Risk Factor Model to Predict Venous Thromboembolism in Hospitalized Medical Patients

Center for Quality of Care Research, Baystate Medical Center, Springfield, Massachusetts, USA.
Journal of Hospital Medicine (Impact Factor: 2.3). 04/2011; 6(4):202-9. DOI: 10.1002/jhm.888
Source: PubMed


The Joint Commission requires that all medical inpatients be assessed for venous thromboembolism (VTE) risk, but available risk stratification tools have never been validated.
We conducted a retrospective cohort study of patients age ≥18 years, admitted to 374 US hospitals in 2004-2005, with a primary diagnosis of pneumonia, heart failure, chronic obstructive pulmonary disease (COPD), stroke, and urinary tract infection, and length of stay ≥3 days. Subjects were randomly assigned (80/20) to a derivation or validation set. We then assessed VTE (International Classification of Diseases, Ninth Revision [ICD-9] code plus diagnostic test plus treatment), patient demographics, 21 potential risk factors, and other comorbidities. We created a VTE risk stratification tool using multivariable regression modeling and applied it to the validation sample.
Of 242,738 patients, 612 (0.25%) patients fulfilled our criteria for VTE during hospitalization, and an additional 440 (0.18%) were readmitted for VTE within 30 days (overall incidence of 0.43%). In the multivariable model, age, sex, and 10 additional risk factors were associated with VTE. The strongest risk factors were inherited thrombophilia (OR 4.00), length of stay ≥6 days (OR 3.22), inflammatory bowel disease (OR 3.11), central venous catheter (OR 1.87), and cancer. In the validation set, the model had a c-statistic of 0.75 (95% CI 0.71, 0.78). Deciles of predicted risk ranged from 0.11% to 1.46% with observed risk over the same deciles from 0.17% to 1.81%.
The risk of symptomatic VTE in general medical patients is low. A risk factor model can identify those at sufficient risk to warrant pharmacologic prophylaxis.

11 Reads

  • No preview · Article · Feb 2012 · Clinical medicine (London, England)
  • [Show abstract] [Hide abstract]
    ABSTRACT: Pulmonary embolism (PE) is a rare but serious event that may occur after spinal surgery. To correlate PE incidence after spinal arthrodesis with surgical approach, region of spine operated, and primary spinal pathology. To identify PE incidence trends in this population. The Nationwide Inpatient Sample was queried using ICD-9 codes (81.01-81.08) for spinal fusion procedures over a 21-year period (1988-2008). Other data points included PE occurrence, surgical approach, spinal region, surgical indication, and mortality. Multivariate and relational analyses were performed. 4,505,556 patients were identified and 9530 had PE (incidence=0.2%). PE patients had higher odds of combined A/P surgical approaches than posterior approaches (OR=1.97; 95% CI=1.66-2.33), and PE incidence was higher in thoracic versus cervical or lumbar fusions (OR=2.54; 95% CI=2.14-3.02). PE was more likely with vertebral fracture (OR=1.85; 95% CI=1.53-2.23) and SCI with vertebral fracture (OR=4.59; 95% CI=3.72-5.70) than without trauma. Between 1988 and 2008, the PE incidence remained stable for patients with intervertebral disk degeneration and scoliosis, but increased for patients with vertebral fracture, and SCI with vertebral fracture. There was greater inpatient mortality with occurrence of a PE (OR=12.92; 95% CI=10.55-14.41). Although the incidence of PE in spinal arthrodesis patients is only 0.2%, there is a higher incidence after combined A/P approaches, thoracic procedures, and trauma surgical procedures. Despite the overall PE incidence remaining stable since 1988, incidence steadily increased among trauma patients. Further research is needed to explain these trends, given the context of changing patient populations and improving surgical techniques and prophylaxis measures. Greater caution and prophylaxis among trauma patients may be warranted.
    No preview · Article · Feb 2012 · Clinical neurology and neurosurgery
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Mortality prediction models can be used to adjust for presenting severity of illness in observational studies of treatment effectiveness. We aimed to determine the incremental benefit of adding information about critical care services to a sepsis mortality prediction model. In a retrospective cohort of 166‚ÄČ931 eligible sepsis patients at 309 hospitals, we developed nested logistic regression models to predict mortality at the patient level. Our initial model included only demographic information. We then added progressively more detailed information such as comorbidities and initial treatments. We calculated each model's area under the receiver operating characteristic curve (AUROC) and also used a sheaf coefficient analysis to determine the relative effect of each additional group of variables. Model discrimination increased as more detailed patient information was added. With demographics alone, the AUROC was 0.59; adding comorbidities increased the AUROC to 0.67. The final model, which took into account mixed (hierarchical) effects at the hospital level as well as initial treatments administered within the first two hospital days, resulted in an AUROC of 0.78. The standardized sheaf coefficient for the initial treatments was approximately 30% greater than that for demographics or infection source. A sepsis disease risk score that incorporates information about the use of mechanical ventilation and vasopressors is superior to models that rely only on demographic information and comorbidities. Until administrative datasets include clinical information (such as vital signs and laboratory results), models such as this one could allow researchers to conduct observational studies of treatment effectiveness in sepsis patients.
    Full-text · Article · May 2012 · Pharmacoepidemiology and Drug Safety
Show more