Skip navigation

The same as yesterday, laddie!: inappropriate laboratory tests

We use laboratories with ever greater frequency to order tests to help make or exclude diagnosis, or help develop a prognosis, or because we always order these tests in these sorts of patients. Sometimes the ordering of those tests is inappropriate.
Bandolier well remembers an occasion, many years ago, when a new house officer darted into a laboratory on a Friday evening at 5.00 pm demanding an "urgent" albumin measurement on a patient. A wise old biochemist (one of Bandolier's heroes) asked
Q: has a test been done previously?
A: Yes, one was done yesterday
Q: What was the result then?
A: 42 (or whatever albumin results are).
So the sample was held up to the sunset, and the wise old biochemist said "It's the same today, laddie, the same today".
The point is that albumin concentrations don't really change that much unless something major has happened, and it hadn't in this case. The requesting of an urgent albumin in this case was inappropriate.


David Naylor and Carl van Walraven have done a heroic job in trying to discover how best to define inappropriateness, and to find out how big the problem of inappropriate laboratory testing is by systematically reviewing laboratory audits [1]. Their paper is long and complicated, but full of interesting thoughts, most of which Bandolier can only skate over.

Their searching was thorough, and they restricted themselves by excluding radiological or pathological tests, and screening tests. They found 44 articles, 34 of which had explicit criteria for appropriateness (and which looked generally to Bandolier to be fair, or even conservative).


The main results from this enormous amount of work are summarised in the Table, where information has been pooled to give an estimate of the overall inappropriate use for different categories. The clear answer is that tests are used inappropriately to a very large extent, which in some cases is over 90%. We do not know the corollary - that is, how often should tests be used but are not, since few studies examined that aspect.
Summary of inappropriate test use
Study Number of reports Number of tests Percent inappropriate Range
Studies with implicit criteria 11 5360 56 11 - 95
General biochemistry & haematology 5 63,030 15 11 - 70
Microbiology 7 4979 46 5 - 95
Cardiac enzymes 2 843 39 38 - 96*
Thyroid function 4 2490 30 17 - 55
Drug monitoring 16 2787 46 5 - 83
*range includes data not used for calculation of overall mean


Anyone with a personal or professional interest in this needs to read the paper in its entirety, probably at least three times. The authors pack the discussion with good sense, including a paradigm for the classification of test appropriateness.

Clearly more work needs to be done. Radiology, pathology and screening tests were omitted. Results there may be different, because radiologists have worked hard to ensure the appropriateness of tests which may cause harm, as exemplified by the Ottawa ankle and knee tests in particular (see Bandolier 12 , 21 , 49 ). The reports almost all emanated from sick patients in teaching hospitals. What difference would there be for out-patients, or primary care, or non-teaching hospitals? So the best guess at the moment is that inappropriate use of laboratory tests is uncomfortably common, something that most laboratory personnel would recognise.

Financial implications

The implications for costs are significant, even though each intervention is in itself small - an example of the technology creep identified in the very first issue of Bandolier . In the UK we spend £40 billion or so on the NHS, and an estimate of costs of laboratories and laboratory testing is 4% of that, or £1.6 billion. Then it's a question of choosing a figure for inappropriate tests. If it is only 10%, that is £160 million wasted, but the figures in the Table suggest that may be conservative. The real answer is probably unknowable, but whatever it is, it is large.

And the UK has one of the best, and cheapest, laboratory services. In Europe and elsewhere laboratory costs are higher as a percentage of total healthcare costs, rising to perhaps twice or more in the USA.

Where do we go from here?

George Lundberg, editor of JAMA, pens an accompanying editorial [2] which emphasises the need for an outcomes research agenda. While he is right, the hope may be a pious one, because however important the issues raised by the review, the fact is that this is not seen as an important issue, neither academically in our Universities, nor managerially in the NHS. We start from a low base and there are no resources to do this work. Lundberg begins his editorial with words he used in 1975 which encapsulate the problem. Lets hope the same words don't have to be repeated in 2021.


  1. C van Walraven, CD Naylor. Do we know what inappropriate laboratory utilization is? A systematic review of laboratory clinical audits. JAMA 1998 280: 550-8.
  2. GD Lundberg. The need for an outcomes research agenda for clinical laboratory testing. JAMA 1998 280: 565-6.

previous or next story in this issue