David Leonhardt and Amanda Cox had an interesting Upshot piece about new research showing that heavier babies do better in school. One implication is that many of the induced births that doctors have performed in recent decades have actually been counterproductive from the standpoint of the health of the child. (Obviously the health of the mother must also be considered.)
This is an interesting finding, although I'll leave it to medical professionals to assess the strength of the evidence here. But it does raise an interesting issue from the standpoint of GDP accounting and measurements of living standards.
Let's assume that this finding is accurate and that many of the c-sections and other methods to hasten child birth have actually been a net negative from the standpoint of the child's health and neutral with respect to the mother's health. All of these procedures get counted in GDP as part of the economy's output. This means that we were counting services that were making us worse off as part of GDP. If we didn't have these procedures, other things equal, our GDP would be lower.
This issue arises in health care all the time for the simple reason that most of us are not in a position to assess the best medical treatment and must rely on the wisdom of our doctors and the medical profession. This differs from something like clothes, where we might think we are the best judges of the clothes that we should wear. (Okay, can the snide comments about my wardrobe.) At the end of the day what we value is our health, not the number of tests, procedures, and drugs we get.
This is why I have always thought that for purposes like constructing cost-of-living indexes, we are best off just pulling out the money we spend on health care and measuring the price increases of non-health care consumption against the income we have left over after paying for health care expenses. This would treat spending on health care like a tax. If we want to then incorporate changes in our health into our assessment of living standards then we look directly at outcome measures (e.g. life expectancy, morbidity rates, self-rated health conditions), not the volume of health services we are consuming.