Health Decision Science
with Nick Bansback (UBC School of Population and Public Health, Centre for Health Evaluation and Outcome Sciences
Why do some people think they are better decision-makers than others?
Are they right in their thinking?
With these questions, Nick began his Evening Rounds presentation on the use of decision science in health care. People working in the health care system make critical decisions every day, yet these decisions may be unknowingly based on poor evidence and not in the best interest of the patient. It is often assumed that new research leads to better decision-making – but research is incomplete and complex, and the human brain has its own limitations, so this assumption is unfounded. Health decision science can help us improve health care by allowing us to make better decisions with the best evidence on-hand.
What is decision science?
Decision science includes three components:
Descriptive: How do people make decisions?
Normative: How should decisions be made?
Prescriptive: How can we help people overcome decision errors?
There are numerous errors and biases in the way that people make decisions. Understanding these types of errors can facilitate overcoming them.
We undervalue the future. We don’t do things now that will positively impact our future. For example, saving money for retirement, keeping physically fit, delaying treatment because of side effects
We don’t know how others value. Value is relative and subjective – we might not understand that other people value things differently than ourselves. For example, physicians only recommending treatment that they would take, not understanding that the patient may place more value on a different treatment.
When it’s cognitively difficult, we give up. If there is too much complexity or too many options, we delay making a choice. For example, deciding which diet to follow, or when to seek health care.
We overestimate the chance of good things and underestimate the chance of bad things. We are not accurate in our perception of the evidence. For example, both patients and clinicians overestimate the benefits of treatment and underestimate the potential for harm.
We base decisions on previous (and underpowered) outcomes. We don’t know the odds of every decision we make, but we know (or can estimate) the outcomes, so our judgements are biased by outcome and disregard the actual evidence. For example, anecdotal information on side effects and treatment outcomes.
Normative decision theory is a branch of decision theory that uses decision analysis to quantitatively deal with uncertainties. Analysis involves building decision trees and assigning probabilities to identify the best decision. Normative decision analysis can be used to inform many aspects of health care, including research, treatment, policy, and even hospital design.
So how do we actually use decision theory to improve decision making?
Educate: Decision analysis requires both decision literacy and understanding of probability theory – not an easy concept for many people. People need to learn about probabilities.
Awareness of bias: Learn to recognize overconfidence, consider opposite arguments and reframe thinking.
Tools: Use decision support tools, decision framing and policies to make it easier for people to make better decisions.