Association between hospital community services and county population health in the USA

Association between hospital community services and county population health in the USA Abstract Objectives Little research has utilized population level data to test the association between community health outcomes and (i) hospital-sponsored community services that facilitate access to care and (ii) hospital-sponsored community building services in the USA. Therefore, the purpose of this study was to examine these relationships. Methods A secondary data analysis of the 2016 County Health Rankings and American Hospital Association databases was conducted via zero-truncated negative Binomial regression. Results Findings indicate a statistically significant difference between the number of community healthcare access services and community building services with county’s rank of health behavior. However, no statistically significant differences were found between the number of community healthcare access services and community building services with county rankings of length of life, quality of life or clinical care. Conclusions Our findings suggest that quality measures of services may play a more important role in community health improvement and that there is opportunity for hospitals to revamp the way in which community health needs assessments are conducted. Additional federal action is needed to standardize hospital sponsored community health service data reporting so that practitioners, hospital administrators and researchers can more specifically define hospitals’ role in public health protection in the USA. health services, population-based and preventative services, social determinants Introduction While the USA outperforms many countries in global rankings, it often lags behind in health outcomes compared to other developed nations.1 Health disparities research indicates that socioeconomic and ethnic/racial inequality in the US disproportionately contributes to poor outcomes.2,3 Thus, national strategies such as the Affordable Care Act (ACA) of 2010 and Healthy People 2020 emphasize the importance for organizations to address structural and cultural inequalities in social determinants of health such as neighborhood and built environments, economic stability, social and community contexts, education and access to healthcare.4,5 In recognition of the role of hospitals in achieving equitable community health outcomes, the ACA requires tax exempt hospitals to complete community health needs assessments (CHNAs) at least every 3 years and to devise implementation plans to address health needs raised by communities. As hospitals have the potential to positively impact the health of the surrounding community,6 many hospitals have expanded the scope of medical services to increase healthcare access in order to better meet neighboring community health needs. Implementation of community services in hospitals, especially in response to community identified needs, provides an avenue to increase equity in access to care, and ultimately, improve health outcomes for populations that are often high risk, underserved, and have unmet needs (e.g. low income and rural). Although we know that the number and types of community-oriented services vary across hospitals, their overall population level health influence has not been clearly described in the extant literature. Given the paucity of research in this area, this article fills this scientific void by investigating the association between hospital sponsored community health services and population level health outcomes in the USA. Method Data sources We used the 2016 Area Health Resource File (AHRF), the 2016 American Hospital Association Dataset (AHA) and the 2016 County Health Rankings National Database for analyses. The AHRF database provides data concerning health professionals, health facilities and population characteristics including demographic and economic aspects all attributed to the county level.7 County location is identified by a two digit State and three digit county Federal Information Processing Standards (FIPS) code. The AHA database includes information on hospital characteristics, demographics, services and expenses from over 6000 hospitals in the USA.8 Each hospital’s location is identified by a FIPS code. The County Health Rankings database, developed by the University of Wisconsin Population Health Institute and the Robert Wood Johnson Foundation in 2016, is a national dataset that ranks county-level health across the USA by state using a multidimensional model that incorporates social determinants defined by four major factors: health behaviors, clinical care, social and economic factors, and physical environment to assess health outcomes—length and quality of life.9,10 County location is identified by a FIPS code, which was utilized to merge the three datasets. Our attempt at understanding the relationship between population factors and hospital community services followed the Rankings conceptual framework. Dependent variables The main dependent variables in this study are community-ranking variables provided by the County Health Ranking Database. These variables include length of life, quality of life, health behaviors and clinical care. We excluded social and economic factors, and physical environment measures as there is not a strong rational indicating hospitals will have a direct influence on either. The elements, which define each of these variables, are listed in Table 1, along with their weightings. Table 1 Community health ranking variable definitions Variable  Measure  Description  Weight (%)  Length of life  Premature death  Years of potential life lost before age 75 per 100 000 population (age-adjusted)  100  Quality of life  Poor or fair health  Percentage of adults reporting fair or poor health (age-adjusted)  20  Poor physical health days  Average number of physically unhealthy days reported in past 30 days (age-adjusted)  20  Poor mental health days  Average number of mentally unhealthy days reported in past 30 days (age-adjusted)  20  Low birth weight  Percentage of live births with low birth weight (<2500 g)  40  Health behaviors  Adult smoking  Percentage of adults who are current smokers  33  Adult obesity  Percentage of adults that report a BMI of 30 or more  17  Food environment index  Index of factors that contribute to a healthy food environment  7  Physical inactivity  Percentage of adults aged 20 and over reporting no leisure-time physical activity  7  Access to exercise opportunities  Percentage of population with adequate access to locations for physical activity  3  Excessive drinking  Percentage of adults reporting binge or heavy drinking  8  Alcohol-impaired driving deaths  Percentage of driving deaths with alcohol involvement  8  Sexually transmitted infections  Number of newly diagnosed chlamydia cases per 100 000 population  8  Teen births  Teen birth rate per 1 000 female population, ages 15–19  8  Clinical care  Uninsured  Percentage of population under age 65 without health insurance  25  Primary care physicians  Ratio of population to primary care physicians  15  Dentists  Ratio of population to dentists  5  Mental health providers  Ratio of population to mental health providers  5  Preventable hospital stays  Number of hospital stays for ambulatory-care sensitive conditions per 1000 Medicare enrollees  25  Diabetic monitoring  Percentage of diabetic Medicare enrollees ages 65–75 that receive HbA1c monitoring  13  Mammography screening  Percentage of female Medicare enrollees ages 67–69 that receive mammography screening  13  Variable  Measure  Description  Weight (%)  Length of life  Premature death  Years of potential life lost before age 75 per 100 000 population (age-adjusted)  100  Quality of life  Poor or fair health  Percentage of adults reporting fair or poor health (age-adjusted)  20  Poor physical health days  Average number of physically unhealthy days reported in past 30 days (age-adjusted)  20  Poor mental health days  Average number of mentally unhealthy days reported in past 30 days (age-adjusted)  20  Low birth weight  Percentage of live births with low birth weight (<2500 g)  40  Health behaviors  Adult smoking  Percentage of adults who are current smokers  33  Adult obesity  Percentage of adults that report a BMI of 30 or more  17  Food environment index  Index of factors that contribute to a healthy food environment  7  Physical inactivity  Percentage of adults aged 20 and over reporting no leisure-time physical activity  7  Access to exercise opportunities  Percentage of population with adequate access to locations for physical activity  3  Excessive drinking  Percentage of adults reporting binge or heavy drinking  8  Alcohol-impaired driving deaths  Percentage of driving deaths with alcohol involvement  8  Sexually transmitted infections  Number of newly diagnosed chlamydia cases per 100 000 population  8  Teen births  Teen birth rate per 1 000 female population, ages 15–19  8  Clinical care  Uninsured  Percentage of population under age 65 without health insurance  25  Primary care physicians  Ratio of population to primary care physicians  15  Dentists  Ratio of population to dentists  5  Mental health providers  Ratio of population to mental health providers  5  Preventable hospital stays  Number of hospital stays for ambulatory-care sensitive conditions per 1000 Medicare enrollees  25  Diabetic monitoring  Percentage of diabetic Medicare enrollees ages 65–75 that receive HbA1c monitoring  13  Mammography screening  Percentage of female Medicare enrollees ages 67–69 that receive mammography screening  13  Table 1 Community health ranking variable definitions Variable  Measure  Description  Weight (%)  Length of life  Premature death  Years of potential life lost before age 75 per 100 000 population (age-adjusted)  100  Quality of life  Poor or fair health  Percentage of adults reporting fair or poor health (age-adjusted)  20  Poor physical health days  Average number of physically unhealthy days reported in past 30 days (age-adjusted)  20  Poor mental health days  Average number of mentally unhealthy days reported in past 30 days (age-adjusted)  20  Low birth weight  Percentage of live births with low birth weight (<2500 g)  40  Health behaviors  Adult smoking  Percentage of adults who are current smokers  33  Adult obesity  Percentage of adults that report a BMI of 30 or more  17  Food environment index  Index of factors that contribute to a healthy food environment  7  Physical inactivity  Percentage of adults aged 20 and over reporting no leisure-time physical activity  7  Access to exercise opportunities  Percentage of population with adequate access to locations for physical activity  3  Excessive drinking  Percentage of adults reporting binge or heavy drinking  8  Alcohol-impaired driving deaths  Percentage of driving deaths with alcohol involvement  8  Sexually transmitted infections  Number of newly diagnosed chlamydia cases per 100 000 population  8  Teen births  Teen birth rate per 1 000 female population, ages 15–19  8  Clinical care  Uninsured  Percentage of population under age 65 without health insurance  25  Primary care physicians  Ratio of population to primary care physicians  15  Dentists  Ratio of population to dentists  5  Mental health providers  Ratio of population to mental health providers  5  Preventable hospital stays  Number of hospital stays for ambulatory-care sensitive conditions per 1000 Medicare enrollees  25  Diabetic monitoring  Percentage of diabetic Medicare enrollees ages 65–75 that receive HbA1c monitoring  13  Mammography screening  Percentage of female Medicare enrollees ages 67–69 that receive mammography screening  13  Variable  Measure  Description  Weight (%)  Length of life  Premature death  Years of potential life lost before age 75 per 100 000 population (age-adjusted)  100  Quality of life  Poor or fair health  Percentage of adults reporting fair or poor health (age-adjusted)  20  Poor physical health days  Average number of physically unhealthy days reported in past 30 days (age-adjusted)  20  Poor mental health days  Average number of mentally unhealthy days reported in past 30 days (age-adjusted)  20  Low birth weight  Percentage of live births with low birth weight (<2500 g)  40  Health behaviors  Adult smoking  Percentage of adults who are current smokers  33  Adult obesity  Percentage of adults that report a BMI of 30 or more  17  Food environment index  Index of factors that contribute to a healthy food environment  7  Physical inactivity  Percentage of adults aged 20 and over reporting no leisure-time physical activity  7  Access to exercise opportunities  Percentage of population with adequate access to locations for physical activity  3  Excessive drinking  Percentage of adults reporting binge or heavy drinking  8  Alcohol-impaired driving deaths  Percentage of driving deaths with alcohol involvement  8  Sexually transmitted infections  Number of newly diagnosed chlamydia cases per 100 000 population  8  Teen births  Teen birth rate per 1 000 female population, ages 15–19  8  Clinical care  Uninsured  Percentage of population under age 65 without health insurance  25  Primary care physicians  Ratio of population to primary care physicians  15  Dentists  Ratio of population to dentists  5  Mental health providers  Ratio of population to mental health providers  5  Preventable hospital stays  Number of hospital stays for ambulatory-care sensitive conditions per 1000 Medicare enrollees  25  Diabetic monitoring  Percentage of diabetic Medicare enrollees ages 65–75 that receive HbA1c monitoring  13  Mammography screening  Percentage of female Medicare enrollees ages 67–69 that receive mammography screening  13  The County Health Ranking dataset provides county ranks based on county quartile ranking of each measure within the state.10 However, it does not provide overall county rankings for the entire USA. As such, there are 50 counties from 50 different states, which obtained the highest rank of 1 within each category. Each category is defined separately, which provides variance between categories with regard to rank. For example, a county may achieve a rank of 1 within clinical care, but may receive a rank of 50 with regard to health behaviors. As such, this analysis focuses on each category separately to determine the associations with each category. Independent variables Two independent variables were operationalized as (i) number of services that facilitate access to care and (ii) community building services offered by hospitals in a county. Services that were implemented to facilitate access to care included enrollment assistance programs, women’s health services, teen outreach services, health screenings and health fairs, indigent care clinics, community