- Email: [email protected]
244 Emergency Department as Usual Source of Care in a Nationally Representative Sample of American Households
Research Forum Abstracts
243
Cost of Medical Care for Low-Risk Emergency Department Patients With Venous Thromboembolism: Comparison of Usual Care With Warfarin versus Home Treatment With Rivaroxaban
Kahler ZP, Beam DM, Kline JA/Indiana University, Indianapolis, IN
Background: Target specific anticoagulants (TSAs) such as rivaroxaban have been recently approved for outpatient treatment of venous thromboembolism (VTE, including DVT and PE), allowing clinicians to immediately discharge patients from the emergency department (ED) after diagnosis of low-risk cases. Study Objectives: To determine the medical costs accrued during the first week and then six months after diagnosis of low risk VTE treated with usual care compared with a TSA based home treatment protocol. Methods: This case-control study quantified the total hospital system costs of patients diagnosed with low-risk VTE (as defined by a modified version of the Hestia criteria) at three hospitals. All patients had image-confirmed VTE and were anticoagulated for six months. Control patients were treated with usual care and in accordance with published guidelines, including low molecular weight heparin then warfarin followed by discharge as soon as was practicable. Case patients were treated with an initial dose of rivaroxaban in the ED followed by same-day discharge home with a prescription for rivaroxaban. Medians were compared with Mann-Whitney U test (MWU). Hospital system costs were estimated from UB04 billing data. This study received IRB exemption. Results: We have identified 50 cases and 47 controls. Groups were well matched according to mean age, and proportions by sex and location of thrombus. For all VTE, median cost of medical care for the first week after diagnosis was $8,080 (IQR $4,364$16,388) for controls, compared with $2,854 (IQR $2,498-$4,723) for cases, MWU P<.001. Median cost of care for six months after diagnosis was $11,128 (IQR $8,110$23,390) for controls compared with $4,787 (IQR $3,042-$7,596) for cases, MWU P<.001. Subgroup analyses based upon clot location (PE or isolated DVT) retained significance, with costs for case PE patients 57% lower than control PE patients (P<.001) and 56% lower for DVT patients (P<.003). Conclusion: Cost of medical care was significantly lower for low-risk PE and DVT patients discharged immediately from the ED with rivaroxaban therapy compared to patients treated with low molecular weight heparin and warfarin.
244
Emergency Department as Usual Source of Care in a Nationally Representative Sample of American Households
Janke AT, Brody A, Overbeek DL, Bedford JC, Levy PD/Wayne State University School of Medicine, Detroit, MI
Study Objectives: To characterize the population of community-dwelling Americans who report the emergency department (ED) as their usual source of care (USC), and to evaluate the potential effect of this on health care utilization patterns. Methods: Publicly available data from the 2010 Health Tracking Household Study, a nationally representative survey of 16,671 community-dwelling Americans, were accessed and patients were grouped by reported USC (ED, non-ED, none). Multivariable logistic regression analysis adjusted for complex survey design was used to evaluate predictors of ED as USC. Pearson’s chi-squared tests and univariable logistic regression were used to evaluate group differences and potential effects of USC on selfreported health care utilization patterns. Results: Approximately 2.5% (95% confidence interval [CI] 2.1% to 2.9%) of community-dwelling Americans (n¼349) considered the ED as USC in 2010. On multivariable logistic regression (Table), household poverty (defined by federal poverty line in 2010), high school education or below, Medicare, Medicaid, lack of insurance, poor self-reported health, African American race, non-married household status, residence in metropolitan (versus rural) regions, and living in the Midwest or South were independent predictors of ED as USC. Our model yielded a c-statistic of 0.81 and a F-statistic of 1.23 (P¼.2701) in Hosmer–Lemeshow goodness-of-fit test, suggesting adequate fit. Among those who reported the ED as USC, 50.2% (95% CI 42.7% to 57.7%) used the ED at least once in the preceding year (versus 20.8% for those without a USC and 23.7% for non-ED USC; P<.00001) and 10.5% (95% CI 6.9% to 15.7%) used the ED four or more times in the preceding year (versus 2.3% for those without a USC and 2.0% with non-ED USC; P<.00001); 31.6% (95% CI 25.6% to 38.3%) of this group reported putting off needed medical care, as compared to 21.5% among those with no USC and 17.9% among those with some other USC (P<.00001). In univariable logistic regression with non-ED USC as the base category,
Volume 64, no. 4s : October 2014
ED as USC status was predictive of delaying needed medical care (odds ratio [OR] 2.11 [95% CI 1.56 to 2.87]). Conclusion: While few reported the ED as USC in 2010, this group is more likely than those with other USC or no USC to utilize the ED frequently, and report delaying medical care. Low-income, underinsured African Americans are more likely to report the ED as USC, suggesting the need for future initiatives aimed at improving care coordination for this vulnerable patient group.
