If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Division of Cardiovascular Diseases and Gonda Vascular Center, Department of Internal Medicine, Mayo Clinic, Rochester, MinnDivision of Epidemiology, Department of Health Sciences Research, Mayo Clinic, Rochester, MinnDivision of Hematology, Department of Internal Medicine, Mayo Clinic, Rochester, Minn
The purpose of this study is to estimate medical costs attributable to venous thromboembolism among patients with active cancer.
Methods
In a population-based cohort study, we used Rochester Epidemiology Project (REP) resources to identify all Olmsted County, Minn. residents with incident venous thromboembolism and active cancer over the 18-year period, 1988-2005 (n = 374). One Olmsted County resident with active cancer without venous thromboembolism was matched to each case on age, sex, cancer diagnosis date, and duration of prior medical history. Subjects were followed forward in REP provider-linked billing data for standardized, inflation-adjusted direct medical costs from 1 year prior to index (venous thromboembolism event date or control-matched date) to the earliest of death, emigration from Olmsted County, or December 31, 2011, with censoring on the shortest follow-up to ensure a similar follow-up duration for each case-control pair. We used generalized linear modeling to predict costs for cases and controls and bootstrapping methods to assess uncertainty and significance of mean adjusted cost differences. Outpatient drug costs were not included in our estimates.
Results
Adjusted mean predicted costs were 1.9-fold higher for cases ($49,351) than for controls ($26,529) (P < .001) from index to up to 5 years post index. Cost differences between cases and controls were greatest within the first 3 months (mean difference = $13,504) and remained significantly higher from 3 months to 5 years post index (mean difference = $12,939).
Conclusions
Venous thromboembolism-attributable costs among patients with active cancer contribute a substantial economic burden and are highest from index to 3 months, but may persist for up to 5 years.
Adjusted mean predicted venous thromboembolism-attributable costs among patients with active cancer from index to 5 years post index are substantial.
•
Venous thromboembolism-attributable costs were greatest within the 3 months after the event date and remained significantly higher from 3 months to 5 years post index.
•
Our findings will inform models that assess the cost-effectiveness of alternative interventions to reduce occurrence and guide reimbursement policy.
Venous thromboembolism is a common complication of active cancer.
Active cancer increases venous thromboembolism risk by four- to sevenfold and accounts for nearly 20% of the entire venous thromboembolism burden occurring in the community.
In addition, patients with active cancer-associated incident venous thromboembolism are at increased risk for recurrent venous thromboembolism, and survival among cancer patients with incident and recurrent venous thromboembolism is significantly reduced.
Rates of initial and recurrent thromboembolic disease among patients with malignancy versus those without malignancy. Risk analysis using Medicare claims data.
Direct medical costs of venous thromboembolism and subsequent hospital readmission rates: an administrative claims analysis from 30 managed care organizations.
Are there any differences in the clinical and economic outcomes between US cancer patients receiving appropriate or inappropriate venous thromboembolism prophylaxis?.
Existing estimates of venous thromboembolism-associated costs among patients with cancer have largely focused on complications of anticoagulation therapy, increased length of hospitalization, and the high frequency of venous thromboembolism recurrence.
Are there any differences in the clinical and economic outcomes between US cancer patients receiving appropriate or inappropriate venous thromboembolism prophylaxis?.
Moreover, venous thromboembolism case ascertainment almost always relied on discharge diagnosis codes obtained from billing or administrative claims data.
Health care costs associated with venous thromboembolism in selected high-risk ambulatory patients with solid tumors undergoing chemotherapy in the United States.
Identifying in-hospital venous thromboembolism (VTE): a comparison of claims-based approaches with the Rochester Epidemiology Project venous thromboembolism cohort.
Health care costs associated with venous thromboembolism in selected high-risk ambulatory patients with solid tumors undergoing chemotherapy in the United States.
To address these limitations, we performed a population-based cohort study to estimate the medical costs attributable to venous thromboembolism in individuals with active cancer that included the entire spectrum of cancer-associated venous thromboembolism occurring in the community.
Methods
Study Setting and Design
Olmsted County, Minn. (2010 census population = 144,248) provides a unique opportunity for investigating the natural history of venous thromboembolism.
Under auspices of the Rochester Epidemiology Project (REP), Mayo Clinic, together with Olmsted Medical Center (OMC) (a second group practice), and their affiliated hospitals, provide over 95% of all medical care delivered to local residents, thereby linking the medical records for community residents at the individual level.
