November 4, 2011
This paper estimates the impact of hospitals’ health information technology (IT) adoption on patient outcomes for four common, high mortality conditions: acute myocardial infarction, congestive health failure, coronary artery bypass graft surgery, and pneumonia. We primarily study the use of electronic medical record and computerized physician order entry systems but also examine other information technologies and their utilization. We link detailed hospital discharge data on all Medicare FFS admissions
(over 6,600,000 observations) from 2003-2007 to hospital-level IT adoption data. We employ a difference-in-differences strategy to identify the parameters of interest while allowing for time invariant unobserved hospital quality differences. For all four conditions we find that health IT adoption reduces 60-day mortality (but not readmission rates or length of stay) for the most severely ill patients. Outcomes for patients with mean severity are not significantly affected by health IT. These results suggest that health IT is particularly effective at gathering and providing information on clinically complex patients while we find limited evidence that health IT improves care coordination across providers. These information and coordination effects appear to outweigh routine reminder functions that are often emphasized in clinical treatment guidelines.
Keywords: Hospitals, Information Technology, Mortality We gratefully acknowledge assistance from The Robert Wood Johnson Foundation’s Changes in Health Care Financing &
Organization and the Agency for Health Research and Quality. JEL Classification:
University of Minnesota
The Carlson School, University of Minnesota
The Wharton School, University of Pennsylvania, and NBER
Clinical medical errors cause at least 44,000 deaths annually. With direct medical costs of
$17 billion annually, these errors impose a substantial burden on both the health care system and society as a whole (IOM, 1999). Public and private sector leaders have responded with calls for technological solutions to the medical errors problem. Health information technology
(IT) has long been regarded as a potential solution to many clinical quality problems (IOM,
1999, 2001). These systems may improve quality by capturing clinical information, providing clinical decision support, and improving coordination and communication across providers.
Presidents from both parties have signed into law bills that seek to increase the diffusion and meaningful use of health IT.1 Consequently, the Congress passed and the President signed the HITECH Act in 2009. This bill provides an estimated $27 billion in subsidies for private health care providers to adopt and utilize health IT systems.
Many studies have demonstrated the potential benefits of health IT (Buntin et al., 2011).
This literature suggests several mechanisms through which health IT may affect quality.
Electronic medical records (EMRs) with computerized provider order entry (CPOE) may be used as a decision support platform. These decision support systems are designed to provide reminders for standardized treatment guidelines. Health IT may also be used to gather and distribute clinical information regarding diagnoses, tests, and prescriptions. This information may be particularly valuable for high-severity patients. Finally, health IT may help coordinate care for patients being treated by multiple providers. This literature also emphasizes the importance of adopting sophisticated systems that provide decision support capabilities along with the meaningful use by providers. While these findings are promising, they are largely based on in single-site studies at large academic medical centers and may be subject to selection bias.
Despite its potential, the value of widespread health IT adoption remains