Article Text


Assessing the appropriateness of prevention and management of venous thromboembolism in Australia: a cross-sectional study
  1. Peter D Hibbert1,2,
  2. Natalie A Hannaford2,3,
  3. Tamara D Hooper1,2,
  4. Diane M Hindmarsh1,
  5. Jeffrey Braithwaite1,
  6. Shanthi A Ramanathan2,4,
  7. Nicholas Wickham5,
  8. William B Runciman1,2,3
  1. 1Australian Institute of Health Innovation, Macquarie University, Sydney, New South Wales, Australia
  2. 2Centre for Population Health Research, University of South Australia, Adelaide, South Australia, Australia
  3. 3Australian Patient Safety Foundation, Adelaide, South Australia, Australia
  4. 4Hunter Valley Research Foundation, Newcastle, New South Wales, Australia
  5. 5Adelaide Cancer Centre, Kurralta Park, South Australia, Australia
  1. Correspondence to Peter D Hibbert; Peter.hibbert{at}


Objectives The prevention and management of venous thromboembolism (VTE) is often at variance with guidelines. The CareTrack Australia (CTA) study reported that appropriate care (in line with evidence-based or consensus-based guidelines) is being provided for VTE at just over half of eligible encounters. The aim of this paper is to present and discuss the detailed CTA findings for VTE as a baseline for compliance with guidelines at a population level.

Setting The setting was 27 hospitals in 2 states of Australia.

Participants A sample of participants designed to be representative of the Australian population was recruited. Participants who had been admitted overnight during 2009 and/or 2010 were eligible. Of the 1154 CTA participants, 481(42%) were admitted overnight to hospital at least once, comprising 751 admissions. There were 279 females (58%), and the mean age was 64 years.

Primary and secondary outcome measures The primary measure was compliance with indicators of appropriate care for VTE. The indicators were extracted from Australian VTE clinical practice guidelines and ratified by experts. Participants’ medical records from 2009 to 2010 were analysed for compliance with 38 VTE indicators.

Results Of the 35 145 CTA encounters, 1078 (3%) were eligible for scoring against VTE indicators. There were 2–84 eligible encounters per indicator at 27 hospitals. Overall compliance with indicators for VTE was 51%, and ranged from 34% to 64% for aggregated sets of indicators.

Conclusions The prevention and management of VTE was appropriate for only half of the at-risk patients in our sample; this provides a baseline for tracking progress nationally. There is a need for national and, ideally, international agreement on clinical standards, indicators and tools to guide, document and monitor care for VTE, and for measures to increase their uptake, particularly where deficiencies have been identified.

  • GENERAL MEDICINE (see Internal Medicine)

This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See:

Statistics from

Strengths and limitations of this study

  • The study is designed to be representative of the Australian population rather than a convenience-based or purposive-based sample.

  • The review of medical records, while costly and difficult, allowed compliance to be measured in a real-world setting.

  • Numbers of participants and/or eligible encounters are low for some indicators.

  • There was a high rate of attrition of potential participants and several sources of possible bias. However, weighting using two methods and five different options made no significant difference to the compliance percentage.


Each year in Australia about 1 in every 1000 people develop a first episode of venous thromboembolism (VTE), manifesting as deep venous thrombosis (DVT) and/or pulmonary embolism (PE).1 ,2 This amounts to about 20 000 cases, of which 80% occur during or soon after an admission to hospital.1 ,2 Including loss of productivity, total costs amount to well over $A1 billion per year.3

There is evidence that the appropriate use of pharmacological and mechanical prophylaxis in orthopaedic, general surgical and medical patients can reduce the incidence of VTE,4–6 although a recent paper has questioned use of pharmacological prophylaxis in lower risk medical patients.7 Clinical practice guidelines (CPGs) have been developed, in Australia8 ,9 and elsewhere,10 ,11 to prevent VTE and to standardise the management of DVT and PE. Several initiatives have been undertaken to promote and facilitate their uptake, including implementation guides,12 templates,13 learning modules,14 hospital medication self-assessments15 and patient information pamphlets.16 However, despite these initiatives and the considerable harm from VTE, much of the care provided for VTE is not in line with CPGs17 in both the developed18 and developing worlds.19

