Venous thromboembolism (VTE), which includes both Deep Vein Thrombosis (DVT) and pulmonary thromboembolism (PE), is one of the significant contributors to the healthcare burden in recent times ^[1]. Venous thromboembolism continues to be the most preventable cause of death for hospitalized patients and is associated with high morbidity and medical costs ^[2]. The incidence of VTE ranges from 10% to 40% in the medical and general surgery population ^[3]. The overall annual incidence of DVT in the general population is 0.5-1/1000, with a higher risk among critically sick patients and patients undergoing major surgeries. The incidence of VTE was reported to be 28% in the south Indian population, and it is comparable with the Western world ^[4]. Deep vein thrombosis (DVT) is a medical condition in which blood clots form within the body's deep veins, predominantly occurring in the lower extremities ^[5].

Preventing DVT is of paramount importance in clinical practice, particularly in patients undergoing surgical procedures or those with predisposing medical conditions. To aid in the early identification of individuals at high risk of developing DVT, various risk assessment models have been developed. Two commonly utilized tools for this purpose are the WELL and PADUA scores. Healthcare professionals must understand and utilize these scores effectively as they enhance their competence and capability in implementing suitable preventive measures.

The Wells Score was developed by Wells et al. in 2000, which was designed to assess the probability of DVT in patients based on specific clinical criteria. This scoring system plays a crucial role in stratifying patients into low, moderate, and high-risk categories, providing a clear risk assessment and guiding clinical decisions ^{[21, 19]}. The score includes factors such as active cancer, recent immobility, localized tenderness, and swelling, among others ^[6].

The Wells Score, a simple and effective tool, which plays a crucial role in predicting DVT in surgical patients. It has high negative predictive value, which provides clinicians  a reliable tool for preoperative assessments ^[7,8]. [Annexure 2]

In contrast, the PADUA score was explicitly developed to assess VTE risk in hospitalized medical patients ^[9]; following the name of the city in which it originated, the PADUA score evaluates factors such as cancer, previous VTE, thrombophilia, and immobilization to predict the likelihood of thromboembolic events in medical inpatients. The PADUA score aims to identify high-risk individuals who may benefit from pharmacological thromboprophylaxis or mechanical measures to prevent DVT. It consists of 11 risk factors, including cancer, previous VTE, and heart or respiratory failure, each assigned a weighted score. The total PADUA score categorizes patients into low-, moderate-, or high-risk groups for thromboembolic events [ANNEXURE 3].

While the PADUA score has shown promise in identifying high-risk medical inpatients, its utility is limited in surgical settings and among non-hospitalized patients. Furthermore, comparative studies evaluating the performance of the PADUA score against other risk assessment models, such as the CAPRINI score and WELL score, could be more extensive. Considering the above facts, this study mainly aims to assess the efficacy of PADUA RAS (Risk Assessment Score) over the already well-established WELL RAM (Risk Assessment Model) in early diagnosis of risk of VTE among surgical patients. Numerous validation studies have confirmed the predictive accuracy and clinical utility of the PADUA score across diverse patient populations and healthcare settings ^[10].

In this study, we seek to address this gap by conducting a thorough comparative analysis of the WELL and PADUA scores to determine the risk of DVT development. We intended to answer the research question: Is the PADUA Score a reliable assessment tool for predicting the development of DVT in patients undergoing major surgery compared to the WELL Score? This was done by documenting the socio-demographic profile, examination, and venous Doppler findings.

A retrospective research of 640 participants evaluated the Caprini and Padua score's ability to predict patients at risk of venous thromboembolism. There were significant differences among risk factors between VTE and non-VTE patients. The Caprini RAM identified more VTE patients as a high-superhigh risk compared to the Padua RAM (70.9 vs 23.4%, p< 0.01). The Caprini RAM has superior sensitivity and positive and negative predictive values compared to the Padua RAM (P<0.05). However, the Caprini RAM has poorer specificity than the Padua RAM (p< 0.01). The Caprini RAM had a greater AUC and Youden index (p< 0.01), while the Padua RAM had a lower Youden index at crucial points 4 and 3 (0.010 vs 0.140, p< 0.05). The author concluded that The Caprini risk assessment model is more effective than the Padua model for identifying hospitalized medical patients at risk for venous thromboembolism ^[3].

