Study design and setting
This study was conducted in the Elisabeth TweeSteden Hospital (ETZ), a large general teaching hospital in Tilburg, the Netherlands. Our previous prospective cohort study in non-surgical patients was extended by adding an extra measurement of guideline adherence two years after implementation of Padua-CDS (7).
Padua-CDS
The Padua-CDS uses an automated calculation of the Padua score in the Electronic Hospital Record (EHR) to determine the VTE risk of hospitalised patients (7). The Padua-CDS uses the following EHR data to determine the VTE risk: patient characteristics (sex, age, weight, BMI), medication list (hormonal treatment and anticoagulants), problem list (e.g. malignancy, VTE in the past, thrombophilia), and the mobility score of the Braden scale (mobility) (2, 9). If a patient has a Padua score ≥ 4 and no anticoagulants in use, the CDS generates a pop-up to the prescriber advising to start an LMWH as thromboprophylaxis, with the instruction to consider contra-indications, including bleeding risk. Thromboprophylaxis can be prescribed with a single click on a button within the pop-up, after which the medication order can be signed immediately.
Outcome measure
The primary outcome was the proportion of hospitalised patients for whom thromboprophylaxis was prescribed according to guidelines two years after the implementation (T2), compared to pre- implementation (T0) and directly after implementation (T1). The secondary outcome was the proportion of patients with a high VTE risk without bleeding risk for whom thromboprophylaxis was prescribed according to guideline adherence in T2, compared to T0 and T1.
Guideline adherence was assessed using the Padua Prediction Score (VTE risk) and Improve score (bleeding risk), as included in Table 1 (2, 10). The VTE risk was considered high at a Padua score of ≥ 4 (2). The bleeding risk was considered high at an Improve score of ≥ 7, or when a patient scored positive on the high-risk factors prior bleeding in the last 3 months, an active gastro-duodenal ulcer or a platelet count < 50x109/L (10). According to (inter)national guidelines, thromboprophylaxis with LMWH should be initiated if patients have a high risk of VTE. If patients also have an increased bleeding risk, mechanical prophylaxis in the form of elastic stockings or intermittent pneumatic compression should be started (5, 6).
Table 1
Padua Prediction Score and Improve Bleeding Score [1, 7].
Padua Prediction Score High risk of VTE: ≥4 | Improve Bleeding Score High risk of bleeding: ≥7, or ≥ 1 of the high-risk factors prior bleeding (< 3 months), active gastric or duodenal ulcer or platelet count less than 50 x 109/L |
Risk factor | Score | Risk factor | Score |
Active cancera | 3 | Moderate renal failure (eGFR 30-50ml/min) | 1 |
Previous VTEb | 3 | Male sex | 1 |
Reduced mobilityc | 3 | 40–84 years | 1.5 |
Thrombophilic conditiond | 3 | Active cancer | 2 |
Recent (≤ 1 month) trauma and/or surgery | 2 | Rheumatic disease | 2 |
Age (≥ 70 years) | 1 | Central venous catheter | 2 |
Heart and/or respiratory failure | 1 | Admission in Intensive Care Unit | 2.5 |
Acute MI or ischemic stroke | 1 | Sever renal failure (< 30 ml/min) | 2.5 |
Acute infection and/or rheumatologic disorder | 1 | Liver insufficiency (INR > 1.5) | 2.5 |
BMI ≥ 30 kg/m2 | 1 | ≥ 85 years | 3.5 |
Hormonal treatment | 1 | Thrombocytopenia (< 50 x 109 cell/L) | 4 |
| Recent (< 3 months) bleeding | 4 |
Active gastro-intestinal ulcer | 4.5 |
VTE venous thromboembolism, MI myocardial infarction, BMI Body Mass Index, eGFR estimated Glomerular Filtration Rate, INR International Normalized Ratio. . aPatients with local or distant metastases and/or in whom chemotherapy or radiotherapy had been performed in the previous 6 months bSuperficial vein thrombosis excluded cBedrest with bathroom privileges (either due to patient’s limitations or on physicians order) for at least 3 days dCarriage of defects of antithrombin, protein C or S, factor V Leiden, G20210A prothrombin mutation, antiphospholipid syndrome. |
Participants
In accordance with our previous study, non-surgical patients, ≥ 18 years of age, admitted to the departments of Neurology, Internal Medicine or Oncology & Haematology, with a hospital stay of ≥ 36 hours were included (7). Patients with orders for comfort measures only were excluded. Covid-19 was added as exclusion criterion, as there were no Covid-19 patients in our previous study and their thromboprophylaxis policy differs from general non-surgical patients.
Data collection
Guideline adherence in T2 was assessed prospectively in March 2022 for each eligible patient between ≥ 36 and ≤ 60 hours after admission. Data in T0 and T1 were from our previous study (collected retrospectively in October 2019 and prospectively in March 2020) (7).
Data analysis
Patient data were coded and processed in Datamanager 5.43.0 (The research manager, Deventer, The Netherlands). Data analysis was performed with IBM SPSS Statistics vs 24 (IBM, New York, USA). We chose to include 85 patients, the same number of patients as in T0 and T1. The difference in guideline adherence between the study periods was analysed by univariate logistic regression, followed by multivariate logistic regression. Immobility, malignancy and VTE in the past (the highest scoring VTE risk factors in the Padua score), age and sex and covariates that were different between the three study periods were included in the multivariate logistic regression, using stepwise regression with backward elimination (cut-off p-value < 0.2) (2). Odds Ratios (OR) and 95% confidence intervals (95%CI) were reported, as well as adjusted Odds Ratios (ORadj) for the multivariate logistic regression.