Provider-to-population ratios are a primary measure of

Transcription

1 RN-to-Population Ratio and Population Health: A Multifactorial Study Jeri L. Bigbee, PhD, RN, FNP-BC, FAAN; Sandra Evans, MA.Ed, RN; Bonnie Lind, PhD; Susan Perez, MPH; Lissette Jacobo, BS; and Estella M. Geraghty, MD, MS, MPH, FACP, GISP The purpose of this study was to explore the relationship between registered nurse (RN)-to-population ratio and population health indices. A cross-sectional secondary analysis of existing national data was conducted, using counties as the unit of analysis. Data based on 1,929,414 RNs in 33 states in 2012 were obtained from the National Council of State Boards of Nursing s Nursys database, and county health data were obtained from 2,016 counties from the 2012 County Health Rankings database. Regression analysis indicated that the RN-to-population ratio along with nurse education (percentage of RNs with a BSN or higher degree) and experience (number of years since graduation) was significantly associated with the self-rated health (percentage of adults reporting fair or poor health), mammography screening rates, and teenage birth rates. The associations were all positive: Greater numbers of nurses per capita were associated with better population health indices. These results are consistent with the Nurse Dose concept and support the need for enhanced recognition and policy changes regarding the contribution of nurses to the health of the population. Implications for nursing regulation include the need for sustained and coordinated efforts to support the health of the population through the recruitment and maintenance of a welleducated and experienced nursing workforce. Provider-to-population ratios are a primary measure of workforce adequacy and planning. Benchmarks for physician-to-population and dentist-to-population ratios are the primary metrics in the federal designation of health professional shortage areas and medically underserved areas and populations (Malloch, Davenport, Milton, & Hatler, 2003). However, no such benchmarks exist for nurses, even though nurses constitute the largest group of health care professionals in the United States and the world (U.S. Department of Health and Human Services [HHS], 2013). According to the National Sample Survey of Registered Nurses conducted in March 2008, the national estimated ratio of employed registered nurses (RNs) per 100,000 population (the nurse-to-population ratio) was 854 up from 825 in 2004 (HHS, 2010). Despite the absence of a benchmark for nurses, many assume that an adequate supply of competent RNs is essential for a healthy population, though evidence is sparse. The Institute of Medicine (IOM) (2010) report, The Future of Nursing, raised national awareness about nursing s contribution as the largest health care profession and recommended improved data collection, analysis, and research related to health care workforce requirements. Addressing the basic question of whether the supply of RNs is associated with the health of communities is critical. This question is particularly relevant to nursing regulators because of their role in public protection. Nurse regulators promote the health of the population by serving as leaders at the local, state, and national levels, advocating for healthy communities by ensuring a quality nursing workforce. This article presents an analysis of the relationship between the number, educational level, expertise, and experience of RNs in a community and the health of the population. Review of the Literature Research examining the association between nurse-to-population ratio and population health is limited, and the results are mixed. A few small studies suggest that nurses have a positive effect on the health of populations (Bigbee, 2008); other studies show no significant effect (Bigbee, Evans, Nagel, & Kenski, 2010; Robinson & Wharrad, 2000). Much of the research is limited by the fact that only bivariate correlational analyses were used because sample sizes were small and adjustment for socioeconomic variables was not possible. However, several studies focusing on physician-topopulation ratios have used larger samples and have adjusted for socioeconomic variables. These studies suggest that primary care physician (PCP)-to-population ratios are associated with positive population health indicators, including self-rated health; life expectancy; low birth weight rates; and lower all-cause cancer, heart disease, stroke, and infant mortality rates (Macinko, Starfield, & Shi, 2007). A limitation of most research is that primarily state or national level data were used, rather than data from smaller geographic units, such as counties or census tracts, which can provide a more detailed assessment of local variations in health. 11