An observational case-control study was conducted to predict the risk of DVT using Caprini, Padua and well's score among 72 subjects conducted at Medan. 54.2% are men, and the mean age is 53.14 years. Wells score; caprini score and Padua score have a sensitivity of 80.6%, 61.1%, and 50%, respectively, specificity of 80.65, 66.7%, and 75%, respectively, and accuracy of 87.5%, 64.3%, and 65.7%, respectively. Wells's score has better sensitivity, specificity, and accuracy than Caprini and Padua's for diagnosing DVT. The author concluded that the Wells score has better sensitivity, specificity, and accuracy than the Caprini and Padua scores for diagnosing deep vein thrombosis ^[4].

A retrospective study done by G.V.Shead and Ramji Narayanan in south India with a sample size of n=50 among > 50 years of age presenting to hospitals in south India was done to assess the development of postoperative DVT, which showed a disparity between the frequent occurrence of postoperative DVT in patients aged 50 years and over the frequency of fatal pulmonary embolism in South India.

A study was conducted to provide a narrative review and critique of the new American College of Physicians (ACP) and American College of Chest Physicians (AT9) guidelines for venous thromboembolism (VTE).- Both the ACP and AT9 guidelines have different methodologies and estimates of the risk-benefit of VTE prophylaxis. - The complexity of the guidelines and lack of consensus on VTE risk assessment contribute to difficulties in implementing the guidelines in practice. - The evidence used by the AT9 guideline to determine VTE risk based on the Padua score may not apply to the US inpatient population, as the study population had less co-morbidity, and a Padua score of 3 was associated with a significant risk of pulmonary embolism. The author concluded that The Padua score for predicting VTE risk in medical inpatients has limitations, and simpler risk assessment models may be superior ^[11].

A retrospective study was conducted to evaluate the effectiveness of the Padua risk assessment model among DVT patients who were analyzed in Beijing Shijitan Hospital between April 1, 2017 and June 30, 2017. The Padua risk assessment scale is used to evaluate the risk of DVT. This study noted a significant association between age, acute infection, prothrombin time, D dimer, erythrocyte sedimentation rate, and platelet count with thrombosis. The specificity of the Padua model was better than the sensitivity (80.7% vs. 50%, p<0.05) to predict DVT in internal medical patients. Similar findings reported among surgical patients with specificity to sensitivity of 87.5% vs. 67.5 % (p<0.05). The AUC of ROC in internal medical patients was more than that in surgical patients. The authors concluded that the Padua model is more specific than sensitive to predicting DVT in hospitalized patients ^[12].

Source of data:

• The study will be conducted on patient who is undergoing major surgery under General Surgery, ENT, Ophthalmology, Neurosurgery, Urology, Gastroenterology, Orthopedic Department in Victoria Hospitals under BMCRI, Bangalore.

Methods of Collection of Data

Study design: Cross-Sectional, analytical study.

Study period: “January 2023 to January 2024”

Place of study: Victoria Hospital, Bangalore Medical College and Research Institute.

Inclusion Criteria                                                            

• Age more than 18 years.
• Patient willing to give informed consent.
• Patients who have more than 72 hours of hospital stay.
• Patient undergoing major surgery ^[9] of more than 2-h duration under general/spinal/epidural/regional anesthesia.

Exclusion Criteria

• Patients who were diagnosed as having DVT at the time of admission.
• Patients on anticoagulant treatment for any reason.
• Patient planned for lower limb vascular interventions.
• Patient in whom anticoagulant therapy was contraindicated due to any reasons.
• Patient with a history of bleeding disorder.

