Saturday, May 27, 2023

GPT4 on QALYs

 In health econmics, much is made about QALYs. But seems like the fundamental source of QALYs is dubious. Can we really translate half as many migraines per month, to QALY Years? Even if you can do surveys and get an average, is the dispersion huge so as to made the average of doubtful value? I am not referring to general criticisms of the use of QALYs, but the source of the actual QALY digits themselves.

ChatGPT

Quality-Adjusted Life Years (QALYs) are a commonly used measure in health economics that attempt to capture both the quantity and the quality of life produced by healthcare interventions. They provide a way of comparing the benefits and costs of different treatments, even if those treatments address very different conditions. The calculation of QALYs involves assigning a utility value, typically between 0 (representing death) and 1 (representing perfect health), to each health state that a person might experience.

You raise an important concern about the source of these utility values. They are usually derived from surveys or questionnaires that ask people to make hypothetical choices between different health states. For example, they might ask how many years of life in perfect health a person would be willing to give up in order to avoid a given number of years with a particular health problem, such as chronic migraines.

However, these surveys do have limitations. Firstly, people's stated preferences may not accurately reflect their true preferences. For example, people might not fully understand the implications of the health states they are being asked to consider, or they might answer in a way that they think is socially desirable rather than reflecting their true feelings. This can introduce bias into the resulting utility values.

Secondly, there is often considerable variability in the utility values that people assign to the same health state. This can reflect genuine differences in people's preferences and tolerance for different health states. For instance, some people might find migraines much more debilitating than others do. But it can also reflect measurement error and random variability in people's responses. This variability can make the average utility value less representative of any individual patient's preferences.

These issues can indeed make the resulting QALY estimates less reliable. It's important to consider them when interpreting QALY-based analyses. However, while imperfect, QALYs remain a useful tool for comparing the benefits of different health interventions in a systematic and consistent way. There are also ongoing efforts to improve the methods used to derive utility values, to try to make them more accurate and representative.

User
For example, I have a mild disability in my left shoulder, and I am quite nearsighted, both of which might be attached to QALYs, yet I'd rate my quality of life as 1.00, without a hesitation.
ChatGPT

Your experience highlights an inherent challenge with the QALY system. The QALY system aims to provide a standardized, measurable, and universally applicable unit for comparing the value of different health interventions. However, individual experiences and attitudes towards health and disability can vary significantly, which is not always fully captured by QALY measurements.

The utility values used to calculate QALYs are designed to represent average preferences across a population. They may not necessarily reflect the preferences or experiences of any individual person. In your case, despite having a mild disability and being nearsighted, you still rate your quality of life as 1.00 (perfect health). This shows that you've adapted well to your health state and/or do not consider these conditions to have a significant negative impact on your quality of life.

On the other hand, some other individuals with similar health conditions might perceive them as significantly reducing their quality of life. Their perception may be influenced by many factors such as their personal experiences, other co-existing health conditions, their coping mechanisms, their support network, and their overall outlook on life.

So while QALYs are a useful tool in health economics, they can't fully capture all the nuances of individuals' experiences. They are most useful as a population-level measure for comparing the relative value of different health interventions, rather than as a measure of any individual person's quality of life. As always, the limitations of any measure must be taken into account when interpreting the results.

User
The concern is that someone who is blind has a year by year QALY of 0.5, yet we wouldn't want to rate that person as half as deserving of dialysis or chemotherapy. That would be appalling, yet the QALY method goes right there.
ChatGPT

You raise an important point, and indeed, this is one of the most serious ethical considerations when applying QALYs in health care decision-making. The QALY method is designed to maximize health benefits across a population given limited resources, but it does not take into account issues of fairness, equity, or individual rights.

If we strictly adhered to a QALY-based approach, it could imply giving preference to treatments for conditions that have a higher utility value (such as improving eyesight) over those with a lower utility value (like providing dialysis or chemotherapy for a person who is already blind). This would indeed be ethically troubling.

To mitigate this concern, many health systems and organizations that use QALYs for health economic evaluations also consider other factors. These can include the severity of the disease, the age of the patient, the potential for improvement, the potential for prevention of future health problems, and societal values around fairness and equity.

In fact, the National Institute for Health and Care Excellence (NICE) in the UK, which uses QALYs in its decision-making process, explicitly states that it may recommend treatments that are not the most cost-effective if there are other compelling reasons to do so, such as if the treatment is for a very severe or life-threatening condition.

The use of QALYs as a decision-making tool is not meant to replace ethical considerations, but rather to provide a quantitative measure of health benefits that can be one part of the decision-making process. Decisions about healthcare resource allocation should always consider a range of ethical, social, and personal factors, not just QALYs.

User
On the other hand, if I may switch teams ,I saw a recent article that assigned QALYs to 20 different stroke treatment quality measures. All were approved and endorsed. Some had huge health benefits and great QALYs, some very bad indeed. Since we have free hand to choose quality measures for hospitals, we should choose the most impactful ones, and the good QALYs simply correlate to a large benefits in health (tempered by cost). So there, QALYs are quite good because you literally are, already, looking at a market basket of measures and knowingly choosing a few measures from the basket for implementation.
ChatGPT

Your point is well taken. Despite the potential limitations and ethical concerns, QALYs can provide valuable information when assessing different health interventions or quality measures. The benefit of QALYs is that they offer a standardized measure that can be used to compare the effectiveness of different treatments or interventions, taking into account both the quantity and quality of life.

In the case of stroke treatment quality measures, for instance, QALYs can help to identify which measures are likely to have the greatest impact on patients' health. This can inform decisions about which measures to prioritize for implementation. By considering the QALYs associated with each measure, health care providers can choose interventions that offer the most health benefit per unit of cost. This can help to improve the efficiency and effectiveness of health care services.

In this context, the use of QALYs can be seen as a way of ensuring that health care resources are used in a way that maximizes health benefits. This can be particularly valuable in situations where resources are limited and difficult choices need to be made about which interventions to implement.

However, it is still important to bear in mind the limitations of QALYs and to interpret them with caution. In particular, it is important to consider other relevant factors, such as ethical considerations and patient preferences, alongside the QALY estimates. Furthermore, when interpreting QALYs, it's crucial to remember that they are based on averages and may not accurately reflect the preferences or experiences of all individuals.