Well, this is a complicated question and I cannot provide a satisfactory answer without a fairly exhaustive empirical study. However, I thought the following offering might be of interest. The OECD provides data on the number of physicians as well as some health outcome variables for its members in its Health at a Glance publication. For the year 2009, I obtained the following variables:
1. Life expectancy at birth, years
2. Infant mortality rates, deaths per 1000 live births
3. Cancer mortality rate. Malignant neoplasms, deaths per 100,000 population (separately for males and females)
4. Practicing physicians, Density per 1,000 population (head counts)
5. Total health expenditure per capita (public and private) in US PPP dollars.
I then proceeded to quickly see what the relationship between these variables and the number of physicians per 1,000 of population might be using a simple LOWESS smooth (0.8 bandwidth). The plots generated are interesting. First, as the number of physicians per 1000 population rises, total health spending per capita also rises though it starts to level off once you reach about 4 physicians per 1,000 (Fig1). More physicians does mean more health spending so recent increases in Canadian medical graduates mean that spending could be poised to take off again in the not too distant future. Second, having more physicians is associated with longer life expectancy at birth but the relationship becomes flat at about 3 physicians per 1,000 population (Fig2). Third, more physicians are associated with a declining infant mortality rate but again the relationship levels off after 3 physicians per 1,000 population is reached (Fig3).
However, when it comes to cancer deaths, the results are a little murkier. For males, when it comes to their cancer mortality, I’m not sure it matters too much what the per capita number of physicians is. There seems to be a slight inverse u-shaped relationship with the cancer death rate first increasing and then declining (Fig4). The inverse u-shape is somewhat more pronounced for females. Does having more physicians per capita at first simply result in many more cases of cancer being diagnosed and treated which then results in higher mortality rates? At 2.4 physicians per 1,000 population, Canada would be just at the point where increasing physician numbers might be associated with declines in cancer mortality rates – assuming of course you would want to base your health policy decision making on a simple bivariate technique using one year of OECD data.
You cannot really draw any firm policy conclusions from the limited data and technique in this blog post. It would be useful to look at more health outcomes – for example there was OECD data for heart disease mortality – as well as break out the physician variable into GPs and Specialists - and of course add more years of data. Running regressions controlling for confounding factors would be the next step. However, I would imagine looking at the relationship between health expenditures, health care resources and health outcomes is a useful area for health policy decision makers looking for evidence on what to spend more – or less – money on.