Middle-Aged and Older Adult Health Care Selection

Abstract

This study assesses the prevalence of primary-care physician (PCP) bypass among rural middle-aged and older adults. Bypass is a behavior where people travel beyond local providers to obtain health care. This article applies a precise Geographic Information System (GIS)-based measure of bypass and examines the role of community and non-health-care-related characteristics on bypass. Our results indicate that bypass behavior among rural middle-aged and older adults is multifaceted. In addition to the perceived quality of local primary care, dissatisfaction with local services, such as shopping, creates an effect that increases the likelihood of bypass, whereas strong community ties decrease the likelihood of bypass. The results suggest that the “outshopping theory,” where respondents select services in larger regional economic centers rather than local “mom and pop” providers, now extends to older adult health care selection.

Keywords

health care selection rural older adult health care health care bypass

In contrast to residents living in urban communities, who generally have access to a range of nearby health care options, people living in rural areas often travel substantial distances to obtain primary health care (Bull, Krout, Rathborne-McCuan, & Shreffler, 2001; Goins, Spencer, & Byrd, 2009). Although not all rural residents seek health care in urban areas, many do because of a greater number and, in general, better quality of health care options. Although it can be a burden, regardless of age, traveling long distances for health care becomes particularly problematic for middle-aged and older adults who have increased age-related health issues that require more frequent care (Mattson, 2011; Spafford, Rudman, Leipert, Klinger, & Huot, 2010). As a result, a substantial number of rural middle-aged and older adults travel considerable distances beyond local providers to obtain primary health care outside their rural communities. This behavior is commonly referred to as bypass (Liu, Bellamy, Barnet, & Weng, 2008; Radcliff, Brasure, Moscovice, & Stensland, 2003).

Bypass behavior may not be problematic initially if it is simply an expression of choice in health care. In fact, a primary reason cited by rural patients for bypassing local providers is the perception that it allows them to receive better quality care (Keeler et al., 1992). However, if traveling for health care becomes prohibitive for financial, health, or other reasons, older patients may have to switch providers (Bull et al., 2001; Morton & Weng, 2013). This loss of continuity in primary care is linked to lower overall quality of health care and increased likelihood of hospitalization (Campbell, Roland, & Buetow, 2000; Knight, Dowden, Worrall, Gadag, & Murphy, 2009). Consequently, should the increased costs associated with bypass behavior become too great in the long run, a change in primary-care physician (PCP) can have a significant effect on the health of rural middle-aged and older adults.

Whereas bypass may result in better health care in the short term for those who are healthy and wealthy enough to travel, it can also penalize other rural middle-aged and older adults. Many rural communities face PCP shortages and struggle to retain physicians, in part because low population densities often fail to provide the demand needed to support local clinics (Radcliff et al., 2003). Bypass behavior lowers the demand for rural PCPs further, which threatens the economic viability of local providers. In the extreme, bypass can contribute to the loss of rural PCPs and result in the creation of “health care deserts,” where already vulnerable rural residents must travel outside their communities to receive basic health care (Liu et al., 2008). To better understand bypass behavior of rural middle-aged and older adults, this research proposed an important change in the current understanding of PCP bypass behavior by arguing that, in addition to perceived satisfaction with local PCPs, the quality of other local services, as well as levels of community attachment, influence PCP selection significantly. This new perspective asserts that bypass is multifaceted, and that health care selection is part of a recent shift in rural consumer habits, where rural residents seek to bundle goods and services for increased convenience. In our analysis, we use the 2010 Montana Health Matters (MHM) data to examine how satisfaction with health care and local services affects rural bypass behavior among middle-aged and older adults. Finally, this research is the first to focus solely on the unique health care needs of rural middle-aged and older adults during the period of life when health care consumption increases significantly.

The Social Context of Bypass Among Middle-Aged and Older Adults

The bypass literature developed from the need to explain the observation that rural patients willingly travel significant distances to receive health care outside their communities. However, studies have focused primarily on dissatisfaction with local health care services and have ignored some of the broader social contexts in which bypass occurs (Basu & Mobley, 2010; Escarce & Kapur, 2009; Radcliff et al., 2003). We argue that health care selection should be viewed as part of the shift in consumer shopping habits, where consumers seek to bundle goods and services (Flora et al., 1992). We used outshopping theory to frame health care selection in our study population as part of this shift.

Outshopping theory was developed to understand the effects of big box retailers on rural patterns of consumption. Outshopping occurs when consumers travel beyond local providers to purchase non-local retail goods and services even when they are available locally (R. B. Brown, Hudspeth, & Odem, 1996). For example, by offering both dry goods and groceries, the added convenience of big box stores affects the economic viability of “mom and pop” retail and grocery stores in surrounding communities (Artz & Stone, 2006; Blanchard & Lyson, 2002). Therefore, according to outshopping theory, if consumers are unhappy with one type of local service, the likelihood of outshopping for all their needs increases, even if they are satisfied with other local services (Arnold & Luthra, 2000; Bebko, 2000; Carpenter & Moore, 2009; Lee, Johnson, & Gahring, 2008; Martens, Dooley, & Florax, 2010). Consequently, if rural older adults view primary health care as a consumer service, they may bypass local PCPs and choose to use non-local health care located where they engage in other consumer activities. Outshopping theory also argues that bypassing local goods and services is less likely when people feel a strong attachment to their communities. Strong levels of community attachment can produce a desire to “buy local,” even when non-local options are perceived to be more convenient and satisfactory. Therefore, health care selection is also a social process in which those with strong community attachments are less likely to bypass local services (R. B. Brown, 1993; R. B. Brown et al., 1996).