Methodology

• After obtaining approval and clearance from the institutional ethics committee the patients fulfilling the inclusion criteria will be enrolled for the study after obtaining informed consent.
• Case record form as mentioned in Annexure – 2 will be used to obtain detailed demographic, clinical, and surgical history of each patient.

Any patient presenting to Department of General surgery at BMC&RI and who meets the inclusion criteria were included in the study after an informed written consent for being part of the study as well as venous Doppler, and treatment by Major Surgery.

• All enrolled subjects underwent pre-operative risk stratification using Well score (table-1) and Padua score (table-2). (Annexure-3)
• All patients (under either or both scores) underwent color Doppler ultrasound post-operatively after 7 days (or earlier, if clinically indicated) for assessment of DVT. Then, the follow-up Doppler was done again at 30 days.
• Comparative analysis of pre-operative WELL and PADUA Score for prediction of post-op DVT was carried out and data was analyzed.

Assessment tools

• Proforma for written informed consent.
• WELL’s score
• PADUA’s prediction score
• Study proforma
• Relevant investigations

Statistical Analysis

The gathered information was entered into Microsoft Excel 2016 and IBM SPSS Statistics for Windows 10, Version 29.0. (Armonk, NY: IBM Corp). For categorical variables, descriptive statistics, frequency analysis, and percentage analysis were used to characterize the data, while the mean and SD were employed for continuous variables. The independent sample t-test was used to find the significant difference between the bivariate samples in independent groups. The Chi-Square test was used similarly to see the significance in qualitative categorical data; if the expected cell frequency is less than 5 in 2x2 tables, then Fischer's Exact was used. ANOVA is used to identify the correlation between more than two subgroups. The probability value 0.05 is considered a significant level in all the statistical tools.

FIG 1: Flow chart depicting distribution of study population among various age groups

SAMPLE SIZE ESTIMATION

Sample size was estimated by using nMaster software Version 2.0 by applying following details in the above formula. Based on the study “Comparison of CAPRINI’s and PADUA’s Risk Assessment Scores in the Prediction of Deep Vein Thrombosis in Surgical Patients at a Tertiary Care Hospital”by VembuAnand and group [Indian Journal of Surgery https://doi.org/10.1007/s12262-022-03405-4] (Thirty-one (10.33%) out of 300 patients developed DVT in the post-operative period). Based on the above parameter with an alpha of 0.05 (2 sided) and precision level of 5%, the estimated sample size using the sample size formula for Single proportion. The above parameter and formula give us a sample size of 142 subjects.

Sample size will be rounded off to 160.

Sample size

Calculation of sample size- based on prevalence

Drop-out rate- 10%

According to the formula

N=Z² _(1-α)/2p (1-p)

                d²

           P      : Expected proportion

d    : Absolute precision

           1-α/2: Desired confidence level

This observational study included 160 patients operated on for various elective and emergency conditions at the Department of General Surgery at hospitals attached to BMCRI between August 2022 and January 2024. Eighty (50%) of these patients were females, and the other eighty (50%) were males, as shown in Table NO 1.

They were between 19 and 88 years old, with a mean age of 56.78 years. The majority (120) patients were in the age group >40 years, as shown in Table No.2. (FIG 3)

TABLE NO 1: Distribution of study participants according to Gender.

Figure 2: Distribution of study participants according to Gender.

TABLE NO 2: Distribution of study participants according to Age.

Figure 3: Distribution of study participants according to Age.

TABLE NO 3: Distribution of study participants according to BMI

Figure 4: Distribution of study participants according to BMI.

In our study, 65% of the study participants have a BMI<25,and the remaining 35% of participants were noted to have a BMI ≥ 25, as shown in Table 3. (FIG 4)

TABLE NO 4: Distribution of study participants according to clinical characteristics.