Using REP resources, we performed a cohort study to study cost attributable to venous thromboembolism among cancer patients. The study was approved by the Mayo Clinic and OMC Institutional Review Boards.
Study Population
All Olmsted County, Minn. residents with incident deep venous thrombosis or pulmonary embolism over the 40-year period, 1966-2005, were identified as previously described.
Incident venous thromboembolism events were recorded by experienced nurse abstractors and were limited to patients residing in Olmsted County for whom this was a first lifetime symptomatic venous thromboembolism.
The present study included all incident venous thromboembolism cases with active cancer (excluding nonmelanoma skin cancer). Active cancer had to have been documented in the 92 days (365/4, or about 3 months) prior to venous thromboembolism event date. Cancer was considered inactive when the patient had undergone curative surgery or chemotherapy or radiotherapy with no evidence of residual disease. Myeloproliferative or myelodysplastic disorders, chronic myelocytic or lymphocytic leukemia, and hematopoietic growth factor therapy for these disorders were considered as always-active cancer. For the few patients with multiple primary cancers, we used the cancer in the 92 days on or prior to the incident venous thromboembolism if one was prior to and one was after venous thromboembolism event. We used the more recent cancer if both were prior to the venous thromboembolism. If both primary cancers were diagnosed on the same day, a hematologist/oncologist (AAA) re-staged all cancer(s) and we used the cancer with the highest stage.
The Mayo Cancer Registry, available since 1972, includes patient demographics at cancer diagnosis and tumor classification using International Classification of Diseases for Oncology, 3rd edition, and also provides enumeration of the Olmsted County population with cancer from 1973 to the present, from which controls can be sampled.
After verifying consent to use of medical records for research and Olmsted County residency, the list of possible cancer controls for each venous thromboembolism case was subset to those Olmsted County residents with cancer whose first cancer diagnosis was within ±5 years of the venous thromboembolism case's cancer diagnosis (Figure).
We further matched on sex, date of birth (±5 years), and year of registration (±5 years). Matching on year of registration ensures a similar duration of medical records. For each case, the list of possible controls was randomly sorted and a control medical visit date after January 1, 1988 was chosen (index date). The control's cancer was confirmed to be active within ±3 months of the index date, and the duration of active cancer to be at least as long as or up to 2 years longer than the duration of active cancer of the case. Medical records were also reviewed to confirm no history of venous thromboembolism prior to or within 3 months after the index date.
FigureFlow diagram on the identification and selection of an Olmsted County, Minn. resident with active cancer and no venous thromboembolism (controls) matched to each Olmsted County resident with active cancer-associated incident venous thromboembolism, 1988-2005, on age, sex, and duration of active cancer.
Through an electronic data-sharing agreement between Mayo Clinic and OMC, patient-level administrative data on health care utilization and associated billed charges incurred at these institutions are shared and archived within the REP Cost Data Warehouse for use in approved research studies. Data are electronically linked, affording complete information on all hospital and ambulatory care delivered by these providers to area residents from January 1, 1987 through December 31, 2011. The REP Cost Data Warehouse includes information on all Olmsted County residents (ie, both sexes, all ages, and all payer types, including the uninsured) and contains line-item detail on date, type, frequency, and billed charge for every good or service provided; long-term care, indirect, and outpatient pharmaceutical costs are not included. Recognizing discrepancies between billed charges and true resource use, the REP Cost Data Warehouse employs widely accepted valuation techniques to generate a standardized inflation-adjusted estimate of the costs of each service or procedure in constant dollars. Cost estimates in this study were adjusted to 2013 dollars.
Because cost data are only available electronically since 1987 and we wished to obtain costs in the year prior to index, the present study was limited to all Olmsted County case-control pairs whose index dates occurred between 1988 and 2005.
Each case and control was followed forward in time for costs from 1 year prior to their respective index date to earliest of death, emigration from Olmsted County, conversion to venous thromboembolism case (controls only), or December 31, 2011 (study end date). We ensured similar periods of observation for each case and matched control by censoring both members of each pair at the shortest length of follow-up for either member.