In summary, we hypothesize that dissatisfaction with local health care and shopping can push rural middle-aged and older adults to bypass local PCPs, whereas strong community ties create a pull on these people to shop locally, thereby counteracting the push of outshopping. Therefore, the bypass of local health care among rural middle-aged and older adults is multifaceted and requires a more detailed analysis of how satisfied they are with local community services as well as their level of community attachment. Based on the outshopping theory, we expect that the level of dissatisfaction with local PCPs and shopping will increase the likelihood of bypass behavior significantly, whereas strong community attachment will decrease the probability of bypass. This application of outshopping theory provides important new insights into how rural middle-aged and older adults select their PCPs.

Method

Data

Data for this analysis came from the 2010 MHM study. MHM gathered self-reported information from Montana residents on their health care use, satisfaction with health care services, health care access issues, and current care-giving arrangements. Measures of family, community, and demographic characteristics provided the social context for the health care experiences reported. The sample was drawn using the U.S. Postal Service’s (USPS’s) computerized Delivery Sequence File (DSF), which contains all known addresses in Montana. Given the address-based sample design, a multi-method, five-wave mail/telephone survey protocol with a US$2 honorarium was used to maximize response to the survey (Dillman, Smyth, & Christian, 2009). All householders were asked to respond. The multi-method design resulted in 3,512 respondents, 3,019 in rural¹, and 493 in urban communities. The response rate for the study was 52% (rural = 54%, urban = 47%). This rate includes an adjustment for 1,200 original DSF addresses that could not be linked to a householder name, and non-residential addresses that were identified using GoogleEarth® and certified mailings (Call, Erickson, & Yorgason Jeremy, 2013). The weighting scheme developed for these data took into account ineligible addresses, the multi-stage cluster sampling design, and survey non-response, making the final sample representative of the population of Montana. Because this study focused on rural, middle-aged and older adults, only respondents residing in a rural community who were age 55 and above were included in the analyses.² We excluded respondents who were enrolled in the Veteran Health Administration (VHA) system because they are motivated financially to use VHA PCP options and most rural veterans travel outside their community to access VHA facilities. The final sample size was 1,588.

Measures

Bypass

The operationalization of bypass in previous research varies. For example, bypass has been identified as seeking health care outside of one’s ZIP code (Buczko, 1994), the straight-line distance from ZIP code of residence to ZIP code of care receipt (Bronstein & Morrisey, 1991), or the distance between the center of the ZIP code of residence and the center of the ZIP code of care receipt (Radcliff et al., 2003). While these studies have provided important insights into bypass, the use of ZIP code–based data is problematic because the USPS assigns ZIP codes using mail delivery patterns, which are not representative of true communities and change from time to time (U.S. Census Bureau, 2013). Other researchers have defined bypass as rural patients using hospitals that are 15 or more miles away from their residence when there is a closer facility (Liu et al., 2008).

To develop an improved measure of bypass, all PCPs in Montana were identified using North American Industry Classification System Codes in data from Esri business analysis and the Centers for Medicare and Medicaid Provider of Service files (CMS, 2010; Esri, 2013). We geo-referenced, or mapped, the exact location of each respondent, his or her PCP, all other PCPs, and health clinics (clinics without an MD on staff were excluded). Respondents were coded as 1 for bypassing if they met the following two criteria: (a) traveled more than 15 road miles to receive primary health care, and (b) another PCP that the respondent did not use was located less than 15 road miles away. The 15-mile boundary was used to be consistent with previous research (Liu et al., 2008; Radcliff et al., 2003). Furthermore, to ensure that our findings were not dependent on the 15 road mile definition, a sensitivity test was also conducted using a 20 road mile definition of the bypass.

Community push

The community push characteristics measured dissatisfaction with local health care and shopping. Both were measured on a 1 to 7 Likert-type scale with 1 being “exceptional” and 7 “badly needs improvement.”

Community pull

Respondents reported the number of close friends they had who lived in the same community with an open-ended item and then the responses were divided by 10. The community fit variable was based on a question that asked respondents how well they felt they fit into their community, and was reported on a scale from 1 (poorly) to 7 (well). Similarly, the “in common with community” variable asked respondents on a scale from 1 (nothing) to 7 (everything), “how much [you] have in common with most people in your community.”

Controls

For the commuting flow variable, respondents were coded as 1 if they were located in a census tract with a primary commuting flow of 30% or greater to a metropolitan, micropolitan, or small town core. Commuting flows to larger regional economic centers were based on Rural–Urban Continuum Area (RUCA) Codes (Economic Research Service, 2013). Next, prior research has demonstrated that rural residents are often attracted to health care in larger communities with more health care options, although the draw is not as strong for residents as the distance between them increases (Guagliardo, 2004). Therefore, a variable known as a gravity score was included to account for attraction. The community PCP gravity score was the number of PCPs in the resident ZIP code multiplied by the number of PCPs in the ZIP code where they received care, divided by the square of the distance traveled for health care. Because no respondents reported receiving health care in a ZIP code with a smaller number of PCPs than the resident ZIP code, gravity scores decreased with distance and higher scores indicated a pull by large groupings of non-local PCPs (Guagliardo, 2004; Luo & Wang, 2003).

