (Apter et al., 2013). Inadequate numeracy was also associated with higher prevalence of comorbidities and prescribed medications in individuals receiving outpatient care at a Veterans Affairs health system (Garcia-Retamero, Andrade, Sharit, & Ruiz, 2015), and more emergency room visits and hospitalizations in a sample of persons with multiple sclerosis (Marrie, Salter, Tyry, Fox, & Cutter, 2014). Oral literacy has been closely associated with dental health outcomes independent of age, gender and socioeconomic status (Batista, Lawrence, & Sousa, 2017; Divaris, Lee, Baker, & Vann, 2011; Lee et al., 2011; Lee, Divaris, Baker, Rozier, & Vann, 2012). No research has investigated how health literacy dimensions are associated with cancer screening behaviors. To address this research gap, the purpose of this study was to examine dimensions of health literacy related to Pap test screening behaviors among Korean immigrant women in the U.S.

Method

Design

This cross-sectional study used baseline data obtained from a community-based health literacy-focused intervention trial to promote breast and cervical cancer screening among 560 KA women in the U.S. (parent study). Detailed information about the parent study can be found in Han et al. (2017). Briefly, KA women 21–65 years of age who were able to read and write English or Korean, had not received age-appropriate mammograms and/or Pap tests for the past two years, and were willing to provide a written consent form to allow the study team to audit medical records for relevant cancer screenings were eligible for the parent study. These women were recruited from Korean ethnic churches on the east coast (Baltimore-Washington metropolitan area) of the U.S. This target region is home to the third-largest Korean immigrants in the U.S. (Zong & Batalova, 2014). KA women were recruited by trained community health workers from the respective churches in the target area. Data were collected at baseline, 3-month and 6-month follow-ups. We used baseline data from 560 enrolled participants for this analysis.

Procedures

The Institutional Review Board approved study protocols. Trained bilingual research staff obtained written informed consent following a brief explanation about the study and the verification of participants’ eligibility. The staff then administered the study survey which included the Assessment of Health Literacy in Cancer Screening (AHL-C) (Han, Huh, Kim, Kim, & Nguyen, 2014). The AHL-C is based on Baker’s conceptualization of health literacy (Baker, 2006) and includes the following domains: print literacy, numeracy, comprehension, and familiarity. Detailed procedures to evaluate a study participant’s health literacy in cancer screening using the AHL-C were reported in Han et al. (2014). In short, the AHL-C is a paper-pencil questionnaire; although all instructions and items were written in Korean, all health literacy-related items were written in English. This is because a previous validation study using the Korean version of the two most popular health literacy instruments—Rapid Estimate Adult Literacy in Medicine (REALM) and the Test of Functional Health Literacy in Adults (TOFHLA)—failed to effectively capture KA women’s health literacy (Han, Kim, Kim, & Kim, 2010). Han et al. (2010) highlighted that Korean is phonetic language, which violates the presupposition of popular literacy instruments such as REALM (i.e., a high association between decoding skills and comprehension). Trained staff used a list of cancer-specific words to evaluate their reading skills and used a nutrition label to assess their arithmetic skills. Each woman received $20 as a token of our appreciation.

Measures

The following sociodemographic characteristics were associated with this study: age, marital status, education, income comfortability, length of stay in the U.S., and English proficiency. Income comfortability was measured on a 5-point Likert scale (1 = very comfortable to 5 = very uncomfortable). Health care system factors such as insurance status and a provider’s recommendation about Pap test use were measured. A provider’s recommendation was assessed using one item, “Has your primary care provider recommended Pap test use within the last 2 years?” These variables were selected because they have been shown to be associated with Pap testing. The outcomes were KA women’s lifetime and triennial Pap test use. Triennial Pap test use was assessed by subtracting a participant’s most recent Pap test use from the date of the baseline data obtained.

