INTRODUCTION AND SPECIFIC AIMS “If burnout only

INTRODUCTION AND SPECIFIC AIMS “If burnout only affected individuals in isolation, it would be far less…devastating in its impact than it is. Burnout in human services agencies is like an infection in hospitals: it gets around. It spreads…to clients.”- Edelwich & Brodsky, 1980 In recent years, healthcare provider burnout has rightfully been described as a health care crisis and an epidemic that must be treated as a national priority (Jha, Ilif, & Chaoui, 2019; Shin, Gandhi, & Herzig, 2016). Alarmingly, over 40% of hospital nurses—who form the front line of healthcare delivery to millions of patients around the clock—suffer from burnout (Aiken, Clarke, Sloane, Lake, & Cheney, 2008). A 2019 national survey found that nearly 75% of Americans are concerned about healthcare provider burnout, and nearly 80% fear that provider burnout diminishes the quality of their care and threatens their safety (The Harris Poll & American Society of Hospital Pharmacists, 2019). And they are not wrong. Hospital nurse burnout has been found to be linked to negative patient outcomes such as decreased patient satisfaction (Leiter, Harvie, & Frizzell, 1998; McHugh, Kutney-Lee, Cimiotti, Sloane, & Aiken, 2011; Vahey, Aiken, Sloane, Clarke, & Vargas, 2004) reports of poor quality of care (Laschinger, Shamian, & Thomson, 2001; Parker & Kulik, 1995; Poghosyan, Clarke, Finlayson, & Aiken, 2010; Van Bogaert, Clarke, Roelant, Meulemans, & Van de Heyning, 2010), safety and overall adverse events (Leiter & Laschinger, 2006; Liu et al., 2018). Over 440,000 people die annually from preventable adverse events, making it the third leading cause of death in the United States (Advisory Board Company, 2013; James, 2013; Makary & Daniel, 2016). This staggering number of preventable deaths is made even worse by the fact that adverse events are known to be underestimated (Classen et al., 2011; U.S. Department of Health and Human Services, 2012). In addition to the grave cost in human lives, these errors cost hospitals over $30 billion dollars annually (Institute of Medicine [IOM], 2000). Globally, one out of every ten hospitalizations results in preventable adverse events (Jha et al., 2013). Nationally, it has been estimated that one out of every three hospital admissions results in patient harm from a preventable adverse event (Classen et al., 2011). The pervasiveness of hospital nurse burnout and preventable adverse events—and their dire consequences—demand a deeper investigation into how, and to what extent, hospital nurse burnout leads to preventable patient adverse events. Background “How well we are cared for by nurses affects our health, and sometimes can be a matter of life or death” Institute of Medicine, 2004, pg. 2 Central to understanding and improving hospital patient outcomes is a thorough assessment of the preceding structures and processes of care delivery by nurses, the primary providers of direct care to hospital patients 24 hours a day (IOM, 2004). Given the pivotal role nurses play in the provision of care, it is necessary to examine factors that influence their performance. Burnout, an occupationally-based, multidimensional phenomenon, has been found to negatively impact job performance (Maslach, Jackson, & Leiter, 1996). The majority of research on nurse burnout has predominantly focused on two areas: predictors and job consequences. Studies have found that interpersonal relationships and organizational characteristics of work environments, such as poor staffing and quality of management, are determinants of nurse (RN) burnout (Aiken, Clarke, Sloane, Sochalski, & Silber, 2002; Vahey et al., 2004; Van Bogaert et al., 2010). Other studies have linked burnout to negative job outcomes, including decreased job satisfaction, absenteeism, intent-to-leave and turnover (Aiken et al., 2001; Maslach, Schaufeli, & Leiter, 2001; McHugh et al., 2011). However, the relationship between provider burnout, job performance and patient outcomes remains relatively understudied (Halbesleben, Rathert, & Williams, 2013). Nurse burnout has been found to be associated with: decreased patient satisfaction (Leiter et al., 1998; McHugh et al., 2011; Vahey et al., 2004), increased urinary tract and surgical site infections (Cimiotti, Aiken, Sloane, & Wu, 2012), altered medication administration practices (Halbesleben et al., 2013), poor overall quality of patient care (Poghosyan et al., 2010; Van Bogaert et al., 2010), safety and overall adverse events (Laschinger & Leiter, 2006; Liu et al., 2018). The findings of these studies warrant further inquiry into the mechanisms by which nurse burnout