Project Details - Ended
- Grant Number:R36 HS018272
- Funding Mechanism:
- AHRQ Funded Amount:$24,863
- Principal Investigator:
- Project Dates:9/1/2009 to 4/30/2011
- Care Setting:
- Health Care Theme:
The influence of health information technology (IT) on where consumers decide to receive hospital inpatient services is largely unknown. The decisions consumers make can affect the cost and quality of those services as well as the market power of the hospitals. The demand analysis completed in this project complements existing supply-side analyses to provide more complete and dynamic estimates of the impact that health IT has on health care markets. The analysis required information about hospitals’ characteristics and information technology as well as patients’ characteristics and hospital choices from 1999 through 2006. Data needed to perform this analysis came from several sources. Hospital characteristics data were obtained from the American Hospital Association (AHA) annual survey. This database contains information on hospitals’ physical and organizational characteristics such as location, number of full-time physicians, and number of beds. The AHA database was linked with the Health Information and Management Systems Society Analytics Database. This dataset contains detailed historical information on the health IT software, hardware, and infrastructure installed in the surveyed hospitals. Three types of health IT were evaluated: electronic medical records, computerized physician order entry, and picture archiving systems. Inpatient Medicare claims data are the source of patient-level characteristics and hospital choices.
The project aims included:
- Measure the effect of hospital adoption of health IT on the demand for inpatient care.
- Estimate the impact of health IT by type of inpatient service.
- Evaluate the effect of changes in patient hospital choices using consumer surplus as a welfare measure.
The demand for hospital inpatient services was estimated using standard econometric choice models that included patient characteristics, hospital characteristics, and observed patient choices. A hospital’s decision to implement health IT was considered a treatment or policy intervention, and the change in the total number of patients using the hospital was the outcome of interest. A discrete choice model was estimated using patient-level data to predict the probabilities of patients choosing each hospital in their choice set. The parameter estimates from these models show how health IT affects a patient’s likely hospital choice. An aggregate level model was employed in situations where the data set was too large to estimate at an individual patient level.
The impact of health IT was small, if it existed, in that it was not found to affect large numbers of patients’ choices or have a large impact on overall hospital demand. The picture archiving systems variables and interaction terms in both the market level and individual choice models are jointly significant and expected consumer surplus is positive. Effects of the other technologies on demand were not significant. Although the value of health IT is positive, for some types of health IT, the effect on market share may not be enough to justify the financial investments.