How Healthcare System Comparisons Are Actually Measured: A Methodology Primer

Step 1: Know what kind of number you're looking at

Health statistics come in a few basic forms, and each behaves differently:

Before comparing two countries, make sure both numbers are the same type, measured the same way, for a similar year.

Step 2: Check the units, every time

Units are the easiest thing to get wrong. One example from this month's research: the infant mortality data on Our World in Data (accessed October 2026) gives the US 2024 value as 0.548 deaths per 100 live births. Most reports quote infant mortality per 1,000 live births, which makes the same figure 5.5. If you put the two side by side without converting, the US looks ten times better than it is.

Maternal mortality is conventionally reported per 100,000 live births, physician supply per 1,000 people, and hospital infections often per 1,000 patient-days. Mixing these in a table without labels is a common reason comparisons mislead.

Step 3: Check the year, country by country

A table labelled with one year often mixes several. In the datasets behind our current articles:

• Health spending as a share of GDP is available for 2023.
• Physician supply is labelled 2023, but the underlying values for Germany and France are from 2022 and for Cuba 2021.
• The maternal mortality series shown on the map is for 2020.
• Infant mortality estimates run to 2024.
• Out-of-pocket shares are a mix of 2023 and 2024.

So a "2026 healthcare comparison" may combine data from four or five different years. That's normal, because statistics take time to collect and verify, but it should be disclosed. A 2020 figure also carries the effects of the first year of the COVID-19 pandemic, which affected health systems and death rates differently from country to country.

Step 4: Ask whether the number was counted or estimated

Some figures are direct counts from registries, such as births and deaths recorded by civil registration. Others are modelled estimates, produced where registration is incomplete or where countries define events differently. Maternal mortality is a measure where this matters, so check whether a series reports registered counts or adjusted estimates before comparing countries.

Methods also change over time. The maternal mortality data we use warns that its US figures for 2003–2017 reflect a shift in how maternal deaths were recorded. A trend line that crosses a methods change can show a rise or fall that is partly an artefact.

Step 5: Understand how money is made comparable

Comparing spending across currencies needs a conversion, and the choice matters. Market exchange rates move with currency markets and can make a country look richer or poorer overnight. Purchasing power parity (PPP) adjusts for what money actually buys locally, which usually gives a fairer picture of how much care a given sum can pay for. Always check which one a comparison uses, and whether figures are in current or inflation-adjusted terms.

Ratios to GDP avoid currency conversion altogether, which is one reason they're popular. The trade-off is that they depend on the size of the economy, as our share-of-GDP comparison explains.

Step 6: Take apart any composite index

Composite rankings combine many indicators into one score. The Commonwealth Fund's Mirror, Mirror 2024 report (September 2024) is a well-documented example, and it shows how the machinery works:

1. Scope: 10 high-income countries and 70 measures.
2. Domains: access to care, care process, administrative efficiency, equity and health outcomes.
3. Data: the Commonwealth Fund's own international surveys of the public, older adults and primary care doctors, plus OECD, WHO and other sources.
4. Normalisation: for each measure, a country's result minus the average of the nine non-US countries is divided by that measure's standard deviation, so different units can be combined.
5. Weighting: each country's domain score is the average of its measures, and the overall score is the average of the five domains, which are weighted equally.
6. Special handling: the US was identified as a statistical outlier and excluded from the averages used to score the other countries.
7. Sensitivity tests: the authors re-ran the ranking with measures randomly removed and survey results shifted slightly, to see how stable the order was.

The result put Australia first, the Netherlands second, the UK third and the US last overall. But the US ranked second on care process. A single overall rank hides that a country can do well on one dimension and poorly on another.

The report is also candid about its limits. It notes that survey answers can reflect different expectations across cultures, that hospital and mental health care are poorly covered by comparable data, that domains are weighted equally by choice, and that middle-ranked countries can swap places with small data changes. Composite indices that publish this level of detail are easier to trust than those that don't. For a comparison of popular index brands, see WHO vs CEOWORLD vs Numbeo.

Step 7: Notice who is included, and who is left out

Mirror, Mirror covers 10 wealthy countries, so Colombia, Mexico and Turkey aren't in it. Other rankings cover only countries where data or survey partners exist. A country's absence from a list says nothing about its performance. It means it wasn't measured.

Facility rankings bring a different problem. Newsweek and Statista's hospital list, for example, warns readers not to set one country's hospital scores against another's, since the inputs differ by country. We cover that in our look at which countries have the best hospitals.

Step 8: Match the measure to your decision

National indicators describe whole populations, including people who never use private hospitals. If you're deciding where to have a specific operation, population-level outcomes are context, not evidence about your surgeon. The more useful measures at that point are procedure volumes, complication and infection data, accreditation status and the hospital's emergency backup. Our guide to reading a hospital quality report covers those.

A checklist you can use on any comparison

• What exactly is the indicator, and what is its denominator?
• What units are used, and are they the same for every country in the table?
• What year is each country's value from?
• Is the figure counted or modelled, and did the method change during the period shown?
• For money: exchange rates or PPP? Current or inflation-adjusted?
• For composites: how many measures, how normalised, how weighted, and how sensitive is the order?
• Which countries or facilities were excluded, and why?
• Does the measure describe a population, or the specific care you would receive?