Demographic/household variables included age (in years), sex (female = 1), marital status (married = 1), level of education (high school diploma or less = 1, some college = 2, and college degree = 3), and Internet access (1 = any type of home access). Because of the significant demographic trends of rural retirement-age in-migration throughout Montana, a dummy variable that controlled for being a new in-migrant was included (Winkler et al., 2013). A respondent was coded 1 if they moved into a new community in the past 10 years and 0 otherwise. Income was the total annual household income measured in units of US$1,000.

Self-reported health was measured on a scale from 1 to 5, with 1 being “poor health” and 5 “excellent health.” For self-reliant travel, respondents who drove themselves to their PCP were coded 0, those who relied on someone they lived with were coded as 1, those dependent on someone who did not live with them were coded as 2, and those who used other forms of transportation, that is, public, community organizations, and so on, were coded as 3. The “location of health care specialist” variable was coded 0 if the respondent saw a specialist in a different ZIP code than his or her PCP, 1 if the specialist was in the same ZIP code, and 2 if she or he did not have a specialist. Health insurance was coded 1 if respondents had Medicare or Medicaid, 2 if they had employer/union or direct purchase health plans, 3 for other types of coverage (e.g., Indian Health Services and federal retired employee health insurances, etc.), and 4 for no health insurance.

Analysis

Missing data were addressed using multiple imputations with chained equations. This regression-based procedure creates multiple copies of the data with different plausible estimates for missing values. Using the chained-equations approach, multiple imputation takes into account the different distributions of variables that need to be imputed. Using the mi impute command in Stata 13, we estimated 50 imputed data sets. We used imputed data sets separated by 100 iterations because graphical diagnostics from preliminary imputations suggested that the imputation model converged well before this point. The imputation model included all the variables used in the logistic regression analysis below. After creating the completed data sets, analyses were performed separately on each imputed data set and results for each data set were combined using Rubin’s rules via Stata’s mi estimate command.

Descriptive statistics are presented first to compare general characteristics of rural middle-aged and older adults who bypassed with those who did not. Next, chi-square tests were used to examine the relationship between middle-aged and older adult PCP bypass and perceived satisfaction with local health care and services. Then, multivariate logistic regression models were estimated to determine which middle-aged and older adults were more likely to bypass their local PCPs. To explore the role of outshopping theory in understanding bypass fully, we estimated an additional model with interactions between the main explanation for bypass in past literature, dissatisfaction with local health care, and the other significant push and pull variables. Those interactions will illustrate whether or not combined dissatisfaction with local health care and shopping amplified the likelihood of bypass, and if the pull of community attachment variables mitigated the probability of bypass when middle-aged and older adults were dissatisfied with local health care.

Results

The descriptive statistics presented in Table 1 compare bypassing and non-bypassing respondents. The results indicate that the averages of the two groups were similar. However, middle-aged and older adults who bypassed were in better health. This difference in self-reported health is consistent with previous research on bypass that found health to be a predictor of bypass (Liu et al., 2008; Radcliff et al., 2003). The greater number of friends among non-bypass middle-aged and older adults was also consistent with outshopping theory and the pull associated with community attachment. Finally, the difference in distance traveled to the PCP revealed that rural middle-aged and older adults who bypass travel, on average, 33.1 miles more each way than do rural middle-aged and older adults who use local PCPs.

Table 1.

Descriptive Characteristics of Bypassing and Non-Bypassing Rural Middle-Aged and Older Adults.

	Bypass		Non-bypass
	M	SD	M	SD
Age	67.4	8.7	67.6	8.3
Self-reported health	3.8	1.1	3.3	1.0
Number of friends	10.7	10.5	14.6	14.2
Distance traveled to PCP	43.2	53.8	10.1	16.8
n	455		1,133

Source. Montana Health Matters, 2010.

Note. PCP = primary-care physician.

Overall, 28.7% of rural middle-aged and older adults in the MHM data bypassed local PCPs. The results in Table 2 illustrate the utility of outshopping theory in understanding bypass. We found that bypass was more likely with increased dissatisfaction with local health care (a push factor): About 74.4% of those most dissatisfied with their local health care bypassed local PCPs, as opposed to 12.1% who were least dissatisfied with local health care and bypassed (p < .001). The same pattern existed for dissatisfaction with local shopping, although the relationship was not quite as strong.

Table 2.

Outshopping Community Characteristics of Rural Middle-Aged and Older Adult Bypass (N = 1,588).

	Number of respondents	Percent bypass	p value
Community push characteristics
Dissatisfaction with local health care
1 = satisfied	151	12.1
2	311	16.7
3	332	20.1
4	293	22
5	176	25.2
6	152	65.7
7 = dissatisfied	173	74.8	<.001
Dissatisfaction with local shopping
1 =satisfied	106	18.8
2	148	24.5
3	219	39.4
4	310	38.2
5	278	44.8
6	248	44
7 = dissatisfied	279	47.2	<.001
Community pull characteristics
Number of friends in community (by quartiles)
0-4	410	42.3
5-10	386	30.4
11-18	392	22.6
19+	400	15.6	.028
Community fit
7 = strong fit	341	21.6
6	345	26.5
5	359	29.3
4	311	41.2
3	123	41.2
2	59	47.2
1 = weak fit	50	52.8	.008
In common with community
7 = everything in common	108	22.6
6	303	26.5
5	440	29.2
4	405	41.2
3	175	42.2
2	95	47.2
1 = nothing in common	62	51.9	.008

Source. Montana Health Matters, 2010.

