Introduction

A persistent issue for immigrant and other minority elders is the underutilization of healthcare services [1–3]. Immigrants may have limited accessibility to US healthcare services [4–6] because of differences in languages [7] and culture [8]. They may also lack an understanding of how the US healthcare systems differ from those of their native countries. Asian immigrant elders are among the most vulnerable groups of elders underutilizing healthcare services [9].

Many studies with Chinese and Korean immigrants have employed the Andersen–Newman model to explain various healthcare service utilization behaviors in the US or Canada [7], such as nursing home use [10], mental health service use [2, 11, 12], and community service use [13]. However, only a few studies have tried using this model to explain physician visits.

Although most studies of Asian immigrant elders have unequivocally found a pattern of underutilization, methodological challenges such as securing a sample led to those studies being conducted in major metropolitan cities with established ethnic communities in, for example, New York City [9], Boston [14], Chicago [15], and Los Angeles [16, 17]. Factors contributing to healthcare service utilization among Asian immigrant elders living in places where systematic institutional support from the ethnic community is not readily available are unknown. In addition, heterogeneity among Asian subgroups has not been studied [18]. The Asian communities in Arizona have a much smaller population size and a shorter immigration history, compared to the communities in New York City or Los Angeles. The Chinese and Korean populations tend to be scattered across all parts of Maricopa County, Arizona. The 2000 Census data show 658 census tracts in Maricopa County. Among these 658 census tracts, 20 of them have Chinese households with a population threshold of 100 or more people, and there was only 1 census tract containing Korean households with a population threshold of 100 or more people. Hence, Asian ethnic communities in Arizona tend to have fewer structured organizations that can support these communities.

In this study, the majority of the Chinese participants immigrated from Taiwan or the Mandarin-speaking region of China. In Arizona, Cantonese-speakers have a longer immigration history than Mandarin-speaking elders [19], which means that these respondents have been in the US for a relatively short time. The Korean immigrant elders in this study have distinctively different backgrounds from the Chinese respondents in terms of both language and culture. This made it imperative to conduct separate analyses, rather than treating respondents as a homogeneous group. The elderly populations of Chinese and Korean Americans in Arizona have grown rapidly over the last two decades. However, no studies have been conducted among these elders. Thus, we do not know how these two Asian groups might differ from those residing in other parts of the United States. The PI of this study was approached by a Taiwanese medical doctor who required a needs assessment with Mandarin speaking Chinese elders. She had been informed that state governments could not provide any services or support without such a needs assessment. Funding limitations meant that the PI could not include more than these two groups in the study.

Theoretical/Conceptual Framework

The current study’s conceptual framework was based on the Andersen–Newman behavioral model of healthcare service use. In this model, an individual’s propensity to use healthcare is influenced by three factors: predisposing, enabling and need factors [20, 21]. In the traditional Andersen–Newman health behavior model, predisposing factors cannot be changed by intervention because these factors have innate characteristics—demographic background, immigrants’ length of stay in the U.S., and age. Enabling factors refer to resources that allow individuals to access the healthcare system, such as financial resources and knowledge about healthcare. Need factors are health conditions requiring individuals to use the healthcare system. The Andersen–Newman model’s adaptability makes it applicable to various groups, including Chinese-Americans [11] and Korean-Americans [16].

The present study’s goal was to explore the contributions of predisposing, enabling, and need factors on healthcare use behaviors among Chinese and Korean immigrant elders living in Arizona. This is the first study with these two groups of Asian elders in Arizona regarding their healthcare service use. It will fill a gap in the current literature by expanding our understanding of healthcare use behaviors among Asian immigrant elders living in areas without established ethnic enclaves.

Methods

Participants and Data Collection

This study utilized a cross-sectional survey design. Employing face-to-face interviews, data were collected from 217 Asian immigrant elders (116 Chinese and 101 Koreans) residing in two counties in the state of Arizona from December 2005 through March 2007.

The inclusion criteria for this study were: (1) 65 years or older; (2) of either Chinese or Korean origin; (3) living in a community setting; and (4) understand and speak Mandarin or Korean.

The institutional review board of the first author’s affiliated institution approved the proposal to ensure minimal risk to study participants. Survey participants were offered $30 in cash to compensate their time and efforts.

This study used a snowball sampling method to recruit participants. Given that there were no resources available for establishing a sampling frame, this [22] may have been one of the best ways of identifying Chinese or Korean immigrant elders who were scattered across Maricopa and Pima Counties. After each interview, respondents were asked to provide names of other potential participants who might meet the inclusion criteria.

One Mandarin-speaking and one Korean-speaking graduate student were trained in survey-interview skills. On average, interviews lasted 90 min.

Measures

Outcome Variables

Healthcare Service Use

Healthcare service use was measured and operationalized as the frequency of physician visits in the past 12 months. The responses ranged from 0 to 30.

Predisposing Factors

Predisposing factors were age, sex, marital status, and length of stay in the United States.

