Family Practice Advance Access originally published online on January 22, 2008
Family Practice 2008 25(1):3-8; doi:10.1093/fampra/cmm075
A new diagnostic rule for deep vein thrombosis: safety and efficiency in clinically relevant subgroups
Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, the Netherlands. Correspondence to Diane B Toll, Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, PO Box 85500, 3508 GA Utrecht, the Netherlands; Email: d.b.toll{at}umcutrecht.nl
Received 6 July 2007; Revised 8 November 2007; Accepted 3 December 2007.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Background. Recently, a new, simple diagnostic rule was introduced to enable GPs to safely refrain from referring a considerable proportion of the patients suspected of having deep vein thrombosis (DVT). The rule (which includes seven patient history and physical examination items plus the result of a D-dimer test) discriminates ‘very low’ risk patients (not to be referred) from patients with an increased risk of DVT (to be referred). However, the rule's ‘efficiency’ (proportion of patients designated by the rule as very low risk) and safety (DVT prevalence among these very low risk patients) may change according to patient characteristics.
Objective. To test the rule's safety and efficiency in clinically relevant subgroups; i.e. across three age groups, in men and women, and in patients with and without a history of DVT, separately.
Methods. We retrospectively analysed data of 2086 primary care patients suspected of DVT, in whom all rule items and the result of the reference (‘gold’) standard (compression ultrasonography) were collected.
Results. The rule's efficiency decreased with age from 38.1% in the relatively young (<50 years) compared to 9.8% in patients aged 
70 years. The percentage of DVT among the very low risk patients was <1.5% in all subgroups. The low efficiency in the elderly could be improved without compromising the safety by increasing the D-dimer threshold.
Conclusion. The rule can safely exclude DVT in primary care patients suspected of DVT, irrespective of age, gender and history of DVT.
Keywords. Diagnosis, prediction rule, primary health care, venous thrombosis.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
The American Academy of Family Physicians and the American College of Physicians recommend that prediction rules to aid in diagnosing deep vein thrombosis (DVT) should be applied in practice.1,2 The prediction rule developed by Wells et al.3 (Wells rule) has been validated frequently in secondary care patients and may therefore be the ‘method of choice’. However, in primary care, this rule does not adequately rule out DVT.4 Hence, we recently developed and externally validated a particular rule for safely excluding DVT in primary care.5,6 This rule, which includes seven patient history and physical examination items plus the result of a D-dimer test, discriminates ‘very low’ risk (non-referral) patients from patients with an increased risk of DVT. Application of this rule reduces the number of unnecessary patient referrals for ultrasound measurements and consequently patient burden, at the cost of an acceptable low proportion of DVT (<1%) in the non-referred patients (Table 1).
|
However, it is well known that the relationship between disease state and the performance of a diagnostic test or rule may change according to characteristics of the patients.7–12 Physicians may thus rightly doubt whether the performance of our rule varies when it is applied in clinically relevant patient subgroups, such as the elderly or patients with a history of DVT and/or pulmonary embolism (PE). Knowledge of these potential modifying factors enhances the implementation of the diagnostic rule.13,14
We had several motives to assume that there might be heterogeneity in the performance of our rule among age groups. First, the incidence of DVT increases significantly with age.15–17 Second, compared to younger patients, elderly patients visiting the GP with signs and symptoms suggestive for DVT commonly suffer from substantial co-morbidity, which can confine the symptoms of DVT. Third, the specificity of the D-dimer test—the most important predictor or test in the rule—decreases with age.18,19 Finally, the specificity of the Wells rule for excluding DVT in secondary care has shown to decrease significantly with age.20
In addition, we hypothesized that the performance of the rule might be different in men and women because some risk factors for DVT are gender specific (e.g. hormones and pregnancy).
