Using clinical guidelines to assess the potential value of laboratory medicine in clinical decision-making

Introduction It is often quoted that 70% of clinical decisions are based on laboratory results, but the evidence to substantiate this claim is lacking. Since clinical guidelines aim to document best-practice decision making for specific disease conditions, inclusion of any laboratory test means that the best available evidence is recommending clinicians use it. Cardiovascular disease (CVD) is the world’s most common cause of mortality, so this study reviewed all CVD guidelines published by five national/international authorities to determine what proportion of them recommended laboratory testing. Materials and methods Five leading CVD guidelines were examined, namely the European Society of Cardiology (ESC), the UK National Institute for Health and Clinical Excellence (NICE), the American College of Cardiology (ACC), the Australian Heart Foundation (AHF) and the Cardiac Society of Australia and New Zealand (CSANZ). Results A total of 101 guidelines were reviewed. Of the 33 individual ESC guidelines relating to CVD, 24/33 made a direct reference to the use of clinical laboratory tests in either diagnosis or follow-up treatment. The same applied to 15/20 of NICE guidelines, 24/32 from the ACC and 15/16 from the AHF/CSANZ. Renal function and blood count testing were the most recommended (39 and 26 times), with lipid, troponin and natriuretic peptide measurement advocated 25, 19 and 19 times respectively. Conclusions This study has shown that laboratory testing is advocated by between 73% and 94% of individual CVD guideline recommendations from five national/international authorities. This provides an index to assess the potential value of laboratory medicine to healthcare.


Introduction
The contribution of laboratory medicine to patient diagnosis, management and follow-up has proven difficult to quantify with systematic evidence of improved patient outcomes scarce (1). The phrase that 'laboratory medicine influences 70% of clinical decisions', or similar, has been published many times but the evidence to substantiate this claim is lacking. An editorial in the Annals of Clinical Biochemistry stated that the 70% figure was first published in 1996 and was based on anecdotal evi-dence and unpublished studies (2). The editorial lists various examples of the use of this phrase, albeit with slight modifications, for example; Lord Carter's report on the UK Pathology service, the First Report of the UK House of Commons Select Committee on Health, and in the UK Department of Health report "Modernizing Pathology Services" (2). A related and also oft-quoted statistic is that 70% of the electronic patient record is composed of laboratory data, but the main limitation of this obser-vation is that the presence of laboratory results in a patient record does not necessarily equate to it being used in any clinical decision-making process (3). Over-requesting of testing is a common feature of many healthcare systems and panels of tests may include many analytes which are unrelated to the patient's clinical condition. A 2016 study of laboratory use by oncologists and cardiologists found that 75% of all their patients underwent laboratory testing, and that this testing led to a substantial clinical decision in 66% of the patients (4).
Over time the 70% claim has apparently gained legitimacy simply due to the number of times that it had been repeated. Partly to examine this claim more closely, the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) established a Task Force on the Impact of Laboratory Medicine on Clinical Management and Outcomes in 2012. Its purpose was twofold: first, to evaluate the available evidence supporting the impact of laboratory medicine in healthcare; and secondly to develop a study design methodology for new retrospective and prospective studies capable of generating evidence to determine the contribution made by laboratory medicine to healthcare.
The IFCC Task Force published a summary of its findings in 2015 which included as one of its suggestions that the contents of authoritative clinical guidelines could provide an objective means of assessing the role of laboratory medicine in the management of specific health conditions (1). Clinical guidelines are documents which aim to guide decisions regarding diagnosis, management and treatment in specific areas of health care. By using the best evidence available the assumption is that each individual guideline recommendation is how clinicians should be making their clinical decisions.
In respect to which health condition should have its guidelines examined, according to the World Health Organization (WHO) cardiovascular disease (CVD) is the world's most prevalent cause of mortality encompassing a large number of diseases including those of coronary heart disease, heart failure, rhythm and valvular abnormalities as well as cerebrovascular diseases (5). The report estimated that 17.9 million people worldwide die from CVDs which equates to 31% of all deaths and, of these, 85% were due to myocardial infarction or stroke (5). The British Heart Foundation report that CVD accounts for almost 170,000 deaths in the UK costing the National Health Service there £6.8 billion in 2012/2013 (6). In Europe, CVD is responsible for 3.9 million (45%) of all deaths annually (7). The American Heart Association has reported that coronary heart disease is the leading cause of death among Americans, accounting for 840,678 (30%) deaths in 2016 (8). According to the Australian Institute of Health and Welfare Alliance, CVD directly contributed to 45,400 deaths in 2015 which accounts for 29% of all deaths that year in Australia (9).
Accordingly, this study has chosen to examine some national and international cardiovascular clinical guidelines in detail from the countries just mentioned in order to determine what proportion of them recommended laboratory testing.

