| Verified against primary record | |
| Definition | Datasets created primarily for administrative purposes by government agencies or bodies acting for them[1] |
|---|---|
| Why limitations are structural | Such sources are not created in response to the need for statistical data[1] |
| The named defect | Alignment of administrative units with the statistical units of interest[1] |
| Records read | Three United Nations statistical guidance documents, 2 October 2026 |
| Independently reported | |
| Control | The collection and processing of administrative data is beyond the control of the statistical office[2] |
| Bands apply only to the rows beneath them. The worked examples below are drawn from records cited in their own entries. | |
Administrative data is the raw material of most official statistics on cross-border healthcare, and understanding its one structural weakness explains most of the errors found in published medical tourism figures. The weakness is not sloppiness. It is that the data was built for a different job.
What it is
Administrative data sources are data sets created primarily for administrative purposes by government agencies or other entities working on behalf of the government.[1] A fuller formulation describes data collected by government institutions as part of their routine operations, which have not originally been collected for production of statistics.[2] The administrative functions that generate it are authorisations, registrations, permits, payments, sanctions and control.[3]
In cross-border healthcare the relevant administrative acts are a prior authorisation granted, a reimbursement claim paid, an entitlement document issued, a hospital admission coded for billing, a visa stamped. Each creates a record. None was designed to answer the question of how many people travelled for treatment.
The four structural limitations
International statistical guidance sets out the limitations in a consistent list, and all four appear in medical tourism figures.
No control over production. Administrative and other data sources are not created in response to the need for statistical data, and statistical agencies have in general no control or influence over the data production process.[1] The collection and processing is beyond the control of the statistical office, and unlike with a survey there is no control over who is reporting their data.[2]
Definitions that belong to the administrative rule. The units and variables an administrative source contains are defined by administrative rules and may therefore not be identical to those required by a statistical office, and concepts, definitions and classifications related to administrative data can differ from the ones needed for statistical purposes.[2] A variable exists because a rule needed it, with the boundary the rule needed, not the boundary an analyst wants.
Coverage error in both directions. Both over-coverage and under-coverage can lead to bias.[2] A reimbursement dataset covers only the reimbursed, which omits the self-paying majority in most medical travel and includes administrative records that never corresponded to a completed treatment.
The units mismatch. This is the central one, and official guidance gives it a name. Among the quality issues to assess when using administrative sources is the alignment of administrative units with statistical units of interest, illustrated with the example of a tax unit standing in for an establishment or enterprise.[1] The same guidance lists redundancy and duplicates as a distinct quality indicator.[1]
Translated into this field, the administrative unit is an authorisation, a claim, an admission or a discharge. The statistical unit of interest is a person who travelled. They are not the same object, and nothing in the dataset converts one into the other.
Three worked examples
The European Commission’s annual collection on cross-border healthcare under Directive 2011/24/EU illustrates the definitions limitation. It records authorisations and reimbursements, and it contains no field for the reason the person travelled and no disease breakdown, because reimbursement entitlement never depended on either. France reports that it cannot separate requests made under the Directive from those made under the Social Security Coordination Regulations, because its administrative process does not distinguish them. Eight member states operate no prior authorisation system at all, so for those countries the authorisation variable is structurally empty rather than zero.
English waiting time data illustrates the units mismatch at national scale. The referral-to-treatment collection counted roughly 7.3 million incomplete pathways against approximately 6.2 million unique patients in May 2026.[4] Neither figure is wrong. One counts administrative pathways and the other counts people, and a person with two referred conditions appears twice in the first.
Hospital discharge databases illustrate coverage and duplication together, and are treated separately in hospital discharge data in medical tourism.
How to use it anyway
None of this makes administrative data unusable. It is usually the only complete enumeration available, it is produced continuously, and it is free of the response bias that affects surveys. What it requires is that the unit be named in every statement derived from it. A sentence that says a country authorised a given number of cross-border treatments is defensible. The same figure presented as a number of patients is not, and the distance between the two is unknown unless the dataset carries a patient identifier.
See also
- Hospital discharge data in medical tourism, the main administrative source in this field
- International patient visits and unique patients, the units mismatch in detail
- Medical tourism surveys, the alternative source type
- Medical tourism statistics, the overview of available measurement
References
- United Nations Expert Group on National Quality Assurance Frameworks. Module for quality assurance when using administrative and other data sources for official statistics. March 2025. Verified against primary record: definitions, limitations and quality indicators read in full. Retrieved 2 October 2026.
- United Nations Statistics Division and UN Women. Guidance and toolkit on quality assurance of administrative data for official statistics. December 2022. Verified against primary record: definition and limitation sections read in full. Retrieved 2 October 2026.
- United Nations Statistics Division. Handbook on management and organization of national statistical systems, chapter 8. April 2023. Verified against primary record: definition and list of administrative functions read. Retrieved 2 October 2026.
- NHS England. Consultant-led referral to treatment waiting times. Verified against primary record: pathway and patient counts read from the published series. Retrieved 29 September 2026.
Sourcing note: the three United Nations statistical guidance documents were opened and read on 2 October 2026 and the waiting time series on 29 September 2026. The phrase describing the alignment of administrative units with statistical units is the formal name given by that guidance to what is loosely described elsewhere as counting episodes instead of people. The European Commission cross-border healthcare findings used as worked examples are cited in full in the entries on patient flows and statistics.