Finally, in various studies on the diagnosis of DVT in suspected patients, patients with a history of DVT and/or PE (i.e. history of venous thromboembolism or VTE) were excluded, usually because the diagnosis in these patients is assumed to be different or perhaps more difficult.3,21–23
The aim of the present study was to determine the applicability of our DVT rule across clinically relevant subgroups defined by age, gender and history of VTE. The applicability of the rule was tested by measuring the ‘efficiency’ of the rule (i.e. proportion of patients designated by the rule as having a very low risk of DVT and thus do not have to be referred) and the rule's ability to ‘safely exclude’ DVT (% missed DVT cases among these very low risk patients). In case of a predefined low efficiency or safety, we attempted to improve the rule by adjusting the D-dimer threshold or the rule's threshold to indicate very low risk.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
We retrospectively analysed the data of 2086 consecutive primary care patients suspected of DVT. The data originated from a large cross-sectional study on the diagnosis of DVT, executed between January 1, 2002 and January 1, 2006. Of this study, the data of the first 1295 patients were used for deriving the rule5, and the data of subsequent 532 patients were used for externally validating the rule6. Hence, for present analysis, the data of the derivation and validation study were combined and supplemented with the data of 259 additional patients from the same cross-sectional study who were included after the derivation and validation study.
The characteristics of the study were described in detail elsewhere.5,6 In short, the study was performed in the Netherlands, among 110 primary care physicians affiliated with three non-academic hospitals. All patients clinically suspected of DVT by the primary care physician were eligible for the study. Suspicion of DVT was based on swelling, redness or pain of the legs that had been present for not more than 30 days. Patients in whom PE was also suspected were excluded. Just before the start of the study, the participating primary care physicians received detailed instruction in a specially organized conference, including workshops dedicated to the logistics of the study. In addition, all GPs received similar information by mail. For motivational purposes, newsletters including information about the progress and inclusion rate of the project and feedback information were forwarded to all participating physicians and their assistants during the first year of the study.
The primary care physician systematically obtained patient history and physical examination and subsequently referred patients to the hospital to undergo D-dimer testing. Finally, real-time compression ultrasonography of the symptomatic leg was used as reference test in all patients. D-dimer level was measured by either the ELISA method (VIDAS, bioMérieux, France) or a latex assay method (Tinaquant, Roche, Germany), depending on the routine of the participating laboratory. In an earlier study, the optimal thresholds of the D-dimer tests were determined; the test was considered abnormal if the assay yielded a D-dimer level 400 ng/ml (Tinaquant assay) or 500 ng/ml (VIDAS).24 DVT was considered present if (at least one of) the proximal deep veins of the legs were not or not completely compressible, as determined with a 5–7.5 MHz linear-array sonographic scanner (system V GE/Sonotion).25 In patients with a normal ultrasound, the test was repeated within 7 days to exclude DVT.
Table 1 shows a short overview of the characteristics of the DVT rule and the performance of the rule in the derivation and validation study. In those analyses, each patient's total rule score was dichotomized at a threshold of 
3 to distinguish the very low risk (not to be referred) patients from patients with an ‘increased risk’ of DVT (score 4; to be referred for objective diagnostics, i.e. compression ultrasonography). Both the derivation and validation study showed that applying the rule, and retaining the very low risk patients in primary care, would result in approximately one-quarter reduction in referrals, at the cost of less than 0.7% missed DVT cases.
Sample size
For the subgroup analyses presented in this study, we had access to a very large database of 2086 patients. To calculate the power to assess the safety of the DVT rule in all seven subgroups, the percentage of missed DVT cases among the very low risk patients was estimated at 0.5% (between the 0.7% found in the derivation population5 and 0% in the validation population6). To be able to exclude a percentage false negatives of 3% or higher (95% upper limit of the confidence interval), taking a type I error of 0.05 (one sided) and a type II error of 0.2, 108 patients with a score 3 are needed in each subgroup.26 Previous studies5,6 showed that approximately 25% of the patients has a score 3, thus each subgroup should contain at least 432 patients to achieve enough statistical power.
Statistical analysis
We calculated for each patient the total score using the DVT rule (Table 1). Applying the score threshold of 3, we estimated the proportion of patients in whom DVT could be excluded according to the rule (efficiency) and the percentage of DVT in this non-referral group (i.e. false negatives = 1—negative predictive value; ‘safety’) with 95% confidence intervals. To study the effect of age, all patients were a priori categorized into three groups: <50 years, 50–70 years and 70 years, while considering group sizes based on the a priori power considerations. Developing DVT before the age of 50 (youngest age group) is relatively rare: <25 per 100 000 patients, while as from the age of 70 years or older (eldest age group), the incidence increases significantly to >232 per 100 000.15 If the age boundaries would have been chosen more extreme, some age groups would have become relatively small, compromising statistical power. Next, the analyses were repeated after the total study population was divided based on gender and finally based on the patient's history of VTE.