outreach, rural health clinics and referral centers, transportation to health services and mobile health services. We classified community building oriented services as those that aim to improve particular public health outcomes and determinants in a community. These services included immunization programs, nutrition programs, tobacco treatment services, support groups, community health education, patient education centers, patient representation services and meals on wheels. The overall number of services offered by hospitals in each of these categories was aggregated to create the total number of access and community building services. Data within the AHA survey does not provide indication concerning service reach (provision to one or more county) nor frequency of services offered. In addition, the hospital community needs assessment process leaves the definition of service community to the hospital in the community needs assessment process. As such, we have identified the county in which the hospital is located as the most likely area in which provision of services is to occur and have assigned all services offered by the hospital to that county as identified through FIPS codes. Control variables To control for differing county characteristics, we used variables previously identified as important for health outcomes. The percent of the population within the county which is younger than 18 or over 65 years of age are operationalized as continuous variables representing two opposite ends of the health spectrum. That is, younger individuals are typically less burdened with chronic disease than older individuals who experience increased acute injuries.11 Additionally, racial and ethnic minority populations have also been indicated as important in the use of health services and, in particular, are identified to have disparities in access and utilization of health services.12 Thus, continuous variables indicating county population percentages of the following groups were included in the analysis: African Americans, American Indians or Alaskan Natives, Asians, Native Hawaiians or other Pacific Islanders, Hispanics and non-Hispanic Whites. We also include a continuous variable which indicates the county population percentage which is English proficient, as this provides insight into barriers to healthcare utilization and need.12 Health outcomes are also different for males and females.13 As such, the percentage of the county population that was female was included as a binary variable. Additionally, given that location (e.g. urban versus rural) is an indicator of service need and utilization, county population percentage within rural locations was included as a continuous variable.14 Furthermore, diabetes prevalence rates provide additional indication of healthcare services needed;15 thus, county diabetes prevalence rate was included as a continuous variable. Food insecurity (i.e. percent of the county population identified as food insecure) was also included as this variable provides an indication of transportation needs and is a predictor of high cost users of healthcare services.16 Finally, percent of the population with limited access to healthcare was included as a continuous variable to further define transportation and health insurance barriers to care which would indicate greater need for community services.17 Analysis To compare descriptive statistics and data included and excluded from the analysis, means, standard deviations and Kruska–Wallis tests were employed. The researchers used a zero truncated negative binomial regression model to explore the bivariate relationship between the outcome, independent and control variables. Negative binomial and Poisson regression were both considered because the dependent variable is count data. Over-dispersion and zero truncation were assessed for each outcome variable. All dependent variables demonstrated over-dispersion (variance greater than the mean) and require the use of negative binomial regression, which corrects for this excess variability in the count data.18 Since there are no counties that receive a zero rank, a zero truncation adjustment was applied to the models.18 Finally, an exposure variable for the number of counties within each State was utilized to control for these differences. STATA 14 was used to run all analyses, models were estimated though maximum likelihood, and P-values were calculated based off of the Z-statistic. Incident rate ratios, standard errors, Z-statistics and significance level are reported. Results Overall, there are 3 156 counties included in the dataset. Due to missing data in the County Health Rankings dataset, a total of 97 counties were excluded. Medians and interquartile ranges of rank data and means and standard deviations of county population characteristics are reported in Table 2. Within the sample, the mean percent of those younger than 18 across all counties was 22.74%, while the percent of individuals aged 65 and over was 17.05%. The majority of the sample were non-Hispanic Whites (77.28%) while African Americans and Hispanics consisted of 9.08 and 8.80%, respectively. Individuals of other ethnicities consisted of <2.5% of the population. Over 98% of the sample was proficient in English. Half of the sample (49.97%) were female. Table 2 Descriptive statistics Dependent variables  Included in analysis  Excluded from analysis    Median  Interquartile range (25–75%)  Pop  Median  Interquartile range (25–75%)  Pop  Significance  Life length  38  (17–65)  3059  65  (59–85)  3  –  Health behaviors  38  (17–65)  3059  71  (63–94)  3  –  Quality of life  38  (17–65)  3059  74  (58–94)  3  –  Clinical care  38  (17–65)  3059  67  (16–79)  3  –  Independent variables  Means  Standard deviation  Pop  Means  Standard deviation  Pop  Significance  Number of community access services  27.88  29.74  3059  20.14  25.89  97  ***  Number of community building services  33.05  35.91  3059  22.71  28.19  97  ***  County population per million  0.10  0.33  3059  0.00  0.00  82  ***  Percent not for profit  0.46  0.32  3059  0.44  0.34  97    Percent of the population younger than 18  22.74  3.35  3059  20.75  4.21  82  ***  Percent of population 65 or over  17.05  4.30  3059  21.28  5.28  82  ***  Percent of population African American  9.08  14.38  3059  1.77  7.93  82  ***  Percent of population American Indian or Alaskan Native  2.21  7.64  3059  2.63  6.31  82  ***  Percent of population Asian  1.35  2.59  3059  1.06  4.67  82  ***  Percent of population Native Hawaiian or Other Pacific Islander  0.11  0.41  3059  0.65  5.27  82  ***  Percent of population Hispanic  8.80  13.35  3059  10.35  15.35  82    Percent of population non-Hispanic White  77.28  19.82  3059  82.12  20.92  82  ***  Percent of population not proficient in English  1.81  2.91  3059  2.10  4.15  82  ***  Percent of population female  49.97  2.22  3059  48.57  2.85  82  ***  Percent of population in rural location  57.53  31.19  3059  99.34  4.39  82  ***  Percent of the population diabetic  11.00  2.32  3059  10.21  1.92  82  ***  Percent of the population which is food insecure  14.79  3.85  3059  12.85  3.63  82  ***  Percent of the population with limited access to healthcare  7.92  7.47  3059  25.59  13.95  82  ***  Dependent variables  Included in analysis  Excluded from analysis    Median  Interquartile range (25–75%)  Pop  Median  Interquartile range (25–75%)  Pop  Significance  Life length  38  (17–65)  3059  65  (59–85)  3  –  Health behaviors  38  (17–65)  3059  71  (63–94)  3  –  Quality of life  38  (17–65)  3059  74  (58–94)  3  –  Clinical care  38  (17–65)  3059  67  (16–79)  3  –  Independent variables  Means  Standard deviation  Pop  Means  Standard deviation  Pop  Significance  Number of community access services  27.88  29.74  3059  20.14  25.89  97  ***  Number of community building services  33.05  35.91  3059  22.71  28.19  97  ***  County population per million  0.10  0.33  3059  0.00  0.00  82  ***  Percent not for profit  0.46  0.32  3059  0.44  0.34  97    Percent of the population younger than 18  22.74  3.35  3059  20.75  4.21  82  ***  Percent of population 65 or over  17.05  4.30  3059  21.28  5.28  82  ***  Percent of population African American  9.08  14.38  3059  1.77  7.93  82  ***  Percent of population American Indian or Alaskan Native  2.21  7.64  3059  2.63  6.31  82  ***  Percent of population Asian  1.35  2.59  3059  1.06  4.67  82  ***  Percent of population Native Hawaiian or Other Pacific Islander  0.11  0.41  3059  0.65  5.27  82  ***  Percent of population Hispanic  8.80  13.35  3059  10.35  15.35  82    Percent of population non-Hispanic White  77.28  19.82  3059  82.12  20.92  82  ***  Percent of population not proficient in English  1.81  2.91  3059  2.10  4.15  82  ***  Percent of population female  49.97  2.22  3059  48.57  2.85  82  ***  Percent of population in rural location  57.53  31.19  3059  99.34  4.39  82  ***  Percent of the population diabetic  11.00  2.32  3059  10.21  1.92  82  ***  Percent of the population which is food insecure  14.79  3.85  3059  12.85  3.63  82  ***  Percent of the population with limited access to healthcare  7.92  7.47  3059  25.59  13.95  82  ***  ***Significant at P = 0.001. **Significant at P = 0.01. *Significant at P = 0.05. Table 2 Descriptive statistics Dependent variables  Included in analysis  Excluded from analysis    Median  Interquartile range (25–75%)  Pop  Median  Interquartile range (25–75%)  Pop  Significance  Life length  38  (17–65)  3059  65  (59–85)  3  –  Health behaviors  38  (17–65)  3059  71  (63–94)  3  –  Quality of life  38  (17–65)  3059  74  (58–94)  3  –  Clinical care  38  (17–65)  3059  67  (16–79)  3  –  Independent variables  Means  Standard deviation  Pop  Means  Standard deviation  Pop  Significance  Number of community access services  27.88  29.74  3059  20.14  25.89  97  ***  Number of community building services  33.05  35.91  3059  22.71  28.19  97  ***  County population per million  0.10  0.33  3059  0.00  0.00  82  ***  Percent not for profit  0.46  0.32  3059  0.44  0.34  97    Percent of the population younger than 18  22.74  3.35  3059  20.75  4.21  82  ***  Percent of population 65 or over  17.05  4.30  3059  21.28  5.28  82  ***  Percent of population African American  9.08  14.38  3059  1.77  7.93  82  ***  Percent of population American Indian or Alaskan Native  2.21  7.64  3059  2.63  6.31  82  ***  Percent of population Asian  1.35  2.59  3059  1.06  4.67  82  ***  Percent of population Native Hawaiian or Other Pacific Islander  0.11  0.41  3059  0.65  5.27  82  ***  Percent of population Hispanic  8.80  13.35  3059  10.35  15.35  82    Percent of population non-Hispanic White  77.28  19.82  3059  82.12  20.92  82  ***  Percent of population not proficient in English  1.81  2.91  3059  2.10  4.15  82  ***  Percent of population female  49.97  2.22  3059  48.57  2.85  82  ***  Percent of population in rural location  57.53  31.19  3059  99.34  4.39  82  ***  Percent of the population diabetic  11.00  2.32  3059  10.21  1.92  82  ***  Percent of the population which is food insecure  14.79  3.85  3059  12.85  3.63  82  ***  Percent of the population with limited access to healthcare  7.92  7.47  3059  25.59  13.95  82  ***  Dependent variables  Included in analysis  Excluded from analysis    Median  Interquartile range (25–75%)  Pop  Median  Interquartile range (25–75%)  Pop  Significance  Life length  38  (17–65)  3059  65  (59–85)  3  –  Health behaviors  38  (17–65)  3059  71  (63–94)  3  –  Quality of life  38  (17–65)  3059  74  (58–94)  3  –  Clinical care  38  (17–65)  3059  67  (16–79)  3  –  Independent variables  Means  Standard deviation  Pop  Means  Standard deviation  Pop  Significance  Number of community access services  27.88  29.74  3059  20.14  25.89  97  ***  Number of community building services  33.05  35.91  3059  22.71  28.19  97  ***  County population per million  0.10  0.33  3059  0.00  0.00  82  ***  Percent not for profit  0.46  0.32  3059  0.44  0.34  97    Percent of the population younger than 18  22.74  3.35  3059  20.75  4.21  82  ***  Percent of population 65 or over  17.05  4.30  3059  21.28  5.28  82  ***  Percent of population African American  9.08  14.38  3059  1.77  7.93  82  ***  Percent of population American Indian or Alaskan Native  2.21  7.64  3059  2.63  6.31  82  ***  Percent of population Asian  1.35  2.59  3059  1.06  4.67  82  ***  Percent of population Native Hawaiian or Other Pacific Islander  0.11  0.41  3059  0.65  5.27  82  ***  Percent of population Hispanic  8.80  13.35  3059  10.35  15.35  82    Percent of population non-Hispanic White  77.28  19.82  3059  82.12  20.92  82  ***  Percent of population not proficient in English  1.81  2.91  3059  2.10  4.15  82  ***  Percent of population female  49.97  2.22  3059  48.57  2.85  82  ***  Percent of population in rural location  57.53  31.19  3059  99.34  4.39  82  ***  Percent of the population diabetic  11.00  2.32  3059  10.21  1.92  82  ***  Percent of the population which is food insecure  14.79  3.85  3059  12.85  3.63  82  ***  Percent of the population with limited access to healthcare  7.92  7.47  3059  25.59  13.95  82  ***  ***Significant at P = 0.001. **Significant at P = 0.01. *Significant at P = 0.05. There were more rural counties (57.53%) than urban counties, and 11% of the sample were diabetic. Other characteristics include 14.79% being food insecure and 7.92% having limited access to healthcare services. Median county rankings for health ranking factors were 38 for life length, health behaviors, quality of life and clinical care. The mean number of community healthcare access services across all counties was 27.88 (SD = 29.74) and 33.05 (SD = 35.91) for community building services. Furthermore, the mean number of community healthcare access services was 20.14 (SD = 25.89) and 22.71 (SD = 28.19) for community building services. In the sample that was excluded from the analysis, the median county rankings for health ranking factors were 65 for life length, 71 for health behaviors, 74 for quality of life and 67 clinical care, which was consistently higher than the included study sample. Table 3 displays the results of the bivariate analyses. These results reveal significant associations between each of our key outcome variables and number of community healthcare access services and building services per million population. There is a lack of association between percent not-for-profit hospitals within the county and the county rankings. In addition, county population per million, percent of the population younger than 18, percent of the population African American, Native Hawaiian or Alaskan Native, Hispanic, not proficient in English and female are not consistently associated with the county ranking independent variables. Table 