Figure 5: Distribution of study participant’s according to clinical characteristics.

Among the study participants, 1.3% had h/o of major surgery within 1 month, 3.1% of study population had varicose vein, and 15% of the study population had bilateral swollen legs.0.6% of the study population had h/o sepsis within 1 month of presenting to hospital, 1.3% had h/o pneumonia within 1 month and 3.1% diagnosed to have COPD at the time of presentation. 0.6% of the study population had h/o DVT, and 4.4% of the population had h/o multiple injuries within one month.

In our study group, 35% of the population was smokers, and 31.3% of the population were alcoholics.

TABLE NO 5: Distribution study participants according to Type of Surgery.

Figure 6: Distribution study participants according to Type of Surgery.

TABLE NO 6: Distribution of study participants according to PADUA RAM

Figure 7: Distribution of study participants according to PADUA RAM

On pre-operative risk assessment using the Padua scale, 16% of the population were high risk, and 83.1% were low risk groups, as shown in Table 6. (FIG 7)

TABLE NO 7: Comparative efficacy of WELL and PADUA scores in the prediction of DVT.

TABLE NO 8: Association between study parameters and the development of DVT among study participants.

There was no significant association noted between the development of DVT and the age and gender of patients, BMI, smoking habits and alcohol consumption and other co-morbidities among the study population.

However significant association noted (p<0.001) noted between development of DVT and duration of procedure (FIG 10) and h/o multiple trauma within 1 month (FIG 9).

Significant association (p<0.05) noted among development of DVT and h/o sepsis without 1 month before presenting to hospital (FIG 8), h/o DVT before surgery, h/o immobilization >72hrs, h/o acute spinal cord injury/paralysis within 1 month of surgery.

Figure 8: Association between H/o sepsis and the development of DVT among study participants

Figure 9: Association between H/o multiple injuries and the development of DVT among study participants.

Figure 10: Association between Duration of surgery and the development of DVT among study participants

• Figure 11: Association between PADUA score and the development of DVT among study participants.

Venous thromboembolism, a significant contributor to the healthcare burden in recent times ^[1], underscores the importance of early detection and prediction of the risk of developing deep vein thrombosis. Implementing thromboprophylaxis for patients at risk can significantly reduce the associated mortality and morbidity.

Various investigative tools are used to evaluate VTE, such as USG, CT angiogram, venous duplex scan, and MR angiography. Different risk assessment scores ^[22-23], have been established to facilitate the early identification of individuals, who are at a high risk of developing DVT. Two commonly utilized tools for this purpose are the WELL and PADUA scores ^[15-18].

The WELL score, which predicts the risk of VTE, includes multiple risk factors that aggravate the risk of DVT. Based on the sum of all the risk factors, the WELL score is categorized into low, moderate, and high risk. Mechanical and pharmacological thromboprophylaxis is indicated based on the risk of developing DVT ^[8].

The PADUA score assesses the VTE risk in hospitalized patients. It consists of 11 risk factors, each credited with a score of 1 ^[9]. If the total score is less than or equal to 3, it is considered low risk, and > 3 is regarded as high risk for developing DVT. Chemical and mechanical thromboprophylaxis are indicated based on the risk ^[24].

Our prospective, single-centred observational study compares WELL and PADUA scores for predicting DVT among patients undergoing major surgery (over 2 hrs). Since it was an observational study, we observed patient's outcomes in terms of the development of DVT. Due to ethical reasons, we could not deprive any patient of the routinely administered thromboprophylaxis. However, to have an accurate prediction assessment of WELL and Padua RAM, we adjusted thromboprophylaxis as a confounding factor. We included patients from various departments to have a broader outlook on the utility of WELL and Padua RAM in predicting DVT.

This prospective observational study was conducted at Victoria Hospital, BMCRI Bengaluru, on 160 patients who underwent major surgery. The WELL RAM, has been validated in both medical and surgical patients. Still, despite being easy to calculate, Padua RAM has yet to be validated for surgical patterns due to a lack of authentic studies and research.