Materials and methods
We conducted a review of all the individual cardiovascular guidelines available on the websites of the UK National Institute for Health and Care Excellence (NICE), the European Society of Cardiology (ESC), the American College of Cardiology (ACC), and Cardiac Society of Australia and New Zealand (CSANZ) and Australian Heart Foundation (AHF) in their online versions which were current on the 23rd April 2020 (10)(11)(12)(13)(14). All the guidelines from the cardiology societies were examined while the NICE guidelines were limited only to those categorised by them as referring to 'cardiovascular conditions'. The AHF and the CSANZ guidelines were combined because a number of their guidelines were shared and, where they were identical, only one was included in the analysis statistics. These specific national/international guidelines were chosen because of their geographical spread and because they have previously been found to be rigorous with, for example, NICE and ESC guidance having the highest Appraisal of Guidelines for Research and Evaluation (AGREE) II scores for cardiovascular risk assessment in a systematic review (15).
Each individual guideline was read through in its entirety to determine if it was documented that laboratory involvement (defined as a test usually associated with one of the Laboratory Medicine disciplines, Histology or Genetics) was required for either an initial diagnosis or the ongoing management of care of the cardiovascular condition referred to in the document. Data was collected separately for tests which were used for initial diagnosis compared to those required for follow-up, and all forms of testing were included, not just analyses traditionally associated with CVD.
The proportion of guidelines containing laboratory medicine recommendations were determined for each national/international authority (AHF and CSANZ regarded as one) and for them combined. These proportions were calculated by simply dividing the number of individual guidelines indicating testing (if used for diagnosis and/or followup this was counted as one) by the total number of guidelines, whether for each national/international authority or for all authorities combined.
For solely the purpose of determining which types of tests were mentioned most frequently by the guidelines, some related tests were categorised together. Thus, renal function related testing included 'renal function tests', 'creatinine' and 'eGFR' in the text; natriuretic peptides included 'natriuretic peptides', 'BNP' and 'NT-proBNP'; 'calcium', if mentioned separately from other electrolytes, was included in an "electrolytes group" and FBC and CBC were regarded as synonyms. Genetic and histology tests were not subcategorised any further.
As this is a review of existing guidelines, no Institutional Review Board approval was required. Table 1 shows the 101 guidelines related to CVD available from NICE, ESC, ACC and the AHF/CSANZ and summarises the number and/or percentage indicating pathology testing for either initial diagnosis, follow up pathology testing or a combination of both. Taken together, in vitro diagnostic testing was required for diagnosing 64% of the conditions alluded to in the cardiovascular guidelines, with a similar proportion (62%) advocating testing for the continued management of such patients. As tests were sometimes recommended for both the diagnosis and follow-up of same condition, it means some form of testing was, on average, required by 77% of individual guidelines. Twenty seven different test categories were explicitly mentioned 276 times within the 78 guidelines requiring testing (Tables 2-5). The commonest were renal function related biochemistry (mentioned 39 times), complete (or 'full') blood counts (26 times), lipids (25 times) and blood/plasma glucose (20 times). More specific to cardiovascular disease were the troponins (19 times) and natriuretic peptides (19 times).        There are both limitations and strengths to using guidelines as a measure of laboratory medicine's contribution to healthcare. One potential limitation is that clinical guidelines are not necessarily followed by all physicians or other healthcare staff. A clinician's own pathway for patients may therefore involve more or less testing than is being recommended. However, the main advantage of our method is that at least the guidelines define an evidence-based recommendation for best practice in each specific clinical scenario. The guidelines also tend to be specific in the test or tests that are recommended, which is advantageous in two ways. First, it helps ensure that the contribution of laboratory medicine to healthcare is not exaggerated by over-requesting -as could be the case if health records were solely being examined -and, secondly, that a clinical decision relevant to the health condition is intended to be taken on the basis of the result.

Results
It should be noted that the guidelines examined in this study were produced in relatively affluent countries with developed healthcare systems and so the use of laboratory medicine testing advocated in other, less wealthy, countries may well differ.
Nonetheless, this does not preclude the same methodology as used here being applied to any alternative CVD guidelines from other countries.
Beyond CVD, there is also the possibility that the same objective approach to guidelines could be applied to determine the contribution of laboratory medicine in other, less prevalent, disease groups. Combining the findings from these different disease groups could potentially provide a more accurate overall assessment of the significance of laboratory medicine to healthcare in general.
In summary, this study has found a requirement for the use of laboratory testing in 73% to 94% of cardiovascular guidelines produced by five different organisations. It does not provide a direct link to improved patient outcomes but does provide an index to assess the potential value of laboratory medicine to healthcare which can complement other metrics.