Thrombosis in Cancer Patients during Hospitalization: Impact on Stays and Costs

Background: Cancer patients represent an at-risk population for Venous Thromboembolic Events (VTE). Following a previous study on two localizations, we intended to determine the incidence and recurrence rate of VTE in breast cancer (BC), colon cancer (CC), lung cancer (LC) and prostate cancer (PC), four of the most frequent cancers in France. We evaluate their impact on the number and the duration of hospitalizations and calculate the additional costs for hospitals compared to cancer patients without VTE.

Methods: The French national hospital database (PMSI) was used to identify BC, CC, LC and PC patients diagnosed in 2010 who were hospitalized for a VTE during the following two years. The number of stays induced by a VTE, the number of patients hospitalized or re-hospitalized and the time before recurrent event were determined using the disease-specific ICD-10 codes, among all VTEs classified as primary or related diagnosis (PD / RD) and significant associated diagnosis (SAD).

Subsequently, we evaluated the extra cost and extra duration of stay when a VTE is managed during a hospitalization. Therefore, we selected stays during witch a VTE is managed but was not the main reason of hospitalization (cancer was classified as primary/related diagnosis and VTE as significant associated diagnosis). Those stays were matched and compared to similar stays without VTE. Costs were estimated from the healthcare system perspective, using the French official tariffs.

Results: Among 194,964 patients newly diagnosed in 2010 with BC, CC, LC and PC in 2010, 1,271 (2.0%) BC patients, 2,866 (6.0%) CC patients, 3,775 (9,6%) LC patients and 997 (2.2%) PC patients, were hospitalized for a VTE or experienced a VTE during a hospital stay, over two years. In total, 12,880 stays related to VTE were identified.

We found that 2.053 cancer patients (23.0% of all VTE patients) were re-hospitalized at least once with a recurrent VTE, representing 3,969 admissions (30.8% of total admissions). In those patients with VTE recurrence, median time between first and recurrent event was 23 days in BC patients, 22 days in CC patients, 19 days in LC patients 25 days in PC patients.

In the second part of our study, the comparison between stays for cancer during which a VTE occurred and stays for cancer without VTE showed that the duration of hospital stay was longer in patients with VTE: in public hospitals, median stay duration raised from 4 to 7 days for BC, from 8 to 16 days in CC, from 2 to 9 days in LC and from 6 to 10 days in PC.

VTE consequently induced a significant extra cost related to hospitalizations over the two years of follow-up: in public hospitals, the median expenditure per stay increased by 37% in BC, (up to Û 5,518), by 61% in CC (up to Û9,878), by 202% in LC (up to Û7,308) and by 22% in PC (up to Û6,200).

Conclusion: VTE appeared to make cancer management much heavier as patients faced increased hospital stays duration whereas healthcare system faced important additional costs. Better prevention and follow-up guidelines could reduce this burden, and benefit both patients and hospitals.