Study: Imaging Improves Health Outcomes
Lower mortality rates are among the benefits delivered by diagnostic imaging, according to a study in the December 2009 issue of the Journal of the American College of Radiology: JACR. The study—by David W. Lee, PhD, of GE Healthcare (Waukesha, Wisconsin), and David A. Foster, PhD, of Thomson Reuters (Ann Arbor, Michigan)—examines the association between the utilization of inpatient diagnostic-imaging services and key hospital outcome measures. The authors find that imaging yields value and therefore should not be curtailed, whether by reductions in payment rates or by limitations to access through certificate-of-need programs and other artificial constraints.
David Lee, PhD
Moreover, Lee and Foster observe that imaging studies contribute to improved patient care—including lower mortality rates—with no attendant increase in health-care costs. The study provides imaging providers with scientific data to counter government efforts to curtail utilization by ratcheting down reimbursement. With regard to health care’s systemic costs, the researchers uncovered nothing to suggest that use of advanced imaging contributes to greater expense. “Actually, we found advanced imaging to be cost neutral in the big picture,” Lee says. He adds, however, that this discovery does not preclude the possibility that advanced imaging can also reduce systemic costs. “Advanced imaging helps diagnose conditions more efficiently, potentially paving the way for better outcomes arrived at earlier and with fewer complications, but we can’t state that with certainty because our study did not examine long-run cost effects, only acute-patient stays.” he says. Methods Described Entitled “The Association Between Hospital Outcomes and Diagnostic Imaging: Early Findings,” the study revolves around Lee’s and Foster’s evaluation of patient-level data. They rely on that particular type of information in order to construct a pair of hospital-specific, risk-adjusted imaging-utilization measures for CT, MRI, ultrasound, and radiography, thereby establishing a straightforward method for gauging whether the patient received a service and, further, for estimating the mean number of services received. The data necessary for this task come from the 2007 edition of the Thomson Reuters Hospital Drug Database, which covers 1.1 million patients treated at 102 US hospitals. Females make up approximately 54% of the cohort; 33% of the women and men combined are between the ages of 45 and 64, with 29.9% between 70 and 84 years old. The smallest of the hospitals has fewer than 200 beds; the largest, in excess of 500 beds. Some are teaching institutions. Many are community hospitals. Most are located in the South. The ratio of rural to urban facilities is about 1:1. Lee says that he wanted to conduct a study of this nature in order to test the validity of the idea that more is better when it comes to diagnostic imaging. Were it valid, “We would expect greater resource use to automatically lead to better clinical and patient reported outcomes, albeit with diminishing returns at the margins,” Lee and Foster write in their article. “The recent US experience—spiraling health care expenditures that now far exceed those of any other industrialized country, and wide variations in resource use across geographic areas and between clinical institutions and health systems—underscores the importance of putting this tenet to the test.” They write, “Policymakers and researchers have both begun questioning this tenet, and a number of empirical studies have demonstrated that geographic areas with high health care utilization have health outcomes and quality of care that are no better, and are sometimes even worse, than areas with less intensive service utilization.” Lee and Foster also characterize their study as a work intended to illustrate what they describe as an important (yet largely unexplored) area of inquiry, given policy shifts toward value-based payment.” Specifically, they contend that one of the greatest obstacles to true value-based payment for imaging is the relative lack of critical evaluations of the relationship between imaging and outcomes. “The growth in diagnostic imaging utilization may reflect the existence of financial incentives within the health care system, a desire to limit professional liability, and an inherent preference for the ‘latest and greatest,’” the authors write. “Population-based, empirical evaluations of the value of imaging have mixed results and provide only a limited context for policy recommendations for use of imaging services.” Scant Reaction The study is significant, in Lee’s estimation, because it introduces “an interesting new piece of evidence that goes to the core of the debate about imaging’s value,” he says. Even so, and much to Lee’s surprise, the study has engendered scant criticism from imaging’s detractors. “There hasn’t been much criticism of our study, which I find puzzling,” he says. “Sometimes, the best way to counter a set of facts that goes against the conventional wisdom is simply to ignore them,” and perhaps that accounts for the silence, he suggests. The study might prove difficult to pick apart; Lee and Foster anticipated attacks, and they structured their research methodology to make it as bulletproof as possible. For example, pertinent factors and variables affecting both the utilization of imaging and patient outcomes from one hospital to the next were taken into account, as were hospital-to-hospital mortality rates. Woven into the study’s conclusion is a call for further exploration of the relationship between resource use and the clinical outcomes associated with imaging. “The next step is to repeat this study, but with a larger sample of hospitals,” Lee says. “This will help us verify that the findings remain robust, and that they can hold up in a broader sample. We’ll also want to drill down so that we can see if these results vary by underlying condition.” A different (but related) line of inquiry that Lee would like to pursue deals with the impact of imaging on costs and quality in a purely ambulatory setting. That, however, promises to be a vastly more complex task because “the endpoint measurements extend so much farther out and are so much more difficult to capture,” he says; “also, the number of confounders in the outpatient environment is significantly greater.” Lee has not yet taken steps to initiate either of these research projects, but he expresses confidence about getting them approved, should he decide to propose them formally. “At GE Healthcare, our entire business strategy is being built around providing valid evidence of the value of our technologies,” he says. “Studies like these fit right into the way we do business.” Rich Smith is a contributing writer for