Dear Ash & Jon,

Thanks for noticing our article. Unfortunately, the provision PDF version currently on the web contains some typos including display of Fig 4 (this should be corrected in the final version of the paper, due to be posted any day now). The meat of the paper is in Fig 4, the correct version copied here:

 

 

 

Fig displays the threshold probability as a function of benefit/harms ratio as derived by cognitive system II (EUT or EBM approach) (solid line). The treatment should be given if the probability of disease is above the threshold, otherwise it should be withheld. Note that if cognitive system I perceives that harms are higher than system I benefits (B_I < H_I), the threshold probability is ALWAYS higher than the EBM threshold (dotted line). However, if B_I > H_I, the threshold probability is ALWAYS lower than the EBM threshold (dashed line).

We thinks this can explain variation in practice, particularly when physicians try to apply research evidence (from the group of patients) to individual patients. We have also developed a model that takes a diagnostic test into consideration (the paper is being written as we speak).

We are currently trying to empirically test our model (using vignette-based approach). If any of the Group’s practicing physicians is intrerested in reviewing a couple of case vignettes, please contact me.

Thank you all

ben