Plain Layout

\end_inset \end_layout \begin_layout Plain Layout \align center \begin_inset Caption \begin_layout Plain Layout \begin_inset CommandInset label LatexCommand label name "tab:Weighting-of-evidence" \end_inset Weighting of evidence \end_layout \end_inset \end_layout \end_inset \end_layout \begin_layout Plain Layout We do however need to permit \begin_... inset Quotes eld \end_inset absence of weighting \begin_inset Quotes erd \end_inset so that we can record \begin_inset Quotes eld \end_inset unspecified \begin_inset Quotes erd \end_inset association between results and evidence. It might be logical to permit \begin_inset Quotes eld \end_inset positive \begin_inset Quotes erd \end_inset and \begin_inset Quotes eld \end_inset negative \begin_inset Quotes erd \end_inset weighting without any strength of association, as particularly retrospective records will contain this sort of qualification. \series bold [Do we need \begin_inset Quotes eld \end_inset uncharacterised \begin_inset Quotes erd \end_inset ?] \end_layout \begin_layout Plain Layout [Is it simply best to use a \begin_inset Quotes eld \end_inset halfway \begin_inset Quotes erd \end_inset score to signal uncharacterised --- I think not, as this gives no evidence of dispersion. Do we also need to record dispersion i.e. precision? Jaynes again??] \end_layout \end_inset \end_layout \begin_layout Subsection The nature of problems \begin_inset CommandInset label LatexCommand label name "sub:problem_nature" \end_inset \end_layout \begin_layout Standard A `problem' is a diagnostic hypothesis. This can cover the whole of Medicine, so coding is potentially tricky. The several issues are: \end_layout \begin_layout Enumerate If we are too vague, then multiple similar entities that have different implications can be lumped together. There are several possible solutions: \end_layout \begin_layout Enumerate Just live with the multiple codes, all on a single level. This will however tend to result in failure to unify e.g. tuberculosis at many sites under \begin_inset Quotes ...Read More