Hospital‐acquired venous thromboembolism (HA‐VTE) is a morbid, often fatal, and — in cases of pulmonary embolism (PE) or lower‐extremity deep venous thrombosis (LE DVT) — pharmacologically potentially preventable condition. Because the Centers for Medicare and Medicaid Services now mandates use of a “present‐on‐admission” flag for all codes, discharge data from a range of hospitals can be assessed automatically for HA‐VTE incidence. However, estimates of HA‐VTE incidence in a medical adult inpatient population and proportion of PE/LE DVT to other thromboses remain poorly characterized.


We analyzed discharge records of patients hospitalized more than 2 days with nonsurgical diagnosis‐related groups (DRGs) in 83 academic medical centers between October 1, 2009, and March 1, 2011, in the UHC database. The incidence and location of HA‐VTE events were based on specific International Classification of Disease, 9th revision (ICD‐9‐CM) codes coupled with the present‐on‐admission indicator equal to “no.” We also assessed comorbidities as well as receipt of VTE prophylaxis on hospital day 1 or 2 within this population.


Among 2,525,068 medical hospitalizations, 12,847 patients (0.51%) were diagnosed with HA‐VTE; of these, just less than half had either a PE (2449; 19.1%) or a LE DVT (3848; 30%). A significant number had an upper‐extremity (2893; 22.5%) or superficial vein (3248; 25.3%) thrombosis, whereas the rest had thrombosis not otherwise specified (n = 409). Among the PE and LE DVT HA‐VTE cases, 34.3% had a diagnosis of cancer, 47.8% spent time in the intensive care unit, and 16.5% died in the hospital. Also among these cases, 54.9% received pharmacologic thromboprophylaxis on hospital day 1 or 2.


In a cohort of medical hospitalizations at 83 UHC academic medical centers, only 0.51% of patients hospitalized ultimately developed HA‐VTE; of these, just half were PE or LE DVT (and thus potentially preventable by pharmacologic prophylaxis); and of these, almost half developed despite use of VTE prophylaxis. The population that developed HA‐VTE (PE or LE DVT) had significant comorbidities as well as a high risk of needing intensive care and a high risk of in‐hospital death. More research is needed to understand the potentially changing incidence and patterns of HA‐VTE and to determine supplemental strategies — beyond pharmacologic prophylaxis — to further reduce HA‐VTE, especially in the significantly ill HA‐VTE pulmonary embolism/lower‐extremity deep venous thrombosis population.