In this study, we determined both the Well and Padua scores pre-operatively for surgical patients and assessed their efficacy in predicting the post-operative development of DVT ^[1]. WELL score was significantly high among the patients who developed DVT, after adjusting for prophylaxis (p<0.001) in our patients, which is similar to the findings of other studies done by Liu et al. ^{[3, 7]}.

In our study, patients are divided into four age groups, each group including 40 patients. The mean age of the study population is 56.78 years, which is not similar to the findings in other studies, which showed the mean age to be around 40-55 ^[3]. This difference might be because of the equal distribution of patients in all age groups.

In our study, no significant correlation was noted between increased BMI and the development of DVT, similar to the survey conducted by Obi AT et al. ^[7]. Still, a positive correlation with a significant p-value was noted in a study by Liu X et al. ^[3].

Various clinical characteristics and co-morbidities were assessed, including BMI, history of surgeries, varicose veins, and smoking/alcohol habits. Interestingly, while factors like BMI and specific co-morbidities showed no significant association with DVT development, others, such as a history of sepsis within one month and duration of surgery, emerged as significant predictors of DVT. This underscores the multifactorial nature of DVT risk, which seeks attention to the importance of thorough preoperative assessment.

In our study, the sensitivity of the WELL and Padua scores in predicting DVT was found to be 76.2% and 47.6%, respectively, compared with the study conducted by Liu X et al. ^[3], where SN was noted to be 70.9% and 23.4% respectively. The WELL score, which includes a broader array of risk factors, including age, BMI, and co-morbidities, showed a stronger correlation with DVT development than the PADUA score. Specifically, the WELL score had a higher sensitivity (76.2%) but lower specificity (69.1%) compared to PADUA (sensitivity: 47.6%, specificity: 83.5%). This indicates that while WELL may identify more cases of DVT, PADUA is more specific in identifying low-risk patients.

Table NO 9: Comparison of Sensitivity, Specificity, PPV and NPV of Caprini and Padua scores between our study and other studies.

COVID-19 infection was considered a potential risk factor for the development of DVT; however, we didn't find any significant statistical significance in this regard. The duration of surgery emerged as a significant predictor of DVT, underscoring the importance of intra-operative management and post-operative monitoring in high-risk patients.

Similar to data reported from more extensive national (National Surgical Quality Improvement Program) ^[47], regional (Michigan Surgical Quality Collaborative) ^[13] and institutional organizations, we found increased DVT risk in the setting of sepsis in our study.

The sensitivity and PPV of WELL RAM are higher than Padua RAM's. The high accuracy of the WELL RAM, can be explained by the fact that, this model has more assessment factors than the Padua model. The low sensitivity of the Padua RAM for DVT may also be attributed to the fact that many patients with multiple high risks were ignored. On the other hand, the specificity of the WELL RAM was lower than the Padua RAM. This result may also be related to fewer assessment factors in the Padua RAM and illustrates a high accuracy in evaluating low-risk patients by the Padua RAM ^[14].

The efficacy of both WELL and Padua RAM in predicting post-operative DVT was significantly high, with the advantage of WELL over Padua being better predictive. Hence, based on our study, the Padua score can be independently applied as a risk assessment tool for predicting post-op DVT and pre-op VTE prophylaxis in patients undergoing major surgery.

Limitations of the study

Our study's limitations include its observational design and the potential presence of unmeasured confounding variables. Additionally, the sample size of the study is relatively small which may affect the generalizability of the results. Future research involving larger cohorts and multi-centre studies could validate these risk assessment models more comprehensively.

Our study assesses the efficacy of both WELL and PADUA risk assessment models in predicting post-operative DVT among surgical patients. While WELL offers detailed risk stratification, PADUA presents a more straightforward alternative with comparable predictive accuracy. Incorporating these tools into clinical practice can help in the early identification of high-risk patients and facilitate targeted thromboprophylaxis interventions.

