FALCOn (Facteur AnthropoLogique Cancer Orl)

An analysis using a classification method will define a typology of therapeutic choices given their aggression and potential risk.

These typologies of prescriptions will then be confronted with the characteristics of physicians in terms of propensity for risk taking and their socio-demographic characteristics using a chi2 test (univariate) and a logistic regression (or polynomial if a binary typology cannot be released (multivariate).

Variability of choice and influence of physician characteristics on these choices will also be analyzed based on patient characteristics

Expected results :

• Impact of socio-professional characteristics on the decision-making in oncology of aerodigestive carcinomas.
• Impact of psychological factors ie tendency to take risk

7 clinical cases will serve as a basis for the creation of clinical vignettes: A standard case (T2N1M0 pelvi-lingual carcinoma, good general condition patient) will be part of these vignettes allowing a validation on the respect of the guidelines.The other 6 clinical cases will be cleavants.

The experimental plan will take the form of a Latin Square to propose (in addition to the standard case), 6 vignettes to each practitioner by ensuring that all the factors will be presented in a balanced and random way.

Practitioners will be asked to give their clinical recommendations for each clinical case presented.

All the health practitioners will be asked to self-evaluate their propensity to take risk using a 11-point Likert scale ranging from "not at all willing to take risks" to "fully prepared to take risks" in 4 different domains including their daily life, their personal finances, their patient's health and their own health. The Likert scale used in this test is a 0-10 scale where 0 means "not at all willing to take risks" and 10 means "fully prepared to take risks"

Individual attitudes toward risk will be obtained through certainty equivalent elicitation of a lottery. Using a scrollbar, respondents will be asked to provide the amount of money that makes them indifferent between gaining it for sure or playing the lottery giving them 500 euros with half a chance and nothing otherwise. The expected gain of this lottery is 250 euros, so when the elicited certainty equivalent is inferior (equal/superior) to 250 euros, the respondent is considered as risk averse (neutral/seeking).

Individual attitudes towards uncertainty will be obtained through certainty equivalent elicitation of a lottery where the probability of gain is unknown. Using the same scrollbar, respondents will be asked to provide the amount of money that makes them indifferent between gaining it for sure or playing the lottery giving them 500 euros with an unknown probability and nothing with an unknown probability.

Individual rationality attitudes will be obtained through an Allais test. The expected utility theory is an axiomatic model of decision under uncertainty and is often considered as a normative model of rationality. Allais and Kahneman and Tversky proposed a set of two binary choices that allow to test the independence axiom of the expected utility theory under risk and therefore the adequacy of an individual to the model.

The Allais task will be assessed with this 2 questions:

Choice 1: Which option do you prefer?

• Option A gives you 100% of chance to win 2000$
• Option B gives you 80% of chance to win 3000$ and 20% of chance to win 0$ Choice 2: Which option do you prefer?
• Option C which gives you 25% of chance to win 2000$ and 75% of chance to win 0$
• Option D which gives you 20% of chance to win 3000$ and 80% of chance to win 0$ Choice pattern AD indicates a violation of rationality. Same line of reasoning applies for the choice pattern BC.