leads to poor patient outcomes. One possible pathway emerges from evidence that patient outcomes are adversely affected when nurses are unable to provide necessary care (Papastavrou, Andreou, & Efstathiou, 2014). In health services research, the provision of care is also referred to as the process of care, or care provided and received (Donabedian, 1988). One process of care measure that has been proposed in the literature, missed care – defined as necessary patient care left undone – is considered to be an error of omission in patient safety literature (Kalisch, Landstrom, & Hinshaw, 2009). In a systematic review of 17 quantitative studies, missed nursing care was found to be associated with decreased patient safety and quality of care (Papastavrou et al., 2014). Studies have also linked missed care to medication errors, pressure ulcers, falls, nosocomial infections (Kalisch, Xie, & Dabney, 2014; Lucero, Lake, & Aiken, 2010; Schubert et al., 2008), patient satisfaction (Lake, Germack, & Viscardi, 2016; Schubert et al., 2008), readmissions (Brooks Carthon, Lasater, Sloane, & Kutney-Lee, 2015; Brooks Carthon, Lasater, Rearden, Holland, & Sloane, 2016), increased inpatient mortality (Ball et al., 2018; Schubert, Clarke, Aiken, & de Geest, 2012), and nurse perceptions of quality of care and patient safety (Ball, Murrells, Rafferty, Morrow, & Griffiths, 2014). The next logical inquiry is whether, and to what extent, missed care serves as a pathway between nurse burnout and patient outcomes. Though not explicitly studied to date, a number of studies suggest that a link may indeed be present. One study assessing nurse burnout in nursing homes found a significant association with missed care (White, Aiken, & McHugh, 2019). Research has shown that burnout is associated with poor job performance, decreased productivity and effectiveness at work (Dewa, Loong, Bonato, Thanh, & Jacobs, 2014; Maslach et al., 2001; Parker & Kulik, 1995), as well as prolonged stress-related health outcomes, including insomnia (Vela-Bueno et al., 2008) and substance abuse (Maslach et al., 2001). The provision of required nursing care such as surveillance, communication, and education may be negatively impacted by burned out nurses who have fewer physical, cognitive and emotional resources (Halbesleben & Rathert, 2008). Lower productivity and efficiency may lead to greater amounts of missed care. Furthermore, nurses must constantly make cognitively complex decisions about patient care involving attention, thought, knowledge, and judgment (IOM, 2004), all of which may be affected by burnout. As burnout leads to withdrawal and distancing from work, burned out nurses may make suboptimal decisions, taking shortcuts and doing the bare minimum as opposed to adhering to best practices (Maslach, 2003). Given this context and body of literature, the purpose of this study is to expand our understanding of how – and to what extent – hospital nurse burnout impacts the delivery of care and patient outcomes, specifically nurse-reported frequent adverse events. This study hypothesizes that hospitals with higher proportions of burned out nurses are more likely to have higher proportions of missed care, which may partially explain the link between hospital nurse burnout and nurse-reported frequent adverse events. Study Overview, Specific Aims, and Hypotheses This study is a secondary data analysis of three linked datasets. Measures of nurse burnout, nurse-reported adverse events, and additional nursing factors were derived from the University of Pennsylvania’s Multi-State Nursing Care and Patient Safety Survey, conducted from 2005-2008 (Aiken et al., 2011). Information on the structural characteristics of study hospitals was obtained from the American Hospital Association’s Annual Survey. Information on patient illness severity was obtained from the Centers for Medicare & Medicaid Services’ Provider Specific File data. To achieve the study’s purpose, two specific aims were addressed: Specific Aim 1: To examine the relationship between hospital nurse burnout and patient outcomes, specifically nurse-reported adverse events. Hypothesis 1: Hospitals with higher proportions of burned out nurses will have increased odds of nurses reporting frequent adverse events, including medication errors, pressure ulcers, falls with injury, hospital-associated urinary tract infections and hospital-associated central line infections. Specific Aim 2: To determine whether missed care partially mediates the relationship between hospital nurse burnout and patient outcomes, specifically nurse-reported adverse events. Hypothesis 2: Hospitals with higher proportions