Overall, the community pull variables were also related to bypass. Rural middle-aged and older adults reported lower levels of bypass as the number of friends they had in their community increased. Furthermore, only 21.6% of those with a strong community fit reported bypassing local health care, which is considerably lower than the 52.8% of those with a weak fit who bypassed (p = .008). Finally, bypass became less common as respondents reported having increasing levels of commonalities with other community members: About 22.6% of middle-aged and older adults with higher levels of commonality with their community bypassed, by comparison with 51.9% with low levels of commonality with community residents who bypassed (p = .008). Again, these significant relationships are consistent with outshopping theory and suggest that push factors increase bypass, whereas pull factors reduce bypass behavior.

In Table 3, we report the determinants of older adult bypass behavior in Montana from logistic regression models. The results of Model 1 indicated that dissatisfaction with local health care was related significantly to bypass in the study population and an additional unit of dissatisfaction increased the odds of bypass by a factor of 1.71. The odds ratio for dissatisfaction with local shopping (1.32) was also significant, indicating that the push of being dissatisfied with local health care and shopping increased the odds of bypass. These results suggest that dissatisfaction with local services beyond health care help push middle-aged and older adults into seeking health care services outside of their community.

Table 3.

Logistic Regression Analysis Predicting Middle-Aged and Older Adult Rural Bypass Behavior (N = 1,588).

	Model 1		Model 2 interactions
	Odds ratio	95% confidence interval	Odds ratio	95% confidence interval
Community push characteristics
Dissatisfaction with local health care	1.71***	[1.65, 1.88]	0.75	[0.41, 1.49]
Dissatisfaction with local shopping	1.32***	[1.27, 1.37]	0.87	[0.68, 1.11]
Community pull characteristics
Number of friends in community	0.94**	[0.89, 0.95]	0.99	[0.95, 1.03]
Community fit	0.86*	[0.76, 0.93]	0.91	[0.72, 1.24]
In common with community	0.84*	[0.76, 0.91]	0.64	[0.47, 0.98]
Controls
Commute flow >30%	2.36***	[1.77, 3.07]	2.29***	[1.54, 3.48]
Community PCP gravity score	1.01	[0.99, 1.07]
Age
55-64	—	—	—	—
65-74	1.09	[0.83, 1.51]	0.99	[0.76, 1.74]
75+	1.10	[0.78, 1.52]	0.85	[0.56, 1.37]
Male	0.97	[0.73, 1.32]	0.79	[0.49, 1.17]
Married	1.07	[0.73, 1.62]	1.02	[0.75, 1.75]
Recent in-migrant	1.41***	[1.13, 1.92]	1.43***	[1.08, 1.91]
Education
High school or less	—	—	—	—
Some college	1.10	[0.88, 1.53]	1.28	[0.89, 2.01]
College or more	1.15	[0.74, 1.29]	1.05	[0.69, 1.72]
Income (in US$1,000)	1.08*	[1.03, 1.15]	1.05*	[1.02, 1.18]
Internet access (no Internet access ref.)	0.80	[0.58, 1.19]	0.72	[0.48, 1.09]
Self-reported health	1.10*	[1.06, 1.28]	1.09*	[1.02, 1.32]
Self-reliant travel
Self	—	—	—	—
Relative/someone living in home	0.94	[0.34, 1.87]	0.95	[0.54, 2.08]
Relative/someone not living in home	0.30**	[0.11, 0.83]	0.51*	[0.13, 0.96]
Other	0.47	[0.09, 2.21]	0.73	[0.44, 3.15]
Location of health care specialist
Different ZIP code	—	—	—	—
Same ZIP code	1.14*	[1.04, 1.37]	1.10*	[1.03, 1.25]
No specialist	0.99	[0.87, 1.09]	0.99	[0.79, 1.16]
Health insurance
Medicare/Medicaid	—	—	—	—
Employee/purchased insurance	1.78	[0.64, 1.98]	1.19	[0.63, 1.34]
Other	1.41	[0.79, 2.28]	1.45	[0.79, 2.66]
No insurance	0.77	[0.47, 1.38]	0.89	[0.77, 1.82]
Dissatisfaction with health care × Dissatisfaction with shopping			1.25***	[1.10, 1.38]
Dissatisfaction health care × Number of friends			0.99	[0.98, 1.05]
Dissatisfaction with health care × Fit			0.89	[0.80, 1.09]
Dissatisfaction with health care × In common			0.86*	[0.74, 0.95]

Source. Montana Health Matters, 2010.

Note. PCP = primary-care physician.

p < .05. **p < .01. ***p < .001.

The results of Model 1 also showed that all three community pull variables were significant: The number of friends in the community, community fit, and how much the respondents had in common with their communities all decreased the odds of bypass behavior significantly. Thus, community attachment creates a strong pull for middle-aged and older adults to use local primary care.

