TABLE 2

Coefficients of different linear regression models and application to testing data

Ross et al. [14] mean (r=0.82)p-valueRoss et al. [14] (r=0.59)p-valueMiyamoto et al. [16] (r=0.68)p-valueZapico et al. [17] (R2=0.48)p-valueNew model (R2=0.48)p-value
Observations n103810832782194
Intercept4.9484.6824.213−21.6261.830.248
6MWD, m0.0230.0250.0260.026<0.0010.031<0.001
Sex4.1030.002
Height, cm0.1740.043
Weight, kg−0.0710.02−0.0230.22
HRR, bpm−0.0150.047
Applied to testing dataset      
 SEE, mL/min/kg3.033.224.363.08
 SEE% mean19%20%27%19%
Ability to classify to low risk (V′O2max >15 mL/min/kg)      
 Sensitivity0.960.960.910.69
 Specificity0.230.230.220.78
 PPV0.630.630.630.82
 NPV0.800.800.640.64

r from correlation between maxima oxygen uptake (VO2max) and 6-min walk distance (6MWD); R2 from linear regression of VO2max against 6MWD. HRR: heart rate response in 6-min walk test (heart rate at maximal exercise – heart rate at rest); NPV: negative predictive value; PPV: positive predictive value; SEE: standard error of the estimate; SEE% mean: SEE as percentage of mean peak oxygen uptake. All coefficients, p-values and R2 for Ross et al. [14], Miyamoto et al. [16] and Zapico et al. [17] are copied from the original publications.