It was arbitrarily decided in advance that if the proportion of patients in whom DVT can be excluded was substantially decreased (<15%) or the percentage of DVT in the non-referral group was unacceptably high (>3%), we would attempt to improve the rule's performance by adjusting either the D-dimer cut-off level or the rule's threshold for this specific subgroup.
Five hundred and sixty-two of the 2086 subjects had missing values for one or more predictors in the rule. Missing values ranged from 0.9% in gender to 12.2% for the D-dimer test result (2.8% in oral contraceptive use, 4.6% trauma, 6.4% calf difference 3 cm, 7.2% presence of malignancy, 7.3% recent surgery and 10.7% vein distension). Missingness of data seldom occurs completely at random. Deleting subjects with a missing value does not only lead to a loss of statistical power, but often also to biased results. Therefore, imputing missing values is generally preferred to complete case analysis.27–30 Missing data were thus (multiple) imputed, using S-PLUS for windows, professional edition, version 6.2. Only when determining the optimal D-dimer threshold for a subgroup of patients with a low efficiency or safety, patients with missing D-dimer levels were excluded.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
The performance of the rule for the whole study population and the subgroups are presented in Table 2.
|
Overall, 416 of the 2086 patients had DVT (prevalence of 19.9%). Applying the rule, DVT could be excluded in 23.1% of the patients (efficiency). Just 0.9% of these patients with a score
3 had DVT (safety).
The prevalence of DVT ranged from 16.7% in patients aged <50 years to 21.7% for patients of 70 years or older. The proportion of patients with a score 3 (efficiency) was 38.1% in the youngest age group, 24.1% in patients aged 50–70 years and 9.8% in patients of 70 years or older. The percentage of DVT in these very low risk patients was 0.6%, 1.4% and 0.7%, respectively.
To improve the rule's efficiency in the eldest age group, we redefined the D-dimer threshold to indicate increased values to 1000 ng/ml (instead of 400 ng/ml for the Tinaquant assay and 500 ng/ml for the VIDAS assay). Applying the same rule cut-off value of 3, the proportion of patients in whom DVT could be excluded then increased from 9.8% to 33.8%, at the cost of only 1.4% missed DVT cases in this very low risk group. Further increasing the D-dimer threshold in elderly beyond 1000 ng/ml revealed an unacceptably high-proportion DVT among the patients with a score 3 (Fig. 1).
|
Applying a rule cut-off level of
4 or 5 instead of 3 in the elderly (without adjusting the D-dimer level) improved the efficiency of the rule only marginally. When a cut-off level of 6 was introduced, the proportion of patients in whom DVT could be excluded increased from 9.8% to 14.6%, while the percentage of missed DVT cases increased to 2.8%. When further increasing the rule cut-off level to 7, the proportion of patients in whom DVT could be excluded increased substantially to 32.8%. However, an unacceptable high proportion of these patients (5.6%) had a DVT (Fig. 1).
The majority (63.3%) of the patients included in the study were female. The prevalence of DVT was 16.7% in females and 25.4% in males. Among the females, 26.6% had a score 3, of whom 1.2% had a DVT. Of the males, just 16.9% had a score 3. The percentage DVT among males with a score 3 was 0.2%.
More than one quarter (27.4%) of the 2086 patients who visited the GP with signs or symptoms of DVT had a history of VTE. The proportion of patients with a score 3 was lower in patients without a previous VTE (22.0%) versus patients who have had a DVT or PE before (25.8%). The percentage of patients with DVT among the non-referral patients was in both patients with and without a previous VTE around 1% (1.2% and 0.8%, respectively).
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Several diagnostic DVT rules have been developed to enhance the diagnostic workup in patients suspected of the disease. This study quantified the performance of a recently developed simple rule to safely exclude DVT in clinically relevant subgroups of primary care patients.5 In the literature, no formal threshold has been postulated regarding the maximum failure rate of diagnostic strategies for DVT (acceptable percentage DVT in patients in whom DVT was ruled out according to the diagnostic strategy). Preferably, the failure rate of a (new) diagnostic strategy is as low as the failure rate of the most frequently used objective diagnostic modality, i.e. serial compression ultrasonography (approximately 1%).31,32 However, it seems reasonable that a considerable gain in efficiency of a new diagnostic strategy may counterbalance a failure rate of 2–3%.
