Computerized medicine: good for quality, but not costs November 22, 2009

A longitudinal study of thousands of US hospitals suggests that increasing the levels of medical IT may modestly improve the quality of treatment, but it doesn’t actually help with costs, and may even make things worse in the short run as the current US healthcare economy is subverting any benefits it might otherwise provide.

By John Timmer | Last updated November 20, 2009 2:20 PM

Electronic medical records and the general digitization of medical data and practices are promoted as a way to slow the rapidly inflating costs in the US healthcare system. The push for expanded medical IT has come from the top, with President Obama extolling its virtues and his administration making funding for EMR deployments part of its stimulus package. But many have pointed out that simply throwing computers at a problem isn’t a solution unless the software and practices are also in place to allow the medical community to leverage the technology efficiently. A study of US hospital data suggests they may not be: computerization only had a mild impact on quality of care, and it didn’t seem to alter costs in any significant manner.

The study will appear online at The American Journal of Medicine Friday. Its authors combined three datasets that collectively track the computerization and outcomes at thousands of US hospitals. Data on the deployment of medical IT systems were obtained from an annual survey performed by the Healthcare Information and Management Systems Society. The survey contains over 20 measures of computerization, including both administrative and clinical functions.

Costs and quality of care were obtained from Medicare and Medicaid data, both obtained directly from the government and from a version compiled by the Dartmouth Health Atlas. The latter contains information such as whether the hospital is for-profit, the type of care delivered (acute, psychiatric, etc.), and its location. Quality of care scores were available for pneumonia, congestive heart failure, and acute myocardial infarction. The authors looked at the period from 2003 to 2007, during which time information was available for roughly 4,000 US hospitals.

During the time in question, there was a large increase in the use of computerized systems. By 2007, a typical hospital had implemented nearly two-thirds of the computerized systems covered in the survey, although there was a bias towards adoption of administrative systems. Less than a quarter of the hospitals, for example, had implemented a computerized ordering system for their physicians.

Despite the rise in computerization, however, administrative costs actually climbed slightly during the entire period. Part of this seems to be the costs of deploying the systems themselves, as hospitals in the midst of a major IT expansion had increased administrative costs during this period. Checking the data using a four-year interval, however, suggested that even once the systems are in place and in use, costs don’t start to decline. Still, none of the statistical tests performed by the authors showed a clear correlation between computerization and administrative costs.

The authors performed bivariate analysis to try to identify the factors most closely associated with costs and quality of care. Hospitals that did best on quality of care tended to be larger, nonprofit, and associated with teaching programs. Computerization tended to increase the quality of care for acute myocardial infarction, but not either of the other problems. Multivariate analysis suggested that the improvement may be correlated with the use of computerized systems that focus specifically on patient care.

"We found no evidence that computerization has lowered costs or streamlined administration," the authors concluded. "More encouragingly, greater use of information technology was associated with a consistent though small increase in quality scores." That’s not exactly a ringing endorsement of healthcare IT, and it’s certainly a far cry from some of the improvements promised by its proponents.

Why the disparity? The authors provide three potential explanations. One is simply that the cost of purchasing and supporting IT equipment and software offsets any savings they produce. The other is that the four-year lag used in their analysis to look for long-term savings simply isn’t sufficient; savings will eventually appear, but only once the systems are in use for long enough for everyone to become proficient with them.

They favor the third possibility: the commercial medical marketplace is simply structured in a way that doesn’t favor optimal solutions. "Coding and other reimbursement-driven documentation might take precedence over efficiency and the encouragement of clinical parsimony," they suggest. "The largest computer success story has occurred at Veterans Administration hospitals where global budgets obviate the need for most billing and internal cost accounting, and minimize commercial pressures."

In other words, the current US healthcare economy is subverting any benefits that computerized healthcare might otherwise provide.

The American Journal of Medicine, 2009. DOI: 10.1016/j.amjmed.2009.09.004