of burned out nurses will have higher proportions of missed care. Missed care will partially explain the relationship between hospital nurse burnout and increased odds of nurses reporting frequent adverse events, including medication errors, pressure ulcers, falls with injury, hospital-associated urinary tract infections and hospital-associated central line infections. Study Significance and Policy Implications To the author’s knowledge, this is the first large-scale study in the U.S. to assess the impact of hospital-level nurse burnout on nurse-reported medication errors, falls, pressure ulcers and central line infections while accounting for potential patient, nurse and hospital confounders. Moreover, this study uniquely advances the existing body of research linking nurse burnout to patient outcomes by being among the first to empirically examine missed care as a possible pathway between nurse burnout and preventable adverse events. Additionally, with a highly representative sample of 23,784 nurses from 587 hospitals in four geographically diverse states that comprise a quarter of the U.S. population, this study’s findings are widely generalizable. Notably, four out of the five outcomes individually assessed in this study are indictors of patient safety targeted by federal and state regulations, reimbursement schemes, and public reporting mandates. Specifically, key changes in policy stemming from the Deficit Reduction Act (DRA) of 2005 and the Affordable Care Act (ACA) of 2010 spurred regulations to incentivize hospitals to improve performance on these outcomes. The Hospital-Acquired Conditions Reduction Program (HAC) of the ACA is one such regulation pursuant to which the Centers for Medicare & Medicaid Services (CMS) reduces hospital payments for hospitals with total HAC scores in the worst performing quartile (CMS, 2019). Since its inception, for poor performance on program metrics, the HAC program has withheld over $350 million from hospitals every year (CMS, 2019b). Therefore, hospital administrators have increased motivation to understand modifiable factors such as nurse burnout that influence their organizational performance on these regulated measures. Additionally, this study is important in light of the U.S. Department of Health and Human Services’ Partnership for Patients Initiative. This national public-private partnership is pursuing the goal of reducing hospital-acquired conditions, including the five study outcomes, by 20% in five years (CMS, 2019a). This widespread collaborative includes over 80% of acute care hospitals in the U.S. (Clarkwest et al., 2014). This study’s findings are especially pertinent to this program by providing empirical evidence of a unique strategy that may contribute to the central aim of this initiative. Further highlighting the significance and timeliness of this study’s findings is its alignment with national efforts aimed at reducing provider burnout. The Action Collaborative on Clinician Well-Being and Resilience, led by the National Academy of Medicine, maintains key goals that were addressed by this study: raising the visibility of clinician burnout and emphasizing evidenced-based solutions to improve patient safety by reducing clinician burnout (National Academy of Medicine, 2019). Furthermore, this study’s results may help inform healthcare stakeholders in the international community as well. It is known that nurse burnout and missed care are both highly prevalent in hospitals around the globe (Aiken et al., 2011; Ausserhofer et al., 2014; Poghosyan et al., 2010; Schubert et al., 2012). With over 40 million adverse events resulting in the loss of over 20 million disability-adjusted life years annually, patient harm as a result of unsafe care in hospitals has become a major policy emphasis globally (Jha et al., 2013). Finally, by linking nurse burnout, missed care, and preventable adverse events, this study arguably provides empirical support to the notion that reducing burnout may be foundational to achieving the Triple Aim—a blue print to optimizing health system performance by the simultaneous pursuit of improving the health of populations, enhancing the patient experience of care, and reducing per-capita costs of healthcare (Whittington, Nolan, Lewis, & Torres, 2015). Healthcare leaders have called for the reduction of provider burnout to be added as a fourth pillar to a new Quadruple Aim—an approach this study conceptually and empirically supports (Bodenheimer & Sinsky, 2014; Sikka, Morath, & Leape, 2015). By assessing a mechanism by which nurse burnout impacts patient outcomes, this study substantially advances existing literature