Summary: In response to a congressional request, GAO reviewed allegations that the Department of Veterans Affairs (VA) altered the design of a patient mortality study to obtain more favorable results, focusing on: (1) whether the Chief Medical Director or any other VA official attempted to give the appearance that VA had fewer quality assurance problems at its medical centers than actually existed; and (2) the decision to use a 99-percent confidence level to calculate summary hospital mortality data.
GAO found that: (1) neither the Chief Medical Director nor any other VA official inappropriately attempted to give the appearance that VA had fewer hospitals with higher-than-expected mortality rates than it did; (2) the Chief Medical Director initially decided to use a methodology that the Health Care Financing Administration (HCFA) used to determine mortality rates for Medicare patients in private-sector hospitals in order to obtain comparative mortality data; (3) the VA Office of Quality Assurance found that because of the differences in the databases and statistical tests used, it could not make any direct comparisons between the VA and HCFA mortality study results; (4) the Chief Medical Director's staff interpreted the Chief's insistence on using the HCFA methodology as instructions to alter the VA mortality study so that the results were similar to the HCFA study; and (5) although VA used both the 95- and 99-percent levels in its mortality study, a VA research specialist chose the 99-percent confidence level because it increased VA confidence that the hospitals identified had differences between the observed and expected mortality rates that VA considered meaningful.
