Health literacy and patient engagement among Korean immigrants in the U.S.: exploring the emotional, cognitive, and behavioral dimensions

While patient engagement is central to the Western model of healthcare, behavior-focused models limit understanding of how health literacy shapes patient engagement in emotional and cognitive dimensions among immigrants. This mixed-method study aims to examine the relationship between health literacy and patient engagement across behavioral, emotional, and cognitive dimensions among Korean immigrants in the U.S. In an explanatory sequential mixed-method design, an online survey (N = 105) and semi-structured interviews (N = 20) were conducted to test the hypothesis and to explore the research question. In the quantitative phase, MANCOVA and multiple regression analysis results demonstrated that when English proficiency is controlled for, health literacy is associated with the behavioral and cognitive but not the emotional dimension of patient engagement. Qualitative data supported and complemented these findings, revealing that feelings such as frustrations were commonly found regardless of health literacy levels, whereas higher health literacy seemed to enable participants to enact their cognitive beliefs about patient engagement behaviorally. Patient engagement among Korean immigrants in the U.S. is a subjective experience encompassing behavioral, cognitive, and emotional dimensions. This study highlights the need for a comprehensive conceptual framework for patient engagement among immigrant populations.

Patient engagement is considered an essential pillar of healthcare delivery in the Western model of healthcare (Dentzer, 2013). Research shows that patients who are more actively engaged in their health care experience better health outcomes such as the quality of care and patient satisfaction, show higher compliance, achieve higher self-efficacy, and have lower health costs (Marzban et al., 2022). In this model, patients are assumed to be well-informed and motivated as well as able and willing to express their needs to providers, mostly through verbal expressions (Street and Millay, 2001).

However, not all patients are equally able and willing to be involved in their healthcare. For example, existing research found that older patients as well as African-American and Hispanic patients were more likely to prefer a less active role in medical decision-making compared to younger patients and non-Hispanic white patients (Levinson et al., 2005). In fact, non-Western ethnic minority patients were found to be less active in medical consultations (Schouten et al., 2007). Patients with a migration background showed lower preference as well as confidence in involvement in medical decision making or general communication with health providers (Hyatt et al., 2017).

Research elucidates potential factors that shape patients' preference and capability of patient engagement among ethnic minorities. Some researchers report how racism and perceived discrimination hamper patient engagement in healthcare delivery (Xiao and Wu, 2025). Cultural dimensions such as high vs. low-context, and individualism vs. collectivism, are discussed as potential factors that impact patient-provider interactions, indicating that cultural norms and beliefs among non-Western patients may conflict with their providers in Western societies (Ishikawa and Yamazaki, 2005). Structural barriers such as language discordance and discrimination, as well as provider intercultural competence were also reported to predict patient-centered communication among Hispanic/Latino patients whereas cultural philosophies such as fatalisimo, respeto, and familia shape their expectations for patient-centered communication (Barrett and Um, 2026). Considering the documented variances in patient-provider communication across different cultures, even within Asia (Pun et al., 2018), it is important to break away from the Western-centered, behavior-oriented framework of patient engagement.

Among ethnic minority patients, lower health literacy is one of the key barriers for migrants to active engagement in health care. Patients with low health literacy may struggle to formulate questions about their conditions and effectively engage in decision-making with providers, suggesting a pathway between health literacy and health outcomes through provider-patient interactions (Edwards et al., 2012; Paasche-Orlow and Wolf, 2007). Lower health literacy among migrants contributes to less active participation in medical communication (Sudore et al., 2009).

However, empirical research is sparse on how immigrants' health literacy might affect their healthcare engagement. Furthermore, the behavioral orientation of patient engagement in the existing frameworks may limit the understanding of patient engagement among immigrants (Park and Ahmed, 2023). When patient engagement is measured primarily in behavioral terms such as verbal production, then immigrants who have different social norms about verbal engagement, fear discrimination when they speak, or lack the ability to formulate expressions, may appear less engaged, when in fact they are engaged in other dimensions.

When patients do not act upon the cognitive or emotional dimensions of patient engagement whether due to reasons including fear or social norms, they may seem reserved or even disinterested during the interaction with healthcare providers. However, they may later seek information or support through interpersonal networks or online resources. Furthermore, understanding the relationship between health literacy and patient engagement as a multi-faceted construct is essential to further understand the role of health literacy in shaping healthcare experiences among immigrants and to avoid Western-centric assumptions about their patient engagement.

Therefore, the current mixed-method study aims to fill this research gap by conceptualizing the emotional, cognitive, and behavioral dimensions of patient engagement in healthcare communication and investigating the relationship between health literacy and these three dimensions of patient engagement among Korean immigrants in the U.S.

Health literacy and patient engagement

Health literacy is defined as the ability to access, understand, evaluate, and apply health information for health promotion and management (Sørensen et al., 2012). Health literacy is associated with various health outcomes such as healthcare utilization, access to care, or even mortality rates (Berkman et al., 2011; Levy and Janke, 2016). One of the pathways that link health literacy and these health outcomes is patient-provider interaction (Paasche-Orlow and Wolf, 2007). Prior research suggests that health literacy is associated with the levels of patient engagement in medical encounters and the quality of patient-provider communication, influencing health outcomes (Paasche-Orlow and Wolf, 2007). Health literacy can empower patients to actively engage in medical decision-making, effectively communicate their health concerns and negotiate treatment options, and seek more information from healthcare providers for their self-management of health. Ishikawa and Yano (2008) argued that limited health literacy influences patient participation, specifically in information-seeking, healthcare decision-making, self-care and adherence, and healthcare service use, calling for more research to examine the relationship between health literacy and patient participation in healthcare process.