In all subgroups, the failure rate was below 1.5%. This suggests that the rule can be used to safely exclude DVT in all patients, regardless of their age, gender or history of VTE. However, the rule's efficiency did vary across subgroups. Women were more often designated by the rule as very low risk compared to men. Next, the efficiency of the rule was somewhat higher among patients with a history of VTE. Most striking was the vast decrease of the rule's efficiency with age. As the mobility of elderly is often limited, the possibility of not needing to refer a part of these patients to a hospital for further diagnostic workup in case of suspicion of DVT is extra appealing. The decrease in efficiency could just partly be explained by the slightly higher prevalence of DVT in the eldest age group. It seemed predominantly be caused by one predictor in the rule: the D-dimer test result. Therefore, we evaluated the consequences when a more tolerant D-dimer test threshold or score threshold was used in the eldest age group. D-dimer levels increase with age. Moreover, older patients more often have pre-existing co-morbidity, which may further elevate the D-dimer concentration.33 Thus, elderly are more likely to have false positive D-dimer test results. We determined that in patients aged 70 years, the proportion of patients in whom DVT can be excluded can be improved from 9.8% to 33.8% without substantially compromising the safety (1.4% missed DVT cases), by raising the D-dimer threshold to 1000 ng/ml. We recommend, however, that this new D-dimer threshold in elderly should be verified prospectively to validate its safety.
Increasing the score threshold in elderly resulted in less optimistic results, either in a very small gain in efficiency or—by increasing the score threshold even further—in too many missed DVT cases among the patients with scores below the threshold. Therefore, it is not advisable to increase the score threshold for the elderly.
The results of this study are in accordance with other studies in secondary care that evaluated the safety and efficiency of a low/moderate pretest clinical score based on the Wells rule and a negative D-dimer test result (D-dimer <500 ng/ml) in different age groups.18,20 Both the primary care rule5,6 tested here and the strategy developed by Wells et al.3 can safely exclude DVT in very low risk patients of all age groups in respectively primary and secondary care. Both, however, also have a significantly lower efficiency in the elderly.20 An increased D-dimer threshold of 1000 ng/ml combined with a low/moderate Wells pretest clinical score increased the efficiency in secondary care elderly patients suspected of DVT from 12% to 28%. However, this was at the cost of 3.6% missed DVT cases.20 The present study suggests it is safe to increase the D-dimer threshold to 1000 ng/ml in primary care elderly (gain in efficiency of 9.8–33.8% at the cost of 1.4% missed DVT cases). We did not find previous studies on the safety and efficiency of (other) DVT diagnostic strategies in the other studied subgroups, by gender and previous VTE.
A few methodological issues should be discussed. First, we used serial compression ultrasonography as reference test. Due to its non-invasive character and its good test characteristics, this method is widely used in clinical practice to include or exclude the diagnosis of DVT. However, it is not truly a ‘gold’ standard because the accuracy of serial compression ultrasonography is not 100%.31,32 Consequently, some minor variation may be present in the ‘true’ performance of our rule (with all patients correctly classified as DVT present or DVT absent) compared to the results obtained in present analysis. However, it is unlikely that the safety of the rule is substantially different than presented because the negative predictive value of serial compression ultrasonography is very high (>99%).31,32 In no more than 1% of all patients clinically suspected of DVT, DVT is falsely excluded by serial compression ultrasonography. As the probability of DVT increases with increasing rule scores, these false negative results are more likely to occur in patients with scores above the threshold. Consequently, the true sensitivity and specificity of the rule might be slightly better than present analysis with serial compression ultrasonography as reference standard showed. Second, it may be argued that the efficiency of the rule may be dependent on the prevalence of DVT. However, the lower the prevalence of DVT among clinically suspected patients, the more inefficient referral of all these patients for additional diagnostic testing becomes, and thus the more efficient such a rule indeed will be. Third, although our present analysis was performed on a very large cohort, validation in other primary care patient groups may consolidate the results of our study.
In conclusion, the primary care DVT rule can be used to safely exclude DVT in primary care patients suspected of the disease, regardless of age, gender and a history of VTE. Depending on the patient subgroup, up to almost 40% of the primary care patients suspected of DVT can safely be spared referral for ultrasound measurements when applying the rule.
|
Declaration |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Funding: Netherlands Organisation for Health Research and Development (ZonMw 945-04-009).
