What is dual-processing theory?

Dual processing theory is currently the most favoured explanation – or set of explanations - of the cognitive processes that characterise human judgement and decision-making. It assumes that our cognitive processes include both system or type I or ‘fast’ thinking (which is intuitive, automatic, fast, narrative, experiential and affect-based) and system or type II or ‘slow(er)’ thinking (which is analytical, slow, verbal, deliberative and logical).

Numerous models have been constructed within this theory. The recent Djulbegovic, et al. dual-processing model of clinician decision making proposes, as a description of what happens, that clinicians arrive at an intuitive answer and then, depending on the circumstances, add an analytical answer (using some procedure such as expected value maximisation), followed by, in the latter case, a ‘weighting’ of the two answers to produce a decision.

How does an MCDA/Annalisa based decision aid relate to such descriptive dual-processing models?

Annalisa is a template that envisages both types of cognition – analysis and intuition - contributing to all decisions, albeit in different proportions, depending on task structure and other considerations, as envisaged in Hammond's cognitive continuum theory, one of the original dual ones. The template is intentionally and unapologetically prescriptive (rather than descriptive) in being formally structured for slower, type or system II, thinking. It requires explicit specification of options and criteria, numerical performance rating for all options on all criteria and numerical importance weighting for all criteria. And it uses embedded expected value calculation as the basis for the integration of the performance ratings and importance weightings into an opinion (defined as the set of option scores).

But within that type II framing Annalisa is designed to facilitate the supply of the necessary components - options, criteria, ratings and weighting - by thinking that can occur at any point on the cognitive continuum from highly intuitive to highly analytical. One can construct an Annalisa in 5 minutes or 5 months.

As a matter of interest the two examples presented in Djulbegovic took less than 10 minutes to enter in an Annalisa. The translationALs for Acute Leukaemia and Pulmonary Embolism are here. The screen captures below show the threshold weights that create equipoise for the options, but it is a key principle of the person-centred care we are pursuing that the trade-off between the criteria is to be made, and made transparently, by the patient. So the ability to vary the two weights visibly and dynamically is an important attribute as opposed to simply finding out whether the patient's weighting is above or a threshold value. (In their examples there are only two criteria, so only one weight is needed. That is not the case when the number is increased.) Interestingly Djulbegovic, et al. speak of this trade-off as being made (by the clinician) in Type 1 thinking, virtually implying that it is not possible to arrive at the required weightings in more analytical ways.

 Djulbegovic, B., Hozo, I., Beckstead, J., Tsalatsanis, A., & Pauker, S. G. (2012). Dual processing model of medical decision-making. BMC medical informatics and decision making, 12, 94. doi:10.1186/1472-6947-12-94  Open access