Indeed, several empirical studies found a weak but significant correlation between health literacy and patient engagement (Brabers et al., 2017; Ishikawa and Yano, 2011). For example, focusing on patient involvement in medical decision-making, Brabers et al. (2017) reported that people with limited health literacy prefer less involvement in medical decision-making. Ishikawa and Yano (2011) also found that communicative health literacy is associated with active participation in medical consultations, which leads to better diabetes control.

Health literacy, English proficiency, and patient engagement for immigrants

Although lower health literacy can be found across demographic groups, it is likely to be more prevalent among those with immigrant status (Hoffman-Goetz et al., 2014). Lower health literacy becomes a major barrier for immigrants to actively participate in healthcare processes. For example, immigrants with lower health literacy tend to prefer less involvement in healthcare (Abdelmessih et al., 2019; Hyatt et al., 2017). Even when they prefer active engagement, immigrants are more likely to experience a lack of communication compared to non-immigrants (Hyatt et al., 2017). Immigrants are likely to have other barriers to patient engagement such as language proficiency, preference for indirect communication, cultural beliefs for patient-provider communication, or perceived discrimination from healthcare providers (Schinkel et al., 2019). Considering that expectations for patient participation can be shaped by the cultural norms, the cultural beliefs about patient-provider communication in the country of origin may also influence the level of patient engagement among immigrants in the U.S (Pun et al., 2018). Furthermore, research on immigrant patients acknowledges that health literacy is not merely individual skills but part of acculturation process, developing over the process of acculturation among immigrants (Qian and Ahmed, 2024). Therefore, linguistic and cultural factors can intersect with limited health literacy to influence immigrants' engagement behaviors, contributing to disparate health outcomes for them (Sentell and Braun, 2012).

One of the proxies for acculturation, aside from the length of residence and language used in the household, is English proficiency. English proficiency was frequently studied in relation to health literacy and patient engagement among immigrant populations. Some scholars reported that English proficiency is an antecedent of health literacy, mediating the effect of demographic factors on health literacy (Lee and Choi, 2012). Others found that adequate health literacy may buffer the negative effect of English proficiency on the quality of patient-provider communication among language concordant dyads (Sudore et al., 2009). English proficiency is conceptually different from health literacy as it pertains to the ability to communicate in English, whereas health literacy encompasses a broader scope of abilities and skills (Sørensen et al., 2012). Nevertheless, these studies demonstrate that English proficiency may have confounding effects between health literacy and patient engagement, making the relationships between health literacy and patient engagement more complex for immigrant populations.

Patient engagement in the behavioral, cognitive, and emotional dimensions

One way to capture the complexities of patient engagement is to conceptualize it across emotional, behavioral, and cognitive dimensions, instead of focusing solely on the behavioral aspect (Graffigna et al., 2014; Park and Ahmed, 2023). Graffigna et al. (2014) propose that the synergy across emotional (“feel”), cognitive (“think”), and behavioral (“act”) dimensions positions patients on a continuum of patient engagement. Based on a health psychological approach, they argue that the challenges experienced by diabetic patients are connected to the three experiential dimensions of patient engagement (Graffigna et al., 2014). By doing so, they frame patient engagement as a subjective experience rather than a solely behavioral trait. While this framework is not specifically focused on patient-provider communication, it offers a broader understanding of how patients engage in their health management. This broader lens is useful for examining patient engagement in medical communication, which has traditionally been oriented toward behavioral and cognitive aspects (Park and Ahmed, 2023). Accordingly, this study adopts Graffigna et al.'s (2014) framework of patient engagement and modifies it to explore how immigrant patients engage in medical communication with providers, emphasizing patient engagement in medical communication as a subjective, context-dependent experience.

To illustrate, Graffigna et al.'s (2014) framework includes a good doctor-patient relationship or feeling “guided” in the emotional dimension. Information on the disease or support for knowledge about the disease belongs to the cognitive dimension. Lastly, access to care, improved doctor-patient communication, or devices to support disease management belong to the behavioral dimension. By the same token, the emotional dimension of patient engagement in this study includes satisfaction with communication with healthcare providers, motivation to communicate with healthcare providers, and self-efficacy in information seeking with medical providers. The cognitive dimension includes expectations for the patient's role in medical communication, such as advocating treatment preferences or participating in medical decision-making. It also includes the knowledge about when or what questions to ask healthcare providers. Lastly, the cognitive dimension includes trust in providers' ability to give information.

Existing literature on patient-provider communication extensively examined the behavioral dimension of patient engagement in healthcare encounters. Patient engagement in healthcare encounters has been conceptualized and operationalized as the frequency of verbal participation, such as asking questions and disclosing information (Street and Millay, 2001; Street et al., 2009). The cognitive dimension of patient engagement has also been addressed through research on patients' beliefs and attitudes towards shared decision making (Büdenbender et al., 2023). Furthermore, research has highlighted the importance of emotional and relational aspects of patient-provider communication, such as patient satisfaction (Jiang, 2017). Therefore, the three dimensions identified in Graffigna et al. (2014) are represented across the broader patient-provider communication literature, while there has been a lack of an integrative framework for theorizing communicative patient engagement as a multi-dimensional concept.