Enabling Factors

Health insurance was a binary variable indicating no health insurance (0) or at least one health insurance plan (1). English language proficiency was assessed with three questions about speaking, reading, and writing skills. Responses ranged from not at all (0) to very well (3). The internal consistency reliability was high (α = 0.91). The cultural gap was assessed with the question, “All in all, how much do you think your opinions about cultural topics differ from those of your children?” Responses ranged from not at all (0) to in very important ways (3). Help from children was assessed by counting the total number of times children assisted respondents with specific tasks (α = 0.82). This was derived from responses to the question, “I would like to know if your (children) ever (helps/help) you in any of the following ways”.

Need Factors

Self-reported health status was derived from participant responses to the question, “How would you rate your overall physical health? Is it excellent (4), very good (3), good (2), fair (1), or poor (0)?” Depressive symptoms were measured using The Geriatric Depression Scale-Short Form (GDS-SF), a 15-item measure that assesses depressive symptoms based on DSM-IV criteria, including a depressed mood, and feelings of hopelessness and worthlessness [23]. The GDS-SF has been reliably used in previous studies of immigrant elders living in New York City [24]. Number of medical conditions referred to the number of health problems participants experienced. The range was 0–15. Number of stressful life events was assessed by asking participants whether they experienced any of 10 stressful events, such as the death of family members or a recent move.

Analyses

Given their heterogeneity, we chose to run separate analyses of the two ethnic groups. All analyses were performed using Stata version 12.1 [25]. Descriptive statistical analyses were used. Independent sample t tests and χ2 statistics compared characteristics between the two ethnic groups. Negative binomial regression analyses modelled the count outcome (the number of physician visits), which showed overdispersion (i.e., variance > mean) [26]. A test of the dispersion parameter alpha confirmed overdispersion patterns for both the Chinese and the Korean samples (see Table 2). The results of a multicollinearity test [27] (tolerance = 0.684, variance inflation factor = 1.501) suggested that collinearity among the predictors was not a concern.

Results

Sample Characteristics

Table 1 shows descriptive statistics for variables used in the analyses by ethnic background, as well as the overall statistics. The sample consisted of 217 Asian immigrant elders, with the Chinese sample significantly older than the Korean sample (t = 5.67, p < 0.001). The majority of the sample was female (55 %) and most of the participants were married (72 %). The Korean sample had lived in the U.S. significantly longer than the Chinese sample (t = −4.62, p < 0.001), an average of 25.5 years and 17.5, respectively. The numbers of physician visits did not differ between the Chinese and Korean samples.

Table 1 Sample characteristics of Chinese and Korean immigrant elders in Arizona (N = 217)
Full size table

All of the enabling factors were similar across ethnic groups, except for the perceived cultural gap (t = −6.33, p < 0.001): Chinese participants reported a significantly lower culture gap than their Korean counterparts. Among the need factors, the Chinese sample reported lower levels of depressive symptoms (t = −4.06, p < 0.001), fewer stressful life events (t = −6.42, p < .001), and rated their current health status higher (t = 2.99, p < 0.01) than the Korean sample. However, there was no significant difference in the number of medical conditions.

Negative Binomial Regression Outcomes

Table 2 shows the findings from the ethnic-specific analyses of the Chinese and Korean samples. Among Chinese immigrant elders in this study sample, females were 30 % less likely to make additional physician visits than their male counterparts. Each additional year in the length of stay in the United States was associated with a 2 % increase in physician visits. A unit added to the English language proficiency score revealed a 6 % increase in physician visits. Conversely, a unit increase in the perceived cultural gap score was related to a 26 % decrease in physician visits. Each incremental step in the Geriatric Depression Scale (GDS) symptom score was associated with 4 % more physician visits. Finally, an additional medical condition was linked to a 13 % rise in the likelihood of additional physician visits.

Table 2 Negative binomial regression predicting physician visits using Anderson–Newman health utilization model: by ethnic groups
Full size table

In the Korean sample, being unmarried was associated with a 74 % increase in physician visits. Having health insurance revealed more than a two-fold (239 %) increase in physician visits. An additional medical condition meant a 17 % increase in physician visits. Finally, a unit increase in the self-reported health status was related to a 24 % decrease in the likelihood of additional physician visits.

Discussion

With the Andersen–Newman behavioral model of healthcare service use, the present study assessed the predictors of healthcare service use among Chinese and Korean immigrant elders.

Similar to prior studies, Asian immigrant elders in the current study reported substantially less healthcare service use (about four times a year), compared with all people in the United States aged 65 and over, who used healthcare services an average of 15 times a year [2, 28]. Interestingly, the self-rated health status was similar to the results from the representative sample from the U.S. general population. About 79 % of our study sample rated their health as good to excellent, while 78 % of non-Hispanic Whites rated theirs in the same way [28]. This may mean that Asian immigrant elders use healthcare services at a much lower rate, despite similar healthcare needs.