3 Bivariate analysis county rankings and independent variables   Life length  Behaviors  Quality of life  Clinical care  Coef. (Std)  P-value  Coef. (Std)  P-value  Coef. (Std)  P-value  Coef. (Std)  P-value  Number of community access services  −0.004 (0.000)  0.000  −0.004 (0.000)  0.000  −0.001 (0.000)  0.006  −0.003 (0.000)  0.000  Number of community building services  −0.003 (0.000)  0.000  −0.003 (0.000)  0.000  −0.001 (0.000)  0.000  −0.003 (0.000)  0.000  County population per million  −0.241 (0.032)  0.000  −0.280 (0.035)  0.000  −0.025 (0.034)  0.457  −0.138 (0.026)  0.000  Percent not for profit  −0.040 (0.039)  0.302  −0.033 (0.039)  0.396  −0.075 (0.040)  0.057  −0.025 (0.039)  0.516  Percent of the population younger than 18  −0.12 (0.004)  0.001  0.007 (0.004)  0.061  −0.005 (0.004)  0.162  0.008 (0.004)  0.021  Percent of population 65 or over  0.037 (0.003)  0.000  0.008 (0.003)  0.010  0.008 (0.003)  0.007  0.021 (0.003)  0.000  Percent of population African American  0.005 (0.001)  0.000  0.006 (0.001)  0.000  0.007 (0.001)  0.000  −0.000 (0.001)  0.963  Percent of population American Indian or Alaskan Native  0.009 (0.002)  0.000  0.009 (0.002)  0.000  0.005 (0.002)  0.002  0.009 (0.002)  0.000  Percent of population Asian  −0.075 (0.005)  0.000  −0.073 (0.005)  0.000  −0.03 (0.005)  0.000  −0.061 (0.005)  0.000  Percent of population Native Hawaiian or other Pacific Islander  −0.129 (0.50)  0.011  −0.052 (0.044)  0.238  −0.072 (0.045)  0.112  −0.05 (0.044)  0.213  Percent of population Hispanic  −0.005 (0.001)  0.000  −0.001 (0.001)  0.131  0.000 (0.001)  0.760  0.003 (0.001)  0.001  Percent of population non-Hispanic White  −0.001 (0.001)  0.075  −0.004 (0.001)  0.000  −0.004 (0.001)  0.000  0.002 (0.001)  0.002  Percent of population not proficient in English  −0.039 (0.004)  0.000  −0.014 (0.004)  0.001  0.006 (0.004)  0.155  0.018 (0.004)  0.000  Percent of population female  0.016 (0.006)  0.004  −0.024 (0.006)  0.000  −0.007 (0.006)  0.197  −0.043 (0.006)  0.000  Percent of population in rural location  0.005 (0.000)  0.000  0.004 (0.000)  0.000  0.001 (0.000)  0.001  0.008 (0.000)  0.000  Percent of the population diabetic  0.103 (0.005)  0.000  0.093 (0.006)  0.000  0.077 (0.005)  0.000  0.073 (0.006)  0.000  Percent of the population which is food insecure  0.050 (0.003)  0.000  0.049 (0.003)  0.000  0.049 (0.003)  0.000  0.028 (0.003)  0.000  Percent of the population with limited access to healthcare  0.006 (0.002)  0.000  0.011 (0.002)  0.000  0.006 (0.002)  0.000  0.006 (0.002)  0.000    Life length  Behaviors  Quality of life  Clinical care  Coef. (Std)  P-value  Coef. (Std)  P-value  Coef. (Std)  P-value  Coef. (Std)  P-value  Number of community access services  −0.004 (0.000)  0.000  −0.004 (0.000)  0.000  −0.001 (0.000)  0.006  −0.003 (0.000)  0.000  Number of community building services  −0.003 (0.000)  0.000  −0.003 (0.000)  0.000  −0.001 (0.000)  0.000  −0.003 (0.000)  0.000  County population per million  −0.241 (0.032)  0.000  −0.280 (0.035)  0.000  −0.025 (0.034)  0.457  −0.138 (0.026)  0.000  Percent not for profit  −0.040 (0.039)  0.302  −0.033 (0.039)  0.396  −0.075 (0.040)  0.057  −0.025 (0.039)  0.516  Percent of the population younger than 18  −0.12 (0.004)  0.001  0.007 (0.004)  0.061  −0.005 (0.004)  0.162  0.008 (0.004)  0.021  Percent of population 65 or over  0.037 (0.003)  0.000  0.008 (0.003)  0.010  0.008 (0.003)  0.007  0.021 (0.003)  0.000  Percent of population African American  0.005 (0.001)  0.000  0.006 (0.001)  0.000  0.007 (0.001)  0.000  −0.000 (0.001)  0.963  Percent of population American Indian or Alaskan Native  0.009 (0.002)  0.000  0.009 (0.002)  0.000  0.005 (0.002)  0.002  0.009 (0.002)  0.000  Percent of population Asian  −0.075 (0.005)  0.000  −0.073 (0.005)  0.000  −0.03 (0.005)  0.000  −0.061 (0.005)  0.000  Percent of population Native Hawaiian or other Pacific Islander  −0.129 (0.50)  0.011  −0.052 (0.044)  0.238  −0.072 (0.045)  0.112  −0.05 (0.044)  0.213  Percent of population Hispanic  −0.005 (0.001)  0.000  −0.001 (0.001)  0.131  0.000 (0.001)  0.760  0.003 (0.001)  0.001  Percent of population non-Hispanic White  −0.001 (0.001)  0.075  −0.004 (0.001)  0.000  −0.004 (0.001)  0.000  0.002 (0.001)  0.002  Percent of population not proficient in English  −0.039 (0.004)  0.000  −0.014 (0.004)  0.001  0.006 (0.004)  0.155  0.018 (0.004)  0.000  Percent of population female  0.016 (0.006)  0.004  −0.024 (0.006)  0.000  −0.007 (0.006)  0.197  −0.043 (0.006)  0.000  Percent of population in rural location  0.005 (0.000)  0.000  0.004 (0.000)  0.000  0.001 (0.000)  0.001  0.008 (0.000)  0.000  Percent of the population diabetic  0.103 (0.005)  0.000  0.093 (0.006)  0.000  0.077 (0.005)  0.000  0.073 (0.006)  0.000  Percent of the population which is food insecure  0.050 (0.003)  0.000  0.049 (0.003)  0.000  0.049 (0.003)  0.000  0.028 (0.003)  0.000  Percent of the population with limited access to healthcare  0.006 (0.002)  0.000  0.011 (0.002)  0.000  0.006 (0.002)  0.000  0.006 (0.002)  0.000  Table 3 Bivariate analysis county rankings and independent variables   Life length  Behaviors  Quality of life  Clinical care  Coef. (Std)  P-value  Coef. (Std)  P-value  Coef. (Std)  P-value  Coef. (Std)  P-value  Number of community access services  −0.004 (0.000)  0.000  −0.004 (0.000)  0.000  −0.001 (0.000)  0.006  −0.003 (0.000)  0.000  Number of community building services  −0.003 (0.000)  0.000  −0.003 (0.000)  0.000  −0.001 (0.000)  0.000  −0.003 (0.000)  0.000  County population per million  −0.241 (0.032)  0.000  −0.280 (0.035)  0.000  −0.025 (0.034)  0.457  −0.138 (0.026)  0.000  Percent not for profit  −0.040 (0.039)  0.302  −0.033 (0.039)  0.396  −0.075 (0.040)  0.057  −0.025 (0.039)  0.516  Percent of the population younger than 18  −0.12 (0.004)  0.001  0.007 (0.004)  0.061  −0.005 (0.004)  0.162  0.008 (0.004)  0.021  Percent of population 65 or over  0.037 (0.003)  0.000  0.008 (0.003)  0.010  0.008 (0.003)  0.007  0.021 (0.003)  0.000  Percent of population African American  0.005 (0.001)  0.000  0.006 (0.001)  0.000  0.007 (0.001)  0.000  −0.000 (0.001)  0.963  Percent of population American Indian or Alaskan Native  0.009 (0.002)  0.000  0.009 (0.002)  0.000  0.005 (0.002)  0.002  0.009 (0.002)  0.000  Percent of population Asian  −0.075 (0.005)  0.000  −0.073 (0.005)  0.000  −0.03 (0.005)  0.000  −0.061 (0.005)  0.000  Percent of population Native Hawaiian or other Pacific Islander  −0.129 (0.50)  0.011  −0.052 (0.044)  0.238  −0.072 (0.045)  0.112  −0.05 (0.044)  0.213  Percent of population Hispanic  −0.005 (0.001)  0.000  −0.001 (0.001)  0.131  0.000 (0.001)  0.760  0.003 (0.001)  0.001  Percent of population non-Hispanic White  −0.001 (0.001)  0.075  −0.004 (0.001)  0.000  −0.004 (0.001)  0.000  0.002 (0.001)  0.002  Percent of population not proficient in English  −0.039 (0.004)  0.000  −0.014 (0.004)  0.001  0.006 (0.004)  0.155  0.018 (0.004)  0.000  Percent of population female  0.016 (0.006)  0.004  −0.024 (0.006)  0.000  −0.007 (0.006)  0.197  −0.043 (0.006)  0.000  Percent of population in rural location  0.005 (0.000)  0.000  0.004 (0.000)  0.000  0.001 (0.000)  0.001  0.008 (0.000)  0.000  Percent of the population diabetic  0.103 (0.005)  0.000  0.093 (0.006)  0.000  0.077 (0.005)  0.000  0.073 (0.006)  0.000  Percent of the population which is food insecure  0.050 (0.003)  0.000  0.049 (0.003)  0.000  0.049 (0.003)  0.000  0.028 (0.003)  0.000  Percent of the population with limited access to healthcare  0.006 (0.002)  0.000  0.011 (0.002)  0.000  0.006 (0.002)  0.000  0.006 (0.002)  0.000    Life length  Behaviors  Quality of life  Clinical care  Coef. (Std)  P-value  Coef. (Std)  P-value  Coef. (Std)  P-value  Coef. (Std)  P-value  Number of community access services  −0.004 (0.000)  0.000  −0.004 (0.000)  0.000  −0.001 (0.000)  0.006  −0.003 (0.000)  0.000  Number of community building services  −0.003 (0.000)  0.000  −0.003 (0.000)  0.000  −0.001 (0.000)  0.000  −0.003 (0.000)  0.000  County population per million  −0.241 (0.032)  0.000  −0.280 (0.035)  0.000  −0.025 (0.034)  0.457  −0.138 (0.026)  0.000  Percent not for profit  −0.040 (0.039)  0.302  −0.033 (0.039)  0.396  −0.075 (0.040)  0.057  −0.025 (0.039)  0.516  Percent of the population younger than 18  −0.12 (0.004)  0.001  0.007 (0.004)  0.061  −0.005 (0.004)  0.162  0.008 (0.004)  0.021  Percent of population 65 or over  0.037 (0.003)  0.000  0.008 (0.003)  0.010  0.008 (0.003)  0.007  0.021 (0.003)  0.000  Percent of population African American  0.005 (0.001)  0.000  0.006 (0.001)  0.000  0.007 (0.001)  0.000  −0.000 (0.001)  0.963  Percent of population American Indian or Alaskan Native  0.009 (0.002)  0.000  0.009 (0.002)  0.000  0.005 (0.002)  0.002  0.009 (0.002)  0.000  Percent of population Asian  −0.075 (0.005)  0.000  −0.073 (0.005)  0.000  −0.03 (0.005)  0.000  −0.061 (0.005)  0.000  Percent of population Native Hawaiian or other Pacific Islander  −0.129 (0.50)  0.011  −0.052 (0.044)  0.238  −0.072 (0.045)  0.112  −0.05 (0.044)  0.213  Percent of population Hispanic  −0.005 (0.001)  0.000  −0.001 (0.001)  0.131  0.000 (0.001)  0.760  0.003 (0.001)  0.001  Percent of population non-Hispanic White  −0.001 (0.001)  0.075  −0.004 (0.001)  0.000  −0.004 (0.001)  0.000  0.002 (0.001)  0.002  Percent of population not proficient in English  −0.039 (0.004)  0.000  −0.014 (0.004)  0.001  0.006 (0.004)  0.155  0.018 (0.004)  0.000  Percent of population female  0.016 (0.006)  0.004  −0.024 (0.006)  0.000  −0.007 (0.006)  0.197  −0.043 (0.006)  0.000  Percent of population in rural location  0.005 (0.000)  0.000  0.004 (0.000)  0.000  0.001 (0.000)  0.001  0.008 (0.000)  0.000  Percent of the population diabetic  0.103 (0.005)  0.000  0.093 (0.006)  0.000  0.077 (0.005)  0.000  0.073 (0.006)  0.000  Percent of the population which is food insecure  0.050 (0.003)  0.000  0.049 (0.003)  0.000  0.049 (0.003)  0.000  0.028 (0.003)  0.000  Percent of the population with limited access to healthcare  0.006 (0.002)  0.000  0.011 (0.002)  0.000  0.006 (0.002)  0.000  0.006 (0.002)  0.000  Table 4 displays the zero-truncated negative binomial regression analysis results. We found no statistically significant differences between the number of community healthcare access services provided and the county’s rank of length of life, quality of life, nor clinical care. However, for every one unit increase in the number of community healthcare access services provided, the expected county ranking of health behaviors is multiplied by a factor of 0.999 (i.e. it decreases) while holding all other variables in the model constant. Similarly, we found no statistically significant differences between the number of community building services provided and the county’s rank of length of life, quality of life, nor clinical care. However, for every one unit increase in the number of community building services provided, the expected county ranking of health behaviors is multiplied by a factor of 0.999 (that is, it decreases) while holding all other variables in the model constant. Table 4 Zero truncated negative binomial regression for county rank by community access services offered by hospitals   Life length  Behaviors  Quality of life  Clinical care  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  Number of community access services  0.999 (0.000)  −1.63 (0.103)  0.999 (0.000)  −2.19 (0.028)  1.00 (0.000)  0.14 (0.885)  1.00 (0.000)  −0.96 (0.336)  Log likelihood  −13 640.327  −13 699.256  −13 826.073  −13 534.151  LR Chi2  716.74  609.37  355.97  942.67  Prob > Chi2  0.0000  0.0000  0.0000  0.0000  Psuedo R2  0.0256  0.0218  0.0127  0.0337    Life length  Behaviors  Quality of life  Clinical care    IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  Number of community building services  0.999 (0.000)  −2.01 (0.045)  0.999 (0.000)  −2.62 (0.009)  1.00 (0.000)  −0.24 (0.811)  1.00 (0.000)  −1.05 (0.295)  Log likelihood  −13 639.653  −13 698.263  −13 826.055  −13 534.066  LR chi2  718.08  611.36  356.01  942.84  Prob > chi2  0.0000  0.0000  0.0000  0.0000  Psuedo R2  0.0256  0.0218  0.0127  0.0337    Life length  Behaviors  Quality of life  Clinical care  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  Number of community access services  0.999 (0.000)  −1.63 (0.103)  0.999 (0.000)  −2.19 (0.028)  1.00 (0.000)  0.14 (0.885)  1.00 (0.000)  −0.96 (0.336)  Log likelihood  −13 640.327  −13 699.256  −13 826.073  −13 534.151  LR Chi2  716.74  609.37  355.97  942.67  Prob > Chi2  0.0000  0.0000  0.0000  0.0000  Psuedo R2  0.0256  0.0218  0.0127  0.0337    Life length  Behaviors  Quality of life  Clinical care    IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  Number of community building services  0.999 (0.000)  −2.01 (0.045)  0.999 (0.000)  −2.62 (0.009)  1.00 (0.000)  −0.24 (0.811)  1.00 (0.000)  −1.05 (0.295)  Log likelihood  −13 639.653  −13 698.263  −13 826.055  −13 534.066  LR chi2  718.08  611.36  356.01  942.84  Prob > chi2  0.0000  0.0000  0.0000  0.0000  Psuedo R2  0.0256  0.0218  0.0127  0.0337  Note: Each model controlled for the following: county population per million, percent not for profit, percent of the population younger than 18, percent of population 65 or over, percent of population African American, percent of population American Indian or Alaskan Native, percent of population Asian, percent of population Native Hawaiian or Other Pacific Islander, percent of population Hispanic, percent of population non-Hispanic White, percent of population not proficient in English, percent of population female, percent of population in rural location, percent of the population diabetic, percent of the population which is food insecure, percent of the population with limited access to healthcare. Table 4 Zero truncated negative binomial regression for county rank by community access services offered by hospitals   Life length  Behaviors  Quality of life  Clinical care  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  Number of community access services  0.999 (0.000)  −1.63 (0.103)  0.999 (0.000)  −2.19 (0.028)  1.00 (0.000)  0.14 (0.885)  1.00 (0.000)  −0.96 (0.336)  Log likelihood  −13 640.327  −13 699.256  −13 826.073  −13 534.151  LR Chi2  716.74  609.37  355.97  942.67  Prob > Chi2  0.0000  0.0000  0.0000  0.0000  Psuedo R2  0.0256  0.0218  0.0127  0.0337    Life length  Behaviors  Quality of life  Clinical care    IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  Number of community building services  0.999 (0.000)  −2.01 (0.045)  0.999 (0.000)  −2.62 (0.009)  1.00 (0.000)  −0.24 (0.811)  1.00 (0.000)  −1.05 (0.295)  Log likelihood  −13 639.653  −13 698.263  −13 826.055  −13 534.066  LR chi2  718.08  611.36  356.01  942.84  Prob > chi2  0.0000  0.0000  0.0000  0.0000  Psuedo R2  0.0256  0.0218  0.0127  0.0337    Life length  Behaviors  