The Assessment of Health Literacy in Cancer Screening (AHL-C) was developed to measure health literacy in the context of breast and cervical cancer screening. While other health literacy instruments are available in the context of breast and cervical cancer (e.g., Breast Cancer Literacy Assessment Tool [B-CLAT] and Cervical Cancer Literacy Assessment Tool [C-CLAT]), the items included in the B-CLAT (Williams, Templin, & Hines, 2013) and C-CLAT (Williams & Templin, 2013) primarily focus on knowledge (e.g., different exams available to detect breast cancer). Modeling on the structure of the REALM and TOFHLA, the AHL-C consists of a word recognition test, a numeracy test, and functional literacy items. This 52-item cancer screening-specific health literacy tool (Han et al., 2014) was developed based on Baker’s conceptualization of health literacy (Baker, 2006) which encompasses prior knowledge (“an individual’s knowledge before reading health-related materials or speaking to a healthcare professional”) and reading fluency (“the ability to mentally process written materials and form new knowledge”). The AHL-C represents the following subscales (Han et al., 2014): prior knowledge includes familiarity (12 items) and comprehension (12 items); reading fluency consists of word recognition (12 items), navigational health literacy (12 items), and numeracy (4 items). The AHL-C validation study in KA women reported acceptable to excellent internal consistency reliability, ranging from 0.70 to 0.92 and demonstrated construct validity by correlating with known covariates including age and education (r = 0.11 to 0.62) (Han et al., 2014). The concurrent validity with the REALM was also strong (r = 0.87) (Han et al., 2014). Total scores ranged from 0 to 52 as responses were coded as correct (1) or incorrect (0).

Statistical analysis

Descriptive statistics were used to summarize means and standard deviations (SDs), and tabulate frequencies and percentages. Following a bivariate analysis, backward stepwise logistic regression model was estimated. Marital status was categorized as married/partnered and separated/divorced/widowed/never married. Education was categorized as ≤ high school education and some college education or higher. Income comfortability was categorized in two groups: very comfortable/comfortable and neutral/uncomfortable/very uncomfortable. English proficiency was categorized in two groups: not fluent and fluent. Health insurance status was categorized as uninsured and insured (private/Medicare/Medicaid/other). The variables were eliminated from a full model if p-value was greater than .10, and the variables were added to the model if p-value was less than .05.

Results

Sample characteristics

Sample characteristics are shown in Table 1. All participants were foreign-born Korean immigrants. The sample was middle-aged (mean age±SD=46.1±8.5 years) and mostly married (85.7%), received at least some college education (64.8%, mean education±SD = 14.5±2.7 years), and had stayed in the U.S. for 10 years or more (mean length-of-stay in the U.S.±SD = 15.4±9.7 years). Only 23.4% reported speaking English fluently. About 38% had health insurance and only 15.7% reported that a physician had recommended Pap testing to them in the past two years. The average score on the AHL-C scale was 19.9 (SD = 12.5). About 49% had received a Pap test in their lifetime and 25.4% reported receiving triennial Pap test screening.

Table 1.

Sample characteristics (N = 560)

Characteristics	n (%)	Mean±SD
Age (range = 21–65), years  <40  40–49  50–59  60+	137 (24.5) 205 (36.6) 185 (33.0) 33 (5.9)	46.0 ± 8.6
Married/Partnered	480 (85.7)
Years of education (range = 4–24)  Some college+	363 (64.8)	14.5 ± 2.7
Income comfortability  Very uncomfortable/neutral  Comfortable/very comfortable	412 (73.6) 148 (26.4)	2.9 ± 1.0
Years in the US (range = 0.1–62.3)  <25% of lifetime  ≥25% of lifetime	230 (41.1) 330 (58.9)	15.4 ± 9.7
English proficiency (range=1–4)  Not at all/Poor/Fair  Fluent	429 (76.6) 131 (23.4)	2.7 ± 0.9
Having health insurance	212 (37.9)
Receiving physician recommendation	88 (15.7)
Health literacy (range=0–52)  Familiarity (range=0–12)  Comprehension (range=0–12)  Word recognition (range=0–12)  Navigational health literacy (range=0–12)  Numeracy (range=0–4)		19.9 ± 12.5 2.1 ± 2.9 3.2 ± 3.7 5.4 ± 3.8 7.1 ± 4.0 2.2 ± 1.4
Lifetime Pap test use	273 (48.8)
Triennial Pap test use	142 (25.4)

