If every event in life had an absolutely certain outcome, making decisions would be easy. If I knew which stocks were going to rise and which would fall, investing in the stock market would be facile. If I knew which car in the lot was going to be a lemon and which wasn’t, buying a used car would be simple.

But life is not like that. We do not know the outcome, in advance, of all the events that we encounter in life. This is especially true of healthcare.

And so, we turn to expected health utility theory (see glossary), which looks at the way in which people make health decisions under conditions of uncertainty, when we want to better understand the choices that people make and the measure or value that individuals assign to particular health states. We can give people a choice between a certain outcome or choosing between two other outcomes, each with different probabilities of happening. Or, it can be something like asking people how many years of life in perfect health they would give up for having less life in compromised health. Sometimes these are decisions about health insurance, sometimes these are decisions about health diagnoses or about health treatments and sometimes these are decisions about health spending or saving. But, no matter what, they are ideally always under conditions of uncertainty.

There are three important issues regarding heath utility theory that we must consider. The first issue is why we even care about people’s health values and preferences. After all, if patients living 1,000 miles away have a choice between a surgical procedure or pharmacological management for their disease, isn’t that their personal business? How does it affect us? Well, the reality is that both private and public entities are increasingly being faced with difficult decisions from a funding and access perspective and the use of aggregated health utilities across large populations of people is being used as a key outcome measure in the economic evaluations (think cost-effectiveness analyses) that ultimately decide which resources are funded.¹

The second issue that is critical to this subject is that health utilities are heterogenous. Different patients can have very different utilities for exactly the same health states. Consider the case of John and Maria.

They both have Parkinson’s disease. John is a 51-year-old neurosurgeon and loves to paint in his spare time. Maria is a retired 77-year-old grandmother who writes children’s stories and bakes brownies from home to support various charitable causes in the community. Given the choice between a definite state of living with Parkinson’s disease and managing symptoms through usual standard-of-care drug therapy or the ‘gamble’ of an experimental surgical intervention that could potentially cure Parkinson’s or result in death, these two individuals will have different expected utilities or preferences. A practising neurosurgeon who loves to paint may be willing to take more of a risk on an experimental surgery even with a high(er) probability of death if it rids him of tremors, stiffness and slowing movement than a retired grandmother who is willing to put up with such motor difficulties given her limited activities of daily living. So, the lesson here is that in considering the patient’s point of view, we cannot simply rely on a handful of patients for input and assume that everyone must have similar preferences.

The third issue pertaining to expected health utility theory and the idea of stated preferences is that we need to account for patients’ utility for specific health states and the amount of time that they spend in those states. It is not good enough to simply say that health state ‘A’ is preferred to ‘B’.

We must know how long a patient spends in a particular state in order to truly elicit a meaningful preference. Similarly, we need to know which years of life are being spent in a particular health state. It’s not just how long. Ten years may be different to someone if the ten-year period we’re talking about is from age 29 to 39 or from age 75 to 85.

As patients and clinicians, we don’t encounter these concepts in the real world. At least not in the way I’ve described them. And certainly not as explicitly as someone asking you which health state you prefer and putting a numerical value on it. But these ideas and constructs are critical to an efficient and fair healthcare system. If we don’t take societal and patient preferences for particular health states into account in our decision-making, we are poorer for it.

¹ Tolley, Keith. “What are health utilities?” Hayward Medical Communications, London (2009).