Letters
Emergency medicine Medical practices
Volume 200 - Issue 1

Changing trends in venous thromboembolism-related imaging in Western Australian teaching hospitals, 2002–2010

Author:  David Mountain

Med J Aust 2014; 200 (1): 26-27. || doi: 10.5694/mja13.10292
Published online: 20 January 2014
To the Editor: The article by Segard and Macdonald is interesting but has significant methodological weaknesses and reports data that do not support all of their conclusions.1 Methodologically, retrospective dataset reviews with little demographic or morbidity data to avoid confounding (it is unclear why these were not available) are problematic. Additionally, death certificates were used as a surrogate for improved outcomes from additional diagnosis. Segard and Macdonald ...
Download PDF (232 KB)
Share: Share on Blue Sky Share on Facebook Share on LinkedIn Share by email

To the Editor: The article by Segard and Macdonald is interesting but has significant methodological weaknesses and reports data that do not support all of their conclusions.1

Methodologically, retrospective dataset reviews with little demographic or morbidity data to avoid confounding (it is unclear why these were not available) are problematic. Additionally, death certificates were used as a surrogate for improved outcomes from additional diagnosis.

Segard and Macdonald used census data to estimate growth of the at-risk population (18.9%) over the study period. However, the almost 70% growth in emergency department (ED) attendances2,3 is arguably as relevant, as most venous thromboembolism (VTE) tests are ordered by the ED (own unpublished data). During this same decade Australian patients, increasingly, were older, more obese and presented with more comorbidities — all of which increase pulmonary embolism (PE) risk.4,5

The data presented contradict the authors’ clear inference that increasing D-dimer referrals drove VTE testing. D-dimer referrals increased from 2002 to 2005 and then plateaued (Box), while computed tomography pulmonary angiography (CTPA) requests only started increasing from 2004, and plateaued from 2007.

An alternative hypothesis is that new tests that were perceived as useful increased investigation rates. Additionally, D-dimer testing probably reduced CTPA testing in low-risk patients, improving case selection and potentially reducing false-positive rates.

Using death certificates to estimate PE mortality is problematic, as up to 70% of PEs are misdiagnosed and PE is rarely certified without a premortem diagnosis.6 Per annum, treatment of 10 to 25 patients with stable PE may have been needed to save one life,7 so significant death rate changes were unlikely to be detected.

This study gives useful and important data, but does not give strong evidence that D-dimer testing or CTPA are unhelpful or harmful. The methodological issues, relatively weak surrogate outcome and temporally disconnected increases in test rates necessitate great care in making significant inferences from this study.

Graph of data showing test usage and imaging rates, 2002–2010*

VQS = ventilation–perfusion lung scintigraphy. CTPA = computed tomography pulmonary angiography. PE = pulmonary embolism. VTE = venous thromboembolism. * Plotted from data in Box of Segard and Macdonald.1

Author

  • David Mountain1
  • 1. Academic Department of Emergency Medicine, University of Western Australia, Perth, WA.

Correspondence: david.mountain@health.wa.gov.au

Competing interests

No relevant disclosures.

References

View this article on Wiley Online Library

More like this

View more