Table 5.

Case Study about Brain Natriuretic Peptides (BNP) in the Diagnosis of Heart Failure*

Results from the SGR: The recommendations in the source guidelines on the use of BNP tests in patients suspected of heart failure showed a major inconsistency (type A): The test was treated in 7/16 guidelines; recommendations differed completely in content (2/7 'not recommended', 3/7 'recommended under certain circumstances', 1/7 'recommended in every case', 1/7 'recommended ruling out CHF before an echocardiogram'), and in grading.
Further research: We conducted a systematic review of the diagnostic accuracy of this test in primary care. We did not find strong evidence in favour of its use in this setting (most studies were undertaken after referral which implies a potential spectrum bias [64,65], a clear cut-off was not defined, study results were inconsistent, in particular for concomitant diseases and medication) [66-69], supported by two subsequently published systematic reviews [70,71]. The consensus panel agreed not to recommend the test in our target guideline.
Discussion: Inconsistent recommendations in source guidelines may be due to (i) methodological shortcomings (recommendations were not setting-specific in 6/7 guidelines that addressed both primary and secondary care; literature searches were stated to be comprehensive in only 2/7 guidelines), or to (ii) potential conflicts of interests (4/7 guidelines did not provide any financial disclosures for the authors). Moreover, (iii) contextual influences may have guided the recommendations, such as availability (1/7 guidelines recommended the test, as echocardiograms are not widely available), access (BNP tests have market approval throughout Europe but costs are reimbursed by public funding – e.g. in the U.K. – or privately, e.g. in Germany), or intended resource allocation (1/7 guidelines restricted access to the more expensive echocardiogram, as CHF had to be ruled out by BNP and/or electrocardiogram before the referral). Last but not least the BNP test is an emerging technology where typically only limited information of its benefit is available, and initial studies show predominantly optimistic results [72]. Variations in the adoption of a new (healthcare) technology from one country to another, and also from one physician to another are shown to be large, and 'inextricably interwoven' with culture [72]. We know that more highly trained and committed physicians in the community tend to be 'early adopters' [72], and that specialists are less conservative than generalists [72,73]. It might be that the selection of guideline developing groups and their attitudes influenced the decision to include the BNP test.

*This test was developed to distinguish between heart failure and other conditions that show typical symptoms and signs in a patient. In this paper we use BNP as a synonym for itself and others such as NT-proBNP.