16 February 2017: Thomas Rouyard
Trade-off between behaviour change and health outcomes: an analysis under prospect theory
Speaker(s): Thomas Rouyard
Date: Thursday, 16 February, 2017
Contact person(s): Teresa Bago d'Uva
Background and aim
Health promotion interventions can be improved with methodology from behavioural economics to identify and target decision-making biases at the individual level. In this context, prospect theory (PT) provides a suitable framework within which people’s decision-making processes can be operationalized. Our study is the first to quantify the full utility function under PT of a patient population, using real-life choices and outcomes that include both dimensions of health utility: life duration and quality of life. Specifically, we focused on decision-making processes underlying the trade-off between behaviour change and health outcomes, as it is urgent to improve interventions encouraging behaviour change of people with ‘manageable’ chronic diseases (i.e. diseases that can be controlled with appropriate self-management behaviour).
Method and study population
Our method follows on Attema et al (2013)’s semi-parametric method of utility elicitation in the health domain. It uses certainty equivalents (CEs) to measure the parameters of a specific utility function (exponential family). CEs are elicited through a series of choices over binary prospects. However, we make choices over prospects slightly more complex by adding a quality of life dimension, so that we can capture preferences based on a broader notion of health utility. Our study population includes 120 patients aged between 30 and 60 years old. 60 patients were diagnosed with a chronic disease (Type-2 diabetes mellitus or T2DM, the most common form of the disease), and 60 patients were healthy. T2DM is a good example of a manageable chronic disease: it often requires multiple drug therapies, frequent monitoring of risk factors and regular health care contacts, and good control is associated with health outcomes (life expectancy and quality of life) more similar to their healthy counterparts.
Results and discussion
We observed risk aversion in both the gain and the loss domains, as well as significant loss aversion. Unexpectedly, people with T2DM were found to be more risk averse than healthy people. We also found a significant age effect on preferences. Improving the understanding of decision-making biases is a crucial first step towards the development of new interventions encouraging behaviour change. Our results establish a valuable starting point for the development of new, tailored interventions based on what could be called ‘debiasing strategies’. For example, risk communication interventions may benefit from taking advantage of loss aversion by highlighting the consequences of potential complications.