2 Conceptual Framework The Nurse Dose concept developed by Brooten and others served as the conceptual framework for this study (Brooten & Youngblut, 2006; Brooten, Youngblut, Deosires, Singhala, & Guido-Sanz, 2012; Manojlovich, Sidani, Covell, & Antonakos, 2011; Sidani, Manojlovich, & Covell, 2010). This concept includes three components: dose, nurse, and host response. In this study, the dose component was operationalized on the population level, using RN-to-population ratio as the primary indicator. The nurse component is defined by three elements: education, expertise, and experience. Extensive previous research has addressed these three elements, but most of this research focused on inpatient settings. The positive effect of nurses educational level on inpatient outcomes, including 30-day mortality and failure-torescue rates, is well established (Aiken, Clarke, Cheung, Sloane, & Silber, 2003; Kane, Shamliyan, Mueller, Duval, & Wilt, 2007; Kutney-Lee, Sloane, & Aiken, 2013; McHugh & Ma, 2013). Previous studies have measured nurse expertise in various ways, including level of licensure and nursing skill mix (the number of RNs proportionate to other nursing staff). Multiple studies have established that nursing staffing patterns with higher RN time per patient are associated with better inpatient outcomes, including lower failure-to-rescue rates, lower inpatient mortality rates, and shorter hospital stays (Dunton, Galewksi, Klaus, & Pierson, 2007; Lang, Hodge, Olson, Romano, & Kravitz, 2004; Shekelle, 2013). Studies also support the positive patient outcomes associated with advanced practice registered nurses (APRNs) (Brooten & Youngblut, 2006; Brooten et al., 2012). Regarding the third characteristic, experience, the evidence is mixed. Several studies failed to demonstrate a significant effect of experience (Aiken et al., 2003; Kutney-Lee et al., 2013; McHugh & Ma, 2013), though other studies supported the positive effect of nursing experience on outcomes (Dunton et al., 2007; Kendall-Gallagher & Blegen, 2009; Kutney-Lee, Lanke, & Aiken, 2009). None of these previous studies have focused on the macro level; that is, they have not evaluated the effect of nurses on the health of the population. The theoretical basis of this study also included the concept of localized social capital (Putnam, 1995), in which trust and communication of norms are developed as a result of informal social interactions among families and social groups. Lauder, Reel, Framer, and Griggs (2006) contend that this type of social capital, which creates forms of association that express trust and norms of reciprocity is fundamental to nursing practice, particularly in small communities (p. 75). They assert that nurses transect social networks as both health care providers and community members, utilizing their social capital in promoting the health of the community through both informal and formal functions. The power of nurses informal health promotive influence lies in their community embeddedness built on a variety of formal and informal networks (Lauder, Reel, Farmer, & Griggs, 2006). Thus, we hypothesized that nurses positively impact the health of local communities through their professional work as well as through more informal interactions and leadership, emphasizing health promotion and prevention, which are foundational to nursing s approach. Purpose of Study The purpose of this study was to explore the relationship between nurse-to-population ratio and population health indices. Two research questions were addressed: Are higher nurse-to-population ratios associated with better levels of population health indices? Do nurse-related factors, including education, expertise (level of licensure), and experience, influence the relationship between nurse-to-population ratio and population health indices? Method A cross-sectional secondary analysis of existing national data was conducted, examining nurse-to-population ratios along with nurse characteristics in relation to population health indices, using counties as the unit of analysis. Data were aggregated at the county level. The nursing data were extracted from the National Council of State Boards of Nursing s Nursys database, which includes state licensure data for RNs from all 50 states and the District of Columbia. Aggregated data from 33 states with the most complete RN data were used in the study, which included 1,929,414 RNs residing in 2,016 counties in The data included the total number of licensed RNs residing in each county, RN education level, and graduation dates. In addition, data regarding the number of APRNs by county (n = 1,507) were collected from individual state boards of nursing. Nurse-to-population ratios were computed using 2010 decennial U.S. Census data for county populations. The county-based population health data were drawn from the 2012 County Health Rankings (CHR) database (University of Wisconsin Population Health Institute, 2012). Variables included health outcomes, health behavior, clinical care, social and economic factors, and environmental indices. The population of the 2,016 counties ranged from 276 to 9,962,789 (mean = 89,100, SD = 330,952) with a mean population density of 147 people/square mile (range = 0.2 to 17,179, SD = 595). Of the 2,016 counties, 624 (31%) were metropolitan, and 1,392 (69%) were nonmetropolitan. A complete description of all indices, data sources, and measures included in the CHR is available at www. countyhealthrankings.org/our-approach. The data were analyzed using SPSS (version 18.0). Initial descriptive statistics, including RN-to-population ratios and nurse characteristics by county, were computed. This descriptive analysis was supplemented with visualization using Geographic 12 Journal of Nursing Regulation