Ethical approval: The study protocol was approved by the medical ethical committee of the University Medical Center Utrecht, the Netherlands.
Conflicts of interest: None.
|
Notes |
|---|
Toll DB, Oudega R, Vergouwe Y, Moons KGM and Hoes AW. A new diagnostic rule for deep vein thrombosis: safety and efficiency in clinically relevant subgroups. Family Practice 2008; 25: 3–8.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
1 Qaseem A, Snow V, Barry P, et al. Current diagnosis of venous thromboembolism in primary care: a clinical practice guideline from the American Academy of Family Physicians and the American College of Physicians. Ann Fam Med (2007) 5:57–62.
2 Qaseem A, Snow V, Barry P, et al. Current diagnosis of venous thromboembolism in primary care: a clinical practice guideline from the American Academy of Family Physicians and the American College of Physicians. Ann Intern Med (2007) 146:454–458.
3 Wells PS, Anderson DR, Rodger M, et al. Evaluation of D-dimer in the diagnosis of suspected deep-vein thrombosis. N Engl J Med (2003) 349:1227–1235.
4 Oudega R, Hoes AW, Moons KG. The Wells rule does not adequately rule out deep venous thrombosis in primary care patients. Ann Intern Med (2005) 143:100–107.
5 Oudega R, Moons KG, Hoes AW. Ruling out deep venous thrombosis in primary care. A simple diagnostic algorithm including D-dimer testing. Thromb Haemost (2005) 94:200–205.[Web of Science][Medline]
6 Toll DB, Oudega R, Bulten RJ, Hoes AW, Moons KG. Excluding deep vein thrombosis safely in primary care. J Fam Pract (2006) 55:613–618.[Web of Science][Medline]
7 Coughlin SS, Trock B, Criqui MH, Pickle LW, Browner D, Tefft MC. The logistic modeling of sensitivity, specificity, and predictive value of a diagnostic test. J Clin Epidemiol (1992) 45:1–7.[CrossRef][Web of Science][Medline]
8 Detrano R, Janosi A, Lyons KP, Marcondes G, Abbassi N, Froelicher VF. Factors affecting sensitivity and specificity of a diagnostic test: the exercise thallium scintigram. Am J Med (1988) 84:699–710.[CrossRef][Web of Science][Medline]
9 Hlatky MA, Pryor DB, Harrell FE Jr, Califf RM, Mark DB, Rosati RA. Factors affecting sensitivity and specificity of exercise electrocardiography. Multivariable analysis. Am J Med (1984) 77:64–71.[CrossRef][Web of Science][Medline]
10 Levy D, Labib SB, Anderson KM, Christiansen JC, Kannel WB, Castelli WP. Determinants of sensitivity and specificity of electrocardiographic criteria for left ventricular hypertrophy. Circulation (1990) 81:815–820.
11 Moons KG, van Es GA, Deckers JW, Habbema JD, Grobbee DE. Limitations of sensitivity, specificity, likelihood ratio, and bayes’ theorem in assessing diagnostic probabilities: a clinical example. Epidemiology (1997) 8:12–17.[Web of Science][Medline]
12 Ransohoff DF, Feinstein AR. Problems of spectrum and bias in evaluating the efficacy of diagnostic tests. N Engl J Med (1978) 299:926–930.[Abstract]
13 Mulherin SA, Miller WC. Spectrum bias or spectrum effect? Subgroup variation in diagnostic test evaluation. Ann Intern Med (2002) 137:598–602.
14 Whiting P, Rutjes AW, Reitsma JB, Glas AS, Bossuyt PM, Kleijnen J. Sources of variation and bias in studies of diagnostic accuracy: a systematic review. Ann Intern Med (2004) 140:189–202.
15 Anderson FA Jr, Wheeler HB, Goldberg RJ, et al. A population-based perspective of the hospital incidence and case-fatality rates of deep vein thrombosis and pulmonary embolism. The Worcester DVT study. Arch Intern Med (1991) 151:933–938.