Model 1 also revealed that, among the control variables, respondents living in communities with high commuter flows to surrounding areas were 2.36 times more likely to bypass local PCPs than were rural residents living in communities with low commuting flows. Income was a significant and positive indicator of bypass. Being an in-migrant within the past 10 years was also related significantly and positively to bypass, indicating that recent migrants were 1.41 times more likely to bypass local health care. Self-reported health was positive and significant (odds ratio [OR] = 1.10), indicating that healthier respondents were also more likely to exhibit bypass behavior. Furthermore, middle-aged and older adults dependent on a relative or friend not living with them were significantly less likely to bypass their local PCPs by comparison with respondents able to drive themselves to their PCPs. Furthermore, having a specialist in the same ZIP code as a respondent’s PCP was significant and positive (OR = 1.14). Other control variables did not have significant effects.

To investigate the relationship between bypass behavior and outshopping more completely, Model 2 included an interaction between the main explanations for bypass in past literature—dissatisfaction with local health care (push factor), and dissatisfaction with local shopping (push factor). This push–push interaction between dissatisfaction with local health care and shopping was positive, with an odds ratio of 1.25 that indicated that the odds of bypass increased disproportionately as both push factors became greater. Figure 1 plots predicted probabilities from the model when middle-aged and older adults’ dissatisfaction with local health care was either low or high. In the graph, one indicates low dissatisfaction with local health care, whereas seven designates high dissatisfaction. The graph indicates that the relationship between dissatisfaction with health care and the likelihood of bypass was not as strong when people were satisfied with local shopping than when they were dissatisfied with it. Thus, as the level of dissatisfaction with local health care increased, dissatisfaction with shopping exerted an additional push effect that increased the odds of bypass.

Figure 1.

Interaction of dissatisfaction with local health care and dissatisfaction with local shopping.

Model 2 also included an interaction between dissatisfaction with local health care (push factor), and number of friends, community fit, and “in common with community” variables (pull factors). Unlike the push–push interactions, the interactions between dissatisfaction with local health care and number of friends and community fit were not significant. However, the interaction between dissatisfaction with local health care and “in common with community” (OR = 0.86) was significant and negative. Figure 2 graphs the predicted probabilities of bypass for middle-aged and older adults with everything and nothing in common with their community across the range of dissatisfaction with local health care. The graph reveals that those with everything in common with their community maintained a low and consistent probability of bypass, even as the level of dissatisfaction with local health care increased from low dissatisfaction (one) to high dissatisfaction (seven). However, for respondents with nothing in common with their community, the likelihood of bypass increased significantly as dissatisfaction with local health care. The results of this interaction indicate that among rural middle-aged and older adults, high levels of community attachment, or community pull, can help to overcome the push effect associated with higher levels of dissatisfaction with local health care.

Figure 2.

Interaction of dissatisfaction with local health care and in common with community.

Discussion

The purpose of this research was to explore whether outshopping theory, and the push–pull dynamic of community services and attachment, provides additional understanding of bypass behavior among rural middle-aged and older adults. The results indicated that rural PCP bypass is one facet of the larger consumption patterns of middle-aged and older adults living in rural communities. Like previous studies, our findings confirmed that the quality of local health care is important in bypass (Bronstein & Morrisey, 1991; Escarce & Kapur, 2009; Liu et al., 2008; Radcliff et al., 2003). However, our findings also indicated that dissatisfaction with other local services increased the likelihood of health care bypass, which suggests that even when a quality local PCP is present, bypass may occur because of poor local shopping or the convenience of bundling health care with non-local shopping. In addition, the significance of the community pull variables, together with the interaction between dissatisfaction with local health care and the variable of commonality with community, supported outshopping theory, which all suggest that the attachment to one’s community influences PCP selection. These findings represent a significant shift in the current understanding of rural older adult PCP selection and demonstrate that bypass behavior is multifaceted. Furthermore, a sensitivity test using a 20 road mile distance threshold to define bypass was performed. In this model, all of the community push and pull variables remained significant, suggesting that our findings are not conditional on a single distance threshold used to define bypass behavior.

The significance of the health care specialist being in the same ZIP code as a respondent’s PCP also indicated that convenience is an important factor in health care selection. Based on the MHM data, respondents who bypassed their local PCPs and had specialists located in the same ZIP code as their PCPs, traveled, on average, 27.3 miles farther each way than did respondents with specialists in the same ZIP code who did not bypass their local PCPs. As a result, rural middle-aged and older adults who need specialist care may seek to bundle services for increased convenience. Moreover, the increased reliance on electronic medical records and patients’ desire to have their records accessible by all of their health care providers indicates that the link between PCPs and health care specialists will continue to be an important factor in older adult health care selection. Although many articles have indicated that PCPs refer patients to specialists (Forrest, 2003; Forrest, Nutting, Starfield, & Von Schrader, 2002; Mehrotra, Forrest, & Lin, 2001), there are some situations where specialists refer patients to PCPs (Barnett, Keating, Christakis, O’Malley, & Landon, 2012). As a result, additional research is needed to understand more fully how location and referrals between PCPs and health care specialists affect health care selection.