Healthcare use in the current study is also substantially lower when compared with the respondents’ counterparts living in a metropolitan area. Only 31.4 % of this study’s participants visited their doctor five or more times a year, compared with Asian immigrant elders living in New York City, where about 50 % of Asian immigrant elders did so [29]. This may suggest that living in a place with a lower density ethnic community may intensify the disparities in healthcare service use, compared to a place where ethnic resources are readily available.

The number of medical conditions, a common need factor, increased the likelihood of utilizing healthcare services. This means that among Asian immigrant elders living where the environment may not be conducive to accessing healthcare, absolute need is the major driving force behind healthcare use. Moreover, the Kaiser Family Foundation (2015) looked at primary care health professional shortage areas (HPSAs). The state of Arizona ranking is below average, meaning they need many more primary care health professionals for the entire population. Logically, the gap will be larger for immigrant elders with limited English proficiency (LEP) [30].

This study found a number of differences in the predictors of healthcare use between these two subgroups of Asian immigrant elders; they only have one common predictor contributing to healthcare service use. This reconfirms the importance of taking cultural/ethnic differences into account when planning interventions to increase healthcare service use by Asian immigrant elders. Given the heterogeneity of the two groups, an overall regression model aggregating subgroups may be limited and even misleading.

For Chinese immigrant elders, the lower level of healthcare service use among females confirmed previous findings [31]. The Chinese sample revealed that two predisposing factors (sex and length of stay in the U.S.), two enabling factors (English language proficiency and perceived cultural gap), and two need factors (depression symptoms and medical conditions) were significantly associated with the number of physician visits. Contrary to a previous study’s findings [32], health insurance did not contribute to healthcare use among the Chinese subgroup. Results may support the acculturation hypothesis: a longer duration of residency in the US, which presumably increases acculturation (e.g., English proficiency), was associated with a higher level of service utilization, while the wider culture gap between elders and children may lower healthcare service use among the older people. Interestingly, levels of depressive symptoms were statistically lower among Chinese elders compared to Koreans, yet this variable only influenced the Chinese.

The Korean sample results revealed that a predisposing factor (marital status), an enabling factor (insurance status), and need factors (medical conditions and self-reported health status) were all significantly associated with the number of physician visits. Being unmarried, having health insurance and a lower rate of self-reported health were significant predictors of service use for Korean immigrant elders [33]. We speculate that the absence of a spouse may compel them to seek out professional help more frequently than their married counterparts, who have an informal support system available in their daily lives. In addition, Korean immigrant elders may be hesitant to use healthcare services when they are uninsured. Previous studies have identified Koreans living in the U.S., across gender and age categories, as the least insured Asian ethnic group [34]. This definitely helps to explain the healthcare access disparity; it may be why Korean immigrant elders had a lower number of physician visits when their perceived health was good. When health insurance is not a necessity in life but an extra benefit, visits to a physician without an urgent need, i.e., preventive visits or routine check-ups, may not be part of healthcare use. Nonetheless, the reason(s) why insurance status and health status factors significantly contribute to healthcare use among Korean immigrant elders, but not their Chinese counterparts, needs to be examined further.

Despite important findings on healthcare use among the Chinese and Korean immigrant elders of this study, the results should be interpreted with caution due to some of the study’s limitations. A causal relationship between significant predictors and healthcare service use based on a cross-sectional design survey cannot be generalized to all Chinese and Korean immigrant elders. Given the non-representative sampling method, the findings can only be interpreted within the limitations of the sample characteristics, Asian immigrant elders residing in non-metropolitan areas where ethnic-specific services are not available. Future studies may need to include more culturally relevant enabling factors, such as the availability and accessibility of culturally-competent healthcare providers, and the comfort level of the participants in seeking services when they feel the need.

New Contributions to the Literature

The findings clearly present the importance of paying attention to the special circumstances and needs of Asian immigrant elders living in non-metropolitan areas. To the best of our knowledge, this is the first comparative study to explore and compare factors contributing to healthcare service use among Chinese and Korean immigrant elders living in such an area. The findings suggested the importance of family support among immigrants in terms of health care service use. However, this study makes a further contribution to the literature: the idea that the perception of an intergenerational cultural gap by immigrant elders is associated with health care service use [35]. Community service providers need to recognize the unique characteristics of this population and make a conscious effort to increase healthcare service use among them. For example, it may be critical to communicate with these two different groups in their native languages, or to better understand their cultural backgrounds as they relate to healthcare service utilization. It would be very helpful for healthcare providers to use trained medical interpreters instead of involving family members or lay persons.

The relationships between health insurance and healthcare service use among immigrant elderly groups showed the difference between the two groups. For Chinese immigrant elders, this is consistent with a previous study that had findings from a nationally representative sample [36]. Seeing a strong association between health insurance and healthcare service use among Korean elders is a new finding. Some Korean elders may have to delay receiving medical attention until they are eligible for Medicare because of high medical costs for those without health insurance.