A new tool for research in prognosis: the ‘‘5-D-model’’ applied in a textbook analysis on prognostic information (PI)
Eychmüller, Steffen, Kantonsspital, St. Gallen, Switzerland, Glare, Paul, RPAH, Sydney, Australia, Christakis, Nicholas, Harvard Medical School, Boston, USA

Objective of study: Oncological textbooks are considered as one important source for PI especially in advanced illness. The Textbook of De Vita et. al. was analysed for metastatic conditions (quantitative and qualitative PI).
Methods: quantitative analysis encompassed PI reflected e.g by survival curves (in relation to other chapters like diagnosis, treatment etc.). The ‘‘5-D model’’ of prognosis was introduced as a tool to build relevant categories for qualitative PI (trigger words for the ‘‘5-D’’ categories: Death, course of Disease, Disability/discomfort, Drug toxicity, Dollars). This text analysis was limited to four major cancer sites (breast, lung, colon, prostate).
Results: Without limitation to metastatic disease, quantitative PI can be found on 7% (lung) to 20% (colon) of all pages dedicated to the cancer site when using the content list for the search. This proportion does not reflect the actual content of PI in the full text: e.g. 75% of all pages in the chapter on breast cancer contain relevant PI (content list: 5%). Breast and lung cancer provide more qualitative PI on the category ‘‘discomfort’’ than prostate and colon 27 and 29% versus 13 and 6% of PI- trigger words). Prognosis of ‘‘Drug toxicity’’ is mostly discussed in colon cancer, whereas the prognosis of ‘‘Dollars’’ is not considered except in lung cancer.
Conclusions: In De Vita’s Textbook, one of the major reference books in oncology, there is plenty PI but scattered throughout the text. PI seems to be mostly understood as ‘‘prognosis regarding length of survival’’. Patients may need also information about other aspects of prognosis (symptoms, disability etc.) for decision making. The ‘‘5-D model of prognosis’’ may serve as a key not only to screen available PI, but possibly to tailor future outcome criteria in research in oncology and palliative care.