Stepwise multiple logistic regression of KA women’s Pap test use

In a bivariate analysis, lifetime and triennial Pap test use were associated with all dimensions of health literacy (p<.001) except numeracy. Backward stepwise logistic regression revealed: Lifetime Pap test use was associated with age (adjusted odds ratio [AOR]: 1.04, 95%CI: 1.02–1.07, p = .001), length of residence in the US (AOR: 0.62, 95% CI: 0.41–0.92, p = .018), provider’s recommendation (AOR: 15.48, 95% CI: 6.47 to 37.02, p<.001), familiarity (AOR: 1.20, 95% CI: 1.11 to 1.31, p<.001), and navigational health literacy (AOR: 1.10, 95% CI: 1.04 to 1.16, p = .001). Triennial Pap test screening was associated with income comfortability (AOR: 1.70, 95% CI: 1.05–2.74, p = .029), a provider’s recommendation (AOR: 13.62, 95% CI: 7.89 to 23.50, p<.001), and comprehension (AOR: 1.08, 95% CI: 1.02 to 1.14, p = .008).

Discussion

In general, the KA women in the study sample had low health literacy in cancer screening. Specifically, familiarity and comprehension subscales were considered particularly challenging with 54% and 36% of the study sample scoring zero on the subscales, respectively. Limited English proficiency was highly relevant to the high prevalence of low health literacy observed among recent immigrants in the U.S. (Berger et al., 2017; Jun & Nan, 2018; Sentell & Halpin, 2006). Indeed, despite a high level of education (about 65% having some college+ education) with an average length of stay in the U.S. being 16.5 years, only less than a quarter (23.4%) of the KA sample indicated their English as being fluent. Limited English proficiency noted in the study sample was consistent with national statistics where more than two thirds of the KA population are noted as being monolingual (Shin, 2003). Prior research revealed that even highly educated individuals without any language barrier are challenged with correctly interpreting numbers and understanding cancer risks and that the challenge becomes even greater for those with limited English proficiency (Fiscella et al., 2002; Jacobson et al., 2016; Lipkus et al., 2001; Sentell et al., 2015; Sentell & Halpin, 2006). The findings suggest that KA women are at high risk for low cancer literacy—the degree to which individuals have the capacity to obtain, process, and understand health information and services needed to make health decisions across the cancer spectrum including prevention (Simmons et al., 2017)—which may lead to cervical cancer disparities borne by this predominantly immigrant group (Innos & Horn-Ross, 2003; Lin et al., 2002; Liu et al., 2003).

Less than half (48.8%) of the KA women reported a lifetime Pap test and 25.4% had a triennial Pap test. The triennial screening rate is far below the national recommendation of 93% (Office of Disease Prevention and Health Promotion, 2014). A recent analysis of pooled data from the California Health Interview Survey (Chawla, Breen, Liu, Lee, & Kagawa-Singer, 2015) found that Asian American women consistently had the lowest rates of cervical cancer screening, with 81.2% reporting a Pap test in 2007 and their Pap test rates did not significantly change over time (77.9% in 2001). In contrast, the overall receipt of mammography among Asian American women increased (75.6% in 2001 vs. 81.8% in 2009). Korean and Chinese women represented two Asian subgroups with the lowest rates of Pap test use (78% and 77.5%, respectively, in 2007; Chawla et al., 2015). The results highlight the importance of targeting Asian American women in the U.S. including KA women in delivering effective Pap smear promotion programs.

Stepwise logistic regression revealed that familiarity, navigational health literacy, and comprehension were particularly relevant dimensions of health literacy to Pap test screening behaviors in KA women. Specifically, the odds of receiving a lifetime Pap test increased by 20% for each 1-point increase in familiarity (AOR: 1.20, 95% CI: 1.11 to 1.31, p<.001) and 10% for each 1-point increase in navigational health literacy (AOR: 1.10, 95% CI: 1.04 to 1.16, p = .001) subscale scores, after controlling for covariates. Likewise, the odds of triennial Pap test increased by 8% with every 1-point increase in comprehension subscale scores (AOR: 1.08, 95% CI: 1.02 to 1.14, p = .008), after controlling for covariates. It is difficult to explain why different health literacy dimensions were associated with the same behavior—Pap test screening—but in different time frames (lifetime vs. triennial). One possible explanation may be that comprehension (i.e., understanding the meaning of essential medical terminologies used in Pap test screening) might have elicited an immediate benefit on the women’s uptake of a recent Pap test even when known covariates of cancer screening (e.g., physician recommendations, insurance) were controlled for. In contrast, one’s assessment of her ability to use relevant medical terminologies proficiently (familiarity) and being able to adequately apply relevant medical terminologies throughout the cancer screening navigational trajectory (navigational literacy) were important but insufficient precursors to sustaining cancer screening behaviors because they failed to yield an increase in recent Pap test use. Taken together, the findings suggest that future intervention should address these key dimensions of heath literacy in promoting Pap test screening among recent immigrant women such as KAs. For example, a recent randomized controlled trial (Han et al., 2017) in which a comprehensive health literacy-focused community health worker intervention was implemented resulted in larger effect sizes than prior community health worker intervention programs using knowledge education with or without navigation assistance (Kim et al., 2016).