Quality of life  Clinical care  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  Number of community access services  0.999 (0.000)  −1.63 (0.103)  0.999 (0.000)  −2.19 (0.028)  1.00 (0.000)  0.14 (0.885)  1.00 (0.000)  −0.96 (0.336)  Log likelihood  −13 640.327  −13 699.256  −13 826.073  −13 534.151  LR Chi2  716.74  609.37  355.97  942.67  Prob > Chi2  0.0000  0.0000  0.0000  0.0000  Psuedo R2  0.0256  0.0218  0.0127  0.0337    Life length  Behaviors  Quality of life  Clinical care    IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  Number of community building services  0.999 (0.000)  −2.01 (0.045)  0.999 (0.000)  −2.62 (0.009)  1.00 (0.000)  −0.24 (0.811)  1.00 (0.000)  −1.05 (0.295)  Log likelihood  −13 639.653  −13 698.263  −13 826.055  −13 534.066  LR chi2  718.08  611.36  356.01  942.84  Prob > chi2  0.0000  0.0000  0.0000  0.0000  Psuedo R2  0.0256  0.0218  0.0127  0.0337  Note: Each model controlled for the following: county population per million, percent not for profit, percent of the population younger than 18, percent of population 65 or over, percent of population African American, percent of population American Indian or Alaskan Native, percent of population Asian, percent of population Native Hawaiian or Other Pacific Islander, percent of population Hispanic, percent of population non-Hispanic White, percent of population not proficient in English, percent of population female, percent of population in rural location, percent of the population diabetic, percent of the population which is food insecure, percent of the population with limited access to healthcare. Discussion Main findings of this study As part of the implementation of the ACA, hospitals across the nation, regardless of tax status, offer an array of healthcare access services and health interventions. Our study findings confirm this and identified that such hospital services do not necessarily improve population level health outcomes. We found either neutral or negative relationships between hospital-sponsored community services and county level health rankings, suggesting an opportunity for hospitals to better operationalize services to improve health outcome. What is already known on the topic? Expanding the role of hospitals in keeping communities healthy and preventing disease has been advocated for on a national stage through the implementation of the ACA in 2010. However, the extent to which hospital community services impact population outcomes is not fully understood. What this study adds The current study represents one of few studies that quantifies the association of hospital sponsored community services and population health outcomes and factors in the USA. Past literature has mainly focused on dimensions of effective hospital partnerships aimed at community health improvement6,19 or on the impact of state community benefit laws and hospital ownership as it relates to health promotion services as noted by a review of the literature by Shortell et al.20 The statistically significant associations identified in our study were small in terms of effect-size; this reveals the lack of overall influence that hospital community health services have on population level. However, such findings present an opportunity for hospitals to increase the impact of community health services, especially for rural and historically marginalized communities. None of the tested independent variables reached statistical significance. Therefore, the number of hospital-sponsored community services offered may not be as important for community health improvement as service quality and reach. Our findings support previous research21 that calls for action from the Federal government to implement a public health framework where non-profit hospitals are charged to work more closely with local actors such as health departments to support ongoing community health initiatives as there may be a misalignment between hospital community services offered and community health needs. This is particularly relevant given that non-profit hospitals are not required to maintain yearly community improvement plans but rather at 3-year intervals. Estimates reported in 2008 suggest that non-profit hospitals tax exemptions yielded ~13 billion dollars;22 however, the extent to which these funds have facilitated community improvements at a population level remains unclear. To maximize the quality and reach of community services, health professionals should collaborate with key stakeholders and community-based entities (e.g. community organizations and coalitions). Such collaborations are most effective when partnerships are established in the planning stages of service or program implementation. These strategies can be particularly useful in communities that experience health inequities as collaborations can help facilitate the establishment of trusting relationships and secure commitments, while overcoming challenges related to buy-in and sustainability.23 Health professionals implementing community health services may also increase health benefits to communities by employing combination and multilevel interventions.24 Combination interventions are programs designed to address health issues at the intersection of social, behavioral and biological factors. In a complimentary manner, multilevel interventions are ‘implemented at the individual, physician, clinic, healthcare organization and/or community level.’25 These approaches have been cited as effective strategies for achieving population level changes in health outcomes while providing health professionals the opportunity to address factors in disease prevention and treatment. Limitations of this study While contributing to the larger literature, this study is not without limitations. The study used a cross-sectional design limited to 1 year, inhibiting the ability to determine causal relationships. Next, the data utilized for this study has limitations worth considering. Currently, there is no comprehensive national database that captures the extent of hospital community health-service data. Therefore, we used multiple databases. The AHA survey relies heavily on self-reported data acquired by asking hospitals to complete the survey, and may present issues of bias and/or social desirability. In addition, the attribution of community services to the county which the hospital resides may be incomplete. Currently, hospitals are able to define the community they serve for the purposes of community benefit, and there is a lack of consistency concerning the duration, intensity, and influence each service provides. Thus, there is great variation in the types of services offered by each hospital. However, inclusion of the county the hospital is located in provided the most consistent and logical method for attribution. The AHRF and County Health Rankings datasets provide a combination of a large number of externally collected datasets. This provides an increase in the likelihood of error and the manipulation of data either through conversion or through defining rankings may arbitrarily change the data or misrepresent county level attributions. Furthermore, the County Health Ranking methods provide a method for creating ranked data, however should other methods be utilized, the results of this study may change. We were unable to control for the potential that services offered occur on county lines or that all community services are the same thus may provide variance unaccounted for within this study. Finally, the current study focuses more directly on the number and types of services offered as opposed to how much time and financial resources the organization spends in delivering those services. Future studies should assess the amount of financial capital an organization spends on providing community services, the duration and intensity of those services, as well as the provision of those services to specific counties or other defined services areas. Additionally, future research may focus on accessing community members’ perceptions of service delivery. These lines of inquiry would provide additional insight into the quality and quantity of services offered as well as perceived barriers and facilitators to service utilization. Our results support previous research that posits hospitals have room for improvement with regard to translating findings from CHNAs into services that better suit the surrounding communities’ needs.26 Furthermore, our findings highlight that hospitals should not be solely relied upon for community health improvement. Therefore, future research should examine the impact of other organizations that are prevalent in communities, such as schools and clinics, to determine the degree to which they impact population health outcomes. These findings indicate a neutral or negative association between length of life, health behaviors, quality of life and clinical care for the general population based on the number of community healthcare access services and community building services offered by hospitals. Although these findings are only first steps towards explicating the population level health benefit of hospital sponsored community services, the results provide indication for the need to better measure variables associated delivery and quality of community health services. Conflicts of interest All authors declare no conflict of interest. References 1 OCED. OCED Better Life Index United States. 2016. Accessed 7 June 2017, 2017. 2 Braveman PA, Cubbin C, Egerter S et al.  . Socioeconomic disparities in health in the United States: what the patterns tell us. Am J Public Health  2010; 100( Suppl 1): S186– 96. Google Scholar CrossRef Search ADS PubMed  3 Lasser KE, Himmelstein DU, Woolhandler S. Access to care, health status, and health disparities in the United States and Canada: results of a cross-national population-based survey. Am J Public Health  2006; 96( 7): 1300– 7. Google Scholar CrossRef Search ADS PubMed  4 United States Department of Health and Human Services. Social Determinants. 2017. https://www.healthypeople.gov/2020/topics-objectives/topic/social-determinants-of-health. 5 Marmot MG, Bell R. Action on health disparities in the United States: commission on social determinants of health. J Am Med Assoc  2009; 301( 11): 1169– 71. Google Scholar CrossRef Search ADS   6 Scutchfield FD, Prybil L, Kelly AV et al.  . Public health and hospitals: lessons learned from partnerships in a changing health care environment. Am J Public Health  2016; 106( 1): 45– 8. Google Scholar CrossRef Search ADS PubMed  7 HHS. Area Health Resource Files (AHRF). 2016. http://ahrf.hrsa.gov/overview.htm. (accessed 12 December 2016). 8 AHA. About. 2016; http://www.ahadataviewer.com/about/. (accessed 12 November 2016). 9 Remington PL, Booske BC. Measuring the health of communities—how and why? J Public Health Manag Pract  2011; 17( 5): 397– 400. Google Scholar CrossRef Search ADS PubMed  10 County Health Rankings. The Ranking Methods. http://www.countyhealthrankings.org/ranking-methods 11 Dall TM, Gallo PD, Chakrabarti R et al.  . An aging population and growing disease burden will require a large and specialized health care workforce by 2025. Health Aff  2013; 32( 11): 2013– 20. Google Scholar CrossRef Search ADS   12 Fiscella K, Franks P, Doescher MP et al.  . Disparities in health care by race, ethnicity, and language among the insured: findings from a national sample. Med Care  2002; 40( 1): 52– 9. Google Scholar CrossRef Search ADS PubMed  13 Bertakis KD, Azari R, Helms LJ et al.  . Gender differences in the utilization of health care services. J Fam Pract  2000; 49( 2): 147– 7. Google Scholar PubMed  14 Laditka JN, Laditka SB, Probst JC. Health care access in rural areas: evidence that hospitalization for ambulatory care-sensitive conditions in the United States may increase with the level of rurality. Health Place  2009; 15( 3): 761– 70. Google Scholar CrossRef Search ADS PubMed  15 Norris SL, Nichols PJ, Caspersen CJ et al.  . The effectiveness of disease and case management for people with diabetes: a systematic review. Am J Prev Med  2002; 22( 4, Supplement 1): 15– 38. Google Scholar CrossRef Search ADS PubMed  16 Fitzpatrick T, Rosella LC, Calzavara A et al.  . Looking beyond income and education: socioeconomic status gradients among future high-cost users of health care. Am J Prev Med  2015; 49( 2): 161– 71. Google Scholar CrossRef Search ADS PubMed  17 Levesque J-F, Harris MF, Russell G. Patient-centred access to health care: conceptualising access at the interface of health systems and populations. Int J Equity Health  2013; 12( 1): 18. Google Scholar CrossRef Search ADS PubMed  18 Long JS, Freese J. Regression Models for Categorical Dependent Variables Using Stata . College Station, TX: StataCorp LP, 2006. 19 Prybil L, Scutchfield FD, Killian R et al.  . Improving Community Health Throught Hospital-Pubic Health Collaboration . Lexington: Commonwealth Center for Governance Studies, 2014. 20 Shortell S, Washington P, Baxter R. The contribution of hospitals and health care systems to community. Ann Rev Public Health  2009; 30: 373– 83. Google Scholar CrossRef Search ADS   21 Pennel CL, McLeroy KR, Burdine JN et al.  . Nonprofit hospitals’ approach to community health needs assessment. Am J Public Health  2015; 105( 3): e103– 13. Google Scholar CrossRef Search ADS PubMed  22 Government Accountability Office. Nonprofit Hospitals: Variation in Standards and Guidance Limits Comparison of How Hospitals Meet Community Benefit Requirements. Publication no. GAO-08-880. 2008; http://www.gao.gov/new.items/d08880.pdf 23 U.S. Department of Health and Human Services. Principles of Community Engagement. 2nd ed. 2011; https://www.atsdr.cdc.gov/communityengagement/pdf/PCE_Report_508_FINAL.pdfUpdated 24 Rimer BK, Conaway M, Lyna P et al.  . The impact of tailored interventions on a community health center population. Patient Educ Couns  1999; 37( 2): 125– 40. Google Scholar CrossRef Search ADS PubMed  25 Cleary PD, Gross CP, Zaslavsky AM et al.  . Multilevel interventions: study design and analysis issues. JNCI Monogr  2012; 2012( 44): 49– 55. Google Scholar CrossRef Search ADS   26 Pennel CL, McLeroy KR, Burdine JN et al.  . Nonprofit hospitals’ approach to community health needs assessment. Am J Public Health  2015; 105( 3): e103– 13. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of Faculty of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Public Health Oxford University Press

Association between hospital community services and county population health in the USA

Loading next page...
 
/lp/ou_press/association-between-hospital-community-services-and-county-population-k2ror2vbw0
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of Faculty of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
1741-3842
eISSN
1741-3850
D.O.I.
