Myth of healthcare statistics

Dear Dr Jeremy Lim,

I refer to your 14 Jun 2013 Straits Times article [1] in which you cited Professor William Haseltine’s new book and Ms Gillian Tett’s subsequent commentary as examples of how Singapore healthcare is being held up as a model for low spending and impressive population health metrics. Sadly, there is a fundamental flaw in Professor Haseltine’s study, a flaw which Ms Tett didn’t detect but which completely changes the rosy picture that he painted.

Professor Haseltine failed to consider old age dependency when he compared Singapore’s healthcare costs to those of Western nations. All else being equal, a more aged population will have higher healthcare costs and since Western nations have more aged populations than Singapore, they naturally will have higher healthcare costs all else being the same. As Singapore’s population ages thirty years down the road, our healthcare costs will go up too even if our healthcare system remains unchanged. So if we were to compare Singapore today versus Singapore 30 years down the road, do we say that Singapore’s cost efficiency has deteriorated over the years or do we say that Singapore’s cost efficiency hasn’t changed but the population has aged and hence incurs higher healthcare costs? Controlling for old age dependency is important without which, there can be no proper or meaningful comparison of healthcare costs across nations.

Old age dependency ratios can be obtained from World Bank. When these are regressed against World Health Organisation’s healthcare costs, a significant relationship is obtained with a p-value of 8.4 × 10-23. The regression relationship is thus a credible predictor of how healthcare costs should vary with old age dependency. Against this predictor, Singapore fares poorly, overspending by 83% compared to old age predicted healthcare spending.

Country 2011 old age dependency ratio 2011 per capita health expenditure 2011 old age dependency predicted per capita health expenditure 2011 % healthcare overspending compared to predicted value
Italy 31.5 $3,436 $4,305 -20%
Korea, Rep. 15.9 $1,616 $1,765 -8%
Spain 25.3 $3,027 $3,290 -8%
UK 25.6 $3,609 $3,340 8%
Germany 31.2 $4,875 $4,251 15%
Finland 27 $4,325 $3,571 21%
Sweden 28.6 $5,331 $3,837 39%
Belgium 26.9 $4,962 $3,551 40%
France 26.4 $4,952 $3,480 42%
New Zealand 20 $3,666 $2,441 50%
Austria 26.4 $5,280 $3,478 52%
Singapore 12.7 $2,286 $1,248 83%
Denmark 25.8 $6,648 $3,384 96%
Netherlands 23.6 $5,995 $3,020 99%
Ireland 17.8 $4,542 $2,079 118%
Canada 20.8 $5,630 $2,570 119%
Australia 20.3 $5,939 $2,487 139%
Switzerland 25.1 $9,121 $3,255 180%
Norway 22.6 $8,987 $2,852 215%
Luxembourg 20.3 $8,798 $2,486 254%
United States 20 $8,608 $2,426 255%

The other issue with Professor Haseltine’s study is his comparison of healthcare costs as a percentage of GDP instead of healthcare costs itself. It is not a particularly meaningful comparison as the following example illustrates.

Both Mr Jones and Ah Huat are of similar age, health and fitness. Mr Jones earns $10,000 a month and spends $200 a month on healthcare. Mr Jones’ healthcare expenditure as a percentage of his salary is therefore 2%. Ah Huat earns $1,000 a month and spends $50 a month on healthcare. Ah Huat’s healthcare expenditure as a percentage of his salary is thus 5%. Therefore, according to the statistics, Ah Huat’s 5% expenditure is higher than Mr Jones’ 2% expenditure even though Ah Huat spends only $50 a month compared to Mr Jones who spends $200 a month.

The notion that Singapore has much better population health metrics than Western nations is also questionable. The following table [2] shows that very little separates Singapore from Western nations in terms of population health metrics.

Country 2011 Life Expectancy (both sexes) 2011 infant survival rate (probability of surviving between birth and age 1 per 1000 live births for both sexes) 2011 adult survival rate (probability of surviving between 15 and 60 years per 1000 population, both sexes) Average
San Marino 100% 100% 100% 100%
Switzerland 100% 100% 100% 100%
Iceland 99% 100% 100% 100%
Japan 100% 100% 99% 100%
Singapore 99% 100% 99% 99%
Italy 99% 100% 99% 99%
Sweden 99% 100% 99% 99%
Israel 99% 100% 99% 99%
Australia 99% 100% 99% 99%
Spain 99% 100% 98% 99%
Luxembourg 99% 100% 98% 99%
Andorra 99% 100% 98% 99%
Cyprus 98% 100% 99% 99%
Canada 99% 100% 98% 99%
Qatar 99% 100% 98% 99%
Norway 98% 100% 99% 99%
Netherlands 98% 100% 99% 99%
Ireland 98% 100% 98% 99%
Monaco 99% 100% 97% 99%
Kuwait 96% 99% 100% 98%
France 99% 100% 97% 98%
New Zealand 98% 100% 98% 98%
Austria 98% 100% 98% 98%
Malta 96% 100% 99% 98%
South Korea 98% 100% 98% 98%
Greece 98% 100% 98% 98%
Germany 98% 100% 98% 98%
Finland 98% 100% 97% 98%
United Kingdom 96% 100% 98% 98%
Bahrain 95% 99% 99% 98%
Belgium 96% 100% 97% 98%
Portugal 96% 100% 97% 98%
Slovenia 96% 100% 96% 98%
Denmark 95% 100% 97% 97%
Chile 95% 99% 96% 97%
Costa Rica 95% 99% 96% 97%
Saudi Arabia 92% 99% 99% 97%
Maldives 93% 99% 97% 96%
Czech Republic 94% 100% 95% 96%
USA 95% 100% 94% 96%

It is these false comparisons have led to Singapore’s health system being wrongly held up as a model for international attention which it rightfully doesn’t deserve. And if as you assert, the controlling hand of our healthcare market is that of our government, then you may have to contend with the reality that the controlling hand has potentially led to 83% overspending in healthcare.

[1] Straits Times, Myth of the invisible hand in health care, 14 Jun 2013, Jeremy Lim

[2]
• 2011 Life Expectancy, Infant mortality and Adult mortality are from World Health Organisation
• For life expectancy, comparison is based on percentage of highest life expectancy which is 83 years for San Marino, Switzerland and Japan
• Infant survival rate is used instead of infant mortality to convert the score to one where the higher it is the better it is. Infant survival rate is 1000 – infant mortality rate. Comparison is based on percentage of highest infant survival rate which is close to 100% for almost all First World nations.
• Adult survival rate is used instead of adult mortality to convert the score to one where the higher it is the better it is. Adult survival rate is 1000 – adult mortality rate. Comparison is based on percentage of highest adult survival rate which is 95% for San Marino, Switzerland and Iceland.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: