When is it valid to assume that heat flux is parallel to B?

It is frequently assumed that heat flow in the plasma scrape-off-layer is everywhere parallel to B, due to the strong anisotropy in electron thermal conductivity. This assumption is convenient but paradoxical. Here are examined three situations where this assumption has sometimes been applied: (1) extrapolating from midplane T_e(R) measurements to divertor heat flux profile, (2) determining the location of the separatrix from measured midplane T_e(R), combined with total heat flux leaving the plasma, and (3) predicting the heat flux to plasma-facing components in the scrape-off-layer of diverted plasmas. Numerical solution of the anisotropic, nonlinear heat equation suggests that the first application is poor, the second well justified, and the third far from accurate. Additional plasma physics effects may mitigate these results, but the simple assumption of dominant parallel heat flow due to anisotropy in electron thermal conductivity is not supported in many important cases.