C. Little et al., USE OF A HOSPITAL DATABASE FOR THE EVALUATION OF CLINICAL BIOCHEMISTRY RESULTS, Journal of Small Animal Practice, 35(4), 1994, pp. 205-210
A database of the clinical chemistry results from unwell animals refer
red to Glasgow University Veterinary School, by general practitioners,
was studied. This database contains over 60,000 records, from more th
an 25,000 animals, obtained over a period of 18 years. A four-year per
iod was selected during which the instruments and methods used for the
se analyses were unchanged. The biochemical analyses made on plasma fr
om those dogs when first presented to the veterinary school were scrut
inised in detail. These data have been summarised in a tabular form us
ing the, first, fifth, loth, 25th, 50th, 75th, 90th, 95th and 99th per
centiles. The distributions of these values differ for different analy
tes and in particular do not conform to a normal curve. These data can
be used to provide estimates of rarity to indicate the frequency with
which extreme values associated with disease can be expected. Thus an
individual datum can be flagged as high, high**, high***, or high***
depending on whether it lies in the top 25 per cent, top 10 per cent
, top 5 per cent or top 1 per cent, respectively, of cases in the hosp
ital database. Low values can be flagged in the same way. This approac
h can complement the use of reference ranges from healthy animals to a
id the interpretation of clinical biochemistry results from a new pati
ent.