The findings that respondents with higher income and better health overall were more likely to bypass suggest that middle-aged and older adults with poorer health and lower incomes potentially have barriers in finances and physical well-being that restrict their access to non-local PCPs. In addition, bypass behavior by the more advantaged middle-aged and older adults in the community may threaten the sustainability of local PCPs. The loss of local PCPs would then jeopardize health care access for middle-aged and older adults who may be too sick or who lack the financial resources needed to travel or move closer to non-local PCPs (D. L. Brown, Bolender, Kulcsar, Glasgow, & Sanders, 2011; Bull et al., 2001). Furthermore, the negative and significant result of those middle-aged and older adults dependent on relatives/friends who do not live with them further highlights the vulnerability of rural middle-aged and older adults. Whereas some middle-aged and older adults may have friends and family nearby who help them obtain health care, others may not. Those dependent on people outside their household for transportation have additional barriers that may inhibit them from receiving basic health care.

Although some bypass could occur because patients are seeking providers who accept their insurance programs, having health insurance, and the type of insurance, did not affect the likelihood of PCP bypass significantly. However, the MHM data are unable to identify a direct link between bypass and the availability of a local PCP that accepts a particular insurance program. It may be that some middle-aged and older adults bypass local PCPs because they are not contracted with their insurance provider, but the non-significance of the health care coverage variable in this, and other bypass research, suggests that this type of bypass behavior is too infrequent to be a significant predictor of rural bypass behavior (Bronstein & Morrisey, 1991; Escarce & Kapur, 2009; Liu et al., 2008).

Another interesting result was the significance of the migrant variable. In recent years, retirement-fueled migration has been an important demographic and community development trend in predominantly rural states, such as Montana. In fact, migrants aged 60 and above represented one of the largest gains in net migration rates in Montana between 2000 and 2010 (Johnson, Winkler, & Rogers, 2013). While some communities have courted the “grey gold” of rural retirement migrants, the results of Model 1 showed that older adult migrants were significantly more likely to bypass a local PCP. This suggests that recent migrants to rural communities may not bring the anticipated economic boon to their new communities. Future research is needed to understand more fully how bypass behavior varies between middle-aged and older adults, and older adult migrants.

The non-significance of the Internet variable is consistent with previous research that has examined rural use of the Internet. Although the Internet is often seen as a technology that overcomes the effects of distance, current patterns in online shopping have not altered the shopping habits of rural residents significantly, particularly for groceries and among middle-aged and older adults (Lennon et al., 2007; Lichter & Brown, 2011) Whereas telemedicine holds great promise in bringing health care to rural middle-aged and older adults, adoption rates have also not met expectations. Technology, regulations, physician buy-in, and patient acceptance all factor into the low use of telemedicine (Call, Erickson, Dailey, et al., 2013; Morgan et al., 2011). Although online shopping and telemedicine can potentially improve access to shopping and health care for rural middle-aged and older adults, current patterns suggest that they will not have a significant effect in the immediate future.

Several limitations need to be considered when interpreting the results of this study. First, demographically, Montana is almost 90% Caucasian (U.S. Census Bureau, 2012). As a result, the MHM and this research were unable to examine how bypass behavior varies between ethnicities. Differences across racial groups may exist in rural older adult bypass behavior. However, the current body of bypass research either cannot examine racial groups, due to similar data limitations, or has found ethnicity to be a non-significant variable (Buczko, 1994; Escarce & Kapur, 2009; Liu et al., 2008; Radcliff et al., 2003; Xu & Borders, 2003). Furthermore, because of the geographic size and relatively sparse population distribution in Montana, the rate of bypass may be even higher for rural middle-aged and older adults in smaller states. Rural residents in states with denser populations potentially have multiple relatively close health care options that can result in increased rates and clustering of bypass behavior. Finally, our measures only allowed us to connect subjective evaluations of local services and not actual shopping behavior. Consequently, we were unable to show directly that middle-aged and older adults combine health care and shopping in a single trip out of their communities. However, research on rural consumer behavior has indicated that many local, rural businesses are being bypassed for the greater selection and convenience found in larger regional shopping areas (Artz & Stone, 2006; Lichter & Brown, 2011). Furthermore, the dramatic increase in retail health clinics located inside big box stores suggests that the health care industry is aware of the significant connection between shopping and health care selection (Bohmer, 2007; Thygeson, Van Vorst, Maciosek, & Solberg, 2008; Wang, Ryan, McGlynn, & Mehrotra, 2010). Consequently, there is some evidence that the subjective evaluations of local shopping quality are connected with real shopping options and therefore, actual shopping behavior.

Finally, the results of this study can help to inform the changes that will come to rural communities through the Affordable Care Act (ACA). In particular, US$1.5 billion over 5 years has been set aside to improve rural health care access through the National Health Service Corps’ (NHSC) loan repayment program (White House, 2013). This program will pay off the loans of recently graduated health care professionals who work for 2 to 3 years in rural clinics. Although this funding is designed to increase health care options for rural patients, the significance of the shopping and community attachment variables suggests that PCP selection is not strictly based on accessibility. The new PCPs may also be bypassed for those near better shopping. Furthermore, the tenure of these appointments may not be long enough to develop a significant community pull or attachment to local PCPs. Therefore, continued research is needed to understand how the ACA and NHSC loan repayment program will affect PCP selection among rural older adults.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Scott R. Sanders, MPA, PhD, is an assistant professor at Brigham Young University and the acting executive director/treasurer of the Rural Sociological Society. His research focuses on rural health care, poverty, and globalization in Southeast Asia.

