The ‘70% claim’: what is the evidence base?

Those involved in the practice of laboratory medicine are convinced of its value to effective and safe patient care, and usually have much anecdotal evidence of instances in which appropriate testing or the provision of interpretative comments ¹ have improved care. However, systematic evidence of the contribution of laboratory medicine to the overall process of diagnosis and management of patients is much harder to obtain – understandably so, in view of the multitude of factors that are involved in reaching a diagnosis and planning treatment for an individual.

Many articles seeking to promote the value of laboratory medicine make use of what has become known as the ‘70% claim’. This presents itself in various forms, most commonly that ‘Laboratory medicine data influences 70% of clinical decisions’, ² or minor variations around this figure (60–80%). The percentage of the electronic medical record that consists of laboratory medicine data is also often adduced in support of the importance of the contribution made by pathologists and clinical scientists to patient care.

The earliest published reference to the claim, and the one most frequently cited, is from 1996, by Forsman at the Mayo Clinic in the USA. ³ The author states: ‘We know that, although the laboratory represents a small percentage of medical center costs, it leverages 60–70% of all critical decisions, e.g. admission, discharge and therapy’. The evidence for this is not provided in the paper, but in an interview with Clinical Laboratory News in 2004, Forsman is quoted as saying: ‘… the statements that have been made take two forms. The first is that the laboratory represents 5% of a health system's costs, yet it affects 95% of the remaining costs. The second statement is that the laboratory contributes 80% of the objective data in the clinical record and influences 60%–70% of critical decision making. Unpublished sources for these statements include conversations with Dr Peter Dysert at Baylor University in Houston, Texas and Dr Michael Becich at the University of Pittsburgh’. ⁴

Becich published an article in 2000 which begins: ‘The involvement of laboratory tests (clinical pathology) or tissue samples (anatomic pathology) in major decisions affecting patients is estimated at 50% to 70%. Initially, these “high” numbers were thought to reflect the importance of pathology tests in larger referral hospitals only, not in the “real medicine” practices of community settings and outpatient clinics. However, in the outpatient setting, pathology tests are involved in more than 80% of the decisions that doctors make that change the lives of patients (Critchfield GC, personal communication)’. ⁵ Becich goes on to say that ‘From anecdotal studies in pathology informatics, we have therefore coined the “70–70” rule. Seventy percent of all the important decisions affecting a patient's life involve a laboratory or pathology test. In addition, pathology data represent an average of 70% of documents currently residing in electronic repositories today, an estimate well documented in electronic medical record literature’. ⁵

It seems likely that the data on which the 70% claim is based represent unpublished studies and anecdotal observations, and cannot be objectively verified at this stage. The assertion in its original form (laboratory medicine data influencing 70% of critical [major] decisions) sounds plausible, but the evidence underpinning it is not available.

Various other forms of the claim have been published, often omitting important qualifying words. Writing in CAP Today in 2000, Foubister stated ‘It's estimated that up to 85% of physicians’ decisions regarding a patient's diagnosis and treatment are based on laboratory test results’. ⁶ No source is given, and note the ‘up to’ 85% and the omission of ‘critical’ or ‘major’ before ‘decisions’ – implying that almost all physicians’ decisions regarding diagnosis or treatment are based on laboratory data, which sounds very questionable. A similarly broad assertion is found in the first report on UK Pathology Services produced by Lord Carter and colleagues: ‘It is estimated that 70–80% of all health care decisions affecting diagnosis and treatment involve a pathology investigation’. ⁷ Again, no source is cited and the words ‘critical’ and ‘major’ are absent.

Much depends on how ‘decisions’ are defined, as Dr Rick Jones has pointed out: ‘We need a much more rigorous approach to understanding the types of decisions made and the impact of different tests in each context. Then we might be able to work up strategies to ensure that where decisions are dependent on results we can provide them in an optimal way, and where they make no difference we can stop wasting money’ (2011, personal communication by email).

An even broader version of the claim occurs in statements by the UK Department of Health in relation to the UK National Health Service (NHS). The First Report of the House of Commons Select Committee on Health (published on 1 May 2002) states that ‘Up to 70% of all diagnoses in NHS patients depend on laboratory tests, hence NHS pathology services are critical for the day-to-day evidence-based care of patients’. ⁸ While the latter part of this statement is undoubtedly true, the generalization of the original 70% claim to diagnoses, even with the vague ‘up to 70%’, appears completely unsupported by anything in the laboratory medicine literature. Considering the whole range of possible medical diagnoses in the NHS, including minor ailments in primary care settings and most of the mental health field, it seems extremely unlikely, and probably represents an unintentional misquotation of the studies cited above. The statement is repeated as ‘60–70% of NHS patients’ diagnoses depend on laboratory tests’ in the UK Department of Health's 2004 paper entitled ‘Modernising Pathology Services’. ⁹

The parallel claim, that 70–80% of the information in the electronic medical record consists of laboratory data, is better supported, though published percentages do vary widely, and the amount of data in the record is a poor proxy for the importance of that data in the care of an individual. ‘A pathological diagnosis of malignancy in a biopsy report takes up very little space in the electronic record but is likely to have a profound effect on treatment’ (Bertholf R, 2011, personal communication by email).

The original data from the Mayo Clinic stated that the relative amount of data on the Mayo Electronic Result Enquiry System was: pathology: 94%, radiology: 3%, patient data: 1%, electrocardiogram: 1% and surgery: 1%. ¹⁰ This was published in 2000, before the advent of digital radiology, which would obviously radically transform the relative amounts of space in the electronic record. A 1984 paper on the computerized intensive care unit found that combined laboratory data (clinical, microbiology and blood gas) made up 38% to 41% of total patient data reviewed, and occupied 16.3% of the database. ¹¹ Schamberg has estimated that laboratory test data information totalled about 82% of the data bits used to store the patient's full health record in the electronic medical record at Aurora Health Systems of Milwaukee, USA. ¹²

The value of laboratory medicine in patient care is unquestioned. A detailed study by the Lewin Group for the Centers for Disease Control and Prevention in the USA concluded ‘Laboratory medicine is an essential element of the health care system. It is integral to many clinical decisions, providing physicians, nurses, and other health care providers with often pivotal information for the prevention, diagnosis, treatment and management of disease’. ¹³ However, demonstration of this value requires more sophisticated measurement than the high-level ‘blanket’ claims discussed here. Until a better evidence base is available, use of the various ‘70% claims’ should be resisted in favour of more specific and evidence-based indices of added value. Further well-designed studies in this area are urgently needed, especially in view of the radical reforms of health care taking place in many countries and the emphasis on added value.

DECLARATIONS

Competing interests: None.

Funding: None.

Ethical approval: Not applicable.

Guarantor: MJH.

Contributorship: MJH researched the literature, and conceived and wrote the manuscript with assistance as acknowledged below.

Acknowledgements: I am grateful to members of the ACB's online mailbase community who provided information and references, including Dr Roger Bertholf, Dr Mike Colley, Mr Mike Collins, Dr Paul Collinson, Dr Danielle Freedman, Prof. Rita Horvath, Mr Dennis Huckerby, Dr Patti Jones, Dr Rick Jones, Dr William Marshall, Dr Sam Vasikaran and Dr Dennis Wright.