3 FIGURE 1 Number of RNs per 100,000 Population Classified by Quantile ,138 1, ,359 3,731 No Data Information Systems analysis (ArcMap version 10.0, ESRI, Redlands, CA). Then, a set of linear regression models was constructed, using self-rated health (the percentage of the county adult population reporting fair or poor health), teen birth rate, and mammography screening rate as dependent variables. RNto-population ratio and nurse characteristics (education, expertise, and experience) were the independent variables. Education was measured using the percentage of RNs in the county holding a BSN or higher degree. Expertise was measured as the level of licensure, using the APRN-to-population ratio. Experience was measured indirectly, using the mean years since graduation by county. For ease of interpreting the Nurse Dose relationships, RN-to-population and APRN-to-population ratios were divided into quartiles. Quartile one (Q1) was the reference group in all models, and quartiles two (Q2), three (Q3), and four (Q4) were entered into the model as separate indicator variables. All models were first adjusted for county levels of income, education, race/ethnicity (coded as percentage non-hispanic white), PCP-to-population ratios, and county population. On the second step, the indicators for quartiles of RN-to-population ratio and nurse characteristics were added to the models. In all models assessing the nurse characteristics, the RN-to-population ratio was included as an adjustment variable in Step 1. Counties with greater than 10% missing data for education were excluded from the education and experience models. Results The descriptive analysis revealed that the mean RN-to-population ratio by county was 1,057 RNs/100,000 population (median = 1,042, SD = 382) with a range of 101 to 3,731/100,000 population. (See Figure 1.) The quartiles of the RN-to-population ratio were as follows: Q1, less than 788; Q2, 788 to 1,041; Q3, 1,042 to 1,296; and Q4, 1,297 and higher. The 10th percentile for RN-to-population ratio was 585/100,000. The number of currently licensed RNs per county ranged from 1 to 73,704 (mean = 958, SD = 2,913). The mean percentage of RNs with a BSN or higher degree by county was 24.9% (range = 0% to 100%, SD = 15.8). The mean APRN-to-population ratio by county (based on 1,507 counties) was 55/100,000 (range = 0 to 411, SD = 38.4). The average number of years since graduation by county was 18.1 (range = 6 to 45; SD = 2.9). (See Table 1.) The regression analysis results assessing the effects of overall RN-to-population ratio are presented in Table 2. After 13

4 TABLE 1 RN Descriptive Findings by County Variable N Mean SD Range RN-to-population ratio 2,016 1, ,731 (per 100,000) Number of licensed 2, , ,704 RNs per county APRN-to-population ratio 1, (per 100,000) Number of licensed 1, ,773 APRNs per county RN education (% BSN 1, or higher) RN experience (average years since graduation) 1, adjustment for the confounding variables, RN-to-population ratio was found to be modestly but significantly associated with the percentage of the population reporting fair or poor health (R 2 = 0.30), mammography screening rates (R 2 = 0.28), and teen birth rates (R 2 = 0.48). In all models, higher RN-to-population ratios were associated with healthier county health indices; compared with Q1, counties in the highest quartile had a 2.8% lower rate of fair or poor health, a 6% higher mammography screening rate, and a 9/1,000 lower rate of teen births (p <.001 for all). After adjustment for the same confounding variables along with the overall RN-to-population ratio, APRN-to-population ratio was a significant predictor of mammography rates and teen birth rates, but not self-rated health. However, the APRN-topopulation ratio did not add substantially to the explanatory power of any of the models. The regression models assessing education and experience indicated that higher percentages of BSNs and more experienced nurses were strongly associated with healthier county indices. (See Table 3.) The effects of the education and experience variables were larger than the overall effect of the RN-to-population ratio. Discussion This exploratory study represents the first major multifactorial investigation of the relationship between RN-to-population ratio and population health in the United States. The findings indicate a significant association