This tool can be used to look up the statistical significance of correlation coefficients for EQ-5D-3L, EQ-5D-5L and SF-6Dv2 value sets.
As explained in the article below, high correlation coefficients for HRQoL value sets derived from different samples – across countries and/or using different valuation techniques – are conventionally interpreted as evidence that the respective samples have similar health-related quality of life preferences. However, value sets contain many inherent rankings of health state values by design.
By calculating correlation coefficients for value sets created from random data, we demonstrate that ‘high’ coefficients are artifacts of these inherent rankings – e.g. median Pearson’s r = 0.783 for the EQ-5D-3L and 0.850 for the EQ-5D-5L instead of zero. Therefore, high correlation coefficients do not necessarily constitute evidence of meaningful associations in terms of similar HRQoL preferences.
Based on simulations, we calculate significance levels and find that many high coefficients reported in the literature are not statistically significant. These ‘high’ but insignificant correlations are in fact spurious.
To cite results in your research:
F Ombler, M Albert & P Hansen (2018), “How Significant Are “High” Correlations Between EQ-5D Value Sets?”, Medical Decision Making 38, 635-45.
Note: if you’d like to test significance for a different model, we may be able to use the significance calculator we created for the above article to generate significance tables for you.