Rural–urban differences in the location of influenza vaccine administration

Vaccine 29 (2011) 5970–5977

Contents lists available at ScienceDirect

Vaccine journal homepage: www.elsevier.com/locate/vaccine

Rural–urban differences in the location of influenza vaccine administration Kevin J. Bennett a,b,∗ , Chaiporn Pumkam a , Janice C. Probst b a b

University of South Carolina, School of Medicine, Department of Family & Preventive Medicine, 3209 Colonial Drive, Columbia, SC 29203, USA University of South Carolina, Arnold School of Public Health, South Carolina Rural Health Research Center, 220 Stoneridge Drive, Suite 204, Columbia, SC 29210, USA

a r t i c l e

i n f o

Article history: Received 16 March 2011 Received in revised form 8 June 2011 Accepted 13 June 2011 Available online 25 June 2011 Keywords: Influenza Vaccination Rural Disparities

a b s t r a c t Background: Influenza vaccination rates remain lower than Healthy People 2010 goals. The lower rates are prevalent in rural areas despite an expansion of services to nontraditional settings. Little is known about disparities in influenza vaccination rates and location of receipt among rural residents. This study seeks to determine if rural residents differ from urban residents in where they obtain an influenza vaccination, and to determine what factors contribute to these differences. Methods: Data from 2002–2005 BRFSS were used and combined with the 2006 Area Resource File (analytic n = 70,468, unweighted, 48,392,455 weighted). Unadjusted analyses examined the proportions of influenza vaccinations obtained in traditional clinical settings vs. others across rurality: Urban, Large Rural and Small Rural. Multivariable logistic regression models were conducted to identify individual and county-level factors associated with the higher rate of vaccinations in clinical settings. Results: Rural residents, particularly in Small Rural counties (80.8%) were more dependent upon clinical settings than Urban residents (69.1%) for influenza vaccinations. In adjusted analyzes, living in a Large or Small Rural county remained significant related to an increased odds of being vaccinated in a clinical setting (OR 1.17, 95% CI 1.06–1.29 and OR 1.45, 95% CI 1.24–1.69 respectively). Other related contributory factors included socioeconomic factors, health status, health condition, and per capita income of the county. Conclusions: Rural residents depend upon traditional, clinical settings when an influenza vaccination is sought. The results can be used for further research and programs to improve access to and delivery of influenza vaccinations for disparate populations. © 2011 Elsevier Ltd. All rights reserved.

1. Introduction Influenza remains a leading cause of hospitalization and death among the United States population. The highest rates of hospitalizations from seasonal influenza are among children aged <6 months [1], persons aged ≥65 years, and those with medical conditions which increase the risk for influenza complications [2]. However, during the 2009 H1N1 pandemic, more hospitalizations were found among adults ≤65 years [3]. As a result, current national guidelines recommend an expansion for influenza vaccination to all persons aged ≥6 months while persons aged ≥50 years and those who have underlying medical conditions are priority groups to be vaccinated [3]. Despite much progress in influenza vaccination coverage, only 60–70% of adults aged ≥65 years and 30–40% of high-risk younger adults are vaccinated each year, causing national

∗ Corresponding author at: University of South, Carolina School of Medicine, Department of Family and Preventive Medicine, 3209 Colonial Drive, Columbia, SC 29203, USA. Tel.: +1 803 434 3611; fax: +1 803 434 8374. E-mail addresses: [email protected] (K.J. Bennett), [email protected] (C. Pumkam), [email protected] (J.C. Probst). 0264-410X/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.vaccine.2011.06.038

influenza vaccination levels far below the Healthy People 2010 objectives of 90% and 60% respectively [4,5]. Low influenza vaccination rates are prevalent among those who are young, nonwhites, rural and do not have insurance or a regular source of care [5–9,4]. To expand influenza vaccination coverage, the use of nontraditional locations (e.g. pharmacies, retail stores, churches, and workplaces) has emerged as a public health strategy to enhance influenza vaccination coverage among population subgroups not covered by the traditional health care system [8,9]. The provision of influenza vaccination services in nontraditional settings has been shown to be cost effective and to improve access [10]. Yet, research has indicated that vaccine access in these nontraditional locations continues to be limited among populations targeted for improvement, such as rural residents, those with chronic health conditions, low income, non-white populations, and younger age groups [11,12]. Rural residents face multiple barriers to vaccine access and delivery. Their overall lower rate of preventive services receipt [13,14] may be due to insufficient services, fewer traditional health care settings, higher rates of poverty and uninsured people, and longer distances to sources of care [14]. With evidence indicating at least 70% of all influenza vaccinations are delivered in traditional

K.J. Bennett et al. / Vaccine 29 (2011) 5970–5977

clinical settings [12], areas with fewer providers, hospitals, or clinics (particularly in many rural locations), will have a more difficult time obtaining such vaccinations. Previous studies have included rural in their analyses, but have not specifically examined disparities in influenza vaccination rates or location of receipt among rural residents, or attempted to understand the factors related to the disparate rates [11–13]. Thus, the purpose of the study reported here is to determine if rural residents differ from urban residents in where they obtain an influenza vaccination, and to determine what factors are associated with these differences. This would help determine which segments of the population and what types of setting need to be targeted to make influenza vaccine more accessible and increase the proportion of adults vaccinated.