Lance D. Erickson, PhD, is an associate professor of sociology at Brigham Young University. He has two main research interests—the intersection of health, aging, and rural communities and the transition to adulthood.

Vaughn R. A. Call is a professor of sociology and the director of the Gerontology Program at Brigham Young University. His current research focuses on aging across the life course, health care access in rural communities, health issues facing older veterans, and longitudinal research methodology.

Matthew L. McKnight is a master’s student at Brigham Young University. His research interests focused on community development, health, and spatial methods.

References

Arnold

S. J.

Luthra

M. N.

(2000). Market entry effects of large format retailers: A stakeholder analysis. International Journal of Retail & Distribution Management, 28, 139-154.

Artz

G. M.

Stone

K. E.

(2006). Analyzing the impact of Wal-Mart supercenters on local food store sales. American Journal of Agricultural Economics, 88, 1296-1303.

Barnett

M. L.

Keating

N. L.

Christakis

N. A.

O’Malley

A. J.

Landon

B. E.

(2012). Reasons for choice of referral physician among primary care and specialist physicians. Journal of General Internal Medicine, 27, 506-512.

Basu

Mobley

L. R.

(2010). Impact of local resources on hospitalization patterns of Medicare beneficiaries and propensity to travel outside local markets. Journal of Rural Health, 26, 20-29. doi:10.1111/j.1748-0361.2009.00261.x

Bebko

C. P.

(2000). Service intangibility and its impact on consumer expectations of service quality. Journal of Services Marketing, 14, 9-26.

Blanchard

Lyson

(2002, November). Access to low cost groceries in nonmetropolitan counties: Large retailers and the creation of food deserts. Paper presented at the Measuring Rural Diversity Conference.

Bohmer

(2007). The rise of in-store clinics—Threat or opportunity? New England Journal of Medicine, 356, 765-768. doi:10.1056/NEJMp068289

Bronstein

J. M.

Morrisey

M. A.

(1991). Bypassing rural hospitals for obstetrics care. Journal of Health Politics, Policy and Law, 16, 87-118.

Brown

D. L.

Bolender

B. C.

Kulcsar

Glasgow

Sanders

S. R.

(2011). Intercounty variability of net migration at older ages as a path-dependent process. Rural Sociology, 76, 44-73.

10.

Brown

R. B.

(1993). Rural community satisfaction and attachment in mass consumer society. Rural Sociology, 58, 387-403.

11.

Brown

R. B.

Hudspeth

C. D.

Odem

J. S.

(1996). Outshopping and the viability of rural communities as service/trade centers. Journal of the Community Development Society, 27, 90-112.

12.

Buczko

(1994). Bypassing of local hospitals by rural medicare beneficiaries. The Journal of Rural Health, 10, 237-246. doi:10.1111/j.1748-0361.1994.tb00237.x

13.

Bull

C. N.

Krout

J. A.

Rathborne-McCuan

Shreffler

M. J.

(2001). Access and issues of equity in remote/rural areas. Journal of Rural Health, 17, 356-359.

14.

Call

V. R.

Erickson

L. D.

Dailey

N. K.

Hicken

B. L.

Rupper

Yorgason

J. B.

Bair

(2015). Attitudes toward telemedicine in urban, rural, and highly rural communities. Telemedicine and e-Health. Advance online Publication. doi:10.1089/tmj.2014.0125

15.

Call

V. R. A.

Erickson

L. D.

Yorgason Jeremy

(2013, November). Older adults living in rural areas: Reducing coverage errors in the 2010 Montana Health Matters Study. Paper presented at the Annual Meeting of the Gerontological Society of America, New Orleans, LA.

16.

Campbell

Roland

Buetow

(2000). Defining quality of care. Social Science & Medicine, 51, 1611-1625.

17.

Carpenter

J. M.

Moore

(2009). Consumer demographics, retail attributes and cross-shopping behavior. Journal of Textile and Apparel, Technology and Management, 6, 1-13.

18.

Centers for Medicare & Medicaid Services (CMS). (2010). Provider of services public use file. Retrieved from https://www.cms.gov/Research-Statistics-Data-and-Systems/Downloadable-Public-Use-Files/Provider-of-Services/POS2010.html

19.

Dillman

D. A.

Smyth

J. D.

Christian

L. M.

(2009). Internet, mail and mixed-mode surveys: The tailored design method. New York, NY: John Wiley.

20.

Economic Research Service. (2013). Rural-urban commuting area codes. Retrieved from http://www.ers.usda.gov/data-products/rural-urban-commuting-area-codes.aspx

21.

Escarce

J. J.

Kapur

(2009). Do patients bypass rural hospitals?: Determinants of inpatient hospital choice in rural California. Journal of Healthcare for the Poor and Underserved, 20, 625-644.

22.

Esri (2013). Methodology statement: 2013 Esri US business locations and business summary data. Redlands, CA: Author.

23.

Flora

C. B.

Flora

J. L.

Spears

J. D.

Swanson

L. E.

Lapping

M. B.

Weinberg

M. L.

(1992). Rural communities: Legacies and change. Boulder, CO: Westview Press.

24.

Forrest

C. B.

(2003). Primary care gatekeeping and referrals: Effective filter or failed experiment? British Medical Journal, 326, 692-695.

25.

