Table 1—

Sensitivity analyses of the ratio of national standardized death rates and the ratio of M/I ratios

ScenariosRatio of national standardized death rate (1994–1998)
Ratio of M/I ratios
U.S. vs. U.K.U.S. vs. Canadavs. U.K.vs. Canada
Scenario 1: as reported (U.S. incidence: 14.8/100,000 population)2.62.03.33.2
Scenario 2: U.S. death rate
    Excess: 10%2.41.83.02.9
    Excess: 20%2.11.62.72.6
    Excess: 50%1.31.01.71.6
Scenario 3: increase U.S. incidence rate to
    a) highest regional rate (17.8/100,000 population) (13)2.82.7
    b) upper 95% CI of highest regional rate (20.3/100,000 population) (13)2.42.4
Scenario 4: scenario 3a + 20% mortality excess1.91.9
  • Scenario 1, no assumption; scenario 2, corrected for assumed overestimation of diabetes mortality in the U.S. by 10, 20, or 50% due to overassignment of diabetes as underlying cause of death; scenario 3, corrected for assumed underestimation of true national incidence rate through averaging regional rates: increase incidence rate to highest regional rate (scenario 3a) or the upper 95% CI of highest regional rate (scenario 3b); and scenario 4, corrected for combination of overestimation of mortality and underestimation of incidence rate.