As healthcare is facing an affordability crisis, there is an urgent need to move towards being able to monitor the appropriateness of care (care in line with evidence-based or consensus-based guidelines).20 The CareTrack Australia (CTA) study was designed to establish baseline estimates of the appropriateness of care delivered, at a population level, by a range of practitioners in real-world settings, and to determine what would be needed to monitor the ongoing appropriateness of care.21 CTA showed that adult Australians received appropriate care for 22 common conditions at 57% of eligible healthcare encounters during 2009 and 2010; VTE compliance was reported at 58%.22 The aim of this paper is to present and discuss the detailed CTA findings for VTE as a baseline for compliance with guidelines at a population level, from which to track progress resulting from future interventions.


The CTA methods have been described in detail elsewhere.21 ,22 Some aspects of relevance to VTE are summarised here.

Development and ratification of indicators

An initial list of 15 indicators (with 54 subcriteria) was sourced from recommendations within the National Health and Medical Research Council guidelines8 ,9 and sent to three practising specialist haematologists who were Heads of Departments, asking them to comment on and rate each on a scale of 1–9 for appropriateness23 in the Australian context during 2009 and 2010. A two-round review process was used and a formal process was employed for managing discrepancies between specialists.21 Opinions of other specialists were not canvassed for logistical reasons. This resulted in 39 indicators being accepted as appropriate: 31 relating to pharmacological and mechanical prophylaxis and eight to risk assessment, discharge care and management of DVT or PE (see table 1).

Table 1

CTA indicators (and aggregated sets of indicators) by compliance

Recruitment of participants and healthcare providers

A sample designed to be representative of the Australian adult population was used. Households were randomly selected from a phone directory (the Telstra White Pages) from defined regions within New South Wales and South Australia and contacted using a Computer-Assisted Telephone Interview (CATI).21 ,22 One adult was randomly selected from each household and was asked to participate. Those who agreed were sent a mail package containing information about the study and a consent form to allow access to their medical records. Participants who provided consent were called back and asked if they had been admitted overnight to a hospital or had one or more of the CTA conditions, and which healthcare providers they had seen for these in 2009 and 2010. Hospitals identified by the participants were contacted and asked to provide their consent for medical record access.22

Review of medical records

Medical record reviews were undertaken for the 1154 consenting participants whose healthcare providers had also provided consent. Healthcare encounters were deemed eligible for scoring of VTE indicators if a participant had been admitted overnight during 2009 and/or 2010.

Experienced registered nurses were recruited and trained as surveyors to conduct the medical record reviews using a web-based tool for onsite encrypted data collection. They were provided with formal training and received a manual with detailed criteria for inclusion, exclusion and scoring of indicators.

Estimates of compliance were measured as the percentage of eligible encounters for the VTE indicators that were answered ‘yes’.21 ,22 The inclusion criteria for the indicators for VTE prophylaxis were specific to particular types of surgery (eg, hip fracture surgery or abdominal surgery) or medical conditions (eg, decompensated cardiac failure or acute on chronic lung disease).8 ,9 As the CTA study was designed to measure the overall appropriateness of the healthcare delivered for 22 conditions and was not powered for significant results at indicator level, the number of eligible encounters for many indicators was low. To address this, indicators were aggregated into broader, clinically meaningful categories. For example, orthopaedic conditions with pharmacological prophylaxis (indicators 45–48) were grouped and included hip arthroplasty, hip fracture surgery, knee arthroplasty and lower limb fractures (see table 1).

Data relating to documentation of VTE risk assessment (indicator 42) was not included in the analysis reported here, as a review of surveyor practices revealed that some had assumed that a risk assessment had been carried out whenever appropriate prophylaxis had been prescribed, whether or not explicit documentation of an assessment was found. This was in breach of the criteria for this indicator, and these data were thus excluded.

