fig8

Figure 8. Hair mass index (HMI) abacus translates clinical pattern and severity (SCS on parted midline) into dynamics of hair regression (TTCC). The scalp coverage scoring performed on the hair parted on the midline is shown relative to the maximum attainable score (y-axis; SCS per cent of maximum) as it correlates with our most synthetic parameter: TTCC (x-axis; days). The latter was obtained after conversion of daily scalp hair productivity (compound index of hair growth area percentage) in the dimension “time”. Circles and squares symbolise females and males, respectively. Plain blue circle and square reflect the maximum values found in healthy controls (y = 97% and x = 21.6 and 19.6 days in female and male controls, respectively) followed by empty symbols for averages found in patient groups split into classes according to pattern and severity by the observer. Neither TTCC nor SCS on parted midline in the least affected patients (NoP, LI, HI and HII) can discriminate patients from healthy gender-matched controls. The x-y intercept and linear correlation coefficient (y = 106.41-0.72x; R² = 0.973) confirm mathematically that the clinician did an acceptable job in terms of classification as well as in estimating SCS. The HMI abacus reflects a compound measure based on TTCC and SCS, albeit potentially not exempt from “classification errors”. Statistical approaches help clinicians to deal with the error associated with most - if not all - measurements. Instead of “jumping by chance” from one class to another because of the binary nature of the scales proposed by Ludwig and Hamilton, HMI is a continuum, and the dispersion of data around the average is detailed for each gender separately in Figure 9. NoP: no pattern.