Although Graffigna et al. (2014) developed their framework in the context of chronic disease management, this study draws on this integrative framework to explore the behavioral, emotional, and cognitive dimensions of patient engagement beyond disease-specific care trajectory. In this study, these dimensions capture the ways people think, feel, and act as they engage with care providers in healthcare interactions. Furthermore, although the original framework does not explicitly include contextual factors such as language and culture, its integration of not only behavioral but emotional and cognitive dimensions of patient engagement allows for a more nuanced understanding of this complex, situated interaction between immigrant patients and non-immigrant providers shaped by multiple contextual factors such as language, culture, and discrimination (Schinkel et al., 2019). By conceptualizing patient engagement as a subjective experience (Graffigna et al., 2014), this framework allows for more room to explore the contextual factors that shape this experience for immigrants, such as perceived discrimination, cultural influences, or linguistic barriers, thereby giving us an opportunity to explain how the pathway between health literacy and patient engagement is influenced by these factors.

However, existing research on health literacy and patient engagement among immigrant populations primarily focused on the behavioral dimension and not enough research has been conducted on more comprehensive experiences of patient engagement among immigrant patients (Park and Ahmed, 2023). Prior studies often did not treat patient engagement as a multi-dimensional concept and did not explicitly differentiate the behavioral and cognitive dimensions even when they included both dimensions in their operationalization of measurement. For example, patient engagement was often conceptualized as an ultimately behavioral manifestation such as verbal occurrence of asking questions, advocating for their preferences, or sharing information (Park and Ahmed, 2023; Sloan and Knowles, 2017; Street and Millay, 2001). Furthermore, given that immigrant patients navigate healthcare interactions with different interactional norms and cultural values, the established link between health literacy and patient engagement may be more nuanced. For example, the relationship between health literacy and emotional dimension of patient engagement may be more dependent on relational aspects and provider-driven factors such as patient-centered communication skills (Jiang et al., 2024).

Health literacy and patient engagement among Korean immigrants: cultural norms

To contextualize cultural norms for healthcare settings among Korean immigrants in the U.S., this section draws on research involving Korean Americans as a broad term for ethnic designation, first-generation Korean immigrants who later immigrated to the U.S., and Koreans living in Korea. As of 2023, Korean Americans, inclusive of both foreign-born, first-generation immigrants and American-born ethnic Koreans, are the fifth-largest Asian population group, accounting for 8% of the U.S.'s Asian population (Im, 2025). Over 50% of Korean Americans are foreign-born immigrants, reflecting Korea's relatively short immigration history. Korean Americans are more likely to be uninsured and experience limited English proficiency and communication challenges in healthcare settings compared to other Asian groups (Jang et al., 2016; Ryu et al., 2001). Korean Americans are also more likely to report experiencing discrimination in a healthcare setting and lack a usual source of care compared to non-Hispanic White people (Le et al., 2023). Although Korean Americans have lower health literacy than Koreans living in Korea (Chung et al., 2021), several studies have also documented low patient participation in patient safety behaviors, such as speaking up and asking questions, among Korean adults in Korea (Hwang et al., 2019; Lee et al., 2020). This suggests that limited patient engagement may reflect cultural beliefs or social norms rather than health literacy alone.

Previous studies have acknowledged that cultural dimensions such as collectivism, Confucianism, and high power distance, which characterize Korean culture, could potentially lead patients to adopt a more passive, deferential role in medical interactions (Ishikawa and Yamazaki, 2005; Meng et al., 2024). In fact, a previous study on Korean cancer patients reported that Koreans often self-regulate their verbal participation and take on a “passive” role due to Korean-specific interactional norms in healthcare settings (Park et al., 2026). These expectations could potentially lead to a more nuanced and complex relationship between health literacy and patient engagement as Confucianism and collectivistic values emphasize social harmony over individual expression, making this study population distinctive.

Focusing on first-generation Korean immigrants in the U.S., the current study aims to investigate the relationship between health literacy and patient engagement by exploring the emotional, cognitive, and behavioral dimensions of their patient engagement. This study aims to contribute to advancing the knowledge about patient engagement as a subjective experience in emotional, cognitive, and behavioral dimensions as well as the relationship between health literacy and patient engagement in these dimensions among an ethnic minority group. To achieve these goals, this study utilizes an explanatory sequential mixed methods research design, which involves quantitative data collection and analysis followed by qualitative data collection and analysis (Creswell and Creswell, 2022).

Based on the pathway model linking health literacy and patient engagement in the literature, we expect that higher health literacy generally predicts higher patient engagement (Ishikawa and Yano, 2011). By specifying the three dimensions of patient engagement, the following hypothesis is proposed and presented in Figure 1. Hypothesis: Health literacy among Korean immigrants is positively associated with patient engagement across emotional, cognitive, and behavioral dimensions, controlling for English proficiency.

Guided by the model of patient engagement across behavioral, emotional, and cognitive dimensions (Park and Ahmed, 2023), this study intends to explore the comprehensive aspects of patient engagement perceived by Korean immigrants. Thus, the following research question is posed. RQ: How do Korean immigrants describe their behavioral, emotional, and cognitive patient engagement in relation to health literacy?

Study design

This study was designed as a two-phased, mixed-methods research using an explanatory sequential design (Creswell and Creswell, 2022; Lee et al., 2020). The purpose of this research design is to examine the hypothesized relationships in the quantitative phase and then to explain any significant findings and explore the observed relationships in more depth in the qualitative phase. Phase 1 consisted of an online survey on health literacy and patient engagement among Korean immigrants in the U.S. Phase 2 involved online or phone interviews about health literacy challenges and patient engagement. Figure 2 illustrates the study procedures.