Statistical analysis

Mean compliance and associated 95% CIs (using a modified version of the Clopper-Pearson (exact) method) were obtained using the SURVEYFREQ procedure in SAS V.9.3 for Windows (SAS Institute, Cary, North Carolina, USA). To address biases arising from the study design (including adjustment for non-response), two different weighting options and five versions of weights (three based on approaches used in the similar US study23) were used to generate weighted estimates of compliances. These were not significantly different to unweighted compliances overall or for any condition (including VTE). Hence, unweighted compliances were used for this analysis.22 Appendix 2 of the CTA study outlines the detailed methodology and overall results.22


Of the 1154 CTA participants, 481(42%) were admitted overnight to hospital at least once, with a total of 751 admissions eligible for assessment against the VTE indicators. There were 279 females (58%), and the mean age was 64 years (6% were aged 18–39, 17% 40–54, 56% 55–74 and 21% were over 70 years of age).

Of the 35 145 CTA encounters (with duplicates and the risk indicator removed), 1078 (3%) were eligible for scoring against VTE indicators; the number of eligible encounters per indicator ranged from 2 to 84. Records were reviewed at 33 hospitals, with 27 having eligible encounters. Eight of the hospitals had 50 or more eligible encounters.

Overall compliance with the VTE indicators was 51% (95% CI 47% to 54%), with results for aggregated sets of indicators ranging from 34% to 64% (table 1). Omission of the risk assessment indicator from the overall score reduced compliance from the 58% reported originally22 to the 51% reported here. For the eight hospitals with 50 or more eligible encounters, compliance ranged from 45% to 70%.


Our analysis of CTA data has shown that a sample of 481 Australian adults in 27 hospitals received appropriate care for VTE during 2009 and 2010 at only 51% of eligible healthcare encounters, in spite of considerable efforts to promote and facilitate the uptake of CPGs in Australia.12–16 Thus, despite the prevalence, cost, morbidity and mortality associated with VTE and PE, prophylaxis and treatment are still in line with CPGs only half the time. This continues to be a problem in both the developed and developing worlds.17–19

CTA patients who had surgery received appropriate pharmacological or mechanical anticoagulant therapy on only 39% and 57% of occasions, respectively (aggregations from table 1). The ENDORSE study, a multinational cross-sectional survey, also examined the proportion of at-risk patients who received effective prophylaxis.17 It found, for 804 patients from eight Australian hospitals studied in 2006–2007, that 82% of at-risk surgical patients received appropriate prophylaxis. This study did not separate pharmacological and mechanical prophylaxis.17 Baseline (preintervention) compliances for surgical patients were also higher than CTA compliances in two single hospital studies (65% and 74% for pharmacological prophylaxis and 89% and 64% for mechanical prophylaxis).24 ,25 Possible reasons for the lower CTA compliances are that CTA was a population-based study at 27 hospitals which were effectively randomly selected, while ENDORSE mainly collected data from prominent teaching hospitals, and the two single hospital studies were about to start an intervention for VTE prophylaxis, and had possibly raised awareness of the problem.

In contrast, the CTA compliance for at-risk medical patients at 46% (aggregations from table 1) was similar to those in the eight Australian ENDORSE hospitals and a regional hospital (51% and 64%, respectively).17 ,25 Lower compliances for medical than surgical patients in the ENDORSE study and the regional hospital are consistent with the more complex indications in medical patients, and have been noted elsewhere.26 No equivalent Australian studies could be found for indicators associated with management of suspected or confirmed DVT or PE (CTA compliance 64%), or patients with a documented discharge plan and a date for cessation of treatment (CTA compliance 46%), but it would seem reasonable to conclude that both of these areas of practice also require attention.

The risk assessment indicator was studied in two Australian single-hospital studies which both found 0% compliance at the preintervention stage, with modest postintervention compliances of 28% and 36%.24 ,27

The poor compliances with VTE indicators in Australia are consistent with the lack of a system-wide approach. Compliance measures or outcomes are not publicly reported at hospital level,28 VTE is not included in national standards,29 nor is it a national healthcare goal.30 Australian clinicians have identified that setting agreement on clinical guidelines and standards (agreement on risk categories, risk assessment tools, mandatory actions and protocols, provision of summaries), decision-support tools, and reporting results are enablers to delivering appropriate VTE care.31 ,32 The fact that compliance ranged from 45% to 70% between hospitals suggests that some facilities are faring better than others at managing VTE appropriately.