between RN-to-population ratio and the population health indices of self-rated health, teen birth rates, and mammography screening rates, which is in contrast with some previous research. The reason for the contrast may be that previous population-based studies failed to examine health measures other than basic mortality and morbidity indices. In addition, few previous studies used counties as the unit of analysis; thus, localized effects that were evident in this study may have been masked by aggregation of data in larger units such as states. The RN-to-population ratios determined in this study differ from those in the latest National Sample Survey of Registered Nurses (HHS, 2010), which was based on sample data. The overall national employed RN-to-population ratio as reported in the 2008 National Sample Survey was 854/100,000 compared with the mean county RN-to-population ratio of 1,057/100,000 in this study. This difference in ratios could result from multiple factors: the possibility that the states in this study have higher RN-to-population ratios than the excluded states; RNs who are licensed but not actively employed; the effect of counting RNs licensed in multiple states more than once; an increase in the RNto-population ratio over time, which is consistent with previous periods; or a possible sampling error in the 2008 National Sample Survey. Continued monitoring of nursing workforce trends over time is clearly indicated. Theoretically, the findings support the Nurse Dose concept, suggesting that greater nurse doses, as measured in this study by RN-to-population ratios, may be associated with healthier populations. However, the associations suggested by the results may not necessarily indicate causation. For example, does the presence of more RNs per capita result in lower teen birth rates, or might RNs choose to live in communities with lower teen birth rates? The results provide insight regarding the possible effects of the nurse characteristics: education, expertise, and experience. The findings indicated that the more BSN- or higher-educational level nurses living in a community, the healthier the population. This finding supports the current consensus that well-educated nurses are essential not only to promoting quality health care, but to the health of the population as a whole. In The Future of Nursing, the IOM recommends that the proportion of U.S. nurses educated at the baccalaureate level be increased to 80% by 2020 (Institute of Medicine, 2010). Our findings indicate that, on average, approximately 25% of RNs hold a BSN or higher degree; others estimate that about 45% of the U.S. RN workforce hold a BSN or higher degree (HHS, 2010). The need to aggressively expand nursing educational opportunities at the baccalaureate and graduate levels is clear. The finding that experience (as measured by years since graduation) is strongly related to positive population health indices contrasts with some earlier studies focusing on inpatient outcomes. This finding supports our conceptual framework related to social capital and particularly the aspect of community embeddedness, in which experienced nurses may be more integrated in their communities, assuming formal and informal leadership roles, and thereby positively impacting the health of the community through health promotion, education, and advocacy. Considering this, further research is needed to more fully explore the contribution of experienced nurses to their communities. 14 Journal of Nursing Regulation

5 The results of this study did not support the substantial effect of the third nurse characteristic, expertise (as measured by APRN level of licensure), on population health. However, these findings are limited by the fact that approximately half of the counties had 10 or fewer APRNs. Given the wealth of previous research demonstrating the positive effects of APRNs on individual health outcomes, further research is indicated to assess the effect of level of licensure on population health. In addition, in future studies, nurse expertise could be measured in greater depth by including licensed practical and vocational nurses and specialty certification rates. Limitations The limitations of this study include the fact that the sample comprised only 33 states; however, the states represented all regions of the country and included predominantly rural and urban states. The fact that RN location was based on county of residence also represents a possible limitation because nurses often practice outside of their local area of residence. Conceptually, though, we contend that RNs may impact the health of their communities through informal interactions and leadership within their local environments consistent with the concept of social capital, in addition to their professional work. Nurses, their families, and their networks may influence the community by modeling positive health status and behaviors. Also, because the RN data were based on state