Phase 1

Participants

Convenience sampling was utilized to recruit the sample. Eligible survey participants were first-generation Korean immigrants aged 18 years or older residing in the U.S., who had visited a healthcare provider in the previous 12 months. Recruitment flyers were posted on Korean ethnic media websites, online forums, and Facebook clubs for Korean immigrants in the U.S. An online survey using Qualtrics was conducted in March and April 2021. Survey participants received a $5 gift card as compensation.

A priori power analysis was conducted using G*Power 3.1 (Faul et al., 2007) to assess whether the sample size was sufficient for the MANCOVA and the multiple linear regression models. Power analysis indicated that a sample of 108 would be required to detect an effect size of 0.15, at a significance level.05, and the desired power of.90 for the MANCOVA. For the multiple linear regression, a sample of 88 would be sufficient to detect an effect size of 0.15 at a significance level of.05 with the desired power of.90. The final sample of 105 was slightly below the required sample for the MANCOVA but still provides sufficient sensitivity to detect a medium-sized effect. The sample was sufficient for the multiple linear regression analysis. The demographics of the sample are summarized in Supplementary Appendix 1.

Measures

Health literacy

Health literacy (M = 8.30, SD = 1.50) was measured by Health Literacy Skills Instrument – Short Form (HLSI-10) (Bann et al., 2012). The HLSI-SF consisted of 10 items measuring health literacy as a comprehensive construct that encompasses print literacy, numeracy, oral literacy, and internet literacy. Each item contributed to a composite score ranging from 0 to 10 with higher scores indicating greater health literacy. The scale developers suggested that a score of 7 or higher is considered adequate health literacy and reported that 63% of their pilot sample for scale development had adequate health literacy (McCormack et al., 2017). Thus, this cut-point is utilized for quantitative data analysis.

The sample showed a higher mean score of HLSI-SF than that of the pilot sample provided by the developers of the scale (McCormack et al., 2017). The mean health literacy score among the participants was 8.30 (SD = 1.50). 88.46% of participants demonstrated an adequate health literacy level, or a score of 7 or higher. Only 11.54% of them had a total score of 6 or less as opposed to 37% in the pilot sample for scale development (McCormack et al., 2017).

English proficiency

English proficiency was assessed using a single self-reported item asking participants to rate their ability to speak English. Responses were recorded on a 5-point scale ranging from Poor to Excellent (M = 3.20, SD = 1.16). The use of a single, self-reported item was consistent with large population-based health surveys, including the nationally representative Health Information National Trends Survey and the California Health Interview Survey. English proficiency was treated as a continuous covariate in the analyses.

Patient engagement

Patient engagement was measured by a scale focusing on communicative aspects of patient engagement by adapting scales from the previous literature (Xu et al., 2018) and guided by the model of patient engagement across behavioral, emotional, and cognitive dimensions (Barello et al., 2015; Park and Ahmed, 2023). The scale is composed of three subscales. The behavioral subscale of patient engagement consisted of five items adapted from Xu et al. (2018) (Cronbach's α =.91, M = 3.87, SD =.91). An example of the behavioral subscale is “I ask healthcare providers for explanations or information about my health condition on my own initiative”. The emotional subscale of patient engagement was developed for this study and had three items including “I am satisfied with the level of communication about my health with healthcare providers” (Cronbach's α =.85, M = 3.66, SD =.94). The cognitive subscale of patient engagement was developed for this study and included 4 items including “I know what to ask and how to ask a question to healthcare providers” (Cronbach's α =.89, M = 3.90, SD =.88). All items were rated on a 5 point Likert scale ranging from 1 (Strongly disagree) to 5 (Strongly Agree).

Quantitative analysis

105 survey responses were analyzed using descriptive statistics, Exploratory Factor Analysis (EFA), and multivariate analysis of covariance (MANCOVA). SPSS version 28 was used for data analysis.

For patient engagement scale, EFA was conducted using Principal Axis Factoring with the oblique promax method based on the theoretical conceptualization that suggests likely correlation among the items. To determine factorability of the data, the inter-item correlation matrix was inspected; all items correlated with at least one other item, r >.30. In addition, the Kaiser-Meyer-Olkin statistic was greater than.8 (KMO =.90) and Bartlett's test of sphericity was significant (χ 2 = 934.00, df = 66, p <.001), both indicating that the data were suitable for factor analysis. Eigenvalues of ≥1 were used to interpret the number of factors in the data set. The rotated solution showed that items loaded on their intended emotional, cognitive, and behavioral dimensions, except for one item (Behavioral 1), which was removed from any further analysis due to cross-loading. Total variance explained by three factors was 69.89%. Communalities, factor loadings, and the factor to which each item belongs are presented in Supplementary Appendix 2.