Neither reading test (word recognition) nor numeracy were significantly associated with Pap test screening behaviors in the study sample. Prior research seems to support the important role of health literacy—in particular numeracy—in certain chronic care contexts such as diabetes (Bailey et al., 2014; Osborn et al., 2010) and asthma (Apter et al., 2013). A possible explanation for the contrasting finding might be that cancer screening is a relatively simple behavior for which numeracy—though still relevant in adequately interpreting cancer risks and making uptake decisions (Lipkus et al., 2001)—may play a limited role in comparison to other health literacy dimensions. In contrast, both diabetes and asthma require the individual’s multistep mathematical skills, such as counting carbohydrates, interpreting blood glucose monitoring, or applying accurate dosing of insulin or bronchodilator (Apter et al., 2013; Shaw, Armin, Torres, Orzech, & Vivian, 2012). Further research using multidimensional health literacy measures such as AHL-C is warranted to better elucidate how different health literacy dimensions relate to salient health behaviors such as cancer screening.

There are several study limitations to note. First, Pap test screening behaviors were self-reported at baseline and these reports may have underestimated or overestimated the true screening rates. We attempted to minimize this potential bias by assuring confidentiality and by using relevant questions (e.g., place where a Pap test was taken or other relevant contexts such as the person, if any, who accompanied, season when the test was taken, or transportation used) to elicit cues to help women remember their past screening behaviors. Second, the data were pulled from a sample in a randomized controlled trial targeting KA women in the U.S. whose latest Pap screening was at least two years old at the time of study enrollment; consequently, it is likely that the sample over-represented KA women who were overdue for a Pap test. Similarly, the sample included young and middle-aged KA women with less than 10% of the study sample being 60 years or older. In addition, just slightly more than one third (38%) of the sample had health insurance (as opposed to 74.5% at the U.S. national level) (Huang, 2013), thereby limiting the generalizability of the study findings. Last, the analysis data were cross-sectional, prohibiting us from making causal inferences.

Conclusions

Self-assessment of one’s ability to use relevant medical terminologies (familiarity) and one’s ability to apply relevant medical terminologies throughout the cancer screening navigational trajectory (navigational literacy) were associated with an increased likelihood of KA women’s lifetime Pap test screening. On the other hand, understanding of cancer screening-related words was most relevant to triennial Pap test use. Future interventions to increase Pap test screening targeted at KA women should address key dimensions of health literacy such as familiarity, navigation, and comprehension.

Table 2.

Stepwise multiple regression analysis to predict Pap test use^┼

	Lifetime Pap test use	Triennial Pap test use
Characteristics	Odds ratio (95% CI)	Odds ratio (95% CI)
Age	1.04 (1.02–1.07)**	--
Marital status	--	--
Education	--	--
English proficiency	--	--
Income comfortability	--	1.70 (1.05–2.74)*
Residence in the US	0.62 (0.41–0.92)*	--
Health insurance	--	--
Provider’s recommendation	15.48 (6.47–37.02)***	13.62 (7.89–23.50)***
Familiarity	1.20 (1.11–1.31)***	--
Comprehension	--	1.08 (1.02 to 1.14)**
Word recognition	--	--
Navigational health literacy	1.10 (1.04–1.16)**	--
Numeracy	--	--
Pseudo R2 (%)	18.58	19.79

^┼Variables were added if p-value was less than 0.05. Variables were eliminated from the full model if p-value was greater than 0.1. Reference groups were as follows: unmarried (separated/widowed/divorced/never married), low education (high school graduate or less), poor English proficiency (not at all/poor/fair); low income comfortability (very uncomfortable/neutral), shorter residence in the US (<25% of lifetime), uninsured, and no provider recommendation.

^*p<.05,

^**p<.01,

^***p<.001.