Steven W. Heim

The D-Dimer Test for Deep Venous Thrombosis: Gold Standards and Bias in Negative Predictive Value

BACKGROUND Because venous ultrasound (US) fails to fully image the calf veins, there is the potential for US gold standard studies to classify patients with calf deep venous thrombosis (DVT) in the nondiseased category, causing bias in test index calculations. A false increase in negative predictive value (NPV) is especially likely because calf DVT false-negative tests will be counted in the numerator along with the true-negative tests in NPV calculations. We verified the presence and magnitude of this bias for the d-dimer test. METHODS We abstracted data on overall (calf and thigh) and thigh-only test sensitivity, specificity, and NPV from the six English language studies published between March 1995 and October 2001 that compared d-dimer to a gold standard (GS) capable of imaging both thigh and calf veins and that also stratified results by thigh and calf location. Thigh specificity and NPV were calculated classifying calf DVT patients as free of disease. RESULTS The six studies included 81-214 participants and provided 26 comparisons of 16 different d-dimer assays to the GS. Thigh sensitivity was higher than overall sensitivity in 22 of 26 comparisons (range, -0.3 to 8.6); thigh specificity was lower than overall specificity in all comparisons (range, -0.7 to -7.8); and thigh NPV was higher than overall NPV in 22 of 26 comparisons and unchanged in 4 comparisons (range, 0.0-9.2). NPV was >95% in 20 of the thigh results but >95% in only 8 of the overall results. CONCLUSIONS Different GS can produce clinically significant differences in test indices. Care must be taken in interpreting DVT studies that evaluate d-dimer as a rule-out test and that use US as a GS, because missed calf DVT can falsely increase the NPV.

Which Hospitals Have Significantly Better or Worse Than Expected Mortality Rates for Acute Myocardial Infarction Patients? Improved Risk Adjustment With Present-at-Admission Diagnoses

Background— Public reports that compare hospital mortality rates for patients with acute myocardial infarction are commonly used strategies for improving the quality of care delivered to these patients. Fair comparisons of hospital mortality rates require thorough adjustments for differences among patients in baseline mortality risk. This study examines the effect on hospital mortality rate comparisons of improved risk adjustment methods using diagnoses reported as present-at-admission. Methods and Results— Logistic regression models and related methods originally used by California to compare hospital mortality rates for patients with acute myocardial infarction are replicated. These results are contrasted with results obtained for the same hospitals by patient-level mortality risk adjustment models using present-at-admission diagnoses, using 3 statistical methods of identifying hospitals with higher or lower than expected mortality: indirect standardization, adjusted odds ratios, and hierarchical models. Models using present-at-admission diagnoses identified substantially fewer hospitals as outliers than did California model A for each of the 3 statistical methods considered. Conclusions— Large improvements in statistical performance can be achieved with the use of present-at-admission diagnoses to characterize baseline mortality risk. These improvements are important because models with better statistical performance identify different hospitals as having better or worse than expected mortality.

Response to Letter Regarding Article, “Which Hospitals Have Significantly Better or Worse Than Expected Mortality Rates for Acute Myocardial Infarction Patients? Improved Risk Adjustment With Present-at-Admission Diagnoses”

BACKGROUND Public reports that compare hospital mortality rates for patients with acute myocardial infarction are commonly used strategies for improving the quality of care delivered to these patients. Fair comparisons of hospital mortality rates require thorough adjustments for differences among patients in baseline mortality risk. This study examines the effect on hospital mortality rate comparisons of improved risk adjustment methods using diagnoses reported as present-at-admission. METHODS AND RESULTS Logistic regression models and related methods originally used by California to compare hospital mortality rates for patients with acute myocardial infarction are replicated. These results are contrasted with results obtained for the same hospitals by patient-level mortality risk adjustment models using present-at-admission diagnoses, using 3 statistical methods of identifying hospitals with higher or lower than expected mortality: indirect standardization, adjusted odds ratios, and hierarchical models. Models using present-at-admission diagnoses identified substantially fewer hospitals as outliers than did California model A for each of the 3 statistical methods considered. CONCLUSIONS Large improvements in statistical performance can be achieved with the use of present-at-admission diagnoses to characterize baseline mortality risk. These improvements are important because models with better statistical performance identify different hospitals as having better or worse than expected mortality.