A multivariate analysis of covariance (MANCOVA) was conducted to examine the associations between health literacy and the three dimensions of patient engagement with English proficiency as a covariate. MANCOVA was selected as the primary analysis because the categorical classification was aligned with our research questions and had both clinical and theoretical relevance. This approach is also consistent with previous health literacy research (e.g., Sentell and Braun, 2012; Sudore et al., 2009). English proficiency was retained as an a priori covariate as previous literature has consistently identified it as a relevant factor for both health literacy and patient engagement in healthcare communication (Lee and Choi, 2012; Yip, 2012). Controlling for English proficiency also allowed us to examine whether health literacy was associated with the three dimensions of patient engagement independently of a participant's linguistic ability. As MANCOVA requires the predictor variable to be categorical, the cut-point suggested by the scale developers was used for grouping the sample in health literacy (Bann et al., 2012). Thus, health literacy adequacy was utilized as a predictor variable. Then, the MANCOVA was conducted with English proficiency as a covariate. No issues related to missing data or outliers were identified. However, a review of the dependent, continuous variables of emotional, cognitive, and behavioral patient engagement identified that they all had a high degree of negative skewness. Thus, to meet the multivariate normality assumption of MANCOVA, Templeton's (2011) two-step approach for transforming continuous variables into a normal distribution was adopted. After transformation, Mardia's test for multivariate skewness and kurtosis was conducted to ensure multivariate normality. Both skewness and kurtosis were not significant, indicating a normal distribution (skewness =.47, p =.58, kurtosis = 15.29, p =.79). Box's M of 4.99 indicated that homogeneity of covariance matrices across groups is assumed [ F (6, 2,069.66) =.746, p =.61]. The same statistical analyses were also conducted on the original, untransformed variables and the substantive findings remained unchanged.

To address the limitations associated with dichotomization, such as a loss of information and statistical power, multiple regression analyses were also conducted as exploratory, supplementary analyses using the original continuous HLSI-SF score. The multiple regression analyses complement the MANCOVA findings by using the original continuous HLSI-SF scores and by providing outcome-specific estimates for each dimension of patient engagement. For the supplementary multiple regression analysis, assumptions were checked and met. Linearity assumption was confirmed using scatterplots, and Durbin-Watson statistics (1.77–1.91) indicated no autocorrelation. VIF and Tolerance were examined to test multicollinearity. All VIF was well below 5 and tolerance was close to 1, indicating no issues with collinearity. Q-Q plots and scatterplots showed no violations of homoscedasticity and normality of residuals. Sensitivity analyses were also conducted with additional demographic and acculturation-related variables such as age, gender, education, income, immigration status, the length of residence in the U.S., and language used in the household.

Phase 2

Participants

Phase 2 participants were recruited through a combination of convenience and purposive sampling within the nested sample of survey participants who agreed on follow-up interviews. Interview participants were screened in two steps. First, health literacy and patient engagement scores were used to create a 2 × 2 matrix with health literacy scores (7 or lower vs. 8 or higher) and patient engagement scores (below and above average). Applying the same cutoff score of 6 was impractical, as very few participants agreed to follow-up interviews among those who fell below this threshold. To enable meaningful variation in the qualitative sample, the cut-off point for the level of health literacy was adjusted to a 7 out of 10, considering that the mean health literacy score among the sample was above 8. To ensure diversity in the sample, interview participants were contacted to balance the number of participants representing each of the 4 groups (Low vs. high HL, low vs. high PE). Within each cell, participants were purposively selected to maximize variation in demographic characteristics, such as age and gender. However, because of the small number of survey respondents who agreed to participate in the follow-up interviews, it was not possible to achieve balance across all demographic characteristics, health literacy, and patient engagement within each cell. The distribution of health literacy and patient engagement scores among the qualitative sample is presented in Table 1. Participants were compensated with a $15 gift card for their participation in the interviews. Demographic information of participants is provided in Supplementary Appendix 1.

Data collection

A total of 20 semi-structured in-depth interviews were conducted via video conference or phone call by the first author. Participants were given the option to use their preferred language between English and Korean for the interview. 19 out of 20 interviews were conducted in Korean. Interviews were conducted between April and June 2021. Interviews lasted between 27 and 50 min with an average of approximately 39 min. Interview questions were guided by the research question about health literacy and patient engagement based on the model of behavioral, emotional, and cognitive dimensions (Park and Ahmed, 2023).

Qualitative analysis

Interview data were first transcribed verbatim in Korean (except for the one interview conducted in English, Participant 15) and translated by the first author, who is a native Korean speaker and professional translator between Korean and English. To ensure accuracy, 10% of the English translation was randomly selected and backtranslated by another professional translator, blinded to the research questions. After translation and backtranslation, a review panel between the first author and the backtranslator discussed any ambiguity and confirmed equivalence before the English translation was used for further analysis by the research team. This study benefited from researcher triangulation (Flick, 2018) as the two authors conducted an iterative process to discuss and corroborate issues related to the theoretical framework, data collection, and analysis. The first author is a Korean researcher who employed a perspective of a cultural insider, while the second author is a non-Korean researcher who employed a perspective of a cultural outsider to analyze the data. This process allowed for the triangulation of analytical perspectives between the researchers.

Reflexive thematic analysis (RTA; Braun and Clarke, 2019) was employed to analyze the interview data. Following Byrne (2022), the authors first discussed the theoretical assumptions underlying qualitative analysis, namely, constructionist epistemology, experiential orientation, and inductive analysis. The authors also agreed to employ semantic and latent coding for data analysis. Independent coding of the data based on the theoretical framework of behavioral, emotional, and cognitive dimensions was followed by an iterative process of discussions between authors to generate, review, and finalize themes from the initial coding.