2. Methods Our analysis used data from four years (2002–2005) of the Behavioral Risk Factor Surveillance System (BRFSS). The BRFSS is a stratified sample survey, conducted via telephone that represents non-institutionalized adults in the US. The dependent variable of interest was the location of influenza vaccination administration. The question, “Where did you go to get your most recent [vaccination],” had several options, ranging from clinical settings (e.g. physician’s office, emergency department, hospital, health department), the workplace, or a retail location. We used this question to create a dichotomous variable for location; clinical setting vs. other setting. Location of vaccine administration was a core question in the 2002 BRFSS, administered to respondents in all states. In 2003–2005, this question was included only in an optional module. To obtain the most recent data available while ensuring geographic variability, we pooled data from all four years (2002–2005). If a state had information in more than one year, only the most recent data was used for that state (see Appendix A). We excluded respondents with missing data on the variables utilized in the analysis. The initial population consisted of 783,809 unweighted (502,234,021 weighted) respondents, and consisted of all adults responding to the BRFSS and having no missing data for any covariate. The subsequent analytic population for influenza vaccination receipt was limited to those over the age of 18 who both reported obtaining an influenza vaccination (either via injection or nasal administration) in the prior year and responded to the BRFSS question regarding location of vaccine administration. We examined all those over the age of 18 since the Centers for Disease Control and Prevention now recommends all those over the age of 18 receive an annual influenza vaccination [15]. After these exclusions, the initial population was reduced to a final sample size of 70,468 observations, representing an estimated US population of 48,392,455 million adults. The independent variables of interest were rurality and race. Rurality was defined at the county level using the 2003 Urban Influence Codes from the US Department of Agriculture Economic Research Service [16], drawn from the 2006 Area Resource File [17]. Rurality was divided into three categories: Urban (codes 1 and 2), Large Rural (rural counties with a town of 10,000–49,999 persons; codes 3, 5, and 8), and Small Rural (rural counties with a town of 2500–9999 persons or towns of 2500 or smaller; codes 4, 6, 7, 9, 10, 11, and 12). An All Rural category was also created, which collapsed the three types of rural counties into one summary category (i.e. codes 3–12). Race/Ethnicity was obtained from the BRFSS, based upon respondent self-report and categorized as Non-Hispanic White, Non-Hispanic African American, Hispanic, and Other. Additional independent variables were selected using Anderson’s Behavioral Model for Health Services Use [18]. This

5971

framework posits that the use of health care services is contingent upon predisposing characteristics that interact with enabling resources and an individual’s perception of need. For this analysis, predisposing characteristics included sex, age group (18–29, 30–44, 45–54, 55–64, 65–74, 75+), educational level (less than a high school diploma, high school diploma or some college, or a college degree or greater), and smoking status (current smoker vs. nonsmoker). The enabling resources included insurance status (insured vs. uninsured), employment status (employed vs. unemployed), marriage status (married vs. not married), income (expressed as an estimated percent of the federal poverty limit), having a usual source of care, and deferring medical care due to cost. The need factors were self-reported health status (excellent / very good / good versus fair / poor), and having one of the three high risk conditions (asthma, diabetes, pregnancy) that are recommended to receive an immunization. We used four county-level variables as descriptors of the local environment. These variables, obtained from the Area Resource File, included: physician availability, defined as the number of active primary care providers per 1000 residents in a county (divided into quartiles, <0.48, 0.48–0.75, 0.76–1.12, and >1.12); the per capita income of the county (divided into quartiles, <$21,169, $21,169–24,024, $24,025–27,407, >$27,407); the percent of the county’s population that was over the age of 65 (divided into quartiles, <12.2%, 12.2–14.5%, 14.6–17.2%, and >17.2%); and the percent of the county’s population that was female (divided into quartiles, <49.8%, 49.8–50.1%, 50.2–51.3%, and >51.3%). 2.1. Statistical analyses We began by estimating the percentage of respondents reporting an influenza vaccination in the previous year, by the characteristics of interest; all subsequent analyses were then subset to those who reported obtaining an influenza vaccination. We then described the study population, by rurality, the characteristics of interest, and the county-level variables. Further bivariate analysis examined the proportion of influenza vaccinations obtained in clinical settings, by rurality and the independent variables. All differences were tested using Wald Chi Square tests; in cases of multiple comparisons, we utilized the Bonferroni method to reduce the likelihood of a Type I error. We performed four separate multivariable ordinary least squares logistic regressions, using location (clinical setting vs. other) as the dependent variable. All models included the survey year as a control for time. The first model estimated the likelihood of receiving an influenza vaccination a clinical setting with only the rurality of the county. The second model included rurality and race/ethnicity. The third model added predisposing, enabling, and need factors to the preceding model. The fourth and final model included all of the previous variables, as well as the four countylevel variables (physician supply, per capita income, percent over the age of 65, and percent female). To account for the complex stratified sample and the weights within the BRFSS data, analyses were conducted using SAS 9.1 and SAS-Callable SUDAAAN 10.0; all analyses generate estimates representative of the US adult population. 3. Results Table 1 displays estimated influenza vaccination rates, by rurality and the characteristics of interest. Overall, 30.4% of adults aged 18 or over reported an influenza vaccination in the previous year. Rates were higher among whites; females; older age groups; those with a college degree; non-smokers; those with insurance; unemployed; married; those with higher incomes; those with a usual

5972

K.J. Bennett et al. / Vaccine 29 (2011) 5970–5977

Table 1 Proportion of adults reporting with an influenza vaccination, by rurality of county of residence, 2002–2005 BRFSS.

