Article type
Year
Abstract
Background: Evidence of diagnostic test accuracy can be used to infer the clinical effectiveness of a test if the effectiveness of treatment for the target condition has already been established. Systematic reviews that compare the accuracy of a new diagnostic test strategy with the existing test strategy provide the most relevant synthesis of this evidence for clinical practice. The Australian Medical Services Advisory Committee (MSAC) seeks this information to determine the clinical value of diagnostic tests being considered for public funding.
Objectives: We evaluated MSAC assessment reports of diagnostic tests in order to identify the major problems encountered when using evidence about test accuracy to infer the clinical value of a test.
Methods: All assessment reports of diagnostic tests published by MSAC during 2002-2003 were studied. Reports were assessed to determine whether decisions about the clinical value of the test were based on evidence from studies evaluating the impact of the test on patient outcomes (direct evidence) or evidence of the effectiveness of treatment linked with evidence of test accuracy (linked evidence). Linked evidence assessments were classified according to whether the evidence was based on one or more studies that: 1. evaluated a direct comparison of the index test strategy versus the existing test strategy; 2. were applicable to the population and clinical setting of interest; and 3. were designed to avoid potential sources of bias.
Results: Assessment reports on 14 diagnostic tests for 24 clinical indications were published during the study period. Evidence of test effectiveness from controlled trials was available for one test indication. Evidence of effective treatment was established for 20 (87%)of the remaining 23 test indications to justify using evidence of test accuracy to infer the clinical value of the test. Studies that compared the accuracy of the index test strategy head-to-head with the testing strategy that it was most likely to replace were available for 16 (80%) test indications. Studies that evaluated the accuracy of the index test in a patient population and clinical setting that fully reflected intended clinical practice were available for 8 (40%) test indications. Quality assessment of the included studies of test accuracy either demonstrated design deficiencies that are potential sources of bias, or was limited due to a lack of reported information. A decision regarding public funding was not possible due to insufficient clinical evidence for 13 of the 24 test indications.
Conclusions: The evaluation of new diagnostic tests is difficult. There is typically a lack of direct evidence of the impact of the test on patient outcomes, and evidence of test accuracy is often drawn from limited study populations. Those planning Cochrane reviews of diagnostic test accuracy should be mindful that synthesis of evidence of comparative test accuracy in the population and clinical setting in which the test is most likely to be used will provide the most useful research evidence for clinical practice.
Objectives: We evaluated MSAC assessment reports of diagnostic tests in order to identify the major problems encountered when using evidence about test accuracy to infer the clinical value of a test.
Methods: All assessment reports of diagnostic tests published by MSAC during 2002-2003 were studied. Reports were assessed to determine whether decisions about the clinical value of the test were based on evidence from studies evaluating the impact of the test on patient outcomes (direct evidence) or evidence of the effectiveness of treatment linked with evidence of test accuracy (linked evidence). Linked evidence assessments were classified according to whether the evidence was based on one or more studies that: 1. evaluated a direct comparison of the index test strategy versus the existing test strategy; 2. were applicable to the population and clinical setting of interest; and 3. were designed to avoid potential sources of bias.
Results: Assessment reports on 14 diagnostic tests for 24 clinical indications were published during the study period. Evidence of test effectiveness from controlled trials was available for one test indication. Evidence of effective treatment was established for 20 (87%)of the remaining 23 test indications to justify using evidence of test accuracy to infer the clinical value of the test. Studies that compared the accuracy of the index test strategy head-to-head with the testing strategy that it was most likely to replace were available for 16 (80%) test indications. Studies that evaluated the accuracy of the index test in a patient population and clinical setting that fully reflected intended clinical practice were available for 8 (40%) test indications. Quality assessment of the included studies of test accuracy either demonstrated design deficiencies that are potential sources of bias, or was limited due to a lack of reported information. A decision regarding public funding was not possible due to insufficient clinical evidence for 13 of the 24 test indications.
Conclusions: The evaluation of new diagnostic tests is difficult. There is typically a lack of direct evidence of the impact of the test on patient outcomes, and evidence of test accuracy is often drawn from limited study populations. Those planning Cochrane reviews of diagnostic test accuracy should be mindful that synthesis of evidence of comparative test accuracy in the population and clinical setting in which the test is most likely to be used will provide the most useful research evidence for clinical practice.