The WELL score demonstrated higher sensitivity (76.2%) than the PADUA score (47.6%), indicating its effectiveness in identifying patients at risk of developing DVT post-operatively. This higher sensitivity suggests that WELL may be more suitable for broader risk assessment despite its lower specificity (69.1%).The PADUA score exhibited higher specificity (83.5%), making it more reliable in identifying low-risk patients who may not require aggressive thromboprophylaxis. This specificity highlights its utility in reducing unnecessary.

This study underscores the ongoing need for refined risk assessment strategies and emphasizes the importance of evidence-based approaches in mitigating post-operative complications such as DVT. As healthcare continues to evolve, healthcare professionals, surgeons, and medical researchers will play a crucial role in integrating validated risk assessment tools like WELL and PADUA scores into clinical practice, thereby improving patient outcomes and quality of care in surgical settings.

1. Anand V, Ramakrishnan D, Mishra A, et al. Comparison of CAPRINI’s and PADUA’s Risk Assessment Scores in the prediction of deep vein thrombosis in Surgical patients at a Tertiary care Hospital. Indian J Surgery. 2022; 84(3):1-6.
2. Mahan CE, Spyropoulos AC. Venous thromboembolism prevention: a systematic review of methods to improve prophylaxis and decrease events in the hospitalized patient. Hospital Practice. 2010 Feb 1; 38(1):97-108.
3. Liu X, Liu C, Chen X, et al. Comparison between CAPRINI and PADUA risk assessment models for hospitalized medical patients at risk for venous thromboembolism. Interactive Cardio-Vascular and Thoracic Surgery. 2016; 23:538-543.
4. Gatot D, Maardia A I. Differences of wells scores accuracy, CAPRINI scores and PADUA scores in deep vein thrombosis diagnosis. IOP Conf. Series: Earth and Environmental Sciences. 2018; 125:012131.
5. Anderson FA, Wheeler HB, Goldberg RJ, Hosmer DW, Patwardhan NA, Jovanovic B, Forcier A, Dalen JE. A population-based perspective of the hospital incidence and case-fatality rates of deep vein thrombosis and pulmonary embolism: the Worcester DVT Study. Archives of internal medicine. 1991 May 1; 151(5):933-8.
6. Wells PS, Anderson DR, Bormanis J, Guy F, Mitchell M, Gray L, Clement C, Robinson KS, Lewandowski B. Value of assessment of pretest probability of deep-vein thrombosis in clinical management. The Lancet. 1997 Dec 20; 350(9094):1795-8.
7. Modi S, Deisler R, Gozel K, Reicks P, Irwin E, Brunsvold M, Banton K, Beilman GJ. Wells criteria for DVT is a reliable clinical tool to assess the risk of deep venous thrombosis in trauma patients. World Journal of Emergency Surgery. 2016 Dec; 11:1-6.
8. Gatot D, Mardia AI. Differences of wells scores accuracy, caprini scores and padua scores in deep vein thrombosis diagnosis. InIOP Conference Series: Earth and Environmental Science 2018 Mar 1 (Vol. 125, No. 1, p. 012131). IOP Publishing.
9. Barbar S, Noventa F, Rossetto V, Ferrari A, Brandolin B, Perlati M, De Bon E, Tormene D, Pagnan A, Prandoni P. A risk assessment model for the identification of hospitalized medical patients at risk for venous thromboembolism: the Padua Prediction Score. Journal of thrombosis and haemostasis. 2010 Nov 1; 8(11):2450-7.