Preindex Comorbid Conditions
To compare index comorbidities between cases and controls, we obtained all International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnoses codes assigned to each individual in REP Cost Data Warehouse 1 year prior to index and categorized every diagnosis code assigned each individual into the 17 ICD-9-CM chapters and 114 subchapters. A summary measure of comorbid medical conditions in the year prior to index was also obtained using Johns Hopkins Adjusted Clinical Groups (ACG) System software (Johns Hopkins University, Baltimore, Md).
The Johns Hopkins ACG® System [Internet]. Baltimore: Johns Hopkins Bloomberg School of Public Health; 2016 Available at: http://www.acg.jhsph.org/. Accessed April 07, 2016.
ACG software categorizes individuals' diagnosis codes into groupings based on persistence, severity, and etiology of the condition, as well as diagnostic certainty and need for specialty care.
The Johns Hopkins ACG® System [Internet]. Baltimore: Johns Hopkins Bloomberg School of Public Health; 2016 Available at: http://www.acg.jhsph.org/. Accessed April 07, 2016.
ACG software was used to assign a Resource Utilization Band (RUB) value to each individual. RUB categories are aggregations of ACGs that have similar expected resource use, with values ranging from 0 (no relevant diagnosis codes) to 5 (diagnosis codes associated with very high use).
The Johns Hopkins ACG® System: Excerpt from Technical Reference Guide, Version 9.0 [Internet]. Baltimore: Johns Hopkins Bloomberg School of Public Health; December 2009. Available at: http://www.healthpartners.com/ucm/groups/public/@hp/@public/documents/documents/dev_057914.pdf. Accessed April 07, 2016.
Statistical testing used the 2-tailed alpha level of 0.05. The principal outcome was direct medical costs associated with venous thromboembolism. We adjusted for costs from 1 year prior to index, and analyzed costs from index to a maximum of 5 years post index. For each subdivided postindex period, analyses were limited to those who were eligible for costs at the start of each interval. Postindex analyses were subdivided into: index-3 months, 3-6 months, 6 months-1 year, 1-2 years, 2-3 years, 3-4 years, and 4-5 years. Models from post index to 5 years and 3 months to 5 years included length of follow-up from index. In initial analyses, the unadjusted costs for each control were subtracted from costs for its paired case in each time period; statistical significance was assessed using Wilcoxon signed-rank test to account for the highly skewed nature of cost data.
To isolate the costs attributable to venous thromboembolism, we used general linear multivariate modeling to examine the extent to which age, sex, RUB measure of preindex comorbidity, cancer type (14 cancer types were compared with a reference group consisting of head and neck [2.3%], liver [1.2%], lung [12.8%], bone [0.1%], skin [1.9%], other genitourinary [0.3%], myeloproliferative syndromes [1.2%], myelodysplastic syndromes [0.1%], other [1.6%], and unknown 4 [0.5%] cancer types) and stage (continuous variable), and preindex costs accounted for postindex cost differences between cases and controls. This adjusted approach employed 2-part models to account for zero costs
when appropriate, and incorporated a generalized linear model with family distribution based on the modified Park test recommended by Manning and Mullahy.
This analytic approach accounts for the skewed cost distribution while enabling coefficients to be directly back-transformed into the original dollar scale.
We analyzed differences in costs between cases and controls using the method of recycled predictions, setting all individuals as cases with venous thromboembolism or as controls without venous thromboembolism, while all other individual characteristics remain as observed.
Mean values and bootstrapped 95% confidence intervals of the mean difference were calculated. Analyses were conducted in SAS version 9.02 (SAS Institute, Cary, NC).
Results
Demographic and Clinical Characteristics at Index
We identified 374 venous thromboembolism cases and matched controls, both with active cancer. The mean ± SD (median; range) patient age for cases and controls was 65 ± 15 (66; 2-96) and 65 ± 15 (67; 1-95) years, respectively (P = .72), and 48% of case/control pairs were female. The venous thromboembolism event type distribution was deep vein thrombosis alone (n = 260; 70%), pulmonary embolism alone (n = 83; 22%), and pulmonary embolism with deep vein thrombosis (n = 31; 8%). The median (interquartile range [IQR]) duration of follow-up post index was 143 (43, 561) days and ranged from 1 day to 17 years. Cancer stage included those with cancer in situ (stage 0) to metastases (stage 4). Both the cancer type distribution and the cancer stage distribution differed significantly among cases and controls (P <.001 for both; Table 1). In the year prior to index, significant differences between cases and controls were observed in 10 of 17 ICD-9-CM chapters (Table 2). The RUB summary measure of preindex comorbidity also differed significantly for cases compared with controls (P <.001). In the year prior to index, 59% (n = 221) of cases had a RUB value indicative of very high resource utilization compared with 39% (n = 145) of controls.