Phase 1

Health literacy and emotional, behavioral, and cognitive patient engagement

Having established the assumptions were valid, a one-way MANCOVA was conducted to test the hypothesis that there would be mean differences in emotional, cognitive, and behavioral patient engagement between adequate and inadequate health literacy groups. The results demonstrate that there was a significant difference between health literacy adequate and inadequate groups in the combined patient engagement variables after controlling for English proficiency, F (3, 100) = 3.645, p <.05, Pillai's trace =.099, partial η 2 =.099. Pillai's trace was used to interpret the results due to its rigorousness for the unequal size of the samples. The test for covariate was significant, indicating that English proficiency is associated with all three dimensions of patient engagement [F(3, 100) = 11.225, p <.001, Pillai's trace =.252, partial η 2 =.252]. A follow-up univariate test demonstrated that when English proficiency is controlled, significant mean differences between adequate and inadequate health literacy groups are found in cognitive and behavioral dimensions of patient engagement in that the group with adequate health literacy has a higher mean of behavioral and cognitive patient engagement. However, no significant mean difference is found in the emotional dimension (Table 2).

A follow-up multiple regression analysis was conducted to further examine the individual contribution of health literacy and English proficiency to each dependent variable. All three models were statistically significant [Behavioral: F (2, 102) = 27.38, p <.001, adjusted R 2 =.34; Cognitive: F (2, 102) = 21.01, p <.001, adjusted R 2 =.28; Emotional: F (2, 102) = 11.53, p <.001, adjusted R 2 =.17]. The results confirmed that health literacy is a significant predictor for behavioral and cognitive dimensions of patient engagement, even when controlling for English proficiency, while no such association was found between health literacy and emotional dimension. In contrast, English proficiency was a significant predictor for emotional dimension. The standardized beta coefficients (β) indicate that English proficiency showed stronger associations with emotional (β =.40, p <.001) and cognitive (β =.43, p <.001) dimensions of patient engagement. Health literacy was not a significant predictor of the emotional dimension (β =.10, p =.30), whereas it was significantly associated with the cognitive dimension (β =.25, p =.004). For behavioral dimension, health literacy (β =.39, p <.001) had a slightly stronger association than English proficiency (β =.37, p <.001). The summary of multiple regression analysis results is presented in Table 3. These results were largely consistent with the MANCOVA results, as the MANCOVA showed that participants with adequate health literacy differed significantly in behavioral and cognitive patient engagement, compared to the inadequate health literacy group, but not in emotional engagement. Complementing these group-based findings, the regression analyses confirmed that across the full score range, higher health literacy scores were associated with greater behavioral and cognitive patient engagement.

Sensitivity analyses with additional variables, such as age, gender, education, income, immigration status, the length of residence in the U.S., and the language used in the household, were repeated for each of the dimensions. None of the added variables predicted any of the three dimensions, and their inclusion did not materially change the magnitude, direction, or statistical interpretation of the main relationships.

Phase 2

Phase 2 of this study focused on complementing the quantitative results and exploring more about how the three dimensions of patient engagement might be influenced by patient health literacy. The list of themes and relevant dimensions, as well as some illustrative quotations, are provided in Table 4.

Emotional dimension of patient engagement

The emotional, behavioral, and cognitive dimensions of patient engagement were explored in more detail in qualitative data. Participants expressed their frustration and tensions about insufficient patient engagement. Descriptions of negative feelings were particularly frequent among those with lower health literacy, although some participants who scored higher than average for health literacy still expressed frustration or tension. This was in line with our quantitative findings that did not show a difference between adequate and inadequate health literacy groups in emotional patient engagement.

When participants expressed frustration, it was often toward the doctors who did not accommodate their language needs. However, some participants also expressed frustration about themselves, sharing that they felt the need to improve their English skills or showing self-reflection about their reliance on accommodation from providers. For example, participant 3 shared a story about when she visited a doctor with her American husband, the doctor mainly talked with her husband at a fast pace using difficult terms. According to her, this specific interaction “revealed the limits of [her] English proficiency”. Her frustration and negative feelings were directed toward herself, showing signs of self-blame.

Participants frequently described that they felt as if there was something left unchecked, which indicates their uncertainty about communication with providers. Because of language barriers, they often felt uncertain or unclear about what was communicated in their medical encounters. Some alluded to feeling negative emotions such as distrust because of these uncertainties. For example, participant 20 shared that because she could not be certain about how well she understood or made herself understood by her provider, she felt like she needed to do extra work to ensure the accuracy of the information exchanged. Participant 14 also shared that when his communication was not satisfactory, he feels uncertain if he can trust the diagnosis from the doctor.

Behavioral dimension of patient engagement

Supporting our quantitative findings, patients with higher health literacy often said they were active in giving and seeking information in medical encounters. Overall, they described that their health literacy enabled them to enact their cognitive beliefs about patient engagement. For example, participant 17 believes she tends to ask more questions because of her health literacy, expressing: “Why would I bother to ask questions if I think that I wouldn't understand anyway?” In other words, participants described that they would ask questions (enactment) when they were capable of understanding and utilizing the information given from healthcare providers. Participants also had experiences of seeking more information from healthcare providers or sharing their circumstances to help their providers understand them better, although this engagement did not always end as a satisfactory encounter depending on providers' reactions.

Participants also mentioned that their patient engagement was often shaped by both themselves and their providers. Participants often described their level of engagement as dependent on how receptive providers were. Some participants cited their English proficiency as a reason why their patient engagement was dependent upon the interpersonal traits of their providers. For example, Participant 19 shared, “When the doctor doesn't understand me well and has an attitude that seems closed, then I don't ask questions, to be honest. I tend to be influenced by doctors”. Participant 10 also explained that after several unsuccessful experiences communicating with providers, he stopped his efforts for patient engagement.