Table. Logistic Regression Model Emergency Department as Usual Source of Care Odds Ratio Age (Reference: <18) 18-24 25-44 45-64 65+ Sex (Reference: Female) Male Race (Reference: Non-Hispanic White) Non-Hispanic African American Non-Hispanic Other Hispanic Household Poverty (Reference: Above Poverty) Yes Education (Reference: > High School) < High School High School Graduate Metro Area (Reference: Rural) Large (>200,000 people) Small (<100,000 people) Region (Reference: Northeast) Midwest South West Insurance Status (Reference: Private, Other) Medicare Medicaid No Insurance Unmarried Household (Reference: Married) Yes, Unmarried Self-Reported Health (Reference: Good) Fair/Poor Constant N (unweighted)
[95% Confidence Interval]
1.33 1.46 1.16 0.53
[0.71 to 2.48] [0.83 to 2.58] [0.67 to 1.99] [0.24 to 1.18]
1.17
[0.87 to 1.58]
2.36 2.32 0.88
[1.60 to 3.49] [1.32 to 4.08] [0.50 to 1.55]
1.36
[0.95 to 1.95]
1.40 1.25
[0.86 to 2.27] [0.84 to 1.84]
1.55 2.18
[1.01 to 2.38] [1.15 to 4.13]
1.82 2.03 1.03
[1.06 to 3.15] [1.23 to 3.33] [0.57 to 1.85]
2.90 4.37 5.19
[1.61 to 5.23] [2.35 to 8.14] [3.12 to 8.62]
1.54
[1.09 to 2.16]
1.71 0.002 3.49
[1.02 to 2.45] [0.000 to 0.003] 16,322
This table gives results from an adjusted multivariable logistic regression model using data from the 2010 Health Tracking Household Survey, a nationally representative survey of 16,671 community-dwelling Americans. Model yielded a o-statistic of 0.81 and a F-statistic of 1.23 (P¼0.2701) in Hosmer–Lemeshow goodness-of-fit test, suggesting adequate fit.
245
Association of Missed Outpatient Appointments and Emergency Department Visits and Inpatient Admissions
Ankeny A, Isenberger K, Westgard B, Stuck L, Wewerka S/Regions Hospital, Saint Paul, MN; HealthPartners, Saint Paul, MN
Study Objectives: Health care cost controls have focused on improving primary care access to avoid preventable and costly emergency department (ED) and inpatient (IP) admissions. Research on missed outpatient appointments has focused on ED and hospital admission follow-up rather than focusing on missed appointments as potentially leading to or being associated with ED and hospital admissions. One objective of this study is to describe the association of missed outpatient appointments with ED and IP stays while adjusting for demographic, payer type, and health status. A
Annals of Emergency Medicine S87
243
Cost of Medical Care for Low-Risk Emergency Department Patients With Venous Thromboembolism: Comparison of Usual Care With Warfarin versus Home Treatment With Rivaroxaban
Kahler ZP, Beam DM, Kline JA/Indiana University, Indianapolis, IN
Background: Target specific anticoagulants (TSAs) such as rivaroxaban have been recently approved for outpatient treatment of venous thromboembolism (VTE, including DVT and PE), allowing clinicians to immediately discharge patients from the emergency department (ED) after diagnosis of low-risk cases. Study Objectives: To determine the medical costs accrued during the first week and then six months after diagnosis of low risk VTE treated with usual care compared with a TSA based home treatment protocol. Methods: This case-control study quantified the total hospital system costs of patients diagnosed with low-risk VTE (as defined by a modified version of the Hestia criteria) at three hospitals. All patients had image-confirmed VTE and were anticoagulated for six months. Control patients were treated with usual care and in accordance with published guidelines, including low molecular weight heparin then warfarin followed by discharge as soon as was practicable. Case patients were treated with an initial dose of rivaroxaban in the ED followed by same-day discharge home with a prescription for rivaroxaban. Medians were compared with Mann-Whitney U test (MWU). Hospital system costs were estimated from UB04 billing data. This study received IRB exemption. Results: We have identified 50 cases and 47 controls. Groups were well matched according to mean age, and proportions by sex and location of thrombus. For all VTE, median cost of medical care for the first week after diagnosis was $8,080 (IQR $4,364$16,388) for controls, compared with $2,854 (IQR $2,498-$4,723) for cases, MWU P<.001. Median cost of care for six months after diagnosis was $11,128 (IQR $8,110$23,390) for controls compared with $4,787 (IQR $3,042-$7,596) for cases, MWU P<.001. Subgroup analyses based upon clot location (PE or isolated DVT) retained significance, with costs for case PE patients 57% lower than control PE patients (P<.001) and 56% lower for DVT patients (P<.003). Conclusion: Cost of medical care was significantly lower for low-risk PE and DVT patients discharged immediately from the ED with rivaroxaban therapy compared to patients treated with low molecular weight heparin and warfarin.