N (wtd N) Overall Race/ethnicity White African American Hispanic Other Sex Male Female Age groups 18–29 30–44 45–54 55–64 65–74 75+ Educational attainment Less than high school H.S. diploma College degree or more Smoking status Current smoker Non-smoker Insurance status Insured Uninsured Employment status Employed Not employed Marital status Married Not married Percent of poverty limit <100 100–299 300–399 400+ Missing Usual source of care Has a usual source of care No usual source of care Deferred care due to Cost Deferred care due to cost Did not defer care Self-reported health status Excellent/very good/good Fair/poor High risk conditions Yes No

All*

Urban*

All rural*

Large rural*

Small rural*

783,809 (502,460,770) 30.4

561,825 (430,758,699) 30.3

221,984 (71,702,071) 31.1

138,572 (4,425,101) 31.4

83,412 (23,276,970) 30.6

32.6 25.2 22.7 28.0

32.7 25.1 22.4 27.5

32.0 25.7 25.8 31.7

32.3 26.0 26.3 31.0

31.2 24.9 24.5 33.1

29.1 31.6

29.2 31.3

28.9 33.2

29.4 33.3

27.8 33.1

18.1 18.7 25.5 38.6 61.5 72.5

18.2 18.9 25.8 38.7 61.7 72.4

18.1 17.6 24.1 37.5 60.9 72.9

17.9 18.1 24.7 38.4 61.3 73.2

18.5 16.6 22.8 35.8 60.0 72.1

29.5 29.5 32.3

28.8 29.5 32.0

32.4 29.7 34.6

32.3 30.2 34.2

32.5 28.6 35.7

21.5 32.8

21.6 32.6

21.1 34.4

21.3 34.6

20.7 33.9

33.2 14.9

32.9 14.9

34.7 15.0

34.7 14.9

34.5 15.2

23.8 41.3

23.9 41.0

22.9 42.7

23.4 43.0

22.0 42.3

31.6 28.7

31.5 28.5

31.9 29.9

32.5 29.6

30.6 30.6

25.0 28.3 32.7 30.0 36.1

24.5 28.3 32.5 29.9 36.2

27.5 28.5 33.9 31.1 35.4

27.3 29.0 34.7 31.1 35.2

27.9 27.5 32.4 31.1 35.7

33.9 15.8

33.8 15.8

34.8 16.0

35.0 16.2

34.3 15.5

19.5 32.0

19.5 31.8

19.6 33.1

19.2 33.4

20.4 32.5

28.4 41.3

28.4 41.1

28.4 42.6

28.8 43.1

27.6 41.7

44.0% 27.9%

43.9% 27.8%

44.5% 28.5%

44.8% 28.7%

44.0% 27.9%

Bold indicates significantly different from Urban, p < 0.05. * Within-group distribution significantly different p < 0.05.

source of care; those that did not defer care due to cost; those in poorer health, and those with high risk conditions. Overall, the immunization rate was higher among Rural residents (31.1%) versus Urban residents (30.3%); this difference persisted across many of the independent factors of interest. Subsequent analyses were restricted to adults (aged 18 or over) who had an influenza vaccination the previous year (see Table 2). Adults reporting influenza vaccination predominantly lived in urban counties (87.6%), with the remainder in Large Rural (9.1%) and Small Rural (3.3%) counties. Rural residents were more likely to be white, female, older, have a lower educational attainment, current smoker, to be uninsured, unemployed, to be married, to have lower incomes, to defer care due to cost, to be in poorer health, and to not have a high risk condition than their urban counterparts. A majority of adults reporting an influenza vaccination lived in counties in the highest quartiles for per capita income, primary care physicians per 1000 residents ratio, and proportion that

was female and in the lowest quartiles for proportion over the age of 65. These proportions differed significantly as rurality increased; the proportion in the highest quartiles for per capita income, physician ratio, and percent female all decreased, while the proportion in the highest quartile for percent over the age of 65 increased. Table 3 displays the proportion of respondents who received their influenza vaccination in a clinical setting. Overall, 70.0% were vaccinated in clinical settings; this was higher among Rural residents (76.8%), with the highest among Small Rural residents (80.7%). The proportion obtaining an influenza vaccination in clinical settings was higher for non-Whites, females, older adults, those with lower educational attainment, non-smokers, the insured, the unemployed, unmarried, those with lower incomes, those with a usual source of care, those in poorer health, and those with high risk conditions. Administration in clinical settings was also higher among counties with higher per capita incomes, physician to resident ratios, with an older population, and with more females. In

K.J. Bennett et al. / Vaccine 29 (2011) 5970–5977

5973

Table 2 Demographic characteristics of adults reporting vaccination receipt in the past year, by rurality of county of residence, 2002–2005 BRFSS.

