Saturday, June 2, 2007 * 3:00 p.m.-6:00 p.m.
Walt Disney World Swan and Dolphin * Orlando
Oceanic 5

The Effect of IT Capital on Hospital Efficiency

Michael Housman
Ph.D. Student
University of Pennsylvania
Health Care Systems Department
3641 Locust Walk
Philadelphia, PA 19104-6218
PH: (215) 681-6955
FX: (215) 573-7278
Email: housman@wharton.upenn.edu

Research Objective:

Health information technology (IT) is widely thought to improve healthcare quality and productivity, and reduce costs. However, supporting empirical evidence is limited. This study aims to corroborate observational evidence through a cost function estimation to examine relationships between IT capital stock and cost efficiency in US hospitals.

Study Design:

Most studies have utilized cross-sectional analyses or small samples of convenience; instead, we analyze 1999 to 2004 data from multiple sources (HIMSS Analytics, American Hospital Association, and Medicare Cost Reports) describing the implementation status of 40 software applications in US hospitals. Consistent with prior work on hospital efficiency, we utilize a translog cost function system (the primal and the factor share equations) to calculate the marginal product of various factor inputs in hospitals (capital, labor, materials) as well as IT capital. By estimating the model with seemingly unrelated regression (SUR) and adding terms to allow for nonlinear relationships between IT capital and hospital efficiency, we examine whether the marginal benefits of IT capital depend on level of investment.

Population:

Our panel dataset includes a majority of all US acute care hospitals. We exclude government hospitals, psychiatric hospitals, and long term care hospitals.

Findings:

The impact of IT capital on hospital productivity follows a non-linear pattern, suggesting that a certain amount of IT investment is necessary before hospitals gain efficiency benefits. We also find that non-profit hospitals appear less efficient than for-profit hospitals in terms of their IT usage, reaching the “tipping point” at higher levels of IT capital. We also find that significant differences exist in the “cost-reducing effects” of applications. Administrative applications show efficiency gains, even at relatively low levels of IT capital investment, with a higher marginal gain in for-profit hospitals. Clinical applications show no efficiency gains in non-profit hospitals but do appear to improve efficiency in for-profit hospitals. We also examined the impact of applications across different time horizons. Clinical applications demonstrate efficiency gains in non-profit hospitals over time. In for-profit hospitals, clinical applications also yield both short- and long-term gains. Administrative IT applications show an efficiency effect in the short- and long-term within for-profit and non-profit hospitals.

Conclusion:

Higher levels of in IT capital are associated with reduced short-term operating costs in acute care hospitals, after a threshold level of investment has been reached. Non-profit hospitals reach this tipping point at higher levels of IT investment than for-profit hospitals. Administrative applications appear to yield greater efficiency gains than clinical applications but the benefits of all software applications increase within longer time horizons.

Implications:

From the perspective of individual hospitals contemplating whether to invest more in information systems, this research tends to corroborate the investment as being worthwhile. We believe that initial increases in IT capital may entail significant ‘start-up' expenses (networking infrastructure, recruitment of IT staff) which increase costs prior to the efficiency gains that IT applications might provide.