Strengths and weaknesses

The key strength of the CTA study is that it is designed to be representative of the Australian population to minimise selection bias, rather than a convenience-based or purposive-based sample. However, an unavoidable consequence of this strategy, coupled with finite research funds, is that the numbers of participants and/or eligible encounters are low for some indicators; 25 of 38 had insufficient data to report. Findings for these must be disregarded or interpreted with caution. The review of medical records, while costly and difficult, allowed compliance to be measured in a real-world setting and avoided the limitations inherent in asking healthcare providers to respond to clinical vignettes33 ,34 or questionnaires.35 Errors arising from measurement (information bias) were within acceptable limits for implicit review.36–38 Accordingly CTA provides some baseline estimates for compliance against which progress on the provision of appropriate care for VTE could be compared and tracked.

The approach used was associated with a high rate of attrition of potential participants and several other sources of possible bias. Although it was not logistically feasible to design sampling so as to eliminate all possible confounders (confusion bias) or have the sample characteristics to exactly match the Australian population, weighting using two methods and five different options made no significant difference to the overall compliance percentage, or that for VTE;22 this is consistent with providers not altering their clinical practices for patients of different ages, gender, or socioeconomic or health literacy status.

Commentators have raised issues with respect to the levels of evidence for and choice of indicators, effects of comorbidities, inter-rater reliability and the possibility of care having been provided but not recorded.36 ,37 These have all been addressed:22 ,38 compliance was shown to be no different for consensus-based and evidence-based recommendations; the CTA indicators were designed to be clinically relevant but not affected by comorbidities; inter-rater reliability was moderate, but was in line with other studies using implicit medical record review;38 and the effect on overall compliance of care received but not documented is thought to be no more than 10%.23 ,39 ,40


Our analysis of the VTE indicators from the CTA study show that compliance is modest at 51%, despite resources and guidelines being available, and the high associated cost and burden of disease. This is consistent with the lack of a system-wide focus on VTE in Australia as is the case in most of the rest of the world. In line with recommendations arising from the overall CTA study and feedback from clinicians, the challenge is to now move towards agreement on national clinical standards and on the development of indicators and tools to guide, document and monitor the appropriateness of care for VTE. An inclusive, national wiki-based process for achieving this has been proposed.20 VTE data could then be monitored at hospital level and the data aggregated at national and, potentially at international levels to track progress and inform policy.


The authors thank the participants, healthcare providers, practice managers, medical records staff and expert reviewers who generously gave their time and expertise.


View Abstract


  • Contributors The authors have contributed substantially to the conception and design of the study, or acquisition of data or analysis and interpretation of the finding. All have also been actively involved in either the drafting of the manuscript or revising it critically for important intellectual content; and have given approval for this version to be published. PDH was Program Manager for the NHMRC Program Grant and also assisted with the analysis and interpretation of findings and made a substantial contribution to reviewing the literature, synthesising results and findings from other relevant studies, and drafting the manuscript. NAH was responsible for the selection, development and ratification of all the indicators used in the study and assisted TDH with surveyor recruitment and training. She also contributed to the revision of the manuscript. TDH was the Project Manager for CareTrack and coordinated the data collection for the entire study including the extraction of patient medical records. She also coordinated all the necessary approvals and managed the training and performance of surveyors and contributed to the revision of the manuscript. DMH was the main Statistician for CareTrack and undertook analysis of the CareTrack data. She also contributed to the interpretation of the statistical information included in the manuscript. JB was involved in the design and conception of CareTrack Australia as the Chief Investigator-A of the NHMRC Program Grant, and extensive editing and revision of the manuscript. SAR developed the initial sampling plan for CareTrack Australia; provided advice for the process of recruiting participants and organised and managed telephone recruiters. NW was involved as a CareTrack Australia clinical expert, assisted in the ratification of the VTE indicators and provided clinical input into the editing and revision of the manuscript. WBR was the primary instigator of CareTrack and responsible for its conception and design. He was also involved in all aspects of the project from data collection to analysis and interpretation of findings. He has been heavily involved in the editing and revision of the manuscript providing invaluable advice and guidance to the corresponding author.

  • Funding This study was funded by National Health and Medical Research Council (NHMRC) Program Grant No. 568612.

  • Competing interests None declared.

  • Ethics approval Hunter New England Local Health District Human Research Ethics Committee (HREC).

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement No additional data are available.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.