Table 1Cancer Type and Stage Distributions Among Olmsted County, Minn with Active Cancer-Associated Incident Venous Thromboembolism, 1988-2005 (Cases), and Among Matched Olmsted County Residents with Active Cancer and No Venous Thromboembolism (Controls)
Table 2Baseline Patient Characteristics and Comorbidities by International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) Code
Unadjusted mean, median (IQR), minimum and maximum direct medical costs for case/control pairs, and mean cost differences between case/control pairs from 1 year prior to index, all 5 years post index, and selected periods within the 5-year postindex interval are provided in the Supplementary Table (available online). During the period index to 5 years post index, 3 matched pairs had 1 member (1 case and 2 controls) who did not accrue any costs even though they were eligible for costs after their index date (alive and in Olmsted County) and so have zero costs. Three pairs (0.8%) did not incur costs due to zero costs in the first interval, index to 3 months. In the year prior to index, both mean and median costs for cases were slightly higher compared with controls. The unadjusted mean difference in preindex annual costs between cases and controls was $17,915 (95% CI, $12,990-$23,538).
Adjusted Cost Comparisons
Adjusted mean predicted direct medical costs for cases and controls, and the adjusted predicted cost difference for the overall time period index to 5 years, and for intervals within that period, are shown in Table 3. After adjusting for group differences in age at index, sex, costs incurred 1 year prior to index, cancer type and stage, and preindex RUB values, the mean predicted costs for cases ($49,351) were significantly higher than those for controls ($26,529), with a mean predicted difference of $22,822 (bootstrapped 95% CI, $14,554-31,472), as compared with the unadjusted difference of $27,164. The adjusted mean cost was significantly higher for venous thromboembolism cases than controls for index to 3 months, and 6 months to 1 year post index. For the time period 3 months to 5 years, the adjusted mean predicted cost for cases ($42,720) was significantly higher than that for controls ($29,781), with a mean predicted difference of $12,939 (bootstrapped 95% CI, $2,675-$23,881).
To further explore potential causes for the observed difference in adjusted mean cost, we compared the distribution of postindex location of medical care (hospital inpatient, hospital outpatient, emergency department, and ambulatory setting) among cases and controls (Table 3). For time period index to 1 year, cases used significantly more hospital inpatient and emergency department care compared with controls, and marginally more hospital outpatient care. In contrast, controls used significantly more ambulatory setting care compared with cases, although the difference was only 7%.
Discussion
This population-based cohort study was conducted due to a shortage of reliable data regarding the extent to which cancer-associated venous thromboembolism contributes to excess medical costs and for how long any observed excess costs occur. The adjusted predicted mean direct medical costs were significantly higher for cases than for controls from index to 5 years post index; the adjusted predicted mean cost for venous thromboembolism cases was 1.9-fold higher ($49,351) compared with controls ($26,529; mean difference $22,822). Venous thromboembolism-attributable costs were highest for the period index to 3 months after index.
There are very few studies of venous thromboembolism-attributable costs among patients with active cancer. In a study of claims data from adult patients undergoing chemotherapy for selected common high-risk solid tumors (lung, colorectal, pancreatic, gastric, bladder, or ovarian), the all-cause total health care costs over a 12-month period among patients with a diagnosis code for venous thromboembolism and controls matched on cancer site and propensity score were $74,959 and $41,691, respectively (difference = $33,268).
Health care costs associated with venous thromboembolism in selected high-risk ambulatory patients with solid tumors undergoing chemotherapy in the United States.
Due to marked differences in study populations, data sources, methods, and length of follow-up, the absolute difference in costs between cases and controls in this study cannot be compared with those in our study, but the costs are consistently higher for venous thromboembolism cases in both studies. Using methods similar to the present study, the costs of venous thromboembolism related to active cancer exceeded costs of venous thromboembolism related to hospitalization for major surgery but were less than costs of venous thromboembolism related to those hospitalized for acute medical illness.