Observations Weighted observations Rurality distribution Flu shot location Clinical setting Other Race/ethnicity White African American Hispanic Other Male Age groups 18–29 30–44 45–54 55–64 65–74 75+ Educational attainment Less than high school H.S. diploma College degree or more Current smoker Insured Employed Married Percent of poverty limit <100 100–299 300–399 400+ Missing Has a usual source of care Deferred care due to cost Self-reported health status Excellent/very good/good Fair/poor High risk conditions Yes No Per capita income quartiles <$21,169 $21,169–24,024 $24,025–27,407 >$27,407 PCP to resident ratio, quartiles <0.48 0.48–0.75 076–1.12 >1.12 Percent >65, quartiles <12.2 12.2–14.5 14.6–17.2 >17.2 Percent female, quartiles <49.8 49.8–50.1 50.2–51.3 >51.3

All

Urban

All rural

Large rural

Small rural

70,468 48,392,455 100

51,704 42,385,931 87.6

18,764 6,006,524 12.4

4,398,764 4,398,764 9.1

6201 1,607,759 3.3

70.0 30.0

69.1 30.9

76.8 23.2

75.3 24.7

80.8 19.2

75.2 8.5 10.1 6.2 46.4

73.8 8.9 11.0 6.4 46.8

85.2 5.2 4.3 5.2 43.9

84.4 5.5 4.9 5.2 43.8

86.2 5.2 3.1 5.5 44.1

12.2 18.8 16.0 16.6 18.5 17.8

12.4 19.3 16.3 16.4 18.1 17.5

11.0 15.4 13.9 18.2 21.4 20.1

11.1 15.7 14.0 17.9 20.9 20.3

11.2 14.6 12.4 18.5 23.0 20.3

10.8 54.2 35.0 14.9 92.9 49.2 61.3

10.4 53.1 36.5 14.6 93.1 50.1 60.9

13.7 62.0 24.3 17.0 91.8 42.6 64.5

13.4 62.2 24.4 17.0 92.0 42.5 64.4

13.7 62.6 23.7 17.2 91.0 41.7 64.7

9.4 19.4 22.6 34.4 14.3 89.4 6.9

9.1 18.6 22.1 36.0 14.2 89.3 6.9

11.0 25.1 26.2 22.6 15.1 90.4 7.3

10.9 24.6 26.3 23.0 15.1 90.1 6.6

10.4 26.0 27.4 21.4 14.9 91.1 9.1

79.0 21.0

79.6 20.4

74.6 25.4

75.5 24.5

73.0 27.0

22.2 77.8

22.0 78.0

23.8 76.2

23.8 76.2

23.8 76.2

1.5 4.1 11.8 82.6

0.8 2.1 8.4 88.6

6.2 18.0 36.0 39.8

3.8 14.2 36.1 45.8

10.9 28.5 39.4 21.1

2.4 11.4 27.7 58.5

1.6 8.9 26.4 63.1

7.9 29.2 36.9 26.0

3.5 30.1 40.9 25.5

21.7 30.8 26.0 21.5

54.2 26.6 11.8 7.5

59.9 26.0 8.8 5.4

14.0 31.0 32.9 22.0

16.2 34.0 32.5 17.3

8.7 24.0 41.6 25.7

14.6 26.8 23.1 35.5

14.2 26.9 21.6 37.3

17.8 26.4 33.6 22.1

15.9 24.7 34.7 24.6

26.9 33.2 26.0 14.0

Bold indicates significantly different from urban, p < 0.05

general, Rural residents were more likely to obtain an influenza vaccination in a clinical setting, across the independent characteristics of interest. Table 4 summarizes the results from the multivariate ordinary least squares logistic regression models. The progression of models is intended to identify those factors that could explain the higher rate of influenza vaccinations in clinical settings across Rural areas. The first two models indicated a strong rural effect, with residents of Large and Small Rural counties both having higher odds of getting vaccinated in a clinical setting than their Urban peers. Adding

the predisposing, enabling, and need factors (Model 3) reduced the odds of residents of Large and Small Rural counties being vaccinated in a clinical setting, yet both remained positive and significant. The final model (Model 4) added the county-level factors to the previous model. These final estimates show that the odds ratios for Large and Small Rural residents decreased further, yet remained significant (OR 1.17, 1.06–1.29 and OR 1.45, 1.24–1.69, respectively). Other factors significantly associated with an increased odds of obtaining an influenza vaccination in a clinical setting in this final, all inclusive model included: non-White race/ethnicity, older

5974

K.J. Bennett et al. / Vaccine 29 (2011) 5970–5977

Table 3 Proportion of vaccinated adults reporting receipt in a clinical setting, by rurality, 2002–2005 BRFSS.

Overall Race/ethnicity White African American Hispanic Other Sex Male Female Age groups 18–29 30–44 45–54 55–64 65–74 75+ Educational attainment Less than high school H.S. diploma College degree or more Smoking status Current smoker Non-smoker Insurance status Insured Uninsured Employment status Employed Not employed Marital status Married Not married Percent of poverty limit <100 100–299 300–399 400+ Missing Usual source of care Has a usual source of care No usual source of care Deferred care due to cost Deferred care due to cost Did not defer care Self-reported health status Excellent/very good/good Fair/Poor High risk conditions Yes No Per capita income quartiles <$21,169 $21,169–24,024 $24,025–27,407 >$27,407 PCP to resident ratio, quartiles <0.48 0.48–0.75 076–1.12 >1.12 Percent >65, quartiles <12.2 12.2–14.5 14.6–17.2 >17.2 Percent female, quartiles <49.8 49.8–50.1 50.2–51.3 >51.3