licensure records, RNs who were registered in more than one state could have been counted more than once, somewhat inflating the overall RN counts. TABLE 2 Regression Results for RN-to-Population Ratio β SE β p F R 2 Model 1 Dependent Variable: Percentage Fair or Poor Health Adjustment variables a RN-to-population ratio Quartile Quartile <.001 Quartile <.001 Model 2 Dependent Variable: Mammography Screening Rate Adjustment variables a RN-to-population ratio Quartile <.001 Quartile <.001 Quartile <.001 Model 3 Dependent Variable: Teen Birth Rate b Adjustment variables a RN-to-population ratio Quartile <.001 Quartile <.001 Quartile <.001 a Adjustment variables: education, income, race/ethnicity (percentage non-hispanic white), county population, primary care physicians per capita. b Teen births per 1,000 females ages 15 to 19. Implications The study has important implications for policy, education, and future research. Establishing a benchmark for nurse-to-population ratio similar to that established for other health professions should be a policy priority to ensure an adequate supply of experienced, well-educated nurses. The specialty areas of public health nursing and school nursing have advocated for specific nurse-to-population ratios for many years with limited success. Preliminary benchmarks suggested by this study, including the median RN-to population ratio of 1,042/100,000 (1 RN/96 people) and the 10th percentile of 585 RNs/100,000 population (1 RN/171 people), could serve as starting points for establishing a national RN-to-population benchmark. In this effort, nursing regulators can join with consumer and professional groups to support a national nurse-to-population benchmark. Regulatory implications include maintaining high-quality nursing education opportunities that are widely accessible, particularly at the baccalaureate and graduate levels. Ensuring continuing competency of licensed nurses promotes the health of the population. Based on the finding regarding the strong effect of nursing experience on the health of communities, continued support by all professional organizations of our most experienced nurses in staying current and active in their communities is imperative. Future nurse-to-population research is needed to assess causation using a multivariate approach. Longitudinal studies could track nurse-to-population ratios over time in relation to population health trends. Comparative interprofessional research that addresses the contributions of nurses and other health care providers to the population s health should also be conducted. Internationally, similar studies are needed in developing nations and other areas for comparison with this study. Given the varying levels of health among local areas such as counties, an alternative hypothesis could be proposed that some basic underlying factors, such as a positive health promotive culture, economy, or political environment, attract more nurses (and other providers) to live and practice in those areas and produce more positive population 15

6 TABLE 3 Regression Results for RN Education and Experience β SE β p F R 2 Model 1 Dependent Variable: Percentage Fair or Poor Health Adjustment variables a RN education & experience % BSN or higher <.001 Mean years since graduation <.001 Model 2 Dependent Variable: Mammography Screening Rate Adjustment variables a RN education & experience % BSN or higher Mean years since graduation <.001 Model 3 Dependent Variable: Teen Birth Rate b Adjustment variables a RN education and experience % BSN or higher <.001 Mean years since graduation <.001 a Adjustment variables: education, income, race/ethnicity (percentage non-hispanic white), county population, primary care physicians per capita, total RNs per capita. b Teen births per 1,000 females ages 15 to 19. health characteristics. Identifying the possible underlying factors not addressed in this study would be useful in both public health promotion as well as workforce planning. Summary The results of this national multifactorial study that explored the relationship between RN-to-population ratio and population health indices suggest a positive association between the number of RNs per capita and measures of population health. The findings suggest that experienced nurses and RNs educated at the baccalaureate or higher level may have the greatest positive effect on the health of their communities. The results support the establishment of a national nurse-to-population ratio benchmark, similar to benchmarks currently in place for other health professions. Establishment of this benchmark is critical in addressing the long-term disparities in RN-to-population ratios affecting underserved populations. If further research supports the suggestion of nurses unique contribution to the health of the population, then policy and regulatory implications related to enhancing the nursing workforce through recruitment, education, and retention efforts are clear. References