10.1093/pubmed/fdy007
Publisher site
See Article on Publisher Site

Abstract

Abstract Objectives Little research has utilized population level data to test the association between community health outcomes and (i) hospital-sponsored community services that facilitate access to care and (ii) hospital-sponsored community building services in the USA. Therefore, the purpose of this study was to examine these relationships. Methods A secondary data analysis of the 2016 County Health Rankings and American Hospital Association databases was conducted via zero-truncated negative Binomial regression. Results Findings indicate a statistically significant difference between the number of community healthcare access services and community building services with county’s rank of health behavior. However, no statistically significant differences were found between the number of community healthcare access services and community building services with county rankings of length of life, quality of life or clinical care. Conclusions Our findings suggest that quality measures of services may play a more important role in community health improvement and that there is opportunity for hospitals to revamp the way in which community health needs assessments are conducted. Additional federal action is needed to standardize hospital sponsored community health service data reporting so that practitioners, hospital administrators and researchers can more specifically define hospitals’ role in public health protection in the USA. health services, population-based and preventative services, social determinants Introduction While the USA outperforms many countries in global rankings, it often lags behind in health outcomes compared to other developed nations.1 Health disparities research indicates that socioeconomic and ethnic/racial inequality in the US disproportionately contributes to poor outcomes.2,3 Thus, national strategies such as the Affordable Care Act (ACA) of 2010 and Healthy People 2020 emphasize the importance for organizations to address structural and cultural inequalities in social determinants of health such as neighborhood and built environments, economic stability, social and community contexts, education and access to healthcare.4,5 In recognition of the role of hospitals in achieving equitable community health outcomes, the ACA requires tax exempt hospitals to complete community health needs assessments (CHNAs) at least every 3 years and to devise implementation plans to address health needs raised by communities. As hospitals have the potential to positively impact the health of the surrounding community,6 many hospitals have expanded the scope of medical services to increase healthcare access in order to better meet neighboring community health needs. Implementation of community services in hospitals, especially in response to community identified needs, provides an avenue to increase equity in access to care, and ultimately, improve health outcomes for populations that are often high risk, underserved, and have unmet needs (e.g. low income and rural). Although we know that the number and types of community-oriented services vary across hospitals, their overall population level health influence has not been clearly described in the extant literature. Given the paucity of research in this area, this article fills this scientific void by investigating the association between hospital sponsored community health services and population level health outcomes in the USA. Method Data sources We used the 2016 Area Health Resource File (AHRF), the 2016 American Hospital Association Dataset (AHA) and the 2016 County Health Rankings National Database for analyses. The AHRF database provides data concerning health professionals, health facilities and population characteristics including demographic and economic aspects all attributed to the county level.7 County location is identified by a two digit State and three digit county Federal Information Processing Standards (FIPS) code. The AHA database includes information on hospital characteristics, demographics, services and expenses from over 6000 hospitals in the USA.8 Each hospital’s location is identified by a FIPS code. The County Health Rankings database, developed by the University of Wisconsin Population Health Institute and the Robert Wood Johnson Foundation in 2016, is a national dataset that ranks county-level health across the USA by state using a multidimensional model that incorporates social determinants defined by four major factors: health behaviors, clinical care, social and economic factors, and physical environment to assess health outcomes—length and quality of life.9,10 County location is identified by a FIPS code, which was utilized to merge the three datasets. Our attempt at understanding the relationship between population factors and hospital community services followed the Rankings conceptual framework. Dependent variables The main dependent variables in this study are community-ranking variables provided by the County Health Ranking Database. These variables include length of life, quality of life, health behaviors and clinical care. We excluded social and economic factors, and physical environment measures as there is not a strong rational indicating hospitals will have a direct influence on either. The elements, which define each of these variables, are listed in Table 1, along with their weightings. Table 1 Community health ranking variable definitions Variable  Measure  Description  Weight (%)  Length of life  Premature death  Years of potential life lost before age 75 per 100 000 population (age-adjusted)  100  Quality of life  Poor or fair health  Percentage of adults reporting fair or poor health (age-adjusted)  20  Poor physical health days  Average number of physically unhealthy days reported in past 30 days (age-adjusted)  20  Poor mental health days  Average number of mentally unhealthy days reported in past 30 days (age-adjusted)  20  Low birth weight  Percentage of live births with low birth weight (<2500 g)  40  Health behaviors  Adult smoking  Percentage of adults who are current smokers  33  Adult obesity  Percentage of adults that report a BMI of 30 or more  17  Food environment index  Index of factors that contribute to a healthy food environment  7  Physical inactivity  Percentage of adults aged 20 and over reporting no leisure-time physical activity  7  Access to exercise opportunities  Percentage of population with adequate access to locations for physical activity  3  Excessive drinking  Percentage of adults reporting binge or heavy drinking  8  Alcohol-impaired driving deaths  Percentage of driving deaths with alcohol involvement  8  Sexually transmitted infections  Number of newly diagnosed chlamydia cases per 100 000 population  8  Teen births  Teen birth rate per 1 000 female population, ages 15–19  8  Clinical care  Uninsured  Percentage of population under age 65 without health insurance  25  Primary care physicians  Ratio of population to primary care physicians  15  Dentists  Ratio of population to dentists  5  Mental health providers  Ratio of population to mental health providers  5  Preventable hospital stays  Number of hospital stays for ambulatory-care sensitive conditions per 1000 Medicare enrollees  25  Diabetic monitoring  Percentage of diabetic Medicare enrollees ages 65–75 that receive HbA1c monitoring  13  Mammography screening  Percentage of female Medicare enrollees ages 67–69 that receive mammography screening  13  Variable  Measure  Description  Weight (%)  Length of life  Premature death  Years of potential life lost before age 75 per 100 000 population (age-adjusted)  100  Quality of life  Poor or fair health  Percentage of adults reporting fair or poor health (age-adjusted)  20  Poor physical health days  Average number of physically unhealthy days reported in past 30 days (age-adjusted)  20  Poor mental health days  Average number of mentally unhealthy days reported in past 30 days (age-adjusted)  20  Low birth weight  Percentage of live births with low birth weight (<2500 g)  40  Health behaviors  Adult smoking  Percentage of adults who are current smokers  33  Adult obesity  Percentage of adults that report a BMI of 30 or more  17  Food environment index  Index of factors that contribute to a healthy food environment  7  Physical inactivity  Percentage of adults aged 20 and over reporting no leisure-time physical activity  7  Access to exercise opportunities  Percentage of population with adequate access to locations for physical activity  3  Excessive drinking  Percentage of adults reporting binge or heavy drinking  8  Alcohol-impaired driving deaths  Percentage of driving deaths with alcohol involvement  8  Sexually transmitted infections  Number of newly diagnosed chlamydia cases per 100 000 population  8  Teen births  Teen birth rate per 1 000 female population, ages 15–19  8  Clinical care  Uninsured  Percentage of population under age 65 without health insurance  25  Primary care physicians  Ratio of population to primary care physicians  15  Dentists  Ratio of population to dentists  5  Mental health providers  Ratio of population to mental health providers  5  Preventable hospital stays  Number of hospital stays for ambulatory-care sensitive conditions per 1000 Medicare enrollees  25  Diabetic monitoring  Percentage of diabetic Medicare enrollees ages 65–75 that receive HbA1c monitoring  13  Mammography screening  Percentage of female Medicare enrollees ages 67–69 that receive mammography screening  13  Table 1 Community health ranking variable definitions Variable  Measure  Description  Weight (%)  Length of life  Premature death  Years of potential life lost before age 75 per 100 000 population (age-adjusted)  100  Quality of life  Poor or fair health  Percentage of adults reporting fair or poor health (age-adjusted)  20  Poor physical health days  Average number of physically unhealthy days reported in past 30 days (age-adjusted)  20  Poor mental health days  Average number of mentally unhealthy days reported in past 30 days (age-adjusted)  20  Low birth weight  Percentage of live births with low birth weight (<2500 g)  40  Health behaviors  Adult smoking  Percentage of adults who are current smokers  33  Adult obesity  Percentage of adults that report a BMI of 30 or more  17  Food environment index  Index of factors that contribute to a healthy food environment  7  Physical inactivity  Percentage of adults aged 20 and over reporting no leisure-time physical activity  7  Access to exercise opportunities  Percentage of population with adequate access to locations for physical activity  3  Excessive drinking  Percentage of adults reporting binge or heavy drinking  8  Alcohol-impaired driving deaths  Percentage of driving deaths with alcohol involvement  8  Sexually transmitted infections  Number of newly diagnosed chlamydia cases per 100 000 population  8  Teen births  Teen birth rate per 1 000 female population, ages 15–19  8  Clinical care  Uninsured  Percentage of population under age 65 without health insurance  25  Primary care physicians  Ratio of population to primary care physicians  15  Dentists  Ratio of population to dentists  5  Mental health providers  Ratio of population to mental health providers  5  Preventable hospital stays  Number of hospital stays for ambulatory-care sensitive conditions per 1000 Medicare enrollees  25  Diabetic monitoring  Percentage of diabetic Medicare enrollees ages 65–75 that receive HbA1c monitoring  13  Mammography screening  Percentage of female Medicare enrollees ages 67–69 that receive mammography screening  13  Variable  Measure  Description  Weight (%)  Length of life  Premature death  Years of potential life lost before age 75 per 100 000 population (age-adjusted)  100  Quality of life  Poor or fair health  Percentage of adults reporting fair or poor health (age-adjusted)  20  Poor physical health days  Average number of physically unhealthy days reported in past 30 days (age-adjusted)  20  Poor mental health days  Average number of mentally unhealthy days reported in past 30 days (age-adjusted)  20  Low birth weight  Percentage of live births with low birth weight (<2500 g)  40  Health behaviors  Adult smoking  Percentage of adults who are current smokers  33  Adult obesity  Percentage of adults that report a BMI of 30 or more  17  Food environment index  Index of factors that contribute to a healthy food environment  7  Physical inactivity  Percentage of adults aged 20 and over reporting no leisure-time physical activity  7  Access to exercise opportunities  Percentage of population with adequate access to locations for physical activity  3  Excessive drinking  Percentage of adults reporting binge or heavy drinking  8  Alcohol-impaired driving deaths  Percentage of driving deaths with alcohol involvement  8  Sexually transmitted infections  Number of newly diagnosed chlamydia cases per 100 000 population  8  Teen births  Teen birth rate per 1 000 female population, ages 15–19  8  Clinical care  Uninsured  Percentage of population under age 65 without health insurance  25  Primary care physicians  Ratio of population to primary care physicians  15  Dentists  Ratio of population to dentists  5  Mental health providers  Ratio of population to mental health providers  5  Preventable hospital stays  Number of hospital stays for ambulatory-care sensitive conditions per 1000 Medicare enrollees  25  Diabetic monitoring  Percentage of diabetic Medicare enrollees ages 65–75 that receive HbA1c monitoring  13  Mammography screening  Percentage of female Medicare enrollees ages 67–69 that receive mammography screening  13  The County Health Ranking dataset provides county ranks based on county quartile ranking of each measure within the state.10 However, it does not provide overall county rankings for the entire USA. As such, there are 50 counties from 50 different states, which obtained the highest rank of 1 within each category. Each category is defined separately, which provides variance between categories with regard to rank. For example, a county may achieve a rank of 1 within clinical care, but may receive a rank of 50 with regard to health behaviors. As such, this analysis focuses on each category separately to determine the associations with each category. Independent variables Two independent variables were operationalized as (i) number of services that facilitate access to care and (ii) community building services offered by hospitals in a county. Services that were implemented to facilitate access to care included enrollment assistance programs, women’s health services, teen outreach services, health screenings and health fairs, indigent care clinics, community outreach, rural