This describes a finding from an as yet unpublished paper. For copyright's sake I will not disclose it until formal publication. It's findings are: For any individual you can plot changes of TSH and FT4 to get curves for each on a graph where they cross over. This crossover point is called the diagnosing sensitivity of FT4 and TSH responses to each other. However the work shows that thyroid and pituitary sensitivities are linked together in individuals in a non-random pattern. This linkage shows one cannot obtain the curves describing these sensitivities from population data and complicates their derivation by physiological experimentation. This limitation could be evolutionally important by minimising the variation inexpression by individuals in a population to facilitate better survival chances of the species.
In short you cannot use population data to describe an individual's response to T4 therapy and diagnose thereby. Even shorter, therefore it is the individual response that should be used in diagnosis, and NOT by reference to so-called population-derived ranges.