Total

Urban

All rural

Large rural

Small rural

70.0

69.1

76.8

75.3

80.7

69.1* 74.2 73.8 69.7

67.9* 74.0 73.6 68.0

76.3* 76.9 77.4 83.9

74.7 76.6 77.0 82.4

80.5 78.0 79.0 88.0

68.8* 71.1

67.8* 70.1

76.3 77.2

75.0 75.6

79.8 81.5

67.5* 53.7 57.4 70.9 83.0 86.0

67.5* 53.2 56.5 69.8 82.0 85.3

67.8* 58.7 64.9 77.6 89.1 90.1

67.6* 56.8 63.4 75.8 87.8 88.8

68.3* 64.2 68.9 82.1 92.1 93.6

83.1* 71.6 63.5

82.4* 70.7 63.0

86.8* 77.5 69.3

86.6* 75.8 68.0

87.2* 82.3 72.8

67.3* 70.5

66.0* 69.6

75.6 77.0

74.2 75.6

79.5 81.0

69.9* 71.3

69.0 70.4

76.8 76.3

75.2 76.5

81.2 75.9

56.1* 83.5

55.6* 82.6

60.4* 88.9

58.6* 87.7

65.4* 92.3

68.3* 72.8

67.2* 72.0

75.3* 79.4

74.0* 77.8

79.0* 84.0

82.9* 74.7 70.9 60.7 76.2

82.2* 74.2 70.1 60.1 75.0

87.5* 77.5 75.9 67.0 83.9

86.1* 76.5 76.1 64.3 81.3

91.2* 80.0 75.6 75.1 91.1

70.8* 63.0

69.9* 62.2

77.6* 69.5

76.0* 69.5

81.9* 69.5

72.6* 69.8

71.6 68.9

79.3 76.6

78.0 75.1

82.0 80.6

66.7* 82.5

65.9* 81.4

72.6* 89.0

71.2* 88.1

76.7* 91.3

78.3* 67.6

77.4* 66.7

84.7* 74.3

83.0* 72.9

89.2* 78.1

74.5* 77.4 75.6 68.8

68.0* 75.3 75.0 68.4

80.6* 79.2 76.7 75.2

81.4* 77.2 75.3 74.3

79.9* 81.8 80.6 80.1

75.1* 70.7 69.8 69.8

71.7 68.0 68.6 69.3

79.9 76.5 75.9 77.4

73.7 76.0 74.2 76.5

82.9 78.0 83.2 79.5

67.1* 72.3 73.4 77.6

66.9* 71.9 71.5 75.8

74.1* 74.8 77.0 80.9

73.2* 74.0 76.0 78.8

79.2 78.4 79.7 83.6

66.9* 69.5 70.3 71.6

65.2* 68.7 68.6 71.0

76.0 75.3 77.7 77.9

73.8 74.1 76.2 76.4

80.2 78.0 82.2 84.2

Bold indicates significantly different from urban, p < 0.05. * Within-group distribution significantly different p < 0.05.

K.J. Bennett et al. / Vaccine 29 (2011) 5970–5977

5975

Table 4 Odds of receiving an influenza vaccination in a clinical setting.

Intercept Rurality Urban Large rural Small rural Race/ethnicity White African American Hispanic Other Male Age group 18–29 30–44 45–54 55–64 65–74 75+ Educational attainment Less than high school H.S. diploma College degree or more Current smoker Insured Employed Married Percent of poverty limit <100% 100–299% 300–399% 400%+ Missing Has a usual source of care Deferred care due to cost Fair/poor health status Has a high risk condition Per capita income quartiles <$21,169 $21,169–24,024 $24,025–27,407 >$27,407 PCP to resident ratio, quartiles <0.48 0.48–0.75 076–1.12 >1.12 Percent >65, quartiles <12.2 12.2–14.5 14.6–17.2 >17.2 Percent female, quartiles <49.8 49.8–50.1 50.2–51.3 >51.3

Model 1

Model 2

Model 3

Model 4

2.86 (2.66, 3.07)

2.72 (2.55, 2.91)

2.54 (1.99, 3.25)

3.03 (2.29, 4.02)

Ref. 1.35 (1.25, 1.46) 1.69 (1.49, 1.91)

Ref. 1.38 (1.27, 1.49) 1.74 (1.53, 1.96)

Ref. 1.23 (1.13, 1.34) 1.61 (1.41, 1.83)

Ref. 1.17 (1.06, 1.29) 1.45 (1.24, 1.69)

Ref. 1.35 (1.18, 1.54) 1.22 (1.04, 1.41) 1.03 (0.88, 1.21)

Ref. 1.34 (1.15, 1.55) 1.28 (1.09, 1.51) 1.25 (1.06, 1.48) 1.06 (0.99, 1.14)

Ref. 1.32 (1.13, 1.54) 1.30 (1.10, 1.53) 1.26 (1.06, 1.49) 1.06 (0.99, 1.14)

Ref. 0.66 (0.57, 0.76) 0.77 (0.67, 0.88) 1.10 (0.95, 1.27) 1.53 (1.30, 1.79) 1.68 (1.42, 1.99)

Ref. 0.66 (0.57, 0.76) 0.76 (0.66, 0.88) 1.09 (0.94, 1.26) 1.50 (1.28, 1.77) 1.66 (1.41, 1.96)

1.26 (1.08, 1.47) 1.03 (0.96, 1.11) Ref. 0.98 (0.89, 1.08) 0.94 (0.79, 1.10) 0.44 (0.40, 0.48) 0.99 (0.93, 1.07)

1.26 (1.08, 1.47) 1.03 (0.96, 1.11) Ref. 0.98 (0.89, 1.08) 0.95 (0.80, 1.11) 0.44 (0.40, 0.48) 0.99 (0.93, 1.07)

