| Verified against primary record | |
| What a published rank is | Dependent on modelling choices, and to be read within a confidence interval rather than at face value[1] |
|---|---|
| Audited practice | Monte Carlo simulation varying several modelling choices at once, yielding intervals on every rank[1] |
| Weights | Regardless of method, weights are essentially value judgements[2] |
| Records read | A statistical audit, the governing methodological handbook and two peer-reviewed critiques, 2 October 2026 |
| Independently reported | |
| Rank volatility | Slight changes in one indicator may cause drastic changes in position[3] |
| Bands apply only to the rows beneath them. No ranking positions for destinations are reproduced in this entry. | |
Medical tourism rankings are ordered lists of countries, cities or hospitals, usually derived from a composite index and published as a league table. The ordering is the part most widely reproduced and the part that carries the least information, for reasons that the auditors of major international rankings state openly about their own work.
A rank is an interval, not a position
The single most important fact about any ranking is stated plainly in the statistical audit attached to one of the better-documented international indexes. Readers are instructed to consider a ranking not only at face value, but also within the 90 per cent confidence intervals, in order to better appreciate the degree to which an economy’s rank depends on modelling choices.[1]
Those intervals are produced by work, not assertion. The audit runs 4,000 Monte Carlo simulations varying three modelling choices simultaneously, and publishes the resulting interval alongside every reported position.[1] An index whose compilers have done this can tell a reader how much of a country’s position is signal and how much is an artefact of the method. An index that has not done it cannot, and the single integer it publishes conceals that fact rather than resolving it.
Why small changes move positions a long way
Ranks are more volatile than the scores beneath them, because converting a continuous score into an ordinal position discards the distances between entities. Where several are closely bunched, a trivial score change reorders them. A study of university rankings records that ranking weights are arbitrary, that slight changes in one of the indicators may cause drastic changes in a university’s position, and that although more than half the percentage of the indicators has been normalised, the issue of procedure had never been documented precisely.[3]
The consequence for reading a medical tourism league table is that a country moving several places between editions is not evidence of change in its healthcare system. It may reflect a revised weight, a changed data source, a replaced indicator, or nothing but the reordering of a cluster of near-identical scores.
The weights are not what they appear to be
Published weights are routinely read as statements of importance. They are not reliably that. Research on composite indicators finds that relative nominal weights are hardly ever found to match relative main effects, that in many cases the declared importance of single indicators and their main effect are very different, and that the data correlation structure often prevents developers from obtaining the stated importance, even when modifying the nominal weights.[4]
The governing methodological handbook adds that regardless of which method is used, weights are essentially value judgements, and that equal weighting does not mean no weights but implicitly implies that the weights are equal.[2] A ranking is therefore always an expression of a particular view of what matters, and the only question is whether that view is disclosed.
Reputation surveys rank perception
Several rankings in this field draw on opinion rather than records, and the weakness of that design is documented in the adjacent literature on universities, where the observation is that most academics know in detail only their internal networking and only a few reputable universities.[3] Respondents with limited direct knowledge default to prominence, and prominence is partly produced by previous rankings.
The point transfers directly. A perception survey can measure what a sampled population believes about a destination. It cannot measure a supply-side fact such as accreditation status, surgical volume or complication rate, and a ranking that mixes the two without saying which indicator came from which kind of source cannot be interpreted.
What to check before citing a ranking
Five things decide whether a position can be quoted. Whether a confidence interval or sensitivity analysis is published for each rank. Whether the indicator list, weights and aggregation formula are published. Whether each indicator’s data source and reference year are named. Whether the inputs are measured facts or survey perceptions, and whose perceptions. And whether any external statistical audit has been published.
Where those are absent, the honest formulation is that a named organisation placed a country at a given position in a named edition of its own ranking. That is a fact about the ranking. It is not a fact about the country, and it does not become one through repetition. The same test applied to the construction of the underlying score is set out in medical tourism indexes.
See also
- Medical tourism indexes, how the scores behind a ranking are built
- Destination markets in medical tourism, why a destination league table cannot be built from official data
- Medical tourism market forecasts, the same transparency test applied to projections
- Medical tourism statistics, the measurement problems underneath the indicators
References
- European Commission Joint Research Centre. Joint Research Centre statistical audit of the Global Innovation Index 2023. Published as an appendix to the Global Innovation Index 2023. Verified against primary record: audit methodology, simulation design and interval guidance read in full. Retrieved 2 October 2026.
- Nardo M, Saisana M, Saltelli A, Tarantola S, Hoffman A, Giovannini E. Handbook on constructing composite indicators: methodology and user guide. OECD and European Commission Joint Research Centre, 2008, ISBN 978-92-64-04345-9. Verified against primary record: weighting and aggregation chapters read. Retrieved 2 October 2026.
- Fauzi MA, Tan CNL, Daud M, Awalludin MMN. University rankings: a review of methodological flaws. Issues in Educational Research, 2020;30(1):79-96. Independently reported: peer-reviewed review. Retrieved 2 October 2026.
- Paruolo P, Saltelli A, Saisana M. Ratings and rankings: voodoo or science? Journal of the Royal Statistical Society Series A, DOI 10.1111/j.1467-985X.2012.01059.x. Independently reported: peer-reviewed research, author manuscript read. Retrieved 2 October 2026.
Sourcing note: the statistical audit, the methodological handbook and the two critiques were opened and read for this entry on 2 October 2026. The audit is cited for its method and its instruction on reading ranks; none of the positions it reports are reproduced here, because this entry makes no claim about any country’s standing. The university ranking literature is used because it examines the same mechanism with better documentation than exists for rankings in this sector.