health clinics and referral centers, transportation to health services and mobile health services. We classified community building oriented services as those that aim to improve particular public health outcomes and determinants in a community. These services included immunization programs, nutrition programs, tobacco treatment services, support groups, community health education, patient education centers, patient representation services and meals on wheels. The overall number of services offered by hospitals in each of these categories was aggregated to create the total number of access and community building services. Data within the AHA survey does not provide indication concerning service reach (provision to one or more county) nor frequency of services offered. In addition, the hospital community needs assessment process leaves the definition of service community to the hospital in the community needs assessment process. As such, we have identified the county in which the hospital is located as the most likely area in which provision of services is to occur and have assigned all services offered by the hospital to that county as identified through FIPS codes. Control variables To control for differing county characteristics, we used variables previously identified as important for health outcomes. The percent of the population within the county which is younger than 18 or over 65 years of age are operationalized as continuous variables representing two opposite ends of the health spectrum. That is, younger individuals are typically less burdened with chronic disease than older individuals who experience increased acute injuries.11 Additionally, racial and ethnic minority populations have also been indicated as important in the use of health services and, in particular, are identified to have disparities in access and utilization of health services.12 Thus, continuous variables indicating county population percentages of the following groups were included in the analysis: African Americans, American Indians or Alaskan Natives, Asians, Native Hawaiians or other Pacific Islanders, Hispanics and non-Hispanic Whites. We also include a continuous variable which indicates the county population percentage which is English proficient, as this provides insight into barriers to healthcare utilization and need.12 Health outcomes are also different for males and females.13 As such, the percentage of the county population that was female was included as a binary variable. Additionally, given that location (e.g. urban versus rural) is an indicator of service need and utilization, county population percentage within rural locations was included as a continuous variable.14 Furthermore, diabetes prevalence rates provide additional indication of healthcare services needed;15 thus, county diabetes prevalence rate was included as a continuous variable. Food insecurity (i.e. percent of the county population identified as food insecure) was also included as this variable provides an indication of transportation needs and is a predictor of high cost users of healthcare services.16 Finally, percent of the population with limited access to healthcare was included as a continuous variable to further define transportation and health insurance barriers to care which would indicate greater need for community services.17 Analysis To compare descriptive statistics and data included and excluded from the analysis, means, standard deviations and Kruska–Wallis tests were employed. The researchers used a zero truncated negative binomial regression model to explore the bivariate relationship between the outcome, independent and control variables. Negative binomial and Poisson regression were both considered because the dependent variable is count data. Over-dispersion and zero truncation were assessed for each outcome variable. All dependent variables demonstrated over-dispersion (variance greater than the mean) and require the use of negative binomial regression, which corrects for this excess variability in the count data.18 Since there are no counties that receive a zero rank, a zero truncation adjustment was applied to the models.18 Finally, an exposure variable for the number of counties within each State was utilized to control for these differences. STATA 14 was used to run all analyses, models were estimated though maximum likelihood, and P-values were calculated based off of the Z-statistic. Incident rate ratios, standard errors, Z-statistics and significance level are reported. Results Overall, there are 3 156 counties included in the dataset. Due to missing data in the County Health Rankings dataset, a total of 97 counties were excluded. Medians and interquartile ranges of rank data and means and standard deviations of county population characteristics are reported in Table 2. Within the sample, the mean percent of those younger than 18 across all counties was 22.74%, while the percent of individuals aged 65 and over was 17.05%. The majority of the sample were non-Hispanic Whites (77.28%) while African Americans and Hispanics consisted of 9.08 and 8.80%, respectively. Individuals of other ethnicities consisted of <2.5% of the population. Over 98% of the sample was proficient in English. Half of the sample (49.97%) were female. Table 2 Descriptive statistics Dependent variables  Included in analysis  Excluded from analysis    Median  Interquartile range (25–75%)  Pop  Median  Interquartile range (25–75%)  Pop  Significance  Life length  38  (17–65)  3059  65  (59–85)  3  –  Health behaviors  38  (17–65)  3059  71  (63–94)  3  –  Quality of life  38  (17–65)  3059  74  (58–94)  3  –  Clinical care  38  (17–65)  3059  67  (16–79)  3  –  Independent variables  Means  Standard deviation  Pop  Means  Standard deviation  Pop  Significance  Number of community access services  27.88  29.74  3059  20.14  25.89  97  ***  Number of community building services  33.05  35.91  3059  22.71  28.19  97  ***  County population per million  0.10  0.33  3059  0.00  0.00  82  ***  Percent not for profit  0.46  0.32  3059  0.44  0.34  97    Percent of the population younger than 18  22.74  3.35  3059  20.75  4.21  82  ***  Percent of population 65 or over  17.05  4.30  3059  21.28  5.28  82  ***  Percent of population African American  9.08  14.38  3059  1.77  7.93  82  ***  Percent of population American Indian or Alaskan Native  2.21  7.64  3059  2.63  6.31  82  ***  Percent of population Asian  1.35  2.59  3059  1.06  4.67  82  ***  Percent of population Native Hawaiian or Other Pacific Islander  0.11  0.41  3059  0.65  5.27  82  ***  Percent of population Hispanic  8.80  13.35  3059  10.35  15.35  82    Percent of population non-Hispanic White  77.28  19.82  3059  82.12  20.92  82  ***  Percent of population not proficient in English  1.81  2.91  3059  2.10  4.15  82  ***  Percent of population female  49.97  2.22  3059  48.57  2.85  82  ***  Percent of population in rural location  57.53  31.19  3059  99.34  4.39  82  ***  Percent of the population diabetic  11.00  2.32  3059  10.21  1.92  82  ***  Percent of the population which is food insecure  14.79  3.85  3059  12.85  3.63  82  ***  Percent of the population with limited access to healthcare  7.92  7.47  3059  25.59  13.95  82  ***  Dependent variables  Included in analysis  Excluded from analysis    Median  Interquartile range (25–75%)  Pop  Median  Interquartile range (25–75%)  Pop  Significance  Life length  38  (17–65)  3059  65  (59–85)  3  –  Health behaviors  38  (17–65)  3059  71  (63–94)  3  –  Quality of life  38  (17–65)  3059  74  (58–94)  3  –  Clinical care  38  (17–65)  3059  67  (16–79)  3  –  Independent variables  Means  Standard deviation  Pop  Means  Standard deviation  Pop  Significance  Number of community access services  27.88  29.74  3059  20.14  25.89  97  ***  Number of community building services  33.05  35.91  3059  22.71  28.19  97  ***  County population per million  0.10  0.33  3059  0.00  0.00  82  ***  Percent not for profit  0.46  0.32  3059  0.44  0.34  97    Percent of the population younger than 18  22.74  3.35  3059  20.75  4.21  82  ***  Percent of population 65 or over  17.05  4.30  3059  21.28  5.28  82  ***  Percent of population African American  9.08  14.38  3059  1.77  7.93  82  ***  Percent of population American Indian or Alaskan Native  2.21  7.64  3059  2.63  6.31  82  ***  Percent of population Asian  1.35  2.59  3059  1.06  4.67  82  ***  Percent of population Native Hawaiian or Other Pacific Islander  0.11  0.41  3059  0.65  5.27  82  ***  Percent of population Hispanic  8.80  13.35  3059  10.35  15.35  82    Percent of population non-Hispanic White  77.28  19.82  3059  82.12  20.92  82  ***  Percent of population not proficient in English  1.81  2.91  3059  2.10  4.15  82  ***  Percent of population female  49.97  2.22  3059  48.57  2.85  82  ***  Percent of population in rural location  57.53  31.19  3059  99.34  4.39  82  ***  Percent of the population diabetic  11.00  2.32  3059  10.21  1.92  82  ***  Percent of the population which is food insecure  14.79  3.85  3059  12.85  3.63  82  ***  Percent of the population with limited access to healthcare  7.92  7.47  3059  25.59  13.95  82  ***  ***Significant at P = 0.001. **Significant at P = 0.01. *Significant at P = 0.05. Table 2 Descriptive statistics Dependent variables  Included in analysis  Excluded from analysis    Median  Interquartile range (25–75%)  Pop  Median  Interquartile range (25–75%)  Pop  Significance  Life length  38  (17–65)  3059  65  (59–85)  3  –  Health behaviors  38  (17–65)  3059  71  (63–94)  3  –  Quality of life  38  (17–65)  3059  74  (58–94)  3  –  Clinical care  38  (17–65)  3059  67  (16–79)  3  –  Independent variables  Means  Standard deviation  Pop  Means  Standard deviation  Pop  Significance  Number of community access services  27.88  29.74  3059  20.14  25.89  97  ***  Number of community building services  33.05  35.91  3059  22.71  28.19  97  ***  County population per million  0.10  0.33  3059  0.00  0.00  82  ***  Percent not for profit  0.46  0.32  3059  0.44  0.34  97    Percent of the population younger than 18  22.74  3.35  3059  20.75  4.21  82  ***  Percent of population 65 or over  17.05  4.30  3059  21.28  5.28  82  ***  Percent of population African American  9.08  14.38  3059  1.77  7.93  82  ***  Percent of population American Indian or Alaskan Native  2.21  7.64  3059  2.63  6.31  82  ***  Percent of population Asian  1.35  2.59  3059  1.06  4.67  82  ***  Percent of population Native Hawaiian or Other Pacific Islander  0.11  0.41  3059  0.65  5.27  82  ***  Percent of population Hispanic  8.80  13.35  3059  10.35  15.35  82    Percent of population non-Hispanic White  77.28  19.82  3059  82.12  20.92  82  ***  Percent of population not proficient in English  1.81  2.91  3059  2.10  4.15  82  ***  Percent of population female  49.97  2.22  3059  48.57  2.85  82  ***  Percent of population in rural location  57.53  31.19  3059  99.34  4.39  82  ***  Percent of the population diabetic  11.00  2.32  3059  10.21  1.92  82  ***  Percent of the population which is food insecure  14.79  3.85  3059  12.85  3.63  82  ***  Percent of the population with limited access to healthcare  7.92  7.47  3059  25.59  13.95  82  ***  Dependent variables  Included in analysis  Excluded from analysis    Median  Interquartile range (25–75%)  Pop  Median  Interquartile range (25–75%)  Pop  Significance  Life length  38  (17–65)  3059  65  (59–85)  3  –  Health behaviors  38  (17–65)  3059  71  (63–94)  3  –  Quality of life  38  (17–65)  3059  74  (58–94)  3  –  Clinical care  38  (17–65)  3059  67  (16–79)  3  –  Independent variables  Means  Standard deviation  Pop  Means  Standard deviation  Pop  Significance  Number of community access services  27.88  29.74  3059  20.14  25.89  97  ***  Number of community building services  33.05  35.91  3059  22.71  28.19  97  ***  County population per million  0.10  0.33  3059  0.00  0.00  82  ***  Percent not for profit  0.46  0.32  3059  0.44  0.34  97    Percent of the population younger than 18  22.74  3.35  3059  20.75  4.21  82  ***  Percent of population 65 or over  17.05  4.30  3059  21.28  5.28  82  ***  Percent of population African American  9.08  14.38  3059  1.77  7.93  82  ***  Percent of population American Indian or Alaskan Native  2.21  7.64  3059  2.63  6.31  82  ***  Percent of population Asian  1.35  2.59  3059  1.06  4.67  82  ***  Percent of population Native Hawaiian or Other Pacific Islander  0.11  0.41  3059  0.65  5.27  82  ***  Percent of population Hispanic  8.80  13.35  3059  10.35  15.35  82    Percent of population non-Hispanic White  77.28  19.82  3059  82.12  20.92  82  ***  Percent of population not proficient in English  1.81  2.91  3059  2.10  4.15  82  ***  Percent of population female  49.97  2.22  3059  48.57  2.85  82  ***  Percent of population in rural location  57.53  31.19  3059  99.34  4.39  82  ***  Percent of the population diabetic  11.00  2.32  3059  10.21  1.92  82  ***  Percent of the population which is food insecure  14.79  3.85  3059  12.85  3.63  82  ***  Percent of the population with limited access to healthcare  7.92  7.47  3059  25.59  13.95  82  ***  ***Significant at P = 0.001. **Significant at P = 0.01. *Significant at P = 0.05. There were more rural counties (57.53%) than urban counties, and 11% of the sample were diabetic. Other characteristics include 14.79% being food insecure and 7.92% having limited access to healthcare services. Median county rankings for health ranking factors were 38 for life length, health behaviors, quality of life and clinical care. The mean number of community healthcare access services across all counties was 27.88 (SD = 29.74) and 33.05 (SD = 35.91) for community building services. Furthermore, the mean number of community healthcare access services was 20.14 (SD = 25.89) and 22.71 (SD = 28.19) for community building services. In the sample that was excluded from the analysis, the median county rankings for health ranking factors were 65 for life length, 71 for health behaviors, 74 for quality of life and 67 clinical care, which was consistently higher than the included study sample. Table 3 displays the results of the bivariate analyses. These results reveal significant associations between each of our key outcome variables and number of community healthcare access services and building services per million population. There is a lack of association between percent not-for-profit hospitals within the county and the county rankings. In addition, county population per million, percent of the population younger than 18, percent of the population African American, Native Hawaiian or Alaskan Native, Hispanic, not proficient in English and female are not consistently