Hospital‐acquired deep venous thrombosis (DVT) and pulmonary embolism (PE) are common yet preventable diagnoses. Beginning in October 2008, the Centers for Medicare and Medicaid Services (CMS) will not reimburse hospitals for several hospital‐acquired conditions, and will consider including DVT/PE on this list in 2009. Historically, hospitals have not documented whether the secondary diagnosis of DVT/PE was acquired during hospitalization or present on hospital admission. This difference can have several important implications with respect to hospital performance profiling, patient safety reporting, and hospital reimbursement.
To report the implementation and preliminary results of a systematic process to analyze an administrative database to determine the number of all secondary diagnoses of DVT/PE and the number of these that were hospital acquired.
The University of Michigan Health System began coding the timing of each ICD‐9‐CM diagnosis using a present‐on‐admission (PoA) variable for all hospital discharges in late 2005. Using an administrative database for adult patients discharged over a period of 12 months, the incidence of DVT/PE as a secondary diagnosis versus as hospital acquired was identified.
From November 11, 2006, to October 31, 2007, there were 982 hospital discharges with a secondary diagnosis of DVT/PE on all medical and surgical services. Of these, only 276 (28.1%) were determined to be hospital acquired, indicating the vast majority were present on hospital admission. Without using this PoA indicator, approximately 700 cases of DVT/PE would have been considered hospital‐acquired and potentially not reimbursed under proposed CMS changes. Our hospital has further utilized this data to preliminarily analyze general surgery cases with acquired postoperative DVT/PE and examine if adequate prophylaxis measures were instituted. There remain limitations to the use of administrative data. The incidence of hospital‐acquired DVT/PE may be understated, as the sensitivity of ICD‐9‐CM coding is likely suboptimal. Additionally, it is well known that many cases of DVT/PE are diagnosed after hospital discharge and may not be captured with this system. Furthermore, the reliability of reported incidence will be influenced by the accuracy of ICD‐9‐CM and PoA coding. This innovation demonstrated an impressive ability to distinguish between DVT/PE present on hospital admission versus that which was hospital acquired. This may prove to have significant implications for hospital reimbursement. With the recent implementation of the CMS rule that requires hospitals nationally to code PoA value for all discharge diagnoses, this systematic analysis can be implemented at other institutions. Additional benefits of such a system may include the standardizing of medical record documentation and coding practices, as well as the identification of deficiencies in DVT/PE prophylaxis practices.
P. J. Grant, none; V. Bahl, none.
To cite this abstract:Grant P, Bahl V. Using Hospital Administrative Data to Determine if DVT/PE Was Present on Hospital Admission or Was Hospital Acquired: Why Your Institution Needs to Know. Abstract published at Hospital Medicine 2008, April 3-5, San Diego, Calif. Abstract 99. Journal of Hospital Medicine. 2008; 3 (suppl 1). https://www.shmabstracts.com/abstract/using-hospital-administrative-data-to-determine-if-dvtpe-was-present-on-hospital-admission-or-was-hospital-acquired-why-your-institution-needs-to-know/. Accessed November 18, 2019.