1
|
Wang J, de Vale JS, Gupta S, Upadhyaya P, Lisboa FA, Schobel SA, Elster EA, Dente CJ, Buchman TG, Kamaleswaran R. ClotCatcher: a novel natural language model to accurately adjudicate venous thromboembolism from radiology reports. BMC Med Inform Decis Mak 2023; 23:262. [PMID: 37974186 PMCID: PMC10652606 DOI: 10.1186/s12911-023-02369-z] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/14/2023] [Accepted: 11/07/2023] [Indexed: 11/19/2023] Open
Abstract
INTRODUCTION Accurate identification of venous thromboembolism (VTE) is critical to develop replicable epidemiological studies and rigorous predictions models. Traditionally, VTE studies have relied on international classification of diseases (ICD) codes which are inaccurate - leading to misclassification bias. Here, we developed ClotCatcher, a novel deep learning model that uses natural language processing to detect VTE from radiology reports. METHODS Radiology reports to detect VTE were obtained from patients admitted to Emory University Hospital (EUH) and Grady Memorial Hospital (GMH). Data augmentation was performed using the Google PEGASUS paraphraser. This data was then used to fine-tune ClotCatcher, a novel deep learning model. ClotCatcher was validated on both the EUH dataset alone and GMH dataset alone. RESULTS The dataset contained 1358 studies from EUH and 915 studies from GMH (n = 2273). The dataset contained 1506 ultrasound studies with 528 (35.1%) studies positive for VTE, and 767 CT studies with 91 (11.9%) positive for VTE. When validated on the EUH dataset, ClotCatcher performed best (AUC = 0.980) when trained on both EUH and GMH dataset without paraphrasing. When validated on the GMH dataset, ClotCatcher performed best (AUC = 0.995) when trained on both EUH and GMH dataset with paraphrasing. CONCLUSION ClotCatcher, a novel deep learning model with data augmentation rapidly and accurately adjudicated the presence of VTE from radiology reports. Applying ClotCatcher to large databases would allow for rapid and accurate adjudication of incident VTE. This would reduce misclassification bias and form the foundation for future studies to estimate individual risk for patient to develop incident VTE.
Collapse
Affiliation(s)
- Jeffrey Wang
- Department of Biomedical Informatics, Emory University School of Medicine, 1462 Clifton Road, Suite 504, Atlanta, GA, 30322, USA.
| | - Joao Souza de Vale
- Department of Biomedical Informatics, Emory University School of Medicine, 1462 Clifton Road, Suite 504, Atlanta, GA, 30322, USA
| | - Saransh Gupta
- Department of Biomedical Informatics, Emory University School of Medicine, 1462 Clifton Road, Suite 504, Atlanta, GA, 30322, USA
| | - Pulakesh Upadhyaya
- Department of Biomedical Informatics, Emory University School of Medicine, 1462 Clifton Road, Suite 504, Atlanta, GA, 30322, USA
| | - Felipe A Lisboa
- Surgical Critical Care Initiative (SC2i), Uniformed Services University of the Health Sciences, Bethesda, MD, 20814, USA
- Department of Surgery, Uniformed Services University of the Health Sciences and Walter Reed National Military Medical Center, Bethesda, MD, 20814, USA
- Henry M. Jackson Foundation for the Advancement of Military Medicine, Inc., Bethesda, MD, 20817, USA
| | - Seth A Schobel
- Surgical Critical Care Initiative (SC2i), Uniformed Services University of the Health Sciences, Bethesda, MD, 20814, USA
- Department of Surgery, Uniformed Services University of the Health Sciences and Walter Reed National Military Medical Center, Bethesda, MD, 20814, USA
- Henry M. Jackson Foundation for the Advancement of Military Medicine, Inc., Bethesda, MD, 20817, USA
| | - Eric A Elster
- Surgical Critical Care Initiative (SC2i), Uniformed Services University of the Health Sciences, Bethesda, MD, 20814, USA
- Department of Surgery, Uniformed Services University of the Health Sciences and Walter Reed National Military Medical Center, Bethesda, MD, 20814, USA
| | - Christopher J Dente