associated with the county ranking independent variables. Table 3 Bivariate analysis county rankings and independent variables   Life length  Behaviors  Quality of life  Clinical care  Coef. (Std)  P-value  Coef. (Std)  P-value  Coef. (Std)  P-value  Coef. (Std)  P-value  Number of community access services  −0.004 (0.000)  0.000  −0.004 (0.000)  0.000  −0.001 (0.000)  0.006  −0.003 (0.000)  0.000  Number of community building services  −0.003 (0.000)  0.000  −0.003 (0.000)  0.000  −0.001 (0.000)  0.000  −0.003 (0.000)  0.000  County population per million  −0.241 (0.032)  0.000  −0.280 (0.035)  0.000  −0.025 (0.034)  0.457  −0.138 (0.026)  0.000  Percent not for profit  −0.040 (0.039)  0.302  −0.033 (0.039)  0.396  −0.075 (0.040)  0.057  −0.025 (0.039)  0.516  Percent of the population younger than 18  −0.12 (0.004)  0.001  0.007 (0.004)  0.061  −0.005 (0.004)  0.162  0.008 (0.004)  0.021  Percent of population 65 or over  0.037 (0.003)  0.000  0.008 (0.003)  0.010  0.008 (0.003)  0.007  0.021 (0.003)  0.000  Percent of population African American  0.005 (0.001)  0.000  0.006 (0.001)  0.000  0.007 (0.001)  0.000  −0.000 (0.001)  0.963  Percent of population American Indian or Alaskan Native  0.009 (0.002)  0.000  0.009 (0.002)  0.000  0.005 (0.002)  0.002  0.009 (0.002)  0.000  Percent of population Asian  −0.075 (0.005)  0.000  −0.073 (0.005)  0.000  −0.03 (0.005)  0.000  −0.061 (0.005)  0.000  Percent of population Native Hawaiian or other Pacific Islander  −0.129 (0.50)  0.011  −0.052 (0.044)  0.238  −0.072 (0.045)  0.112  −0.05 (0.044)  0.213  Percent of population Hispanic  −0.005 (0.001)  0.000  −0.001 (0.001)  0.131  0.000 (0.001)  0.760  0.003 (0.001)  0.001  Percent of population non-Hispanic White  −0.001 (0.001)  0.075  −0.004 (0.001)  0.000  −0.004 (0.001)  0.000  0.002 (0.001)  0.002  Percent of population not proficient in English  −0.039 (0.004)  0.000  −0.014 (0.004)  0.001  0.006 (0.004)  0.155  0.018 (0.004)  0.000  Percent of population female  0.016 (0.006)  0.004  −0.024 (0.006)  0.000  −0.007 (0.006)  0.197  −0.043 (0.006)  0.000  Percent of population in rural location  0.005 (0.000)  0.000  0.004 (0.000)  0.000  0.001 (0.000)  0.001  0.008 (0.000)  0.000  Percent of the population diabetic  0.103 (0.005)  0.000  0.093 (0.006)  0.000  0.077 (0.005)  0.000  0.073 (0.006)  0.000  Percent of the population which is food insecure  0.050 (0.003)  0.000  0.049 (0.003)  0.000  0.049 (0.003)  0.000  0.028 (0.003)  0.000  Percent of the population with limited access to healthcare  0.006 (0.002)  0.000  0.011 (0.002)  0.000  0.006 (0.002)  0.000  0.006 (0.002)  0.000    Life length  Behaviors  Quality of life  Clinical care  Coef. (Std)  P-value  Coef. (Std)  P-value  Coef. (Std)  P-value  Coef. (Std)  P-value  Number of community access services  −0.004 (0.000)  0.000  −0.004 (0.000)  0.000  −0.001 (0.000)  0.006  −0.003 (0.000)  0.000  Number of community building services  −0.003 (0.000)  0.000  −0.003 (0.000)  0.000  −0.001 (0.000)  0.000  −0.003 (0.000)  0.000  County population per million  −0.241 (0.032)  0.000  −0.280 (0.035)  0.000  −0.025 (0.034)  0.457  −0.138 (0.026)  0.000  Percent not for profit  −0.040 (0.039)  0.302  −0.033 (0.039)  0.396  −0.075 (0.040)  0.057  −0.025 (0.039)  0.516  Percent of the population younger than 18  −0.12 (0.004)  0.001  0.007 (0.004)  0.061  −0.005 (0.004)  0.162  0.008 (0.004)  0.021  Percent of population 65 or over  0.037 (0.003)  0.000  0.008 (0.003)  0.010  0.008 (0.003)  0.007  0.021 (0.003)  0.000  Percent of population African American  0.005 (0.001)  0.000  0.006 (0.001)  0.000  0.007 (0.001)  0.000  −0.000 (0.001)  0.963  Percent of population American Indian or Alaskan Native  0.009 (0.002)  0.000  0.009 (0.002)  0.000  0.005 (0.002)  0.002  0.009 (0.002)  0.000  Percent of population Asian  −0.075 (0.005)  0.000  −0.073 (0.005)  0.000  −0.03 (0.005)  0.000  −0.061 (0.005)  0.000  Percent of population Native Hawaiian or other Pacific Islander  −0.129 (0.50)  0.011  −0.052 (0.044)  0.238  −0.072 (0.045)  0.112  −0.05 (0.044)  0.213  Percent of population Hispanic  −0.005 (0.001)  0.000  −0.001 (0.001)  0.131  0.000 (0.001)  0.760  0.003 (0.001)  0.001  Percent of population non-Hispanic White  −0.001 (0.001)  0.075  −0.004 (0.001)  0.000  −0.004 (0.001)  0.000  0.002 (0.001)  0.002  Percent of population not proficient in English  −0.039 (0.004)  0.000  −0.014 (0.004)  0.001  0.006 (0.004)  0.155  0.018 (0.004)  0.000  Percent of population female  0.016 (0.006)  0.004  −0.024 (0.006)  0.000  −0.007 (0.006)  0.197  −0.043 (0.006)  0.000  Percent of population in rural location  0.005 (0.000)  0.000  0.004 (0.000)  0.000  0.001 (0.000)  0.001  0.008 (0.000)  0.000  Percent of the population diabetic  0.103 (0.005)  0.000  0.093 (0.006)  0.000  0.077 (0.005)  0.000  0.073 (0.006)  0.000  Percent of the population which is food insecure  0.050 (0.003)  0.000  0.049 (0.003)  0.000  0.049 (0.003)  0.000  0.028 (0.003)  0.000  Percent of the population with limited access to healthcare  0.006 (0.002)  0.000  0.011 (0.002)  0.000  0.006 (0.002)  0.000  0.006 (0.002)  0.000  Table 3 Bivariate analysis county rankings and independent variables   Life length  Behaviors  Quality of life  Clinical care  Coef. (Std)  P-value  Coef. (Std)  P-value  Coef. (Std)  P-value  Coef. (Std)  P-value  Number of community access services  −0.004 (0.000)  0.000  −0.004 (0.000)  0.000  −0.001 (0.000)  0.006  −0.003 (0.000)  0.000  Number of community building services  −0.003 (0.000)  0.000  −0.003 (0.000)  0.000  −0.001 (0.000)  0.000  −0.003 (0.000)  0.000  County population per million  −0.241 (0.032)  0.000  −0.280 (0.035)  0.000  −0.025 (0.034)  0.457  −0.138 (0.026)  0.000  Percent not for profit  −0.040 (0.039)  0.302  −0.033 (0.039)  0.396  −0.075 (0.040)  0.057  −0.025 (0.039)  0.516  Percent of the population younger than 18  −0.12 (0.004)  0.001  0.007 (0.004)  0.061  −0.005 (0.004)  0.162  0.008 (0.004)  0.021  Percent of population 65 or over  0.037 (0.003)  0.000  0.008 (0.003)  0.010  0.008 (0.003)  0.007  0.021 (0.003)  0.000  Percent of population African American  0.005 (0.001)  0.000  0.006 (0.001)  0.000  0.007 (0.001)  0.000  −0.000 (0.001)  0.963  Percent of population American Indian or Alaskan Native  0.009 (0.002)  0.000  0.009 (0.002)  0.000  0.005 (0.002)  0.002  0.009 (0.002)  0.000  Percent of population Asian  −0.075 (0.005)  0.000  −0.073 (0.005)  0.000  −0.03 (0.005)  0.000  −0.061 (0.005)  0.000  Percent of population Native Hawaiian or other Pacific Islander  −0.129 (0.50)  0.011  −0.052 (0.044)  0.238  −0.072 (0.045)  0.112  −0.05 (0.044)  0.213  Percent of population Hispanic  −0.005 (0.001)  0.000  −0.001 (0.001)  0.131  0.000 (0.001)  0.760  0.003 (0.001)  0.001  Percent of population non-Hispanic White  −0.001 (0.001)  0.075  −0.004 (0.001)  0.000  −0.004 (0.001)  0.000  0.002 (0.001)  0.002  Percent of population not proficient in English  −0.039 (0.004)  0.000  −0.014 (0.004)  0.001  0.006 (0.004)  0.155  0.018 (0.004)  0.000  Percent of population female  0.016 (0.006)  0.004  −0.024 (0.006)  0.000  −0.007 (0.006)  0.197  −0.043 (0.006)  0.000  Percent of population in rural location  0.005 (0.000)  0.000  0.004 (0.000)  0.000  0.001 (0.000)  0.001  0.008 (0.000)  0.000  Percent of the population diabetic  0.103 (0.005)  0.000  0.093 (0.006)  0.000  0.077 (0.005)  0.000  0.073 (0.006)  0.000  Percent of the population which is food insecure  0.050 (0.003)  0.000  0.049 (0.003)  0.000  0.049 (0.003)  0.000  0.028 (0.003)  0.000  Percent of the population with limited access to healthcare  0.006 (0.002)  0.000  0.011 (0.002)  0.000  0.006 (0.002)  0.000  0.006 (0.002)  0.000    Life length  Behaviors  Quality of life  Clinical care  Coef. (Std)  P-value  Coef. (Std)  P-value  Coef. (Std)  P-value  Coef. (Std)  P-value  Number of community access services  −0.004 (0.000)  0.000  −0.004 (0.000)  0.000  −0.001 (0.000)  0.006  −0.003 (0.000)  0.000  Number of community building services  −0.003 (0.000)  0.000  −0.003 (0.000)  0.000  −0.001 (0.000)  0.000  −0.003 (0.000)  0.000  County population per million  −0.241 (0.032)  0.000  −0.280 (0.035)  0.000  −0.025 (0.034)  0.457  −0.138 (0.026)  0.000  Percent not for profit  −0.040 (0.039)  0.302  −0.033 (0.039)  0.396  −0.075 (0.040)  0.057  −0.025 (0.039)  0.516  Percent of the population younger than 18  −0.12 (0.004)  0.001  0.007 (0.004)  0.061  −0.005 (0.004)  0.162  0.008 (0.004)  0.021  Percent of population 65 or over  0.037 (0.003)  0.000  0.008 (0.003)  0.010  0.008 (0.003)  0.007  0.021 (0.003)  0.000  Percent of population African American  0.005 (0.001)  0.000  0.006 (0.001)  0.000  0.007 (0.001)  0.000  −0.000 (0.001)  0.963  Percent of population American Indian or Alaskan Native  0.009 (0.002)  0.000  0.009 (0.002)  0.000  0.005 (0.002)  0.002  0.009 (0.002)  0.000  Percent of population Asian  −0.075 (0.005)  0.000  −0.073 (0.005)  0.000  −0.03 (0.005)  0.000  −0.061 (0.005)  0.000  Percent of population Native Hawaiian or other Pacific Islander  −0.129 (0.50)  0.011  −0.052 (0.044)  0.238  −0.072 (0.045)  0.112  −0.05 (0.044)  0.213  Percent of population Hispanic  −0.005 (0.001)  0.000  −0.001 (0.001)  0.131  0.000 (0.001)  0.760  0.003 (0.001)  0.001  Percent of population non-Hispanic White  −0.001 (0.001)  0.075  −0.004 (0.001)  0.000  −0.004 (0.001)  0.000  0.002 (0.001)  0.002  Percent of population not proficient in English  −0.039 (0.004)  0.000  −0.014 (0.004)  0.001  0.006 (0.004)  0.155  0.018 (0.004)  0.000  Percent of population female  0.016 (0.006)  0.004  −0.024 (0.006)  0.000  −0.007 (0.006)  0.197  −0.043 (0.006)  0.000  Percent of population in rural location  0.005 (0.000)  0.000  0.004 (0.000)  0.000  0.001 (0.000)  0.001  0.008 (0.000)  0.000  Percent of the population diabetic  0.103 (0.005)  0.000  0.093 (0.006)  0.000  0.077 (0.005)  0.000  0.073 (0.006)  0.000  Percent of the population which is food insecure  0.050 (0.003)  0.000  0.049 (0.003)  0.000  0.049 (0.003)  0.000  0.028 (0.003)  0.000  Percent of the population with limited access to healthcare  0.006 (0.002)  0.000  0.011 (0.002)  0.000  0.006 (0.002)  0.000  0.006 (0.002)  0.000  Table 4 displays the zero-truncated negative binomial regression analysis results. We found no statistically significant differences between the number of community healthcare access services provided and the county’s rank of length of life, quality of life, nor clinical care. However, for every one unit increase in the number of community healthcare access services provided, the expected county ranking of health behaviors is multiplied by a factor of 0.999 (i.e. it decreases) while holding all other variables in the model constant. Similarly, we found no statistically significant differences between the number of community building services provided and the county’s rank of length of life, quality of life, nor clinical care. However, for every one unit increase in the number of community building services provided, the expected county ranking of health behaviors is multiplied by a factor of 0.999 (that is, it decreases) while holding all other variables in the model constant. Table 4 Zero truncated negative binomial regression for county rank by community access services offered by hospitals   Life length  Behaviors  Quality of life  Clinical care  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  Number of community access services  0.999 (0.000)  −1.63 (0.103)  0.999 (0.000)  −2.19 (0.028)  1.00 (0.000)  0.14 (0.885)  1.00 (0.000)  −0.96 (0.336)  Log likelihood  −13 640.327  −13 699.256  −13 826.073  −13 534.151  LR Chi2  716.74  609.37  355.97  942.67  Prob > Chi2  0.0000  0.0000  0.0000  0.0000  Psuedo R2  0.0256  0.0218  0.0127  0.0337    Life length  Behaviors  Quality of life  Clinical care    IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  Number of community building services  0.999 (0.000)  −2.01 (0.045)  0.999 (0.000)  −2.62 (0.009)  1.00 (0.000)  −0.24 (0.811)  1.00 (0.000)  −1.05 (0.295)  Log likelihood  −13 639.653  −13 698.263  −13 826.055  −13 534.066  LR chi2  718.08  611.36  356.01  942.84  Prob > chi2  0.0000  0.0000  0.0000  0.0000  Psuedo R2  0.0256  0.0218  0.0127  0.0337    Life length  Behaviors  Quality of life  Clinical care  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  Number of community access services  0.999 (0.000)  −1.63 (0.103)  0.999 (0.000)  −2.19 (0.028)  1.00 (0.000)  0.14 (0.885)  1.00 (0.000)  −0.96 (0.336)  Log likelihood  −13 640.327  −13 699.256  −13 826.073  −13 534.151  LR Chi2  716.74  609.37  355.97  942.67  Prob > Chi2  0.0000  0.0000  0.0000  0.0000  Psuedo R2  0.0256  0.0218  0.0127  0.0337    Life length  Behaviors  Quality of life  Clinical care    IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  Number of community building services  0.999 (0.000)  −2.01 (0.045)  0.999 (0.000)  −2.62 (0.009)  1.00 (0.000)  −0.24 (0.811)  1.00 (0.000)  −1.05 (0.295)  Log likelihood  −13 639.653  −13 698.263  −13 826.055  −13 534.066  LR chi2  718.08  611.36  356.01  942.84  Prob > chi2  0.0000  0.0000  0.0000  0.0000  Psuedo R2  0.0256  0.0218  0.0127  0.0337  Note: Each model controlled for the following: county population per million, percent not for profit, percent of the population younger than 18, percent of population 65 or over, percent of population African American, percent of population American Indian or Alaskan Native, percent of population Asian, percent of population Native Hawaiian or Other Pacific Islander, percent of population Hispanic, percent of population non-Hispanic White, percent of population not proficient in English, percent of population female, percent of population in rural location, percent of the population diabetic, percent of the population which is food insecure, percent of the population with limited access to healthcare. Table 4 Zero truncated negative binomial regression for county rank by community access services offered by hospitals   Life length  Behaviors  Quality of life  Clinical care  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  Number of community access services  0.999 (0.000)  −1.63 (0.103)  0.999 (0.000)  −2.19 (0.028)  1.00 (0.000)  0.14 (0.885)  1.00 (0.000)  −0.96 (0.336)  Log likelihood  −13 640.327  −13 699.256  −13 826.073  −13 534.151  LR Chi2  716.74  609.37  355.97  942.67  Prob > Chi2  0.0000  0.0000  0.0000  0.0000  Psuedo R2  0.0256  0.0218  0.0127  0.0337    Life length  Behaviors  Quality of life  Clinical care    IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  Number of community building services  0.999 (0.000)  −2.01 (0.045)  0.999 (0.000)  −2.62 (0.009)  1.00 (0.000)  −0.24 (0.811)  1.00 (0.000)  −1.05 (0.295)  Log likelihood  −13 639.653  −13 698.263  −13 826.055  −13 534.066  LR chi2  