244
Emergency Department as Usual Source of Care in a Nationally Representative Sample of American Households
Janke AT, Brody A, Overbeek DL, Bedford JC, Levy PD/Wayne State University School of Medicine, Detroit, MI
Study Objectives: To characterize the population of community-dwelling Americans who report the emergency department (ED) as their usual source of care (USC), and to evaluate the potential effect of this on health care utilization patterns. Methods: Publicly available data from the 2010 Health Tracking Household Study, a nationally representative survey of 16,671 community-dwelling Americans, were accessed and patients were grouped by reported USC (ED, non-ED, none). Multivariable logistic regression analysis adjusted for complex survey design was used to evaluate predictors of ED as USC. Pearson’s chi-squared tests and univariable logistic regression were used to evaluate group differences and potential effects of USC on selfreported health care utilization patterns. Results: Approximately 2.5% (95% confidence interval [CI] 2.1% to 2.9%) of community-dwelling Americans (n¼349) considered the ED as USC in 2010. On multivariable logistic regression (Table), household poverty (defined by federal poverty line in 2010), high school education or below, Medicare, Medicaid, lack of insurance, poor self-reported health, African American race, non-married household status, residence in metropolitan (versus rural) regions, and living in the Midwest or South were independent predictors of ED as USC. Our model yielded a c-statistic of 0.81 and a F-statistic of 1.23 (P¼.2701) in Hosmer–Lemeshow goodness-of-fit test, suggesting adequate fit. Among those who reported the ED as USC, 50.2% (95% CI 42.7% to 57.7%) used the ED at least once in the preceding year (versus 20.8% for those without a USC and 23.7% for non-ED USC; P<.00001) and 10.5% (95% CI 6.9% to 15.7%) used the ED four or more times in the preceding year (versus 2.3% for those without a USC and 2.0% with non-ED USC; P<.00001); 31.6% (95% CI 25.6% to 38.3%) of this group reported putting off needed medical care, as compared to 21.5% among those with no USC and 17.9% among those with some other USC (P<.00001). In univariable logistic regression with non-ED USC as the base category,
Volume 64, no. 4s : October 2014
ED as USC status was predictive of delaying needed medical care (odds ratio [OR] 2.11 [95% CI 1.56 to 2.87]). Conclusion: While few reported the ED as USC in 2010, this group is more likely than those with other USC or no USC to utilize the ED frequently, and report delaying medical care. Low-income, underinsured African Americans are more likely to report the ED as USC, suggesting the need for future initiatives aimed at improving care coordination for this vulnerable patient group.
Table. Logistic Regression Model Emergency Department as Usual Source of Care Odds Ratio Age (Reference: <18) 18-24 25-44 45-64 65+ Sex (Reference: Female) Male Race (Reference: Non-Hispanic White) Non-Hispanic African American Non-Hispanic Other Hispanic Household Poverty (Reference: Above Poverty) Yes Education (Reference: > High School) < High School High School Graduate Metro Area (Reference: Rural) Large (>200,000 people) Small (<100,000 people) Region (Reference: Northeast) Midwest South West Insurance Status (Reference: Private, Other) Medicare Medicaid No Insurance Unmarried Household (Reference: Married) Yes, Unmarried Self-Reported Health (Reference: Good) Fair/Poor Constant N (unweighted)
[95% Confidence Interval]
1.33 1.46 1.16 0.53
[0.71 to 2.48] [0.83 to 2.58] [0.67 to 1.99] [0.24 to 1.18]
1.17
[0.87 to 1.58]
2.36 2.32 0.88
[1.60 to 3.49] [1.32 to 4.08] [0.50 to 1.55]
1.36
[0.95 to 1.95]
1.40 1.25
[0.86 to 2.27] [0.84 to 1.84]
1.55 2.18
[1.01 to 2.38] [1.15 to 4.13]
1.82 2.03 1.03
[1.06 to 3.15] [1.23 to 3.33] [0.57 to 1.85]
2.90 4.37 5.19
[1.61 to 5.23] [2.35 to 8.14] [3.12 to 8.62]
1.54
[1.09 to 2.16]
1.71 0.002 3.49
[1.02 to 2.45] [0.000 to 0.003] 16,322
This table gives results from an adjusted multivariable logistic regression model using data from the 2010 Health Tracking Household Survey, a nationally representative survey of 16,671 community-dwelling Americans. Model yielded a o-statistic of 0.81 and a F-statistic of 1.23 (P¼0.2701) in Hosmer–Lemeshow goodness-of-fit test, suggesting adequate fit.
245
Association of Missed Outpatient Appointments and Emergency Department Visits and Inpatient Admissions
Ankeny A, Isenberger K, Westgard B, Stuck L, Wewerka S/Regions Hospital, Saint Paul, MN; HealthPartners, Saint Paul, MN
Study Objectives: Health care cost controls have focused on improving primary care access to avoid preventable and costly emergency department (ED) and inpatient (IP) admissions. Research on missed outpatient appointments has focused on ED and hospital admission follow-up rather than focusing on missed appointments as potentially leading to or being associated with ED and hospital admissions. One objective of this study is to describe the association of missed outpatient appointments with ED and IP stays while adjusting for demographic, payer type, and health status. A
Annals of Emergency Medicine S87