Healthcare-associated infections (HAIs) remain common, risking patient lives, adding to health care costs, and contributing to the major public health problem of antibiotic-resistant infections. Clinical interventions have been found effective in preventing HAIs under controlled conditions but not under usual care conditions. Infection prevention interventions have not focused on improving hospital-wide nurse work environments as recommended by the Institute of Medicine. The primary aim of this study is to determine whether nursing resources such as work environments and other modifiable features of nursing, including nurse staffing, skill mix, workforce stability, and education are associated with HAIs, with a goal of identifying promising hospital level strategies to facilitate infection reduction. Measures of these modifiable features of nursing will be derived by aggregating, to the hospital level, unique survey data collected in 2015-2016 from 27,319 nurses employed in 583 hospitals in four large states: California, Florida, New Jersey, and Pennsylvania. The hospitals in these large states account for close to a quarter of hospital discharges nationally. These hospital-level measures, together with measures of hospital size, teaching status, and technology derived from American Hospital Association Annual Survey data, will then be merged, separately, with infection data from two different sources; 1) patient-level data that combine information from the Medicare Provider and Analysis Review files (MEDPAR), the Medicare Outpatient Standard Analytic File, and the Medicare Carrier File (Provider Part B) and includes information essential for risk adjustment and enhanced measures of infections, and 2) hospital- level standardized infection ratios (SIR) using data from the Centers for Disease Control and Prevention (CDC) that are publicly reported on the Centers for Medicare and Medicaid Services (CMS) Hospital Compare website. Multilevel logit models and hierarchical linear models (for the patient level data) will be used to estimate the effects of the different nursing resource measures on the likelihoods of different groups of medical and surgical patients acquiring different infections, before and after adjusting for other hospital and patient characteristics. Ordinary least squares regression models (for the hospital level data) will be used to estimate these same nursing effects on the risk-adjusted standardized infection ratios, before and after controlling for other hospital characteristics. The large sample of representative (rather than volunteer) hospitals, the use of the Medicare Part B data to uncover infections in patients discharged to their home or a care facility that were not recognized during their hospital stay, and the use of unique data on work environments and other nursing characteristics will make it possible to capture significant effects of nursing which prior investigations have been unable to explore. The use of multiple data sources and multiple measures of infections, and the use of patient-level data which brings risk-adjustment under the investigators control, will also allow for the triangulation of results and identification of actionable recommendations with promise for reducing HAIs.
By studying nurses and hospitals, we can learn how nursing care can vary across hospitals and if this variation has contributed to the successful prevention of healthcare-associated infections. Understanding the link between nursing care and transmission of infections will inform healthcare facilities on how to best allocate their nursing resources in an effort to eliminate healthcare-associated infections.
