Decision making about healthcare-related tests and diagnostic test strategies. Paper 4: International guidelines show variability in their approaches

J Clin Epidemiol. 2017 Dec:92:38-46. doi: 10.1016/j.jclinepi.2017.09.004. Epub 2017 Sep 14.

Abstract

Objectives: The objective of the study was to describe and compare current practices in developing guidelines about the use of healthcare-related tests and diagnostic strategies (HCTDS).

Study design and setting: We sampled 37 public health and clinical practice guidelines about HCTDS from various sources without language restrictions.

Results: Detailed descriptions of the systems used to assess the quality of evidence and develop recommendations were challenging to find within guidelines. We observed much variability among and within organizations with respect to how they develop recommendations about HCTDS. Twenty-four percent of the guidelines did not consider health benefits and harms but based decisions solely on test accuracy. We did not identify guidelines that described the main potential care pathways involving tests for a healthcare problem. In addition, we did not identify guidelines that systematically assessed, described, and referenced the evidence that linked test accuracy and patient-important outcomes.

Conclusion: There is considerable variability among the processes used and factors considered in developing recommendations about the use of tests. This variability may be the cause for the disagreement we observed in recommendations about testing for the same condition.

Keywords: Clinical practice guidelines; Decision-making; Diagnosis; Evidence-based medicine; Guidelines; Tests.

Publication types

  • Comparative Study
  • Review

MeSH terms

  • Decision Making
  • Diagnostic Tests, Routine / methods
  • Diagnostic Tests, Routine / standards*
  • Evidence-Based Medicine
  • Female
  • Health Services / standards*
  • Humans
  • Internationality
  • Male
  • Practice Guidelines as Topic*
  • Public Health / standards
  • Quality of Health Care*