- Surgical Critical Care Initiative (SC2i), Uniformed Services University of the Health Sciences, Bethesda, MD, 20814, USA
- Emory Department of Surgery, Emory University School of Medicine, Atlanta, GA, USA
- Grady Memorial Hospital, Atlanta, GA, USA
| | - Timothy G Buchman
- Surgical Critical Care Initiative (SC2i), Uniformed Services University of the Health Sciences, Bethesda, MD, 20814, USA
- Emory Department of Surgery, Emory University School of Medicine, Atlanta, GA, USA
- Emory Critical Care Center, Atlanta, GA, USA
| | - Rishikesan Kamaleswaran
- Department of Biomedical Informatics, Emory University School of Medicine, 1462 Clifton Road, Suite 504, Atlanta, GA, 30322, USA
- Surgical Critical Care Initiative (SC2i), Uniformed Services University of the Health Sciences, Bethesda, MD, 20814, USA
| |
Collapse
|
2
|
Neeman E, Liu V, Mishra P, Thai KK, Xu J, Clancy HA, Schlessinger D, Liu R. Trends and Risk Factors for Venous Thromboembolism Among Hospitalized Medical Patients. JAMA Netw Open 2022; 5:e2240373. [PMID: 36409498 PMCID: PMC9679881 DOI: 10.1001/jamanetworkopen.2022.40373] [Citation(s) in RCA: 24] [Impact Index Per Article: 12.0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Indexed: 11/23/2022] Open
Abstract
IMPORTANCE While hospital-associated venous thromboembolism (HA-VTE) is a known complication of hospitalization, contemporary incidence and outcomes data are scarce and methodologically contested. OBJECTIVE To define and validate an automated electronic health record (EHR)-based algorithm for retrospective detection of HA-VTE and examine contemporary HA-VTE incidence, previously reported risk factors, and outcomes. DESIGN, SETTING, AND PARTICIPANTS This cohort study was conducted using hospital admissions between January 1, 2013, and June 30, 2021, with follow-up until December 31, 2021. All medical (non-intensive care unit) admissions at an integrated health care delivery system with 21 hospitals in Northern California during the study period were included. Data were analyzed from January to June 2022. EXPOSURES Previously reported risk factors associated with HA-VTE and administration of pharmacological prophylaxis were evaluated as factors associated with HA-VTE. MAIN OUTCOMES AND MEASURES Yearly incidence rates and timing of HA-VTE events overall and by subtype (deep vein thrombosis, pulmonary embolism, both, or unknown), as well as readmissions and mortality rates. RESULTS Among 1 112 014 hospitalizations involving 529 492 patients (268 797 [50.8%] women; 75 238 Asian [14.2%], 52 697 Black [10.0%], 79 398 Hispanic [15.0%], and 307 439 non-Hispanic White [58.1%]; median [IQR] age, 67.0 [54.0-79.0] years), there were 13 843 HA-VTE events (1.2% of admissions) occurring in 10 410 patients (2.0%). HA-VTE events increased from 307 of 29 095 hospitalizations (1.1%) in the first quarter of 2013 to 551 of 33 729 hospitalizations (1.6%) in the first quarter of 2021. Among all HA-VTE events, 10 746 events (77.6%) were first noted after discharge. In multivariable analyses, several factors were associated with increased odds of HA-VTE, including active cancer (adjusted odds ratio [aOR], 1.96; 95% CI, 1.85-2.08), prior VTE (aOR, 1.71; 95% CI, 1.63-1.79), and reduced mobility (aOR, 1.63; 95% CI, 1.50-1.77). Factors associated with decreased likelihood of HA-VTE included Asian race (vs non-Hispanic White: aOR, 0.65; 95% CI, 0.61-0.69), current admission for suspected stroke (aOR, 0.73; 95% CI, 0.65-0.81), and Hispanic ethnicity (vs non-Hispanic White: aOR, 0.81; 95% CI, 0.77-0.86). HA-VTE events were associated with increased risk of readmission (hazard ratio [HR], 3.33; 95% CI, 3.25-3.41) and mortality (HR, 1.63; 95% CI, 1.57-1.70). CONCLUSIONS AND RELEVANCE This study found that HA-VTE events occurred in 1.2% of medical admissions, increased over time, and were associated with increased adverse outcomes. These findings suggest that approaches designed to mitigate occurrence and outcomes associated with HA-VTE may remain needed.
