Frequently asked questions
Data
How much is “a significant amount" of difference?
(See PG 11.08) Health accounting is as much an art as a science, so it is not really possible to give a specific answer to the question of how much is “a significant amount” of difference. The acceptable level of difference depends partly on the type of data being dealt with and partly on the size of the total estimate.
For example, when dealing with reliable audited expenditure data presented according to strict public accounting rules, one should be very suspicious of any differences and look for errors in data entry. On the other hand, when dealing with data that are known to be estimates, a more tolerant attitude can be adopted. For example, discrepancies of up to 50% could be accepted when dealing with estimates of traditional providers, subject to the following comment:
- Whatever the type of data, the other important aspect is the size of the estimate itself; For example, a 15% discrepancy in a cell accounting for half of total spending is more serious than a 100% discrepancy in a cell contributing only 3% of the total. As a rule of thumb, there is cause for concern if a discrepancy between two estimates amounts to 2% of the total figure for national health expenditure.
- As experience grows in dealing with data sources and in working with the health accounts, health accountants will also develop an intuitive sense of when a discrepancy is important.
Which table should I start with: ‘Financing agents by Functions’ or ‘Providers by Functions’?
(See PG 13.06)Developing the distributional tables requires combining expenditure data on payers, providers, functions, and specific distribution-related characteristics of people using or receiving health goods and services. Two NHA statistical tables are important sources of information: those showing the financing agents by functions (FAxF, See Table 5.4 of the Producers’ guide) and the providers by functions (PxF, See Table 5.3 of the Producers’ guide). Which table –– FAxF or PxF –– is of greater policy relevance is a matter for local decision. In some countries, the policy emphasis is on where various services are provided; in such cases, the PxF table is useful. In others, the emphasis is on who pays for various services; here, the FAxF table is useful. Operationally, however, it is likely that one table cannot be populated without working on the other as well, and both may be needed for distributional analyses.
Experience in various countries suggests that preparing these tables is not a straightforward task. If the payment systems mostly pay by item of service (usually where social insurance is predominant), and if corresponding data are available, then direct estimation of the FAxF table may be feasible. More typically, however, public sector budgets are not allocated or reported by function. Rather, fixed amounts are allocated to providers (sometimes at the input level of budget, as for pharmaceuticals or salaries). In such settings, direct estimation of the FAxF table is only possible for part of the total expenditures – and even then for a relatively small part. If this is the case, then FAxP may be easier.
To populate the PxF and FAxF tables, the best course of action is probably the following sequence of steps:
• to break down as much as possible each financing agent’s payments by function;
• to estimate a table of providers by functions (PxF);
• to construct the financing agents by functions (FAxF) table by combining and reconciling the results of these two estimations.
What should be used at the municipal level to estimate health expenditures as a share of total municipal expenditures?
Health expenditures encompass those activities whose primary intent is to restore maintain and restore health (PG 3.02 and 3.03). Any and all of the health and health-related functions described in the SHA framework can be carried out or financed at municipal level and should be included. At the conceptual level one must decide which activities to include; the PG offers some guidance (see PG 3.06-3.11), but ultimately the decision falls upon the judgment of the health accounts team and their advisors.
The practical problem is that delivery and financing of health activities may be intermingled with non-health activities. Typically this is a problem common to all types of actors in the health system but especially troublesome at the municipal level. Specific advice is hard to give here, as the best solution varies according to the type of data and information available to the health accounts team. Typically some form of judgment is needed; in some cases, one may be able to use time and motion studies or other similar analysis (see PG 13.21-13.26), or the informed judgment of key staff familiar with the programmes in question.
The choice of denominator in the ratio of health expenditure as a share of total municipal expenditure depends upon what message is to be conveyed. A municipal equivalent to GDP is not necessarily meaningful even where it is available. One might try to develop regional GDP estimates using proxy or key measures such as wages to prorate national GDP, or use data from a household survey to develop regional shares of total national household consumption. However, these measures remain at best a crude approximation of regional/municipal domestic product.
However, if the analytic question is more focused there may be a better measure available. For example, if one wants to know how health rates as a municipal priority, one could construct a ratio of municipal public health expenditures to total municipal government revenues or expenditures, and compare the ratio across municipalities or against the national average. Similarly, one could compare household out-of-pocket spending to total household spending at the subnational level and compare this ratio across municipalities and against the national figure. In other words, by restricting the analysis to a particular group of actors one may be able to find a valid and reliable denominator that reflects the total resources available to those actors.
Mexico has done some interesting analysis looking at per capita expenditures on health by state, compared to an index of health status for each state. This analysis is useful for assessing health expenditures relative to need and was used by the Ministry of Health to guide transfers from the central level to various states.
How does one treat expenditures associated with “medical tourism”? Consider two situations:
1. An individual (Person X) from Country A is a resident in Country B. How should expenditures on health in Country B for this person be treated?
2. Person X from Country A goes to Country B for a medical procedure. How should expenditures in Country B for that person be treated?
In some countries medical tourism can be a major issue and hence Country B would like to count expenditures by foreigners (temporarily visiting their country) against their overall health expenditures. For example it has been suggested that in one of the countries in the MENA region total health expenditures excluding medical tourism was $800 million and medical tourism was estimated to be $600 million.
In situation 1, the health expenditures of Person X should be counted in the NHA of Country B because X is a resident in that country. This is a fairly easy situation to handle from a health accounts perspective.
Situation 2 requires more work from the health accountants. This case involves what is essentially a matter of export and import. Country B clearly produced the medical care but it is treated as an export to Country A. For Country A it is a clear case of imported health care, and should be included in Country A’s final consumption.
Thus, health accountants in Country B (the exporters) ought to make an estimate of the amount spent to treat Person X and deduct that amount from their NHA figure. Further, they should deduct that spending from the total payments by Country B financing agents (FAs) on behalf of Person X. For example, if X received free treatment at a municipal clinic, the value of that treatment should be deducted from the total expenditures made through the clinic. To preserve the total expenditure figure, Country B health accountants may wish to present an exhibit table or addendum table showing the value of medical tourism and known financing agents.
Health accountants in Country A (the importers) should include this expenditure in their health accounts. This would be done using the regular classification schedules for health functions and financing agents. The appropriate provider classification is HP.9 (rest of world), and any payments made by Country B financing agents would show as HF.3 (rest of world). The amounts expended would be expressed in currency units of Country A.
However, data limitations may preclude the treatment from being implemented – in either or both countries. Thus a default treatment of leaving the expenditure on person X in Country B's accounts is the result of "flawed" implementation rather than flawed concept. How much time and effort is spent to identify these expenditures depends upon the resources available to the health accounts team and upon the estimated magnitude of medical tourism (either from the perspective of importation or exportation).