ABSTRACT & COMMENTARY
Are Hospitalists Associated with Improvements in Quality of Care?
By Kenneth P. Steinberg, MD, FACP
Professor of Medicine, University of Washington School of Medicine, Seattle, WA
Dr. Steinberg reports no financial relationships in this field of study.
SOURCE: Jungerwirth R, et al. Association of Hospitalist Presence and Hospital-Level Outcome Measures Among Medicare Patients. J Hosp Med 2014;9:1-6.
Summary
The number of hospitalists in the United States has grown rapidly and it has been estimated that more than 80% of hospitals now employ hospitalists. Reasons for the rapidly increasing number of hospitalists include the need to increase efficiency and shorten length of stay, to cover for the decreased number of work hours now allowable for resident physicians, to allow primary care providers to spend more time in their offices, and to allow hospitalists to focus on the increasingly complex inpatient care environment. Hospitalists have been shown to increase efficiency and shorten length of stay, but the data that suggest hospitalists improve quality of care as measured by decreased mortality and readmission rates remain unclear. Many of the previous studies looking at this issue were single-institution studies with relatively small sample sizes.
Using two national databases, the authors of this current study were interested in examining the association between hospitals that utilize hospitalists and two measures of quality of care: 30-day all-cause mortality and 30-day readmission rates. They chose to study these variables in three patient conditions: heart failure (HF), acute myocardial infarction (AMI), and pneumonia (PNA). The investigators used 2008 data from the Centers for Medicare & Medicaid Services (CMS); these were hospital-level data of case mix-adjusted, risk-standardized, 30-day all-cause predicted excess mortality and readmission rates, as measured from the first day of the index inpatient admission. Hospitals with fewer than 25 admissions for a given condition are excluded from the CMS database.
These CMS hospital-level data were then linked to data from the 2008 American Hospital Association Annual Survey Database that provides characteristics for approximately 6500 U.S. hospitals, including whether or not the hospitals employed hospitalists to provide care within the hospital. The primary independent variable was whether or not the hospital used hospitalists. Several other covariates were used including other hospital demographic, geographic, and organizational data.
In the final analyses, there were 3029 U.S. hospitals of which 59.3% reported employing hospitalists. In the bivariate analyses, decreased mortality for all three conditions (HF, AMI, PNA) was associated with the presence of hospitalists and decreased readmission was seen with hospitalists for HF and AMI. However, in the multivariate regression analysis (taking into account other hospital characteristics), there was no statistically significant association between mortality and hospitalist care for any of the three conditions. In contrast, risk-standardized readmission rates were significantly lower for all three conditions in hospitals that employed hospitalists. Thus, the presence of hospitalists was not associated with a reduction in case mix-adjusted, risk standardized, 30-day all cause predicted excess mortality. But hospitalists were associated with a decrease in adjusted and standardized 30-day excess readmissions for all three conditions: HF, AMI, and PNA.
Commentary
This study demonstrates an association between the use of hospitalists by a hospital and a reduction in 30-day excess readmission rates (actual readmissions ÷ predicted readmissions; excess readmissions were calculated by looking at the difference in this ratio for a hospital compared to the national average of hospitals with similar case mix) for three important and common clinical conditions. The study only demonstrates an association and does not demonstrate causation between the presence of hospitalists and an effect on outcomes. In other words, there could be many other reasons why hospitals that employ hospitalists might have a lower excess predicted readmission rate unrelated specifically to the care provided by hospitalists. This study used hospital-level data from two large national databases and complex statistics to make these observations, and the authors do a good job of acknowledging the limitations of the study. The strengths, though, include the size and national bases of the study.
Hospitalists have extensive discharge experience and are uniquely situated to help to affect the care, policies, and culture at institutions surrounding these transitions, even for patients for whom they do not provide direct care. I would be cautious about over-interpreting this study and hopefully these findings will be replicated in other studies. Demonstrating that hospitalists are associated with a reduction in hospital readmissions, an important quality metric, will be yet further evidence of the importance of utilizing physicians specializing in hospital medicine. I agree with the authors that although the use of hospitalists creates another hand-off in the transition between inpatient and outpatient settings, this risk might be overcome by the beneficial effects of hospitalists on the other various determinants of readmission, leading to an overall improvement in quality of care in this domain.