Cognitive dimension of patient engagement

Participants often cited their cognitive beliefs for their engagement behaviors. Many participants, especially those with higher health literacy, expressed expectations for equal partnership through active engagement from both sides. While participants with low health literacy generally focused on patients' role of complying with doctors' instructions or provider's role of giving information to patients, participant 2 mentioned that healthcare providers need to ensure that patients understand their options to make informed healthcare decisions: “I think when we as patients have equitable information to medical providers, that's when our relationship can be equal”. Highlighting the information asymmetry between patients and providers, participant 2 emphasized that equal partnership is possible when providers initiate it.

Some participants explained that active patient engagement is not a social norm in Korean culture and that a paternalistic view of provider-patient relationships is more prevalent. Participants frequently mentioned that due to cultural differences, they feel they are made to be more engaged in healthcare communication in the U.S. Participants seemed to recognize the cultural differences in medical practices between the two countries. For example, participant 20 shared, “[Unlike in Korea] here in the U.S., they tend to leave it up to us to make decisions. They wouldn't pick one option for you. And it's quite hard for me to make decisions. It's a lot of pressure on me”. At the same time, participants were concerned about whether providers had sufficient cultural competence to understand these cultural differences. Participant 13 shared, “I know a doctor should have ideas about their own solutions, but when a patient suggests a different idea, I think they should not be judgmental about it but be open to them and ask them why they think that way, leaving room for negotiation and consultation for an intervention”. Cultural competence was mentioned by many participants, referring to examples of medical practices or genetic differences such as the use of tests or medication, the prevalence of certain diseases (e.g., Type 2 diabetes), or healthcare insurance.

By utilizing a mixed method design, this study aimed to test the relationship between health literacy and behavioral, cognitive, and emotional dimensions of patient engagement and broaden the understanding of patient engagement across these three dimensions.

Quantitative findings revealed that adequate and inadequate health literacy groups showed differences in their behavioral and cognitive patient engagement when their English proficiency was controlled. A multiple regression analysis also demonstrated that while health literacy is a significant predictor for behavioral and cognitive patient engagement, no association was found between health literacy and the emotional dimension. Overall, the findings were consistent with the literature in that high health literacy predicts high patient engagement (Brabers et al., 2017; Ishikawa and Yano, 2011). Considering the focus of the literature on behavioral and cognitive aspects in measuring patient engagement (Graffigna et al., 2014; Park and Ahmed, 2023), it is expected that health literacy predicts these dimensions of patient engagement. However, notably, no such relationship was found between health literacy and the emotional dimension of patient engagement. Our qualitative findings supported this finding as both higher and lower health literacy groups commonly described feelings such as frustrations or tensions.

English proficiency was found to be a significant covariate, which is associated with all three dimensions of patient engagement. Furthermore, multiple regression analysis showed that English proficiency had a stronger association with the cognitive dimension and was the only significant predictor for the emotional dimension of patient engagement. In line with the literature where English proficiency is suggested to play a greater role than health literacy in interpersonal contexts whereas health literacy becomes more significant in mediated contexts (Sentell and Braun, 2012; Yip, 2012), English proficiency carried a greater association with cognitive dimension of patient engagement than health literacy. This finding indicates that the emotional dimension, such as feeling motivated, confident, or satisfied with patient engagement, is greatly determined by English proficiency rather than health literacy. This finding is aligned with the previous literature suggesting that English proficiency may lead to stronger self-efficacy in navigating the healthcare system or communicating with healthcare providers, serving as an important mediator for health literacy (Lee and Choi, 2012; Lee et al., 2020; Yip, 2012).

Qualitative data provided rich details about the dimensions of patient engagement among the sample. As described earlier, many participants shared negative emotions about insufficient patient engagement in medical encounters such as not being prompted to ask questions or express their opinions, regardless of their health literacy levels. These emotions were often internalized as shame or stigma for lower health literacy or English proficiency. Feelings of uncertainty were also commonly expressed, which could affect patient compliance, trust in doctors, or medical decision-making with providers.

Participants often described their feelings as positive when they felt like their doctors understood their needs, even when they did not articulate them well. This finding could indicate that lower patient engagement in behavioral terms does not necessarily lead to negative emotions. By the same token, higher patient engagement in behavioral terms does not necessarily mean positive emotions. For example, when patient engagement was driven by unsuccessful communication attempts between patients and providers, it was not translated into positive feelings. In other words, although their patient engagement was active in behavioral terms, active behaviors did not result in positive emotions. These qualitative findings support the lack of association between health literacy and patient engagement in emotional dimension. Together with the quantitative findings, participant narratives suggest that emotional engagement may operate independently of a patient's individual literacy skills, likely relying more on relational and provider-driven factors. In fact, a previous study reported that immigrant and refugee patients need to feel that they are in a safe space and encouraged by providers, acknowledging the importance of therapeutic alliance before they are motivated to co-produce healthcare (Radl-Karimi et al., 2022). Healthcare models that emphasize the relational and provider-driven factors such as patient-centered communication have consistently been associated with patient empowerment and trust, which seem to be related to the emotional dimension of patient engagement (Jiang, 2017; Jiang and Street, 2019).

Participants described that health literacy allowed for the enactment of cognitive beliefs about patient engagement. At the same time, they also felt their patient engagement behavior was influenced by differences in medical practices or healthcare systems. However, aside from these external factors, interpersonal factors such as providers' personality, attitudes, or cultural competency were often brought up as a factor that influenced patient engagement behaviors. These accounts highlighted that patient-provider communication is a situated interaction and that their actual behavior is shaped by not only their own traits but also external factors such as healthcare systems or provider factors such as cultural competency (Street, 2003). Although participants demonstrated a spectrum of passive to active views about patients' role in healthcare communication, findings indicated that these cognitive beliefs are not always enacted as behaviors due to language barriers, providers’ reactions, or perceived discrimination (Xiao and Wu, 2025).

