Saturday, July 5, 2008

Questions for the Authors

Use this section to ask questions of the authors about any areas of difficulty you are having with Harald's Algorithm and we will try to answer them. Just pose your question as a Comment by clicking the Comment box and type in your question. It will be automatically sent to each of the authors, one of whom should answer shortly.

7 comments:

Greg Canning said...

8 clues or 13 clues???

The front page of this blog lists 8 clues to melanoma , whilst the derm 101 presentation lists 13 clues to melanoma ? why the difference , which should we use?

Dr Cliff Rosendahl said...

Thanks Greg. The front page of the blog is totally up to date (as is the Harald Poster) and reflects the current state of the algorithm. The longer list was condensed down to 8 clues prior to the Mastercass and one of those clues was modified during the Masterclass ( White lines reticular was changed to white edmzlines). This is a dynamic evolving algorithm and Harald has openly invited suggestions which will all be considered.

Dr Cliff Rosendahl said...

sorry that should read "white lines reticular was changed to white lines"

Unknown said...

Hi Greg,
I will update DERM101 as soon as I have more time! Currently, I do not know where to start! But you are right!

Unknown said...

I promise there will be no more updates from my side after this. The simplified rule: >1p or >1c and 1 c=m
is relatively new and was not incorporated in the original version!
13 was collapsed to 8 to make it easier, but if you look closely, the clues are the same, only the package is different!

James Maher said...

I notice that 5 or 6 colours isn't a clue for melanoma in the Algorithm. Is this because it isn't required as there are usually other clues, or is there another reason?

Dr Alan Cameron said...

You are absolutely correct James that a melanoma with 5+ colours will pretty much always have another clue. In fact, the vast majority of 5 colour lesions will not even require a dermatoscope for diagnosis.

This is not the main reason though James. Harald is insistent that any judgement we are required to make in an algorithm should be simpe and easily reproducible across observers. It's not hard to get agreement on whether a lesion has one colour or more than one. Past this point however, in a room full of people there is great variation in the counting of colours (would you really want your life to hang on me deciding whether your lesion had one or two shades of brown?) Harald contends (and i absolutely agree!) that a clue that different people cannot consistently agree on is not very useful.