The National Health Service number.

Abstract

Successful record linkage is dependent upon adequate identification of individuals. Acheson (1967) has stated the five fundamental criteria that should be fulfilled by an identification system: it should provide unique, permanent identification of each member of the population; it should be available on each of any pair of records to be linked, and it should be economical, i.e., it should consist of no more characters of information than necessary. He reviewed numbering systems currently in use, and one that fulfils three of his five criteria is the National Health Service number. It provides unique, perma nent identification which covers virtually the whole population of the United Kingdom. It is no less available than any of the other allocated numbers in current use, and its availability could, perhaps, be increased by administrative changes without undue cost. Although it is doubtful whether more than a small proportion of the population could memorize their National Health Service numbers, this disadvantage applies to any other allocated number. A number derived from nominal identifying data, such as the Hogben number (Hogben, Johnstone, and Cross, 1948), does not have this limitation; but Acheson (1967) has shown that discrepancies commonly occur in successive records derived by clerks from the same person when asked to give his or her sur name (or maiden name), forenames, and date of birth. Perhaps the principal drawback of the use of the National Health Service number as a means of record linkage is the heterogeneity of its structure. All National Health Service numbers consist of alpha and numeric characters. The alpha sequence is of variable length and is always followed by at least one numeric sequence, also of variable length. If there is more than one numeric sequence, they are separated by a slash (/); this is the only legitimate punctuation mark in the National Health Service numbers (excepting the Scottish birth registration number) and its correct location is an essential feature of the number. The numeric sequences are often separated by a full stop, a comma, or a dash; these should all be interpreted as a slash. Spaces between the alpha and numeric sequences, or within either, have no significance. Thus there are several series of numbers in use in England, Scotland, and Wales which collectively are referred to as the National Health Service number. Some of these series involve a sufficiently small number of people to make it feasible to consider renumbering them in line with one of the major series. In view of the advantages, mentioned above, of the National Health Service number over other number ing systems currently in use, it is pertinent to examine these in detail.