Collapse
Affiliation(s)
- Elad Neeman
- Department of Hematology and Oncology, The Permanente Medical Group, San Rafael, California
| | - Vincent Liu
- Division of Research, Kaiser Permanente Northern California, Oakland
| | - Pranita Mishra
- Division of Research, Kaiser Permanente Northern California, Oakland
| | - Khanh K. Thai
- Division of Research, Kaiser Permanente Northern California, Oakland
| | - James Xu
- Internal Medicine Residency Program, Kaiser Permanente Northern California, San Francisco
| | - Heather A. Clancy
- Division of Research, Kaiser Permanente Northern California, Oakland
| | | | - Raymond Liu
- Division of Research, Kaiser Permanente Northern California, Oakland
- Department of Hematology and Oncology, The Permanente Medical Group, San Francisco, California
| |
Collapse
|
3
|
Sepassi A, Chingcuanco F, Gordon R, Meier A, Divino V, DeKoven M, Ben-Joseph R. Resource utilization and charges of patients with and without diagnosed venous thromboembolism during primary hospitalization and after elective inpatient surgery: a retrospective study. J Med Econ 2018; 21:595-602. [PMID: 29480088 DOI: 10.1080/13696998.2018.1445635] [Citation(s) in RCA: 3] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Indexed: 10/18/2022]
Abstract
AIMS To assess incremental charges of patients experiencing venous thromboembolisms (VTE) across various types of elective inpatient surgical procedures with administration of general anesthesia in the US. METHODS The authors performed a retrospective study utilizing data from a nationwide hospital operational records database from July 2014 through June 2015 to compare a group of inpatients experiencing a VTE event post-operatively to a propensity score matched group of inpatients who did not experience a VTE. Patients included in the analysis had a hospital admission for an elective inpatient surgical procedure with the use of general anesthesia. Procedures of the heart, brain, lungs, and obstetrical procedures were excluded, as these procedures often require a scheduled ICU stay post-operatively. Outcomes examined included VTE events during hospitalization, length of stay, unscheduled ICU transfers, number of days spent in the ICU if transferred, 3- and 30-day re-admissions, and total hospital charges incurred. RESULTS The study included 17,727 patients undergoing elective inpatient surgical procedures. Of these, 36 patients who experienced a VTE event were matched to 108 patients who did not. VTE events occurred in 0.2% of the study population, with most events occurring for patients undergoing total knee replacement. VTE patients had a mean total hospital charge of $60,814 vs $48,325 for non-VTE patients, resulting in a mean incremental charge of $11,979 (p < .05). Compared to non-VTE patients, VTE patients had longer length of stay (5.9 days vs 3.7 days, p < .001), experienced a higher rate of 3-day re-admissions (3 vs 0 patients) and 30-day re-admissions (7 vs 2 patients). CONCLUSIONS Patients undergoing elective inpatient surgical procedures with general anesthesia who had a VTE event during their primary hospitalization had a significantly longer length of stay and significantly higher total hospital charges than comparable patients without a VTE event.
