A few thoughts on reflective testing

The term ‘re£ective testing’ was proposed by Paterson and Paterson in 2004 to describe the process by which a Clinical Biochemist in the laboratory adds additional tests to a received request, based on biochemical results, available clinical information, demographic data, clinical experience or a combination of these factors. In their original paper, Paterson and Paterson distinguished re£ective testing from re£ex testing, the process by which additional tests are automatically added to a received sample, according to a set of logical rules. While the term re£ective testing is new, it describes a well-established aspect of the practice of clinical biochemistry in the UK and many other countries. However, as Paterson and Paterson pointed out, the value of the process had not been previously assessed. Intuitively, many would take the view that re£ective testing is a valuable part of the practice of clinical biochemistry and an important way for the laboratory to provide added value to clinicians. The Clinical Biochemist reviews many more laboratory results than other clinicians, and is likely to have a much clearer understanding of the positive and negative predictive value of laboratory tests in a given situation. It may be apparent to the Clinical Biochemist that a requesting clinician has not asked for the best test or combination of tests to answer a speci¢c question, and that the addition of further tests may aid the diagnostic process. In addition, an abnormal result may be identi¢ed on analysis, about which further important clinical information could be generated by performing additional biochemical tests on the received sample. The original paper by Paterson and Paterson assessed the value of adding vitamin D measurement when other results suggested the possibility of vitamin D de¢ciency, or total iron binding capacity and transferrin saturation when the original results suggested the possibility of haemochromatosis. They showed that this pattern of re£ective testing resulted in the diagnosis of vitamin D de¢ciency or haemochromatosis in a signi¢cant number of cases. Further consideration, however, suggests aspects of re£ective testing that may cause concern. Inevitably, the pattern and extent of re£ective testing will vary between Clinical Biochemists, depending on factors such as experience, aptitude, and the time and resources available. The availability of information about the reason for requests is potentially an important in£uence on re£ective testing. So long as individual practice falls within the range considered reasonable by the clinical biochemistry community, inter-individual variation should not give cause for too much concern; individual clinical practice and outcomes varyacross all clinical disciplines.There is, however, little information available about the range of re£ective testing practised in diierent laboratories, making it di⁄cult to characterize the range of normal clinical practice. It would be helpful to benchmark re£ective testing to allow individual practitioners to assess how they compare with their peers. In addition, if re£ective testing is a signi¢cant part of practice, it should be the subject of external quality assurance, as increasingly occurs for the practice of writing interpretive comments on reports. There are also concerns about the ethics of re£ective testing. Some have questioned whether it is ethical to add additional tests to a sample without the patient’s consent. In the UK, patients are assumed to give implied consent for the performance of appropriate tests on any sample that are sent to the laboratory from them. In this context, re£ective testing should not cause any ethical di⁄culties, although exceptions include any test for which it is accepted that speci¢c written consent is required from patients (such as testing of HIV status or genetic tests). For such tests, re£ective testing should never be carried out. In this issue of the Annals, two papers address the acceptability of re£ective testing, one looking at the views of requesting clinicians and the other at the opinions of patients. Overall, the results suggest that both clinicians and patients view the concept favourably, although, notably, a minority of both do not always support re£ective testing, preferring to be speci¢cally consulted before additional tests are performed. In the case of clinicians, support for re£ective testing varies signi¢cantly according to the type of test.While