Analysis of a Newspaper Reserach Article

Analysis of a Newspaper Research Report

Cathy Jaco

University of Phoenix

NUR 438: Statistical Applications

Valerie Opher, BSN, RN, MAS, DM

March 11, 2007

Analysis of a Newspaper Research Report

The well-known saying "There are three kinds of lies: lies, damned lies, and statistics," attributed to Benjamin Disraeli and popularized in the U.S. by Mark Twain, illustrates the extent to which the pubic can be confused by newspaper articles highlighting research statistics. (Rumsey, 2003, p. 17). Simply by opening a newspaper, readers find results of studies, announcements of breakthroughs, statistical reports, forecasts, projections, charts, graphs and summaries of research. Although Disraeli's and Twain's comment may not be directed solely at newspaper research reports, the importance of detecting errors and possible exaggerations in the media cannot be overlooked. This paper will describe a research study funded by the Centers for Medicare & Medicaid Services (CMS) and published in The New York Times on January 25, 2007. Observations will be made about the statistical procedures used, the study findings and conclusions, and the appropriateness of these findings.

Background of Research Study

In March 2003, 613 hospitals involved in a quality-benchmarking database, known as Perspective, agreed to participate in the CMS-Premier Hospital Quality Incentive Demonstration (HQID), a multiyear collaborative of hospitals with the goal to "determine if economic incentives are effective at improving quality of inpatient care" (Lindenauer et al., 2007, p. 3). The impentus for the study was based on both CMS' and Congress' fundamental concern over the current payment system. The intent of the study was to measure how public reporting and pay-for-performance accelerate improvements in hospital care. The study measured changes in adherence to10 individual and four composite measures of quality over a two-year period with additional comparisons between publicly reporting hospitals and publicly reporting plus pay-for-performance volunteer hospitals. Participating hospitals were required to collect and report data on 10 quality measures in relation to three clinical conditions: heart failure, acute myocardial infarction, and pneumonia. For example, hospitals in the pay-for-performance group received additional payment for providing educational materials for discharged patients with congestive heart failure that helped avoid a readmission.

Statistical Procedures Used

This observational study compared hospitals (N=406) engaged in public reporting alone (control group) with hospitals engaged in both public-reporting and pay-for-performance programs (N=207) (intervention group). Using the New England Journal of Medicine article, Public Reporting and Pay for Performance in Hospital