Collapse
Affiliation(s)
- Aryana Sepassi
- a Skaggs School of Pharmacy and Pharmaceutical Sciences , University of California , San Diego , La Jolla, CA , USA
| | - Francine Chingcuanco
- b Health Economics and Outcomes Research, Millennium Health , San Diego , CA , USA
| | - Ronald Gordon
- c Department of Anesthesiology , UC San Diego Health System , San Diego , CA , USA
| | - Angela Meier
- c Department of Anesthesiology , UC San Diego Health System , San Diego , CA , USA
| | - Victoria Divino
- d Health Economics and Outcomes Research, IQVIA , Fairfax , VA , USA
| | - Mitch DeKoven
- d Health Economics and Outcomes Research, IQVIA , Fairfax , VA , USA
| | - Rami Ben-Joseph
- b Health Economics and Outcomes Research, Millennium Health , San Diego , CA , USA
| |
Collapse
|
4
|
Garcia MJ, Lookstein R, Malhotra R, Amin A, Blitz LR, Leung DA, Simoni EJ, Soukas PA. Endovascular Management of Deep Vein Thrombosis with Rheolytic Thrombectomy: Final Report of the Prospective Multicenter PEARL (Peripheral Use of AngioJet Rheolytic Thrombectomy with a Variety of Catheter Lengths) Registry. J Vasc Interv Radiol 2015; 26:777-85; quiz 786. [DOI: 10.1016/j.jvir.2015.01.036] [Citation(s) in RCA: 80] [Impact Index Per Article: 8.9] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/20/2014] [Revised: 01/28/2015] [Accepted: 01/29/2015] [Indexed: 11/29/2022] Open
|
5
|
Nelson RE, Grosse SD, Waitzman NJ, Lin J, DuVall SL, Patterson O, Tsai J, Reyes N. Using multiple sources of data for surveillance of postoperative venous thromboembolism among surgical patients treated in Department of Veterans Affairs hospitals, 2005-2010. Thromb Res 2015; 135:636-42. [PMID: 25666908 PMCID: PMC4453876 DOI: 10.1016/j.thromres.2015.01.026] [Citation(s) in RCA: 20] [Impact Index Per Article: 2.2] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/18/2014] [Revised: 01/10/2015] [Accepted: 01/20/2015] [Indexed: 12/11/2022]
Abstract
BACKGROUND There are limitations to using administrative data to identify postoperative venous thromboembolism (VTE). We used a novel approach to quantify postoperative VTE events among Department of Veterans Affairs (VA) surgical patients during 2005-2010. METHODS We used VA administrative data to exclude patients with VTE during 12 months prior to surgery. We identified probable postoperative VTE events within 30 and 90 days post-surgery in three settings: 1) pre-discharge inpatient, using a VTE diagnosis code and a pharmacy record for anticoagulation; 2) post-discharge inpatient, using a VTE diagnosis code followed by a pharmacy record for anticoagulation within 7 days; and 3) outpatient, using a VTE diagnosis code and either anticoagulation or a therapeutic procedure code with natural language processing (NLP) to confirm acute VTE in clinical notes. RESULTS Among 468,515 surgeries without prior VTE, probable VTEs were documented within 30 and 90 days in 3,931 (0.8%) and 5,904 (1.3%), respectively. Of probable VTEs within 30 or 90 days post-surgery, 47.8% and 62.9%, respectively, were diagnosed post-discharge. Among post-discharge VTE diagnoses, 86% resulted in a VA hospital readmission. Fewer than 25% of outpatient records with both VTE diagnoses and anticoagulation prescriptions were confirmed by NLP as acute VTE events. CONCLUSION More than half of postoperative VTE events were diagnosed post-discharge; analyses of surgical discharge records are inadequate to identify postoperative VTE. The NLP results demonstrate that the combination of VTE diagnoses and anticoagulation prescriptions in outpatient administrative records cannot be used to validly identify postoperative VTE events.
