Systematic Review
Decision analysis to complete diagnostic research by closing the gap between test characteristics and cost-effectiveness

https://doi.org/10.1016/j.jclinepi.2009.01.008Get rights and content

Abstract

Objective

The lack of a standard methodology in diagnostic research impedes adequate evaluation before implementation of constantly developing diagnostic techniques. We discuss the methodology of diagnostic research and underscore the relevance of decision analysis in the process of evaluation of diagnostic tests.

Study Design and Setting

Overview and conceptual discussion.

Results

Diagnostic research requires a stepwise approach comprising assessment of test characteristics followed by evaluation of added value, clinical outcome, and cost-effectiveness. These multiple goals are generally incompatible with a randomized design. Decision-analytic models provide an important alternative through integration of the best available evidence. Thus, critical assessment of clinical value and efficient use of resources can be achieved.

Conclusion

Decision-analytic models should be considered part of the standard methodology in diagnostic research. They can serve as a valid alternative to diagnostic randomized clinical trials (RCTs).

Introduction

What is new?

  • In diagnostic research, decision models form a valuable alternative when randomized clinical trials (RCTs) are infeasible by integrating the best available evidence.

  • Decision-analytic models should be increasingly used to evaluate clinical outcome including cost-effectiveness of fast developing diagnostic techniques.

To date, consensus on the methodology to evaluate new diagnostic tests is lacking [1]. Moreover, with rapid technical advances, especially in the field of imaging, diagnostic techniques undergoing evaluation may be already outdated before diagnostic and clinical values are established [2], [3].

Similar to therapeutic research, a hierarchy can be discerned within diagnostic research [1], [2], [4], [5], [6], [7]. The first step is the assessment of test characteristics of a new test. The next step is the evaluation of its added value [8]. The third step is the assessment of the effect on clinical outcome, and the final step comprises a cost-effectiveness analysis [7]. For evaluation of clinical outcome and cost-effectiveness of tests, randomized clinical trials (RCTs) are often not feasible.

We will shortly review each step of diagnostic research and discuss decision analysis as a useful alternative methodology for critical assessment of clinical value and efficient use of resources.

Section snippets

The hierarchy in diagnostic research

We will illustrate the different phases of diagnostic research following the example of carotid artery stenosis, a well-known risk factor for stroke. Carotid endarterectomy reduces this risk in selected patients [9], [10]. The standard technique for grading stenosis used to be digital subtraction angiography (DSA), an invasive imaging modality. Later on, magnetic resonance angiography (MRA) and duplex ultrasound (DUS) were introduced as noninvasive alternatives [11].

Evaluation of decision models

Examples of validated decision-analytic models showed that the results of decision modeling were consistent with observed data from cohort follow-up studies [27], [28], [29]. For two of these studies, the input parameters of the Markov model originated from a cohort different from that with which the results were compared [27], [28]. Another study used the short-term follow-up data of an observational study cohort for a Markov model. The expected value of the model was compared with the

Conclusions

Evaluation of a diagnostic strategy requires a phased approach. For assessment of outcome in addition to test characteristics frequently, an RCT is neither feasible nor warranted. Decision-analytic models can integrate the best available evidence, including economic data, and should be part of the standard methodology in diagnostic research.

Acknowledgments

The funding for the study was provided by Netherlands Organization for Scientific Research (NWO), the Netherlands Organization for Health Research and Development (ZonMw), 945-04-310.

Conflicts of interests for all authors: none declared.

References (31)

  • D.L. Sackett et al.

    The architecture of diagnostic research

    BMJ

    (2002)
  • K.G. Moons et al.

    Test research versus diagnostic research

    Clin Chem

    (2004)
  • European Carotid Surgery Trialists' Collaborative Group. Randomised trial of endarterectomy for recently symptomatic carotid stenosis: final results of the MRC European Carotid Surgery Trial (ECST)

    Lancet

    (1998)
  • H.J. Barnett et al.

    Benefit of carotid endarterectomy in patients with symptomatic moderate or severe stenosis. North American Symptomatic Carotid Endarterectomy Trial Collaborators

    N Engl J Med

    (1998)
  • P.J. Nederkoorn et al.

    Duplex ultrasound and magnetic resonance angiography compared with digital subtraction angiography in carotid artery stenosis: a systematic review

    Stroke

    (2003)
  • Cited by (0)

    View full text