and will be of interest to hospital administrators, payers, federal and state agencies, policymakers, patient advocacy organizations, researchers, providers and patients alike. By highlighting negative patient outcomes of a globally pervasive phenomenon, this study seeks to inform leaders in healthcare, spur political advocacy, and empirically bolster the rationale for allocating resources towards reducing hospital nurse burnout—which may be a unique and necessary strategy for mitigating preventable harm to patients, reducing costs, and improving patient outcomes. METHODOLOGY The study assessed the relationship between nurse burnout, missed care and patient outcomes. The design was a secondary analysis of linked cross-sectional data obtained from three data sources: the 2005-2008 Multi-State Nursing Care and Patient Safety Survey, the 2006-2007 American Hospital Association’s Annual Survey of Hospitals, and the 2006-2007 Centers for Medicare and Medicaid Services (CMS) Provider Specific File. Data Sources Multi-State Nursing Care and Patient Safety Survey The Multi-State Nursing Care and Patient Safety Study is the parent study upon which the secondary data analysis is based. This survey of nurses was conducted from 2005 – 2008 by the Center for Health Outcomes and Policy Research at the University of Pennsylvania, supported by the National Institute of Nursing Research (RO1NR04513) and led by Principal Investigator, Dr. Linda Aiken (Aiken et al., 2011). The primary goal of the survey was to obtain a broad range of information on nurses and organizational features of their work environments to conduct health services research, inform public policy, and improve patient care and safety. The survey included questions related to nurse demographics, nursing education, staffing levels, work hours, missed nursing care and the quality of the work environment. The survey also included items related to nurse job outcomes, including satisfaction, burnout, and intent to leave, as well as patient outcomes, including nurse reports of adverse events, quality of care and patient safety (Aiken et al., 2011). Over 272,783 nurses were randomly sampled from state nurse licensure lists from four states (Aiken et al., 2011). Fifty percent of all nurses licensed in New Jersey (52,545 nurses), 25% of nurses in Florida (52,545 nurses) and 40% of all nurses in Pennsylvania (64,321 nurses) and California (106,532 nurses) directly received the survey by mail at their home address (Aiken et al., 2011; McHugh et al., 2011). Survey data were collected in New Jersey, Pennsylvania, and California between September 2005 and August 2006, and in Florida from November 2007 through April 2008 (Aiken et al., 2011). A modified Dillman approach was utilized, which included follow-up postcard reminders and an additional mailing of the survey to non-respondents (Dillman, 1978; Dillman, Smyth, & Christian, 2014). To ensure confidentiality, the survey included a perforated area for nurses to remove their identifiable personal information prior to returning the survey. The survey asked nurses to provide the name of their employer. In this way, the names of hospitals were obtained without compromising the privacy of respondents. The survey specifically stated that the employing institutions would never be identified by name in subsequent research. The hospital names were necessary to be able to aggregate the nurse responses to the hospital level, link to existing data, and ultimately assess organizational burnout, performance and outcomes. The initial survey response rate was 39%, and to assess non-response bias that could negatively impact validity, a second random sample of 1300 non-respondents was conducted with phone calls and monetary incentives (Aiken et al., 2011; McHugh et al., 2011). This sample of non-respondents yielded a 91% response rate and it was found that there were no statistically significant differences among the reports of respondents and non-respondents on the relevant variables in this study (Smith, 2009). This double- sampling approach to address nonresponse bias in surveys of front-line nurses used as informants of organizational quality and safety has been shown to yield representative, unbiased samples (Lasater et al., 2019). Furthermore, by surveying nurses directly in their homes as opposed to collecting primary data in hospitals, hospital response bias, a threat to validity, is reduced (Aiken et al., 2011; Kutney-Lee, Lake, & Aiken, 2009). In all, about 100,000 nurse-respondents provided information on nine out of every ten hospitals in all four states (Aiken et al., 2011). Literature supports the approach