Forrest

C. B.

Nutting

P. A.

Starfield

Von Schrader

(2002). Family physicians’ referral decisions. Journal of Family Practice, 51, 215-222.

26.

Goins

R. T.

Spencer

S. M.

Byrd

J. C.

(2009). Research on rural caregiving: A literature review. Journal of Applied Gerontology, 28, 139-170.

27.

Guagliardo

M. F.

(2004). Spatial accessibility of primary care: Concepts, methods and challenges. International Journal of Health Geography, 3, 3.

28.

Johnson

K. M.

Winkler

Rogers

L. T.

(2013). Age and lifecycle patterns driving U.S. migration shifts. Durham: Carsey Institute, University of New Hampshire.

29.

Keeler

E. B.

Rubenstein

L. V.

Kahn

K. L.

Draper

Harrison

E. R.

McGinty

M. J.

Brook

R. H.

(1992). Hospital characteristics and quality of care. JAMA, 268, 1709-1714.

30.

Knight

J. C.

Dowden

J. J.

Worrall

G. J.

Gadag

Murphy

M. M.

(2009). Does higher continuity of family physician care reduce hospitalizations in elderly people with diabetes? Population Health Management, 12, 81-86.

31.

Lee

S.-E.

Johnson

K. K. P.

Gahring

S. A.

(2008). Small-town consumers’ disconfirmation of expectations and satisfaction with local independent retailers. International Journal of Retail & Distribution Management, 36, 143-157. doi:10.1108/09590550810853084

32.

Lennon

S. J.

Kim

Johnson

K. K. P.

Jolly

L. D.

Damhorst

M. L.

Jasper

C. R.

(2007). A longitudinal look at rural consumer adoption of online shopping. Psychology & Marketing, 24, 375-401.

33.

Lichter

D. T.

Brown

D. L.

(2011). Rural America in an urban society: Changing spatial and social boundaries. Annual Review of Sociology, 37, 565-592.

34.

Liu

J. J. X.

Bellamy

Barnet

Weng

(2008). Bypass of local primary care in rural counties: Effect of patient and community characteristics. Annals of Family Medicine, 6, 124-130. doi:10.1370/Afm.794

35.

Luo

Wang

(2003). Measures of spatial accessibility to health care in a GIS environment: Synthesis and a case study in the Chicago region. Environment and Planning B, 30, 865-884.

36.

Martens

B. J.

Dooley

F. J.

Florax

R. J. G. M.

(2010). A spatial analysis of the effect of entry by Supercenter and Warehouse Club Retailers on grocery sales. Journal of Food Distribution Research, 41, 46-57.

37.

Mattson

J. W.

(2011). Aging and mobility in rural and small urban areas: A survey of North Dakota. Journal of Applied Gerontology, 30, 700-718.

38.

Mehrotra

Forrest

C. B.

Lin

C. Y.

(2001). Dropping the baton: Specialty referrals in the United States. Milbank Quarterly, 89, 39-68.

39.

Morgan

D. G.

Crossley

Kirk

McBain

Stewart

N. J.

D’Arcy

. . . Basran

(2011). Evaluation of telehealth for preclinic assessment and follow-up in an interprofessional rural and remote memory clinic. Journal of Applied Gerontology, 30, 304-331.

40.

Morton

L. W.

Weng

C.-Y.

(2013). Health and healthcare among the rural aging. In Glasgow

Berry

E. H.

E. J. V.

(Eds.), Rural aging in 21st century America (pp. 179-194). New York, NY: Springer.

41.

Radcliff

T. A.

Brasure

Moscovice

I. S.

Stensland

J. T.

(2003). Understanding rural hospital bypass behavior. Journal of Rural Health, 19, 252-259. doi:10.1111/j.1748-0361.2003.tb00571.x

42.

Spafford

M. M.

Rudman

D. L.

Leipert

B. D.

Klinger

Huot

(2010). When self-presentation trumps access: Why older adults with low vision go without low-vision services. Journal of Applied Gerontology, 29, 579-602.

43.

Thygeson

Van Vorst

K. A.

Maciosek

M. V.

Solberg

(2008). Use and costs of care in retail clinics versus traditional care sites. Health Affairs, 27, 1283-1292. doi:10.1377/hlthaff.27.5.1283

44.

U.S. Census Bureau. (2012). US Census Montana State and County QuickFacts. Retrieved from http://quickfacts.census.gov/qfd/states/30000.html

45.

U.S. Census Bureau. (2013). ZIP code statistics. Retrieved from http://www.census.gov/epcd/www/zipstats.html

46.

Wang

M. C.

Ryan

McGlynn

E. A.

Mehrotra

(2010). Why do patients seek care at retail clinics, and what alternatives did they consider? American Journal of Medical Quality, 25, 128-134. doi:10.1177/1062860609353201

47.

White House. (2013). Health reform for rural Americans: The Affordable Care Act gives rural Americans greater control over their own health care. Retrieved from https://www.whitehouse.gov/files/documents/health_reform_for_rural_americans.pdf

48.

Winkler

Johnson

K. M.

Cheng

Beaudoin

Voss

Curtis

(2013). Age-specific net migration estimates for US counties, 1950-2010. Retrieved from http://www.netmigration.wisc.edu/

49.

T. K.

Borders

T. F.

(2003). Characteristics of rural elderly people who bypass local pharmacies. Journal of Rural Health, 19, 156-163.