10. Caprini JA, Botteman MF, Stephens JM, Nadipelli V, Ewing MM, Brandt S, Pashos CL, Cohen AT. Economic burden of long‐term complications of deep vein thrombosis after total hip replacement surgery in the United States. Value in Health. 2003 Jan; 6(1):59-74.
11. Maynard G, Jenkins IH, Merli GJ. Venous thromboembolism prevention guidelines for medical inpatients: mind the (implementation) gap. Journal of hospital medicine. 2013 Oct; 8(10):582-8.
12. Chen XL, Pan L, Wang Y. Validity of Padua risk assessment scale for assessing the risk of deep venous thrombosis in hospitalized patients. Zhonghuaneikezazhi. 2018 Jul 1; 57(7):514-7.
13. Khorana AA, Kuderer NM, Culakova E, Lyman GH, Francis CW. Development and validation of a predictive model for chemotherapy-associated thrombosis. Blood, The Journal of the American Society of Hematology. 2008 May 15; 111(10):4902-7.
14. Bahl V, Hu HM, Henke PK, Wakefield TW, Campbell Jr DA, Caprini JA. A validation study of a retrospective venous thromboembolism risk scoring method. Annals of surgery. 2010 Feb 1; 251(2):344-50.
15. Hung DD, Hiep NH, et al. Frequency and risk factor of lower-limb Deep Vein Thrombosis after Major Orthopedic surgery in Vietnamese Patients. Journal of Medical Sciences, 2019; 7(24):4250-4254.
16. Righini M, Le Gal G, De Lucia S, Roy PM, Meyer G, Aujesky D, Bounameaux H, Perrier A. Clinical usefulness of D-dimer testing in cancer patients with suspected pulmonary embolism. Thrombosis and haemostasis. 2006; 95(04):715-9.
17. Baliyan V, Tajmir S, Hedgire SS, Ganguli S, Prabhakar AM. Lower extremity venous reflux. Cardiovascular diagnosis and therapy. 2016 Dec; 6(6):533.
18. Kahn SR, Lim W, Dunn AS, Cushman M, Dentali F, Akl EA, Cook DJ, Balekian AA, Klein RC, Le H, Schulman S. Prevention of VTE in nonsurgical patients: antithrombotic therapy and prevention of thrombosis: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines. Chest. 2012 Feb 1; 141(2):e195S-226S.
19. Wells PS, Anderson DR, Rodger M, Ginsberg JS, Kearon C, Gent M, Turpie AG, Bormanis J, Weitz J, Chamberlain M, Bowie D. Derivation of a simple clinical model to categorize patients probability of pulmonary embolism: increasing the models utility with the SimpliRED D-dimer. Thrombosis and haemostasis. 2000; 83(03):416-20.
20. Geersing GJ, Zuithoff NP, Kearon C, Anderson DR, Ten Cate-Hoek AJ, Elf JL, Bates SM, Hoes AW, Kraaijenhagen RA, Oudega R, Schutgens RE. Exclusion of deep vein thrombosis using the Wells rule in clinically important subgroups: individual patient data meta-analysis. Bmj. 2014 Mar 10; 348.
21. Wells PS, Anderson DR, Bormanis J, Guy F, Mitchell M, Gray L, et al. Value of assessment of pretest probability of deep-vein thrombosis in clinical management. Lancet [Internet]. 1997; 350(9094):1795–8.
22. Pannucci CJ, Laird S, Dimick JB, Campbell DA, Henke PK. A validated risk model to predict 90-day VTE events in postsurgical patients. Chest. 2014 Mar 1; 145(3):567-73.
23. Rossetto V, Barbar S, Vedovetto V, Milan M, Prandoni P. Physicians’ compliance with the Padua Prediction Score for preventing venous thromboembolism among hospitalized medical patients. Journal of Thrombosis and Haemostasis. 2013 Jul 1; 11(7):1428-30.
24. Klen J, Horvat G, Blinc A. Perioperative prevention of venous thromboembolism in abdominal surgery patients based on the Caprini or the Padua risk score—a single centre prospective observational study. Life. 2022 Nov 11; 12(11):1843.