of surveying nurses to obtain accurate, valid and reliable information on hospital organizations and outcomes (Aiken, Sochalski, & Lake, 1997; Lasater et al., 2019; McHugh & Stimpfel, 2012), and many studies have used nurse-reported information as an outcome measure (Aiken et al., 2001; Kutney-Lee et al., 2009; Lucero et al., 2010; Schubert et al., 2008). When compared to other process and outcomes data sources, such as administrative and patient-reported data, nurse-reported data have been found to be a valid and reliable measure of patient care quality and safety outcomes (Cina-Tschumi, Schubert, Kressig, De Geest, & Schwendimann, 2009; Gerolamo, 2008; McHugh & Stimpfel, 2012; Smeds-Alenius, Tishelman, Lindqvist, Runesdotter, & McHugh, 2016). American Hospital Association’s Annual Survey of Hospitals The American Hospital Association (AHA) Annual Survey of Hospitals is a national survey of all hospitals in the United States and has been conducted by the AHA since 1946 (American Hospital Association, 2014). The survey contains hundreds of items of information, including organizational structure, expenses and geographic indicators on over 6,300 hospitals (AHA, 2014). The AHA Annual Survey data from 2006-2007 was used to acquire hospital identification numbers, bed size, teaching status, technological capacity, and ownership—that were used as control variables in the analysis. The AHA Annual Survey data from 2006 was linked to the nurse survey data from New Jersey, Pennsylvania, and California, whereas AHA data from 2007 was linked to the nurse survey data from Florida (Aiken et al., 2011). The Centers for Medicare and Medicaid Services Provider Specific File Similarly, the CMS Provider Specific File data from 2006-2007 was the source for the variable used to control for patient severity, case mix index (CMI). CMI is a hospital-level value indicating the average clinical complexity and resource needs of all patients who received care in a hospital (CMS, 2016; U.S. Department of Health and Human Services [DHHS], 2018a). A higher CMI indicates a more complex patient population. The value is calculated by summing the Medicare Severity-Diagnosis Related Group (MS-DRG) weight of each discharge and dividing by the total number of discharges (DHHS, 2018a). Patients are assigned to one MS-DRG based on multiple factors including the principal and secondary diagnoses, age, sex, procedures performed, comorbidities, complications and discharge status. Although designed by CMS, the MS- DRG weights apply to all patient discharges across all payers (DHHS, 2018a). Nurses Sample The nurse sample, obtained from the Multi-State Nursing Care and Patient Safety Survey, included registered nurses who provided direct patient care as staff nurses in adult, non-federal, acute care hospitals. The rationale for selecting inpatient nurses in acute care settings is that they are the primary providers of care to patients, optimally positioned on the frontlines, 24/7. Furthermore, nurses that reported being assigned 20 or more patients were excluded as these nurses are most likely not staff nurses providing direct bedside care (McHugh et al., 2011). The final study sample included 23,784 registered nurses. Hospitals Hospitals included in the study are adult, non-federal, acute care hospitals from the four states included in the Multi-State Nursing Care and Patient Safety Survey: New Jersey, Pennsylvania, Florida and California. The rationale for excluding federal hospitals, including the Veterans Health Administration (VHA) hospitals, is that these hospitals have uniquely differing patient populations and characteristics, administrative protocols, survey sampling, and reporting measures (U.S. Veterans Health Administration, 2012). Hospitals with less than 10 nurse respondents from the nurse survey were excluded. Aggregating responses of multiple informants, and specifically 10 informants, has been found to be a reliable measure of organizational performance (Marsden et al., 2006). Furthermore, previous studies using the same nurse survey data have found this number to be sufficient for providing reliable information on hospital organizations and patient outcomes (Aiken et al., 2011; Stimpfel, Sloane, & Aiken, 2012). The final study sample included 587 hospitals. Study hospitals had an average of 68 nurse survey respondents, ranging from 10 to 245 nurses per hospital. Study Variables and Measurement The following section describes the study variables, and how they were operationally defined and measured. The primary independent variable was nurse burnout. The study outcomes were nurse-reported, hospital-associated frequent