Aiken, L. H., Clarke, S. P., Cheung, R. B., Sloane, D. M., Silber, J. H. (2003). Educational levels of hospital nurses and surgical patient mortality. JAMA, 290, Bigbee, J. L. (2008). The relationships between nurse and physician-topopulation ratios and state health rankings. Public Health Nursing, 25, Bigbee, J. L., Evans, S., Nagel, J., & Kenski, D. (2010). Using county health rankings to assess population health. Northwest Public Health, 27(2), Brooten, D., & Youngblut, J. M. (2006). Nurse dose as a concept. Journal of Nursing Scholarship, 38(1), Brooten, D., Youngblut, J. M., Deosires, W., Singhala, K., & Guido-Sanz, F. (2012). Global considerations in measuring effectiveness of advanced practice nurses. International Journal of Nursing Studies, 49(7), Dunton, N., Galewksi, B., Klaus, S., & Pierson, B. (2007). The relationship of nursing workforce characteristics to patient outcomes. The Online Journal of Issues in Nursing, 12(3), Manuscript 3. Institute of Medicine. (2010). The future of nursing: Leading change, advancing health. Washington, DC. Retrieved from Reports/2010/The-Future-of-Nursing-Leading-Change-Advancing- Health.aspx Kane, R. L., Shamliyan, T., Mueller, C., Duval, S., & Wilt, T. J. (2007). The association of registered nurse staffing levels and patient outcomes. Medical Care, 45(12), Kendall-Gallagher, D., & Blegen, M. A. (2009). Competence and certification of registered nurses and safety of patients in intensive care units. American Journal of Critical Care, 18(2), Kutney-Lee, A., Lanke, E. T., & Aiken, L. H. (2009). Development of the hospital nurse surveillance capacity profile. Research in Nursing and Health, 32, Kutney-Lee, A., Sloane, D. M., & Aiken, L. H. (2013). An increase in the number of nurses with baccalaureate degrees is linked to lower rates of post-surgery mortality. Health Affairs, 32(3), Lang, T. A., Hodge, M., Olson, V., Romano, P. S., & Kravitz, R. L. (2004). A systematic review on the effects of nurse staffing on patient, nurse employee, and hospital outcomes. Journal of the Nursing Administration, 34(7/8), Lauder, W., Reel, S., Farmer, J., & Griggs, H. (2006). Social capital, rural nursing and rural nursing theory. Nursing Inquiry, 13(1), Macinko, J., Starfield, B., & Shi, L. (2007). Quantifying the health benefits of primary care physician supply in the United States. International Journal of Health Services, 37(10), Malloch, K., Davenport, S., Milton, D., & Hatler, C. (2003). Nursing workforce management: Using benchmarking for planning and outcomes monitoring. Journal of Nursing Administration, 33(10), Manojlovich, M. Sidani, S., Covell, C. L., & Antonakos, C. L. (2011). Linking staffing variables to adverse patient outcomes. Nursing Research, 60(4), Journal of Nursing Regulation

7 McHugh, M. D., & Ma, C. (2013). The effect of hospital nursing on 30- day readmission among Medicare patients with heart failure, acute myocardial infarction, and pneumonia. Medical Care, 51, Putnam, R. (1995). Bowling alone: America s declining social capital. Journal of Democracy, 6, Robinson, J., & Wharrad, H. (2000). Invisible nursing: Exploring health outcomes at a global level: Relationships between infant and under-5 mortality rates and the distribution of health professionals, GNP per capita, and female literacy. Journal of Advanced Nursing, 32(1), Shekelle, P. G. (2013). Nurse-patient ratios as a patient safety strategy: A systematic review. Annals of Internal Medicine, 158(5), Sidani, S., Manojlovich, M., & Covell, C. (2010). Nurse dose: Validation and refinement of a concept. Research and Theory for Nursing Practice, 24(3), University of Wisconsin Population Health Institute. (2012). County health rankings Madison, WI. Retrieved from www. countyhealthrankings.org U.S. Department of Health and Human Services. (2010). The Registered Nurse Population: Findings from the 2008 National Sample Survey of Registered Nurses. Washington, DC. Retrieved from gov/healthworkforce/rnsurveys/rnsurveyfinal.pdf U.S. Department of Health and Human Services. (2013). Negotiated rulemaking committee on designation of MUPs and HSAs. Washington, DC. Retrieved from Jeri L. Bigbee, PhD, RN, FNP-BC, FAAN, is an adjunct professor, Betty Irene Moore School of Nursing, University of California, Davis. Sandra Evans, MA.Ed, RN, is Executive Director, Idaho Board of Nursing. Bonnie Lind, PhD, is a biostatistician, WhiteCloud Analytics. Susan Perez, MPH, is a doctoral candidate, Betty Irene Moore School of Nursing, University of California, Davis. Lissette Jacobo, BS, is a research assistant, Betty Irene Moore School of Nursing, University of California, Davis. Estella M. Geraghty, MD, MS, MPH, FACP, GISP, is Associate Professor of General Internal Medicine, School of Medicine, University of California, Davis. The authors acknowledge the funding support from the National Council of State Boards of Nursing Center for Regulatory Excellence (Grant #40001) along with the assistance of Janice Bell, PhD, MPH, MN, and Jesse King. 17