Lectures
Lecture 1: Production of health
3 parts in the course:
- Health, health care and income
- Equity
- Efficiency
1. Worldwide trends in income and health: over time, across countries and within countries
World income distribution> very unequally distributed across the world
Mean hides distribution what’s behind it.
Distribution of poverty> portion that is below the poverty line
Income and health are strongly associated
2. Health production functions and the income health relationship
HCS aim at improving health distributions> distributions: reduce inequality, disparities. So not about
improving the mean.
Both the level and distribution matter.
HC is only one of many determinants.
Health and economic development are related.
Measurement of marginal health product of HC and income is hampered by:
- Population health measurement> how to measure population health properly, which
indicator can be used?
- Estimation of marginal contribution is not total contribution
- Reverse causality and confounding
Health can affect economy, but economy can also influence health.
Health production function:
- Model the effect of specific types of medical care on population health.
- Population health= f(HC, lifestyle, schooling, environment, human biology, genes, etc.)
- Diminishing returns to scale: marginal product of health investments decreases.
- Maybe even to a level of diminishing returns that there is flat-of-the-curve medicine. Means
that further investments do not lead to further improvements on population health.
- Correlation or causality? Reverse? Spurious?
Curve that describes relationship between input and health is concave> means that an extra unit of
that HC input leads to ever higher health status levels but at a decreasing rate. First units generate
more health than the later inputs.
Marginal product of health on the y-axis will show a decreasing curve.
Marginal product of health= extra health generated by the first unit of medical care until the last unit
of care.
The last units often generate very low levels of additional health. The larger the scale of investment,
the lower the scale achieved because of the diminishing marginal product.
3. Income and health growth over time: some historical evidence
Health and development: demographic and epidemiological transition
Demographic transition:
- Mostly a consequence of a mortality effect and a fertility effect
Mortality effect, simply leading to longer lives.
Fertility effect, fewer births.
, These effects combined lead to an age-structure shift> there will be fewer young and more
older people in the population.
First mortality decline mostly due to technological progress, improvements in medical
technology and public health innovations
Fertility decline is mostly by behavioural change> couples have fewer children.
- In high-income countries, the simultaneous effects of both lead to modest growth of
population (2%).
- Developing countries: adoption of new technologies went much faster than the behavioural
shift which lead to an average population growth of 4%.
- Consequence> there are much higher population growth rates in LDCs, but ageing
consequences in others. Fertility effects leads to aging consequences in developed countries.
Epidemiological transition:
The causes of death and diseases are changing as well. The competing risks of various diseases of
where people die or where people can live with.
1. Expensive diseases drive out cheap diseases> expensive diseases are diseases who do not kill as
much people but do require lifelong treatment
2. Relative increase in old-age diseases
This means you get a shift in the population structures, visible in the pyramids of age groups. Most of
the deaths used the be concentrated under 4 years and currently it is concentrated in the high age
groups, this affects the population health.
Mortality decline was not uniform by age:
Expected age by death and life expectancy at different ages grew most among the newborns. There
were still improvements in other ages, but with a different extent.
Main determinants of the mortality decline, there were various stages>
- Late 18th century> improved nutrition
Half of all mortality decline due increased caloric intake and growth in height.
- Start of 20th century> public health
Water and foodborne diseases. Water purification explains half of mortality reduction.
- Before introduction of vaccines
Many declines in infectious diseases mortality occurred before the introduction of effective
vaccines. (better hygiene and water purification)
- Medical treatment
Two-thirds of cardiovascular mortality decline due to medical advances
Summarizing the historical evidence in a nutshell:
- Phase 1: nutrition with economic growth
1750-1850
- Phase 2: public health, mainly cleaner water in cities and better sanitation
1850-1930
- Phase 3: medical care, first vaccines and antibiotics, later big medicine> technological
treatments of cancer and cardiovascular
1930-2000
Countries life expectancies converge and diverge in different periods.
4. Income and health in poor countries
Very strong association between income and health, but what does this mean for policy?
What is influencing what?
Does this mean that we have to await income growth and development for improved population
health?
Or that they go hand in hand for other reasons? It is possible that both income and health go hand in
hand for other reasons.
,Does all income growth result in better health?
In other words: is wealthier always healthier?
Life expectancy and income has a concave relationship, but also shifts upward. Not so much the case
that countries are growing along the curve, but the curve itself seems to be shifting upward.
Why is that curve shifting upwards? Because technological improvements make health products
cheaper. Upward shift of the Preston curve is only possible later in years due to the technological
improvements.
Empirical analysis of income-health relationship: beyond the bivariate case
- Relative marginal productivity of income, assuming that income comes along with higher
healthcare investments as well.
Relatively low marginal productivity of income at aggregate level.
People started estimating models like this:
hi log(y i ) i
Y is the national income or the measure of national income
H is a measure of population health
B, both are in log terms, measures the elasticity. If both are in log terms, measuring the
change in age as a result of the change in Y but it is measuring the percentage change in
health as a percentage change in income.
Kakwani found strong positive income elasticities of population health but also found them
to be falling with rising income>concavity.
Later, this analysis was extended by including, apart from income, other explanatory
variables (z variables) in the regression equation. When accounting for these Z variables,
keeping them constant while estimating the B, they found that in particular, things like
poverty rate and how that income is used and turned into expenditure, like the share of
public expenditure, became more important. Even to the extent that income became
insignificant. So it’s not just income, it also depends on how that income is used.
For some reasons, some countries could have higher or lower health outcomes, irrespective of their
income.
Delta T measures for each of the years, a trend that is common in all the countries.
You want to know the causal effect of income on population health and you want to interpret that B,
which is elasticity since both are in log. For that reason you want to look for instrumental variables,
those are variables who influence why and in and of themselves do not have an effect on health.
For example the balance between export/import does have an effect on income but not a direct
effect on health. By using this, they estimate the long term causal income elasticity.
Inverted preston curve from Pritchett and summers
Not a concave curve because it is an indicator of bad health, so you see a convex relationship.
, Preston curve> shows that the relationship between income and health is stronger in low income
countries compared to high income countries. Still an increase in health but at smaller increments up
to a level where one starts worrying if there is still any health gain.
Summary of article on the determinants of mortality> article from cutler, deaton et al.
- Income and health are strongly associated both between and within countries
- However: income growth is neither necessary, nor sufficient condition for health
improvement. You can have health improvements without income growth and you can have
income growth without health improvements. So wealthier does not always imply healthier.
- What is key: knowledge, science and technology are key.
Especially the adoption of them. This adoption depends on availability of resources but also
availability of a healthcare system which can accommodate these new developments.
- Accelerated production of new knowledge may even increase the health gap by income,
because of lagged adoption
5. Income an health in rich countries
flat-of-the-curve medicine
1. Are there still gains in mortality or are most of the gains associated with that income rather in
terms of quality of life or morbidity improvements, rather than mortality gains.
We shouldn’t only be focusing on mortality and survival.
Alternative things which you can do:
2. Wilkinson: income inequality hypothesis: not the level but the (re)distribution of income within
country becomes more important.
3. You shouldn’t focus on all mortality, but only on mortality that is potentially avoidable, or
potentially amenable to healthcare.
Avoidable mortality still falls with rising health spending
Recent European study by mackenback et al
- Focus on mortality from conditions amenable to medical care.
Avoidable mortality comprises deaths from certain conditions that should not occur in the
presence of timely and effective healthcare. Things which you can do something about.
- Negative association with both income (GDP) and with health expenditure (% of GDP)
- No such negative relationship for non-amenable mortality.