adverse events, including medication errors, pressure ulcers, falls, urinary tract and central line infections. Missed care was examined to assess whether it acts as a partial mediator in the relationship between nurse burnout and patient outcomes. The covariates of the study are also described in detail below. The section ends with a detailed table of study variables that includes the data source, level of analysis, type of variable, as well as how the variable is defined, categorized and measured in the study. Nurse burnout and missed care were analyzed at the hospital level; adverse events were analyzed at the nurse level. The study’s conceptualization of burnout as a collective, organizational feature supports the hospital level measurement of burnout. This collective experience of burnout measured at the organizational-level is empirically supported (Halbesleben & Leon, 2014). Additionally, the fact that hospital care is provided to patients by multiple nurses over time, and often across units, further supports hospital-level analyses of the independent variables that are thought to impact patient outcomes (Kutney-Lee et al., 2009). Key Study Variables Nurse burnout. The primary explanatory variable in the study, nurse burnout, was measured using the Emotional Exhaustion (EE) subscale of the Maslach Burnout Inventory- Human Services Survey (MBI-HSS) tool (Maslach et al., 1986) that is embedded within the Multi-State Nurse Survey. The MBI is a reliable and valid instrument and is considered the gold standard for measuring provider burnout (Halbesleben & Buckley, 2004). Rationale for the measurement of nurse burnout using the EE subscale is explicated in Chapter 2 and is also consistent with previous studies (Cimiotti et al., 2012; Halbesleben et al., 2013; Maslach et al., 1996; McHugh et al., 2011). Furthermore, internal consistency was assessed across the EE subscale questions and was found to be reliable with a Cronbach’s alpha of 0.92. The EE subscale consists of nine items that use a 7-point Likert scale to assess the frequency of negative feelings comprising burnout, ranging from never to every day (Maslach et al., 1996). EE subscale items include declarative statements to gauge, for example, how often respondents feel they are working too hard on the job and feel burned out at work. Each of the nine items of the EE subscale are given a numerical value from zero to six depending on the frequency the respondent selected, with zero being never and six being every day. The number-value assigned to each of the nine EE subscale items are summed to create a subscale score. The range of theoretical scores on the EE subscale is zero to 54. Based on normative distributions from a prior study conducted on 1,104 physicians and nurses, the numerical cut-off point for high emotional exhaustion for medical professionals has been established as a score of ≥27 (Maslach & Jackson, 1981). This cut-off number was calculated from the upper third of the normative distribution of the study (Maslach & Jackson, 1981). In this study, nurse burnout, defined as high emotional exhaustion, was initially measured from the nurse survey as a dichotomous variable, with any nurse respondent scoring ≥27 considered burned out. This method of using the numerical cut-off for high emotional exhaustion to distinguish burned out nurses is consistent with previous studies (Aiken et al., 2008; Cimiotti et al., 2012; Halbesleben et al., 2013; McHugh et al., 2011). Furthermore, similar to the previously established cut-off, the distribution of MBI scores was examined and 27 was found to be the cut-off for the upper third of this study’s sample. At the nurse level, burnout was measured as a dichotomous variable, being coded “1” for nurses who scored ≥27 on the EE subscale. As the study conceptualizes nurse burnout as a hospital-level phenomenon, nurse survey data was aggregated to the hospital level, where nurse burnout was defined as a continuous variable measuring the proportion of burned out nurses in a hospital. Aggregation of individual burnout scores to obtain a group-level mean burnout has been validated in other studies (Bakker, Emmerik, & Euwema, 2006; Moliner, Martínez-Tur, Peiró, Ramos, & Cropanzano, 2005). Furthermore, the studies linking nurse burnout to patient outcomes have aggregated nurse-level burnout to unit-level (Vahey et al., 2004; Van Bogaert et al., 2010) and hospital-level measures of burnout (Cimiotti et al., 2012; McHugh et al., 2011). Missed care. Data on the mediator assessed in the study, missed care, was obtained from the Multi-State Nursing Care and Patient Safety Survey. In the survey, missed