1.70 (1.41, 2.06) 1.21 (1.09, 1.35) 1.10 (1.01, 1.19) Ref. 1.12 (1.01, 1.25) 1.26 (1.11, 1.44) 1.02 (0.88, 1.17) 1.30 (1.18, 1.42) 1.41 (1.30, 1.53)

1.68 (1.39, 2.03) 1.20 (1.08, 1.34) 1.09 (1.01, 1.19) Ref. 1.11 (0.99, 1.24) 1.25 (1.09, 1.42) 1.02 (0.88, 1.18) 1.29 (1.18, 1.42) 1.41 (1.30, 1.53)

age, lower educational attainment, unemployment, lower incomes, deferring care due to cost, fair or poor health status, having a high risk condition, or living in a county with lower per capita incomes. 4. Conclusions Expansion of vaccine delivery sites beyond clinical settings has been advocated as a way to improve population vaccination rates [19]. We found that only 30.4% of adults reported receiving an influenza vaccination during 2002–2005, with most of these individuals (70.0%) receiving immunization in a clinical setting, consistent with the previous evidence for all vaccinations [12]. In both unadjusted and adjusted analyses, Rural residents were more dependent upon clinical settings for their influenza vaccinations than Urban residents, and residents of Small Rural counties

1.06 (0.85, 1.33) 1.23 (1.04, 1.46) 1.15 (1.04, 1.28) Ref. 0.99 (0.83, 1.18) 0.89 (0.81, 0.98) 0.94 (0.87, 1.02) Ref. 0.86 (0.74, 1.01) 0.97 (0.84, 1.12) 0.88 (0.76, 1.03) Ref. 0.85 (0.77, 0.95) 0.97 (0.88, 1.06) 0.95 (0.88, 1.03) Ref.

were more dependent upon clinical locations than those in Large Rural counties. We hypothesized that including factors associated with the location of vaccination administration, particularly county characteristics, would help to explain these Urban/Rural differences; the final model, however, continued to show a significant and positive effect of living in a rural area upon obtaining an influenza vaccination in a clinical setting. Our results were similar to previous studies that found a dependence upon traditional clinical settings for vaccine delivery, particularly among individuals in poorer health and older age groups [12]. These results also highlight the disparate influenza vaccination rates across certain demographic groups. We identified several socio-demographic factors that were associated with being vaccinated; these included race/ethnicity, gender, age, educational attainment, insurance status, employment status, and income. In

5976

K.J. Bennett et al. / Vaccine 29 (2011) 5970–5977

the adjusted analysis, these factors were also associated with being vaccinated in clinical settings; for example, African Americans and Hispanics were both more likely to be vaccinated in clinical settings (see Table 4), yet had lower influenza vaccination rates when compared to whites (see Table 1). Similar trends are evident for those with lower educational attainment and those with lower incomes. There is a need to investigate the area-level characteristics associated with levels of vaccine receipt, to determine true supply and access. Specifically, if influenza vaccination supply and availability in non-clinical settings is lower for some locales, particularly in these rural areas, vaccinations would be sought from clinical settings. For example, studies that have explored the impact of pharmacies providing influenza vaccinations have shown that states that allow pharmacists to administer vaccines have significant higher rates of uptake [20]. These studies do not, however, control for micro-level access. Evidence suggests that there are fewer non-clinical outlets available in rural settings (such as pharmacies [21–23], retail clinics [24,25], and worksites [26]); thus the vaccination supply would naturally flow to clinical settings. This would also result in an overall reduction in supply and access to these groups, reducing overall uptake. Providers may also be hesitant to order large number of doses, due to the financial risk involved [27,28]. We would encourage future research to attempt to explore these relationships further. The availability of alternate immunization administration locations is largely unknown. Aggregate studies that track and analyze where non-medical sites for administration are located have not yet been done, so it is unclear if a maldistribution of such sites influences our findings. We can assume, however, that rural counties and counties with lower area incomes may not support either retail or worksite locations as readily as larger or more resource intensive areas. We do know that rural communities rely more heavily upon independently owned pharmacies for their prescription needs; recent estimates identified nearly 1900 such organizations as the sole pharmacy in their community. Unfortunately, from 2006 to 2008, these communities suffered a net loss of more than 500 pharmacies [29]. Patient preference is also an important factor in vaccination seeking behavior. It is well documented that certain demographic groups seek vaccinations at different rates, due to differing preferences and health beliefs [30,31]. Patient attitudes about vaccinations, rather than conventional factors such as income and education, may lead to the underuse of influenza vaccination [32]. Although the influenza vaccine has been proven to be safe and effective, belief in the lack of effectiveness of the vaccine and belief in complications of the vaccine can prevent those who could benefit from it from not receiving it [33,34]. Previous evidence and our results show that the presence of healthcare professionals is strongly associated with the probability of receiving an influenza vaccination [34,35]. Primary care professionals can play an important role in discussing the benefits and risks of the vaccine with patients, raising their awareness about influenza vaccination, and eliminating their negative attitude toward the vaccine. In addition, primary health care promotion measures such as the use of recall-reminder systems, patient education, home visits, clinicbased advertising are suggested to enhance the uptake of the vaccine [36,37]. Also, future research should focus on other factors in vaccine-seeking behavior. Our findings regarding of a higher likelihood of non-whites seeking immunizations in clinical settings also mirrors that of other authors [11]. Non-whites were no more likely, however, to receive an immunization at the workplace; rather, the odds of receiving the immunization within retail settings were significantly lower than that of whites. This may indicate either a preference among these groups for clinical settings, or a preference to not seek care in retail settings.