16 Oger E. Incidence of venous thromboembolism: a community-based study in Western France. EPI-GETBP Study Group. Groupe d'Etude de la Thrombose de Bretagne Occidentale. Thromb Haemost (2000) 83:657–660.[Web of Science][Medline]
17 Silverstein MD, Heit JA, Mohr DN, Petterson TM, O'Fallon WM, Melton LJ 3rd. Trends in the incidence of deep vein thrombosis and pulmonary embolism: a 25-year population-based study. Arch Intern Med (1998) 158:585–593.
18 Aguilar C, del Villar V. Diagnostic performance of D-dimer is lower in elderly outpatients with suspected deep venous thrombosis. Br J Haematol (2005) 130:803–804. author reply 805.[CrossRef][Web of Science][Medline]
19 van der Graaf F, van den Borne H, van der Kolk M, de Wild PJ, Janssen GW, van Uum SH. Exclusion of deep venous thrombosis with D-dimer testing—comparison of 13 D-dimer methods in 99 outpatients suspected of deep venous thrombosis using venography as reference standard. Thromb Haemost (2000) 83:191–198.[Web of Science][Medline]
20 Schutgens RE, Haas FJ, Biesma DH. Reduced efficacy of clinical probability score and D-dimer assay in elderly subjects suspected of having deep vein thrombosis. Br J Haematol (2005) 129:653–657.[CrossRef][Web of Science][Medline]
21 Kahn SR, Joseph L, Abenhaim L, Leclerc JR. Clinical prediction of deep vein thrombosis in patients with leg symptoms. Thromb Haemost (1999) 81:353–357.[Web of Science][Medline]
22 Wells PS, Anderson DR, Bormanis J, et al. Value of assessment of pretest probability of deep-vein thrombosis in clinical management. Lancet (1997) 350:1795–1798.[CrossRef][Web of Science][Medline]
23 Wells PS, Hirsh J, Anderson DR, et al. Accuracy of clinical assessment of deep-vein thrombosis. Lancet (1995) 345:1326–1330.[CrossRef][Web of Science][Medline]
24 Oudega R, Toll DB, Bulten RJ, Hoes AW, Moons KG. Different cut-off values for two D-dimer assays to exclude deep venous thrombosis in primary care. Thromb Haemost (2006) 95:744–746.[Web of Science][Medline]
25 Fraser JD, Anderson DR. Deep venous thrombosis: recent advances and optimal investigation with US. Radiology (1999) 211:9–24.
26 Lachin JM. Introduction to sample size determination and power analysis for clinical trials. Control Clin Trials (1981) 2:93–113.[CrossRef][Web of Science][Medline]
27 Donders AR, van der Heijden GJ, Stijnen T, Moons KG. Review: a gentle introduction to imputation of missing values. J Clin Epidemiol (2006) 59:1087–1091.[CrossRef][Web of Science][Medline]
28 Greenland S, Finkle WD. A critical look at methods for handling missing covariates in epidemiologic regression analyses. Am J Epidemiol (1995) 142:1255–1264.
29 Little RJA. Regression with missing X's: a review. J Am Stat Assoc (1992) 87:1255–1264.
30 van der Heijden GJ, Donders AR, Stijnen T, Moons KG. Imputation of missing values is superior to complete case analysis and the missing-indicator method in multivariable diagnostic research: a clinical example. J Clin Epidemiol (2006) 59:1102–1109.[CrossRef][Web of Science][Medline]
31 Birdwell BG, Raskob GE, Whitsett TL, et al. The clinical validity of normal compression ultrasonography in outpatients suspected of having deep venous thrombosis. Ann Intern Med (1998) 128:1–7.
32 Cogo A, Lensing AW, Koopman MM, et al. Compression ultrasonography for diagnostic management of patients with clinically suspected deep vein thrombosis: prospective cohort study. Br Med J (1998) 316:17–20.
33 Sohne M, Kamphuisen PW, van Mierlo PJ, Buller HR. Diagnostic strategy using a modified clinical decision rule and D-dimer test to rule out pulmonary embolism in elderly in- and outpatients. Thromb Haemost (2005) 94:206–210.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  K. J.M. Janssen, Y. Vergouwe, A. R. T. Donders, F. E. Harrell Jr., Q. Chen, D. E. Grobbee, and K. G.M. Moons Dealing with Missing Predictor Values When Applying Clinical Prediction Models Clin. Chem., May 1, 2009; 55(5): 994 - 1001. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||