Collapse
Affiliation(s)
- Richard E. Nelson
- Veterans Affairs Salt Lake City Health Care System, USA
- University of Utah Department of Internal Medicine, Salt Lake City, UT, USA
| | - Scott D. Grosse
- National Center on Birth Defects and Developmental Disabilities, Office of the Director, Centers for Disease Control and Prevention, Atlanta, GA, USA
| | | | - Junji Lin
- Veterans Affairs Salt Lake City Health Care System, USA
- University of Utah Department of Pharmacotherapy, Salt Lake City, UT, USA
| | - Scott L. DuVall
- Veterans Affairs Salt Lake City Health Care System, USA
- University of Utah Department of Internal Medicine, Salt Lake City, UT, USA
- University of Utah Department of Pharmacotherapy, Salt Lake City, UT, USA
| | - Olga Patterson
- Veterans Affairs Salt Lake City Health Care System, USA
- University of Utah Department of Internal Medicine, Salt Lake City, UT, USA
| | - James Tsai
- Office on Smoking and Health, Centers for Disease Control and Prevention, Atlanta, GA, USA
| | - Nimia Reyes
- National Center on Birth Defects and Developmental Disabilities, Division of Blood Disorders, Centers for Disease Control and Prevention, Atlanta, GA, USA
| |
Collapse
|
6
|
|
7
|
Lyman GH, Eckert L, Wang Y, Wang H, Cohen A. Venous thromboembolism risk in patients with cancer receiving chemotherapy: a real-world analysis. Oncologist 2013; 18:1321-9. [PMID: 24212499 DOI: 10.1634/theoncologist.2013-0226] [Citation(s) in RCA: 137] [Impact Index Per Article: 12.5] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 02/03/2023] Open
Abstract
UNLABELLED The occurrence of malignant disease increases the risk for venous thromboembolism (VTE). Here we evaluate the risk for VTE in a large unselected cohort of patients with cancer receiving chemotherapy. METHODS The United States IMPACT health care claims database was retrospectively analyzed to identify patients with a range of solid tumors who started chemotherapy from January 2005 through December 2008. International Classification of Diseases, 9th revision, Clinical Modification Codes were used to identify cancer location, presence of VTE 3.5 months and 12 months after starting chemotherapy, and incidence of major bleeding complications. Health care costs were assessed one year before initiation of chemotherapy and one year after initiation of chemotherapy. RESULTS The overall incidence of VTE 3.5 months after starting chemotherapy was 7.3% (range 4.6%-11.6% across cancer locations) rising to 13.5% at 12 months (range 9.8%-21.3%). The highest VTE risk was identified in patients with pancreatic, stomach, and lung cancer. Patients in whom VTE developed had a higher risk for major bleeding at 3.5 months and at 12 months (11.0% and 19.8% vs. 3.8% and 9.6%, respectively). Health care costs were significantly higher in patients in whom VTE developed. CONCLUSION Those undergoing chemotherapy as outpatients are at increased risk for VTE and for major bleeding complications. Thromboprophylaxis may be considered for such patients after carefully assessing the risks and benefits of treatment.
Collapse
Affiliation(s)
- Gary H Lyman
- Duke University School of Medicine and the Duke Cancer Institute, Durham, North Carolina, USA
| | | | | | | | | |
Collapse
|
8
|
Boulet SL, Amendah D, Grosse SD, Hooper WC. Health care expenditures associated with venous thromboembolism among children. Thromb Res 2012; 129:583-7. [DOI: 10.1016/j.thromres.2011.08.006] [Citation(s) in RCA: 21] [Impact Index Per Article: 1.8] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/02/2011] [Revised: 08/01/2011] [Accepted: 08/02/2011] [Indexed: 11/27/2022]
|
9
|
Beckman MG, Grosse SD, Kenney KM, Grant AM, Atrash HK. Developing public health surveillance for deep vein thrombosis and pulmonary embolism. Am J Prev Med 2011; 41:S428-34. [PMID: 22099369 DOI: 10.1016/j.amepre.2011.09.011] [Citation(s) in RCA: 11] [Impact Index Per Article: 0.8] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 04/08/2011] [Revised: 08/14/2011] [Accepted: 09/07/2011] [Indexed: 11/15/2022]
Abstract
Deep vein thrombosis (DVT) and pulmonary embolism (PE), collectively known as venous thromboembolism (VTE), are an important and growing public health issue, associated with considerable morbidity and mortality. Presently, there is no national surveillance for DVT and PE. This article provides a summary of an expert workgroup meeting convened January 12, 2010, by the CDC. The purpose of the meeting was to inform CDC on the development of U.S. population-based public health surveillance activities for DVT/PE. Topics discussed included: (1) stakeholders, needs, gaps, and target populations; (2) requirements of surveillance systems; (3) challenges, limitations, and potential barriers to implementation of surveillance activities; and (4) integration of research and education with surveillance activities.
Collapse
Affiliation(s)
- Michele G Beckman
- Division of Blood Disorders, National Center on Birth Defects and Developmental Disabilities, CDC, Atlanta, Georgia 30333, USA.
| | | | | | | | | |
Collapse
|