The Impact of the Deficit Reduction Act of 2005 on the Length of Hospital Stay

ABSTRACT Introduction Hospital-acquired infections (HAIs) are the most common complications of hospital care in the United States, and at least 1 in 10 patients admitted to the hospital will develop one. These unnecessary complications lead to about 99,000 premature deaths per year and add approximately $35–$45 billion of direct costs. Most of the cost increase is due to increases in hospital length of stay (LOS). This study examines the effect of a federal policy change on LOS among people with an HAI using the Donabedian quality of health care model as a framework. Specifically, this study examines the effect of the nonpayment policy of the Deficit Reduction Act of 2005 (DRA-2005) on LOS among patients with a specific type of HAI, catheter-associated urinary tract infections (CAUTI). The DRA-2005 penalizes hospitals financially by mandating them to treat patients who develop certain HAIs without compensation from Medicare and Medicaid. The policy change is regarded as a structural variable and the hospital length of stay is the primary outcome. The hypothesis is that LOS will decrease post DRA-2005. Methods This study employs alongitudinal study design to examine the effect of the DRA-2005 on LOS among patients with CAUTI while controlling for other structural and patient characteristics that also may impact LOS. The data are obtained from National Inpatient Sample (NIS), which is a database in the Healthcare Cost and Utilization Project (HCUP). The study population was restricted to adults who acquired CAUTI during their stay in acute care hospitals. Four years of NIS data are used; two years before the implementation of the DRA-2005 (2006 and 2007) and two years afterward (2010 and 2011). Other independentvariables include hospital structural characteristics and patient characteristics. The study hypothesis is tested with logistic and linear regression models. Results The initial bivariate analyses find that LOS significantly decreased by over a half-day after the DRA-2005 took effect. This effect remained in linear regression models that also controlled for hospital and patient characteristics, with an overall 11.3% decrease in the mean LOS post DRA-2005. Logistic regression analysis is used to examine the effect that the DRA-2005 had on the odds of having a very high (top 5%) LOS. Patients post-DRA-2005 had significantly decreased odds (.61 [CI: .54,.71]) of having a very high LOS when controlling for hospital and patient characteristics. In general, the patient and hospital characteristics that also affected LOS are patients’ health and nurse staffing. Conclusion These findings are significant within the context of spiraling healthcare costs and government interventions to halt this trend as they show that government policy can be effective. Even a modest decrease in LOS translates into substantial cost savings to Medicaid/Medicare and increases the availability of hospital beds. This increase in hospital beds is critical in view of the current COVID-19 pandemic and the possibility of future pandemics. Introduction Hospital-acquired infections (HAIs) are the most common complications of hospital care in the United States, and at least 1 in 10 patients admitted to the hospital will develop one. These unnecessary complications lead to about 99,000 premature deaths per year and add approximately $35–$45 billion of direct costs. Most of the cost increase is due to increases in hospital length of stay (LOS). This study examines the effect of a federal policy change on LOS among people with an HAI using the Donabedian quality of health care model as a framework. Specifically, this study examines the effect of the nonpayment policy of the Deficit Reduction Act of 2005 (DRA-2005) on LOS among patients with a specific type of HAI, catheter-associated urinary tract infections (CAUTI). The DRA-2005 penalizes hospitals financially by mandating them to treat patients who develop certain HAIs without compensation from Medicare and Medicaid. The policy change is regarded as a structural variable and the hospital length of stay is the primary outcome. The hypothesis is that LOS will decrease post DRA-2005. Methods This study employs alongitudinal study design to examine the effect of the DRA-2005 on LOS among patients with CAUTI while controlling for other structural and patient characteristics that also may impact LOS. The data are obtained from National Inpatient Sample (NIS), which is a database in the Healthcare Cost and Utilization Project (HCUP). The study population was restricted to adults who acquired CAUTI during their stay in acute care hospitals. Four years of NIS data are used; two years before the implementation of the DRA-2005 (2006 and 2007) and two years afterward (2010 and 2011). Other independentvariables include hospital structural characteristics and patient characteristics. The study hypothesis is tested with logistic and linear regression models. Results The initial bivariate analyses find that LOS significantly decreased by over a half-day after the DRA-2005 took effect. This effect remained in linear regression models that also controlled for hospital and patient characteristics, with an overall 11.3% decrease in the mean LOS post DRA-2005. Logistic regression analysis is used to examine the effect that the DRA-2005 had on the odds of having a very high (top 5%) LOS. Patients post-DRA-2005 had significantly decreased odds (.61 [CI: .54,.71]) of having a very high LOS when controlling for hospital and patient characteristics. In general, the patient and hospital characteristics that also affected LOS are patients’ health and nurse staffing. Conclusion These findings are significant within the context of spiraling healthcare costs and government interventions to halt this trend as they show that government policy can be effective. Even a modest decrease in LOS translates into substantial cost savings to Medicaid/Medicare and increases the availability of hospital beds. This increase in hospital beds is critical in view of the current COVID-19 pandemic and the possibility of future pandemics