care was assessed using a that asks nurses to report which of 12 listed nursing activities were necessary but left undone due to a lack of time on their most recent shift. The 12 tasks include: patient surveillance; skin care; teaching/counseling patients and families; administering medications on time; adequately documenting nursing care; coordinating patient care; pain management; oral hygiene; treatments and procedures; preparing patients and families for discharge; developing or updating care plans; and comforting/talking with patients. These tasks were specifically included in the missed care by nurse researchers and survey methodologists to capture essential nursing care activities (Lake et al., 2017). The specifies that the tasks be necessary for patient care; as such, it was hypothesized that necessary care that is missed would impact patient outcomes. This measure of missed care has been used widely and its predictive validity has been established in studies assessing the impact of nursing factors on patient outcomes (Ball et al., 2014; Brooks Carthon et al., 2015; Lake et al., 2016). Furthermore, internal consistency of the measure was assessed and was found to be acceptable with a Cronbach’s alpha of 0.82. Similar to nurse burnout, missed care was analyzed at the hospital level. This measurement fits within the theoretical framework that there is an organizational-level of burnout and missed care impacting outcomes. Additionally, the aggregation of nurse reports to assess the hospital-level impact of missed necessary care on patient outcomes has been validated empirically (Brooks Carthon et al., 2015; Lake et al., 2016). In the nurse survey data, missed care was first dichotomized, being coded “1” for nurses who reported leaving at least one or more necessary tasks undone, and “0” for nurses who reported leaving no necessary care undone. For analytic purposes, the missed care variable was aggregated to the hospital level and defined as the proportion of nurses leaving at least one necessary task left undone. Research has shown that, at the hospital level, on average, even an increase in 1 unmet nursing care task is significantly associated with a 7 to 9 point increase in the proportion of nurse-reported frequent medication errors, falls and nosocomial infections (Lucero et al., 2010). Outcomes Nurse-reported adverse events. On the Multi-State Nursing Care and Patient Safety Survey, nurses reported the frequency of the five adverse outcomes of this study using a 7-point Likert scale, ranging from never to every day (specifically: never, a few times a year or less, once a month or less, a few times a month, once a week, a few times a week, and every day). The five adverse events included as outcomes of interest in the study were: patient received wrong medication or dose; pressure ulcers developed after admission; patient falls with injury after admission; hospital-associated urinary tract infections; and hospital-associated central-line bloodstream infections. In the study, all of the nurse-reported adverse events were measured and analyzed at the nurse-level as dichotomized variables. Nurses reporting that the adverse events occurred a few times a year or less were defined as infrequent (coded as “0”), while nurses reporting the events occurring more than a few times a year up through every day were defined as frequent (coded as “1”). This dichotomized measurement of nurse- reported adverse events is consistent with previous studies (Kelly, Kutney-Lee, Lake, & Aiken, 2013; Kutney-Lee et al., 2009; Lucero et al., 2010; Olds & Clarke, 2010). This dichotomization helped with ease of interpretation and presentation of study findings, as well as providing results that are comparable across multiple studies. In sensitivity analyses, these outcomes were assessed at the hospital level. Aggregated to the hospital level, nurse-reported adverse events were the average percentage of nurses who reported the adverse event occurring frequently across hospitals. Covariates In order to assess the impact of the hypothesized predictor (nurse burnout) and mediator (missed care) on patient adverse events, the effects of other measured variables known to impact adverse events were statistically accounted for. Research has linked nursing and hospital structural characteristics, as well as organizational features, such as nurse practice environments and staffing, to adverse events (Brennan et al., 1991

******CLICK ORDER NOW BELOW AND OUR WRITERS WILL WRITE AN ANSWER TO THIS ASSIGNMENT OR ANY OTHER ASSIGNMENT, DISCUSSION, ESSAY, HOMEWORK OR QUESTION YOU MAY HAVE. OUR PAPERS ARE PLAGIARISM FREE*******."