Participants referred to health literacy in all three dimensions of patient engagement when they shared experiences of facing difficulties with understanding medical terms, navigating different healthcare systems, or formulating questions. These accounts pictured the relationships between the three dimensions of patient engagement and health literacy in more depth, calling for a more comprehensive conceptualization of patient engagement that acknowledges not only behavioral but also cognitive and emotional dimensions. A more comprehensive framework of patient engagement will inform research on the pathways between health literacy and health outcomes as it enables us to identify specific domains that require a different approach than health literacy alone. For example, given the interdependence of the behavioral, cognitive, and emotional dimensions (Graffigna et al., 2014), improved patient engagement in the behavioral domain may also be related to the cognitive dimension, such as expectations for shared decision-making.

Several limitations should be considered when interpreting the findings of this study. First, because a convenience sampling technique was used, the sample of this study may not be representative of the overall Korean immigrant population in the U.S. The sample was skewed toward people with higher education, female participants, and those with high health literacy. Second, the survey was conducted in English, which may not capture a subgroup of Korean immigrants with limited English proficiency. Although English proficiency was controlled in the quantitative analysis of this study, future research could benefit from translating and validating the survey items in Korean and understanding health literacy in their primary language. The patient engagement scale findings should be interpreted as exploratory as the EFA was conducted on the same sample that was used for hypothesis testing, and factor structure was not confirmed on an independent sample through CFA. Although supplemented with multiple regression analyses and qualitative findings, unequal group sizes in MANCOVA may have reduced statistical power and precision. Also, English proficiency was measured with a single self-report item. Lastly, the data used for this study are cross-sectional. Although health literacy was conceptualized as a predictor of patient engagement, their relationship could be reversed as well.

Despite these limitations, this study provides valuable contributions to understanding the subjective experience of patient engagement in medical encounters among Korean immigrants in the U.S. by conceptualizing it across behavioral, emotional, and cognitive dimensions. This study's contribution lies in using the new conceptualization and empirically testing it, demonstrating that the established relationship between health literacy and patient engagement is in fact more nuanced. This study highlights how existing literature on patient engagement as a purely behavioral trait runs the risk of ignoring the important aspect of patient engagement, that is, emotional dimension, and lastly emphasizing how the pathway between health literacy and patient engagement should incorporate the issue of language and culture when it is studied among immigrant patients. This expanded concept of patient engagement emphasizes the ways in which the pathway between health literacy and patient engagement is influenced by contextual factors, calling for more research in this area. Although prior research acknowledged the importance of language proficiency for immigrants in accessing healthcare services, this study clarifies that health literacy interventions can be useful in facilitating patient engagement in the behavioral sense, whereas language-based interventions should be prioritized for patient satisfaction.

Additionally, the findings have some practical implications. Practitioners who work with immigrant patients should recognize that patient engagement of immigrants can be shaped by not only health literacy but also other important factors such as cultural competency of the providers, linguistic assistance, and internalized discrimination or self-blame. Interventions designed to improve patient engagement of immigrant populations should incorporate not only health literacy, as health literacy alone may not be enough to improve the emotional dimension of patient engagement among immigrants. In other words, health literacy interventions may be helpful for people to adjust their expectations and their own role in medical communication as well as their enactment of active participation in communication, but linguistic assistance or cultural competence of providers may be more useful in improving patients' satisfaction, motivation, and self-efficacy in communicating with healthcare providers. An integrated approach that encompasses cultural sensitivity trainings and linguistic assistance as well as health literacy interventions for patients can be beneficial.

It should be noted that although this study focused on patient health literacy, health literacy has been studied not only as a patient-side factor but also as a provider-side factor. For example, Healthy People 2030 proposes that organizational health literacy refers to “the degree to which organizations equitably enable individuals to find, understand, and use information and services to inform health-related decisions and actions for themselves and others” (U.S. Office of Health and Human Services, n.d.). Furthermore, the National Standards for Culturally and Linguistically Appropriate Services (CLAS) suggest that the provision of health care services be tailored to the cultural and linguistic needs of diverse populations (U.S. Department of Health and Human Services, n.d.). Therefore, the implications of the current study can be effectively addressed by implementing the policy and systematic recommendations.

This study investigated the relationship between health literacy and the emotional, cognitive, and behavioral dimensions of patient engagement among Korean immigrants in the U.S. Quantitative findings indicated that Korean immigrants with higher health literacy have active patient engagement across its behavioral and cognitive dimensions when English proficiency is controlled. However, no such relationship was found between emotional dimension and health literacy. Qualitative data supported this finding by revealing the more complex nuances of patient engagement in emotional, behavioral, and cognitive dimensions and their relationships with health literacy. Although participants with higher health literacy generally demonstrated stronger attitudes and behaviors regarding patient engagement, common emotions underlying such behavior were present regardless of health literacy levels. Taken together, the findings suggested that although the health literacy of the participants shapes patient engagement across behavioral and cognitive dimensions, emotional state of these immigrants could be independent of their behavioral and cognitive patient engagement, revealing the limitations of behavior-oriented frameworks and implications for healthcare experiences for ethnic minority patients.