Recent evidence suggests that costs savings and a reduction in wait times are drivers of seeking care outside of traditional settings [28]. However, the rate of influenza vaccination obtained from nontraditional settings is low. Pharmacy-based vaccination programs in states where pharmacists legally administer influenza vaccines have been shown to be effective [20] but the effect of the program on the rate varies by year and state. Besides supply and distribution problems that should be addressed, additional research that explores more on this topic as a means of improving access and uptake of influenza vaccinations among disparate groups is recommended. The present study has several limitations. First, the BRFSS is a landline telephone survey; by definition, it does not include those who have only cell phones, or those who have no phone at all. This could bias the survey and its results for specific groups and behaviors [38,39]. Second, the information in the BRFSS is all selfreported. This may result in a bias due to over-reporting of desirable outcomes (such as preventive services), or inaccurate recall of the details of such reports. Next, the surveys did not ask about all the high risk conditions or recommended indications for influenza vaccination, so we were not able to fully adjust for these factors [4]. We were able to control for some of these conditions (asthma, diabetes, and pregnancy), but a broader inclusion would reduce the bias. Finally, the BRFSS sample is selected to represent the noninstitutionalized population. The rate of influenza vaccination in non-institutional community settings could be different from that in the institutionalized population. Evidence suggests, however, that the validity of the BRFSS survey is robust [40]. Our analysis sought to investigate influenza vaccination-receipt behaviors, and particularly whether the location of service differed among rural residents. Our results indicating that rural residents depend upon clinical settings when an influenza vaccination is sought can be used for further research as well as programs and interventions to improve access to and delivery of influenza vaccinations for disparate groups.

Appendix A. State inclusions, by BRFSS survey year 2002

2003

2004

2005

Alabama Delaware Dist. of Columbia Illinois Indiana Iowa Kentucky Louisiana Maryland Massachusetts Michigan Mississippi Missouri Montana Nebraska Nevada New Mexico North Dakota Oklahoma Oregon Tennessee Utah Virginia Washington

Colorado Hawaii New Jersey

Florida Georgia New York North Carolina Rhode Island Texas

Arizona Arkansas California Connecticut Idaho Kansas Maine Minnesota New Hampshire Ohio Pennsylvania South Carolina South Dakota Vermont West Virginia Wisconsin Wyoming

References [1] Dawood FS, Fiore A, Kamimoto L, Bramley A, Reingold A, Gershman K, et al. Burden of seasonal influenza hospitalization in children United States, 2003 to 2008. J Pediatr 2010;157(5):808–14.

K.J. Bennett et al. / Vaccine 29 (2011) 5970–5977 [2] Mullooly JP, Bridges CB, Thompson WW, Chen J, Weintraub E, Jackson LA, et al. Influenza- and RSV-associated hospitalizations among adults. Vaccine 2007;25(5):846–55. [3] Fiore AE, Uyeki TM, Broder K, Finelli L, Euler GL, Singleton JA, et al. Prevention and control of influenza with vaccines: recommendations of the Advisory Committee on Immunization Practices (ACIP), 2010. MMWR Recomm Rep 2010;59(RR-8):1–62. [4] Centers for Disease Control Prevention (CDC). State-specific influenza vaccination coverage among adults aged > or =18 years – United States, 2003–04 and 2005–06 influenza seasons. MMWR Morb Mortal Wkly Rep 2007;56(37):953–9. [5] Lu P, Bridges CB, Euler GL, Singleton JA. Influenza vaccination of recommended adult populations, U.S., 1989–2005. Vaccine 2008;26(14):1786–93. [6] Centers for Disease Control Prevention (CDC). Racial/ethnic disparities in influenza and pneumococcal vaccination levels among persons aged >65 years – United States, 1989–2001. MMWR Morb Mortal Wkly Rep 2003;52(40):958–62. [7] Pleis JR, Gentleman JF. Using the National Health Interview Survey: time trends in influenza vaccinations among targeted adults. Eff Clin Pract 2002;5(3 Suppl.):E3. [8] Amburgh JAV, Waite NM, Hobson EH, Migden H. Improved influenza vaccination rates in a rural population as a result of a pharmacist-managed immunization campaign. Pharmacotherapy 2001;21(9):1115–22. [9] Postema AS, Breiman RF, National Vaccine Advisory Committee. Adult immunization programs in nontraditional settings: quality standards and guidance for program evaluation. MMWR Recomm Rep 2000;49(RR-1):1–13. [10] Prosser L, O’Brien M, Molinari NA, Hohman KH, Nichol KL, Messonnier ML, et al. Non-traditional settings for influenza vaccination of adults: costs and cost effectiveness. Pharmacoeconomics 2008;26(2):163–78. [11] Lee BY, Mehrotra A, Burns RM, Harris KM. Alternative vaccination locations: who uses them and can they increase flu vaccination rates? Vaccine 2009;27(32):4252–6. [12] Singleton JA, Poel AJ, Lu PJ, Nichol KL, Iwane MK. Where adults reported receiving influenza vaccination in the United States. Am J Infect Control 2005;33(10):563–70. [13] Bennett KJ, Olatosi B, Probst JC. Health disparities: a rural–urban chartbook. Technical report provided to Health Resources and Services Administration/Office of Rural Health Policy; June 2008. Available from: http://www.ruralhealthresearch.org/topics/100000064/publications az.php. [14] Harris R, Leininger L. Preventive care in rural primary care practice. Cancer 1993;72(3):1113–8. [15] Centers for Disease Control Prevention (CDC). Recommended adult immunization schedule – United States, 2011. MMWR Morb Mortal Wkly Rep 2011;60(4):1–4. [16] Economic Research Service. Urban influence codes. U.S. Department of Agriculture; June 2008. Available from http://www.ers.usda. gov/Briefing/Rurality/NewDefinitions/. [17] Health Resources and Services Administration. Bureau of Health Professions. Area Resource File (ARF) 2008. Rockville, MD: U.S. Department of Health and Human Services, Health Resources and Services Administration, Bureau of Health Professions; 2008. [18] Anderson RM. Revisiting the behavioral model and access to medical care: does it matter? J Health Soc Behav 1995;36(1):1–10. [19] Singleton JA, Poel A, Lu P, Nichol K, Iwane M. Where adults reported receiving influenza vaccination in the United States. Am J Infect 2005;33(10):563–70. [20] Steyer T, Ragucci K, Pearson W, Mainous A. The role of pharmacists in the delivery of influenza vaccinations. Vaccine 2004; 22:1001–6.