The observed increased costs for cases compared with controls within 5 years after index could reflect incremental costs for management of venous thromboembolism complications and venous thromboembolism recurrence. Over the full 5-year postindex time period, 92 (25%) of 374 cases had recurrent venous thromboembolism with a median time to recurrence of 76 days (IQR 22.5 days-299 days). From index to 5 years post index, the adjusted predicted mean cost for these 92 cases was $86,638, vs $38,835 for their matched controls (mean difference $47,803: 95% CI, $23,236-$78,250). The adjusted predicted mean cost of the 282 cases without recurrent venous thromboembolism was $37,466, compared with $22,407 for their matched controls (mean difference $15,059: 95% CI, $8,637-$21,628). Survival after the active cancer-associated incident venous thromboembolism did not differ significantly among those with and without recurrent venous thromboembolism (log rank test P = .22). In the year post index, cases differed significantly from controls in the distribution of location of medical care, with cases using significantly more hospital inpatient and emergency department care. Possible explanations include diagnosis and management of acute venous thromboembolism and venous thromboembolism complications, and differences in the management of cancer, cancer complications, and comorbidities among cases and controls.
Our study has important limitations. The cost estimates are for a single geographic population, which was 83% white in 2010. While no single geographic area is representative of all others, the under-representation of minorities may compromise the generalizability of our findings to different racial/ethnic groups. While costs associated with medications were not included in this analysis, the incremental costs of venous thromboembolism treatment likely would increase the cost difference between cases and controls. Cost estimates were limited to direct medical care costs and did not include indirect or long-term care costs. Finally, while we adjusted for age, sex, costs in the year prior to index, cancer type and stage, and RUB, we cannot exclude that some of the observed cost difference was due to incomplete adjustment for comorbidities and other unmeasured covariates.
In conclusion, venous thromboembolism contributes a substantial economic burden to patients with active cancer. Our findings will inform models that assess the cost-effectiveness of alternative interventions to reduce venous thromboembolism occurrence and guide reimbursement policy.
Acknowledgments
We gratefully acknowledge Catherine L. Brandel, RN, Diadra H. Else, RN, Jane A. Emerson, RN, and Cynthia L. Nosek, RN for excellent data collection and Cynthia E. Regnier, RN, as research project manager.
Supplementary Material
Supplementary TableUnadjusted Direct Medical Costs Attributable to Cancer Associated Venous Thromboembolism Case/Control Pairs by Time Period
Note: 3 matched pairs (1 case and 2 controls) had 1 member who did not accrue any costs even though they were eligible for costs after their index date (alive and in Olmsted County). Because of the selection of minimum follow-up of either member of the pair for the follow up of the pair they did not accrue any costs in the minimum follow-up time. The 1 case with no costs was followed 7 days after index date. The 2 controls with no costs were followed 3 and 4 days after index date.
Rates of initial and recurrent thromboembolic disease among patients with malignancy versus those without malignancy. Risk analysis using Medicare claims data.
Direct medical costs of venous thromboembolism and subsequent hospital readmission rates: an administrative claims analysis from 30 managed care organizations.
Are there any differences in the clinical and economic outcomes between US cancer patients receiving appropriate or inappropriate venous thromboembolism prophylaxis?.
Health care costs associated with venous thromboembolism in selected high-risk ambulatory patients with solid tumors undergoing chemotherapy in the United States.
Identifying in-hospital venous thromboembolism (VTE): a comparison of claims-based approaches with the Rochester Epidemiology Project venous thromboembolism cohort.
The Johns Hopkins ACG® System [Internet]. Baltimore: Johns Hopkins Bloomberg School of Public Health; 2016 Available at: http://www.acg.jhsph.org/. Accessed April 07, 2016.
The Johns Hopkins ACG® System: Excerpt from Technical Reference Guide, Version 9.0 [Internet]. Baltimore: Johns Hopkins Bloomberg School of Public Health; December 2009. Available at: http://www.healthpartners.com/ucm/groups/public/@hp/@public/documents/documents/dev_057914.pdf. Accessed April 07, 2016.
Funding: Research reported in this publication was supported in part by grants from the National Heart, Lung, and Blood Institute under Award Numbers R01HL66216 and K12HL83141 (a training grant in Vascular Medicine [KPC]) to JAH, and was made possible by the Rochester Epidemiology Project (Award Number R01AG034676 of the National Institute on Aging, National Institutes of Health). Research support also was provided by Mayo Foundation. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Conflict of Interest: None.
Authorship: All authors had a role in writing the manuscript and had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analyses.
30287-X&id=gr1.jpg){kind=link}