Physicians and Workload: A Very Simple International Comparison

The Canadian Medical Association has been having its annual meetings this week in Ottawa and in honor of the event, let me put out another international comparison on physicians using data from the OECD Health Statistics 2013.  The first chart (Figure 1) is a basic resource availability measure showing the number of physicians per 1000 population for the OECD countries circa 2011.  The second chart (Figure 2) is a workload measure showing doctor consultations per capita circa 2011. 

As Figure 1 shows, amongst the OECD countries Canada is well below the OECD average and in the bottom third when it comes to physicians per 1,000 of population.  We are between Japan and the United States when it comes to this measure.  The highest number of physicians, relative to population, is in Greece, Austria and Italy.

Figure 2 presents what could be termed a workload measure – physician/doctor consultation per capita.  Here Canada is above the OECD average and in the top third lying between Belgium and Spain.  What is interesting here is that Canada has a number of physicians per 1000 of population that is almost identical to the United States and yet when it comes to consultations per capita, we are nearly double those of the United States.  This would suggest that Canadian physicians are working harder than physicians in the United States.  Of course, we need to factor in not just the number of consultations but the number of consultations relative to the resource base.

So, Figure 3 tries to present a crude “workload intensity” measure by presenting the ratio of consultations per capita to physicians per 1000 population for each of these OECD countries.   This is a crude measure because these numbers do not differentiate between GPs or specialists or types of specialty or composition of the cases or demographic differences across the populations in these countries that might account for differences in consultation numbers.  Moreover, we do not know if these consultations are just five minutes or span a considerably longer period of time.  Moreover, ideally one would want to look at physician workloads while controlling for many of these other confounding factors.

When it comes “workload intensity”, Canadian physicians are above the OECD average and again in the top third – this time between the Czech Republic and Slovenia.  We are again well above the United States in this measure.  What I also found interesting is that Greece has the highest number of physicians per 1000 population and is the lowest in the “workload intensity” measure.

So the final question one has to ask is if is there is any relationship between per capita health spending in these countries and the “workload intensity” of physicians – as measured by these crude and albeit imperfect numbers.  Physicians are the gatekeepers for a lot of health spending so one might expect that there is a relationship between workload and the ultimate health care bill in these countries.

Figure 4 plots the physician workload variable from Figure 3 against total per capita health expenditure in US PPP$ for these OECD countries.  A linear fit is also plotted. Higher workloads are associated with lower per capita total health spending while lower workloads are associated with higher per capita spending.  There seem to be some outliers here – Korea, Japan and the United States.  Korea and Japan have pretty high workloads and fairly low spending. The United States has a pretty low physician workload and very high health spending.  Removing these three observations still results in a negative slope (not shown).

Based on Figure 4, I would venture that these crude numbers suggest that a given stock of physicians seeing more patients is in the long run associated with lower rather than more health spending.  Could it be more frequent consultations are better for monitoring patient health and save money in the long term by keeping on top of medical condition?  Is a factor in the US health system’s high spending the fact that doctors see patients less often so that when they finally decide to see a physician they are in much worse shape and more expensive to treat? Fascinating stuff.

Related articles

What Does Health Spending Buy You?

Are More Physicians Preferred to Less?

Comments

[Nathan W]

I guess you need a lot of doctors if they don't work much. This is probably really expensive given how much it costs to train a doctor.

Perhaps it would be cheaper in the long run for Greece to pay doctors more, even per consultations, to work more?

[genauer]

@ Nathan,

Greece is just a very special place, with respect to nearly everything.

Half the time, OECD and other numbers for this country are missing, PISA scores not published, home ownership and value all wildly distorted. And where numbers exist, like hours worked per year, you can not believe them. Just eliminate this country from any analysis of "normal".

On doctor consultations in Germany (18 per capita per year, some serious disconnect to Figure 3 above !!!!!!, Livio, please check !).

We have a well known, but not much published problem of old folks using that for social interaction. Some attempts like "Praxisgebühr" actually caused the opposite effect.

And some well meaning programs are also shaped in a way, getting you every quarter to your GP and some "specialist". You have to work actually to avoid it.

[genauer]

I checked the OECD database (http://stats.oecd.org/index.aspx?DataSetCode=HEALTH_STAT) , and Livios Figure 2 shows the data (Health Care Utilisation, Consultations) there.

The number for Germany there is 9.7. The numbers in circulation in Germany are 16.3 - 18.1(e.g. http://de.wikipedia.org/wiki/%C3%84rztehopping the english wiki Doctor_hopping is about something completely different !!) 16.3 - 18.1

Greece numbers are missing since 2006, the usual. I interpret this as people do not believe the data before either.

"Deutsche Primär- (Haus-) Ärzte haben im Schnitt 243 Patienten pro Woche, in anderen in unten genannter Studie untersuchten Industrieländern sind es meist zwischen 102 und 154. Die Zeit pro Patientenkontakt lag international im Mittel zwischen 11 und 19 Minuten, in Deutschland bei unter acht Minuten. Eine der Ursachen der verbreiteten Unzufriedenheit deutscher Hausärzte könnte in ihrer Belastung durch die höhere Zahl von kürzeren Patientenkontakten liegen.["

Basically if you need any drug, you have to go once a quarter to the physician, get the prescription, often you do not even see him. A classical case of "making the numbers" and collecting some treatment fees.

And often, when you see him, it is for a very short time (the average 8 minutes above) and many do not listen, no wonder given the circumstances.

Example:

I did stop taking certain prescriptions, he and his colleague really urged me on to take, 2 years ago, and did not diet since that, and my body shows that clearly : - ) I mention this at the followup meeting (both the sampling and the meeting triggered by me) with all blood values ok, and blood pressure too, .... he nods, looks a little in my patient data and utters:

"Fascinating what a strict diet like yours can achieve"

Ieeeek.

I also have to go now 2 times a year to the dentist, otherwise I dont get a 10% repayment on my payments. But a 5 minute check for dark spots does not need any preparation /personal knowledge on his side, so I assume my insurances knows what they are doing. And I typically only have to wait about 5 mintues at the dentist. He is busy but very well organized.

[Livio Di Matteo]

Genauer:That is interesting about the number circulating in Germany as it is certainly much higher than that reported in the OECD. The OECD numbers are for 2011 or nearest year (which in the Greek case is 2006 for consultations per capita).

[steve]

I would have serious doubts about the methodology if I were you. I have been practicing for nearly 30 years and I run a 70 provider group. I have many Canadian friends who are physicians. We vacation in Canada and I have met with physicians there. We inevitably compare hours and workload. They aren't that far apart. I would look at coding differences and at differences in the number of specialties, especially primary care. Heaven knows we do many more procedures than you guys do, so just using consults seems like a very limited approach.

Steve

[Normand Leblanc]

Doctor "A" earn 100$/hr. Doctor "B" earn 200$/hr. I'm sure that a certain number of doctor B will be working less hours (maybe half time) because an equivalent salary of 100$/hr working full time is all they need.

That means that the more you pay them, the less they work and it fit very well with the last graph with a negative slope.