5977

[21] Knapp KK, Quist RM, Walton SM, Miller LM. Update on the pharmacist shortage: national and state data through 2003. Am J Health Syst Pharm 2005;62(5):492–9. [22] Gadkari A, Mott D, Kreling D, Bonnarens J. Characteristics of unmet demand for pharmacists: a survey of rural community pharmacies in Wisconsin. J Am Pharm Assoc 2003;48(5):598–609. [23] Shambaugh-Miller M, Vanosdel N, Mueller K. Reliance on independently owned pharmacies in rural America. Rural Policy Brief 2007;PB2007-6:1–4. [24] Rudavsky R, Pollack CE, Mehrotra A. The geographic distribution, ownership, prices, and scope of practice at retail clinics. Ann Intern Med 2009;151(5):315–20. [25] Rudavsky R, Mehrotra A. Sociodemographic characteristics of communities served by retail clinics. J Am Board Fam Med 2010;23(1):42–8. [26] Probst J, Moore C, Glover S, Samuels M. Person and place: the compounding effect of race/ethnicity and rurality on health. Am J Public Health 2004;94:1695–703. [27] Orenstein WA, Schaffner W. Lessons learned: role of influenza vaccine production, distribution, supply, and demand – what it means for the provider. Am J Med 2008;121(7 Suppl. 2):S22–7. [28] Layton CM, Robinson T, Honeycutt A. Influenza vaccine demand: the chicken and the egg, issue brief, 2005 prepared for the assistant secretary for planning and evaluation, US Department of Health And Human Services, under contract no. 100-03-0001; December 2010. Available from: http://aspe.hhs.gov/pic/fullreports/06/8476-4.doc 2005. [29] Klepser D, Xu L, Ullrich F, Mueller K. Independently owned pharmacy closures in rural America. Rural Policy Brief 2008;2(July):1–4. [30] Hebert PL, Frick KD, Kane RL, McBean AM. The causes of racial and ethnic differences in influenza vaccination rates among elderly Medicare beneficiaries. Health Serv Res 2005;40(2):517–37. [31] Chen J, Fox S, Cantrell C, Stockdale S, Kagawa-Singer M. Health disparities and prevention: racial/ethnic barriers to flu vaccinations. J Commun Health 2007;32(1):5–20. [32] Lindley MC, Wortley PM, Winston CA, Bardenheier BH. The role of attitudes in understanding disparities in adult influenza vaccination. Am J Prev Med 2006;31(4):281–5. [33] Armstrong K, Berlin M, Schwartz J, Propert K, Ubel P. Barriers to influenza immunization in a low-income urban population. Am J Prev Med 2001;20(1):21–5. [34] Winston CA, Wortley PM, Lees KA. Factors associated with vaccination of Medicare beneficiaries in five U.S. communities: results from the racial and ethnic adult disparities in immunization initiative survey, 2003. J Am Geriatr Soc 2006;54(2):303–10. ˜ I, Pérez-Farinós N, Sarría-Santamera A. Factors associated with [35] Pena-Rey influenza vaccination among elderly Spanish women. Public Health 2004;118(8):582–7. [36] Kellerman R, Allred C, Frisch L. Enhancing influenza immunization. Postcard and telephone reminders and the challenge of immunization site shift. Arch Fam Med 2000;9(4):368–72. [37] Ompad D, Galea S, Vlahov D. Distribution of influenza vaccine to high-risk groups. Epidemiol Rev 2006;28:54–70. [38] Blumberg SJ, Luke JV. Coverage bias in traditional telephone surveys of lowincome and young adults. Public Opin Q 2007;71(5):734–49. [39] Blumberg SJ, Luke JV, Cynamon ML. Telephone coverage and health survey estimates: evaluating the need for concern about wireless substitution. Am J Public Health 2006;96(5):926–31. [40] Nelson D, Holtzman D, Bolen J, Stanwyck C, Mack K. Reliability and validity of measures from the Behavioral Risk Factor Surveillance System (BRFSS). Soc Prev Med 2001;46(Suppl. 1):S3–42.