718.08  611.36  356.01  942.84  Prob > chi2  0.0000  0.0000  0.0000  0.0000  Psuedo R2  0.0256  0.0218  0.0127  0.0337    Life length  Behaviors  Quality of life  Clinical care  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  Number of community access services  0.999 (0.000)  −1.63 (0.103)  0.999 (0.000)  −2.19 (0.028)  1.00 (0.000)  0.14 (0.885)  1.00 (0.000)  −0.96 (0.336)  Log likelihood  −13 640.327  −13 699.256  −13 826.073  −13 534.151  LR Chi2  716.74  609.37  355.97  942.67  Prob > Chi2  0.0000  0.0000  0.0000  0.0000  Psuedo R2  0.0256  0.0218  0.0127  0.0337    Life length  Behaviors  Quality of life  Clinical care    IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  IRR (Std)  Z score (P-value)  Number of community building services  0.999 (0.000)  −2.01 (0.045)  0.999 (0.000)  −2.62 (0.009)  1.00 (0.000)  −0.24 (0.811)  1.00 (0.000)  −1.05 (0.295)  Log likelihood  −13 639.653  −13 698.263  −13 826.055  −13 534.066  LR chi2  718.08  611.36  356.01  942.84  Prob > chi2  0.0000  0.0000  0.0000  0.0000  Psuedo R2  0.0256  0.0218  0.0127  0.0337  Note: Each model controlled for the following: county population per million, percent not for profit, percent of the population younger than 18, percent of population 65 or over, percent of population African American, percent of population American Indian or Alaskan Native, percent of population Asian, percent of population Native Hawaiian or Other Pacific Islander, percent of population Hispanic, percent of population non-Hispanic White, percent of population not proficient in English, percent of population female, percent of population in rural location, percent of the population diabetic, percent of the population which is food insecure, percent of the population with limited access to healthcare. Discussion Main findings of this study As part of the implementation of the ACA, hospitals across the nation, regardless of tax status, offer an array of healthcare access services and health interventions. Our study findings confirm this and identified that such hospital services do not necessarily improve population level health outcomes. We found either neutral or negative relationships between hospital-sponsored community services and county level health rankings, suggesting an opportunity for hospitals to better operationalize services to improve health outcome. What is already known on the topic? Expanding the role of hospitals in keeping communities healthy and preventing disease has been advocated for on a national stage through the implementation of the ACA in 2010. However, the extent to which hospital community services impact population outcomes is not fully understood. What this study adds The current study represents one of few studies that quantifies the association of hospital sponsored community services and population health outcomes and factors in the USA. Past literature has mainly focused on dimensions of effective hospital partnerships aimed at community health improvement6,19 or on the impact of state community benefit laws and hospital ownership as it relates to health promotion services as noted by a review of the literature by Shortell et al.20 The statistically significant associations identified in our study were small in terms of effect-size; this reveals the lack of overall influence that hospital community health services have on population level. However, such findings present an opportunity for hospitals to increase the impact of community health services, especially for rural and historically marginalized communities. None of the tested independent variables reached statistical significance. Therefore, the number of hospital-sponsored community services offered may not be as important for community health improvement as service quality and reach. Our findings support previous research21 that calls for action from the Federal government to implement a public health framework where non-profit hospitals are charged to work more closely with local actors such as health departments to support ongoing community health initiatives as there may be a misalignment between hospital community services offered and community health needs. This is particularly relevant given that non-profit hospitals are not required to maintain yearly community improvement plans but rather at 3-year intervals. Estimates reported in 2008 suggest that non-profit hospitals tax exemptions yielded ~13 billion dollars;22 however, the extent to which these funds have facilitated community improvements at a population level remains unclear. To maximize the quality and reach of community services, health professionals should collaborate with key stakeholders and community-based entities (e.g. community organizations and coalitions). Such collaborations are most effective when partnerships are established in the planning stages of service or program implementation. These strategies can be particularly useful in communities that experience health inequities as collaborations can help facilitate the establishment of trusting relationships and secure commitments, while overcoming challenges related to buy-in and sustainability.23 Health professionals implementing community health services may also increase health benefits to communities by employing combination and multilevel interventions.24 Combination interventions are programs designed to address health issues at the intersection of social, behavioral and biological factors. In a complimentary manner, multilevel interventions are ‘implemented at the individual, physician, clinic, healthcare organization and/or community level.’25 These approaches have been cited as effective strategies for achieving population level changes in health outcomes while providing health professionals the opportunity to address factors in disease prevention and treatment. Limitations of this study While contributing to the larger literature, this study is not without limitations. The study used a cross-sectional design limited to 1 year, inhibiting the ability to determine causal relationships. Next, the data utilized for this study has limitations worth considering. Currently, there is no comprehensive national database that captures the extent of hospital community health-service data. Therefore, we used multiple databases. The AHA survey relies heavily on self-reported data acquired by asking hospitals to complete the survey, and may present issues of bias and/or social desirability. In addition, the attribution of community services to the county which the hospital resides may be incomplete. Currently, hospitals are able to define the community they serve for the purposes of community benefit, and there is a lack of consistency concerning the duration, intensity, and influence each service provides. Thus, there is great variation in the types of services offered by each hospital. However, inclusion of the county the hospital is located in provided the most consistent and logical method for attribution. The AHRF and County Health Rankings datasets provide a combination of a large number of externally collected datasets. This provides an increase in the likelihood of error and the manipulation of data either through conversion or through defining rankings may arbitrarily change the data or misrepresent county level attributions. Furthermore, the County Health Ranking methods provide a method for creating ranked data, however should other methods be utilized, the results of this study may change. We were unable to control for the potential that services offered occur on county lines or that all community services are the same thus may provide variance unaccounted for within this study. Finally, the current study focuses more directly on the number and types of services offered as opposed to how much time and financial resources the organization spends in delivering those services. Future studies should assess the amount of financial capital an organization spends on providing community services, the duration and intensity of those services, as well as the provision of those services to specific counties or other defined services areas. Additionally, future research may focus on accessing community members’ perceptions of service delivery. These lines of inquiry would provide additional insight into the quality and quantity of services offered as well as perceived barriers and facilitators to service utilization. Our results support previous research that posits hospitals have room for improvement with regard to translating findings from CHNAs into services that better suit the surrounding communities’ needs.26 Furthermore, our findings highlight that hospitals should not be solely relied upon for community health improvement. Therefore, future research should examine the impact of other organizations that are prevalent in communities, such as schools and clinics, to determine the degree to which they impact population health outcomes. These findings indicate a neutral or negative association between length of life, health behaviors, quality of life and clinical care for the general population based on the number of community healthcare access services and community building services offered by hospitals. Although these findings are only first steps towards explicating the population level health benefit of hospital sponsored community services, the results provide indication for the need to better measure variables associated delivery and quality of community health services. Conflicts of interest All authors declare no conflict of interest. References 1 OCED. OCED Better Life Index United States. 2016. Accessed 7 June 2017, 2017. 2 Braveman PA, Cubbin C, Egerter S et al.  . Socioeconomic disparities in health in the United States: what the patterns tell us. Am J Public Health  2010; 100( Suppl 1): S186– 96. Google Scholar CrossRef Search ADS PubMed  3 Lasser KE, Himmelstein DU, Woolhandler S. Access to care, health status, and health disparities in the United States and Canada: results of a cross-national population-based survey. Am J Public Health  2006; 96( 7): 1300– 7. Google Scholar CrossRef Search ADS PubMed  4 United States Department of Health and Human Services. Social Determinants. 2017. https://www.healthypeople.gov/2020/topics-objectives/topic/social-determinants-of-health. 5 Marmot MG, Bell R. Action on health disparities in the United States: commission on social determinants of health. J Am Med Assoc  2009; 301( 11): 1169– 71. Google Scholar CrossRef Search ADS   6 Scutchfield FD, Prybil L, Kelly AV et al.  . Public health and hospitals: lessons learned from partnerships in a changing health care environment. Am J Public Health  2016; 106( 1): 45– 8. Google Scholar CrossRef Search ADS PubMed  7 HHS. Area Health Resource Files (AHRF). 2016. http://ahrf.hrsa.gov/overview.htm. (accessed 12 December 2016). 8 AHA. About. 2016; http://www.ahadataviewer.com/about/. (accessed 12 November 2016). 9 Remington PL, Booske BC. Measuring the health of communities—how and why? J Public Health Manag Pract  2011; 17( 5): 397– 400. Google Scholar CrossRef Search ADS PubMed  10 County Health Rankings. The Ranking Methods. http://www.countyhealthrankings.org/ranking-methods 11 Dall TM, Gallo PD, Chakrabarti R et al.  . An aging population and growing disease burden will require a large and specialized health care workforce by 2025. Health Aff  2013; 32( 11): 2013– 20. Google Scholar CrossRef Search ADS   12 Fiscella K, Franks P, Doescher MP et al.  . Disparities in health care by race, ethnicity, and language among the insured: findings from a national sample. Med Care  2002; 40( 1): 52– 9. Google Scholar CrossRef Search ADS PubMed  13 Bertakis KD, Azari R, Helms LJ et al.  . Gender differences in the utilization of health care services. J Fam Pract  2000; 49( 2): 147– 7. Google Scholar PubMed  14 Laditka JN, Laditka SB, Probst JC. Health care access in rural areas: evidence that hospitalization for ambulatory care-sensitive conditions in the United States may increase with the level of rurality. Health Place  2009; 15( 3): 761– 70. Google Scholar CrossRef Search ADS PubMed  15 Norris SL, Nichols PJ, Caspersen CJ et al.  . The effectiveness of disease and case management for people with diabetes: a systematic review. Am J Prev Med  2002; 22( 4, Supplement 1): 15– 38. Google Scholar CrossRef Search ADS PubMed  16 Fitzpatrick T, Rosella LC, Calzavara A et al.  . Looking beyond income and education: socioeconomic status gradients among future high-cost users of health care. Am J Prev Med  2015; 49( 2): 161– 71. Google Scholar CrossRef Search ADS PubMed  17 Levesque J-F, Harris MF, Russell G. Patient-centred access to health care: conceptualising access at the interface of health systems and populations. Int J Equity Health  2013; 12( 1): 18. Google Scholar CrossRef Search ADS PubMed  18 Long JS, Freese J. Regression Models for Categorical Dependent Variables Using Stata . College Station, TX: StataCorp LP, 2006. 19 Prybil L, Scutchfield FD, Killian R et al.  . Improving Community Health Throught Hospital-Pubic Health Collaboration . Lexington: Commonwealth Center for Governance Studies, 2014. 20 Shortell S, Washington P, Baxter R. The contribution of hospitals and health care systems to community. Ann Rev Public Health  2009; 30: 373– 83. Google Scholar CrossRef Search ADS   21 Pennel CL, McLeroy KR, Burdine JN et al.  . Nonprofit hospitals’ approach to community health needs assessment. Am J Public Health  2015; 105( 3): e103– 13. Google Scholar CrossRef Search ADS PubMed  22 Government Accountability Office. Nonprofit Hospitals: Variation in Standards and Guidance Limits Comparison of How Hospitals Meet Community Benefit Requirements. Publication no. GAO-08-880. 2008; http://www.gao.gov/new.items/d08880.pdf 23 U.S. Department of Health and Human Services. Principles of Community Engagement. 2nd ed. 2011; https://www.atsdr.cdc.gov/communityengagement/pdf/PCE_Report_508_FINAL.pdfUpdated 24 Rimer BK, Conaway M, Lyna P et al.  . The impact of tailored interventions on a community health center population. Patient Educ Couns  1999; 37( 2): 125– 40. Google Scholar CrossRef Search ADS PubMed  25 Cleary PD, Gross CP, Zaslavsky AM et al.  . Multilevel interventions: study design and analysis issues. JNCI Monogr  2012; 2012( 44): 49– 55. Google Scholar CrossRef Search ADS   26 Pennel CL, McLeroy KR, Burdine JN et al.  . Nonprofit hospitals’ approach to community health needs assessment. Am J Public Health  2015; 105( 3): e103– 13. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of Faculty of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

Journal

Journal of Public HealthOxford University Press

Published: Jan 27, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off