More on ‘universal’versus‘selected’ screening for thrombophilia: the hidden costs of false-positive diagnosis
Article first published online: 26 MAY 2006
British Journal of Haematology
Volume 134, Issue 2, pages 239–240, July 2006
How to Cite
Favaloro, E. J. (2006), More on ‘universal’versus‘selected’ screening for thrombophilia: the hidden costs of false-positive diagnosis. British Journal of Haematology, 134: 239–240. doi: 10.1111/j.1365-2141.2006.06138.x
- Issue published online: 16 JUN 2006
- Article first published online: 26 MAY 2006
- laboratory testing;
- false positives
Last year, I read with some interest the report by Wu et al (2005) who performed a meta-analysis and cost-effectiveness analysis for screening for thrombophilia in some selected high-risk situations. They concluded that ‘selective thrombophilia screening based on previous personal and/or family history of venous thromboembolism was more cost-effective than universal screening in all the patient groups evaluated’. Although I felt that some study limitations were evident at the time, I tended to agree with their conclusion and hence declined to bring these limitations to notice. The issue was recently raised again by Keeling (2006) who, quite logically, noted that, in the case of pregnancy, a strong case for ‘universal screening’ could be argued on the basis of cost in ‘selected’ cases.
In each of these reports, the argument for ‘universal’ screening is in fact ‘selected universal’ screening for given high-risk situations. The distinction is important because inexperienced clinicians may take these arguments for ‘universal’ screening as encompassing low risk (or even no-risk) situations. One significant limitation in the study of Wu et al (2005) is the case of false-positive diagnosis, which appeared to be lacking in their analysis. Their study was, as far as I could tell, based on the differential costs between managing true cases of laboratory-identified thrombophilia (i.e. factor V Leiden, prothrombin G20210A, antithrombin deficiencies, protein C deficiencies, protein S deficiencies, lupus anticoagulants and anti-cardiolipin antibodies and thus preventing future thrombotic events by clinical interventions) versus the costs of not managing these true cases of laboratory-identified thrombophilia (i.e. not identified, because of limited test screening). Accordingly, their study was limited to the premise that the clinical ordering and laboratory screening would only identify true positive cases. There was no assessment or costing for situations where the clinical ordering and laboratory screening identified false-positive cases.
In a recent simplified analysis, we reported that, in a best case scenario for screening patients with unselected VTE, for every true positive case of protein C deficiency identified by laboratory screening, the laboratory will also identify one false-positive case of protein C deficiency (Favaloro et al, 2005). The situation is similar for protein S deficiency. In addition, in a best case scenario for ‘universal’ screening of patients without selection, for every true positive case of protein C (or protein S) deficiency identified by laboratory screening, the laboratory will also identify up to 10 false-positive cases of protein C (or protein S) deficiency. These are ‘best-case’ scenarios. In worst case scenarios, the risk of ‘false positive:true positive’ identification for protein C or S deficiencies is around 4:1 and 40:1 respectively for screening of patients with unselected VTE and ‘universal’ screening of patients without selection. I do not know the cost of treating these patients clinically as ‘true’ thrombophilias, but I do know that such potential false diagnoses do carry considerable cost, both financially and psychologically (Walker et al, 2001; Machin, 2003; Favaloro et al, 2005; Favaloro, 2005). I am also cognisant of the movement towards a more universal (‘non-selected’) thrombophilia screening model in our geographical region (Favaloro et al, 2002), where the detection rate for factor V Leiden was recorded at a high of 25% of all laboratory factor V Leiden tested cases in 1996, but gradually fell to <15% in 2001. Current figures suggest continuing decline consistent with a move towards less selective thrombophilia screening. When the figure reaches c. 5%, we will be, in effect, testing the general population.
I would invite Wu and colleagues (or others) to attempt some additional appropriate costings based on our estimates (or their own estimates as appropriate) of false-positive identification. In the interim, I include myself on the list of those who firmly believe in selective thrombophilia screening (see also Walker et al, 2001; Machin, 2003) or perhaps in ‘universal screening’ in very selected clinical situations. Another alternative would be to restrict the ordering of such thrombophilia tests to clinical specialists who understand the limitations of laboratory screening of thrombophilia and the laboratory tests used in such thrombophilia screening. Clinicians who order such tests should also be able to understand, manage and counsel patients who are positively identified with thrombophilia markers, both those true- and false-positive cases.
- 2005) Diagnostic issues in Thrombophilia: a laboratory scientist's view. Seminars in Thrombosis and Hemostasis, 31, 11–16. (
- 2002) A nine-year retrospective assessment of laboratory testing for activated protein C resistance: evolution of a novel approach to thrombophilia investigations. Pathology, 34, 348–355. , , & (
- 2005) Laboratory identification of familial thrombophilia: do the pitfalls exceed the benefits? A reassessment of ABO-blood group, gender, age and other laboratory parameters on the potential influence on a diagnosis of protein C, protein S and antithrombin deficiency and the potential high risk of a false positive diagnosis. Laboratory Haematology, 11, 174–184. , , , & (
- 2006) Universal or selected screening for thrombophilia. British Journal of Haematology, 133, 106–107. (
- 2003) Pros and cons of thrombophilia testing: cons. Journal of Thrombosis and Haemostasis, 1, 412–413. (
- 2001) Investigation and management of heritable thrombophilia. British Journal of Haematology, 114, 512–528. , & (
- 2005) Screening for thrombophilia in high-risk situations: a meta-analysis and cost-effectiveness analysis. British Journal of Haematology, 131, 80–90. , , , , , , , , , & (