ORIGINAL ARTICLE
Magnitude of Anemia at Discharge Increases 30-Day Hospital Readmissions Colleen G. Koch, MD, MS, MBA,*† Liang Li, PhD,‡ Zhiyuan Sun, MS,‡ Eric D. Hixson, PhD,§ Anne Tang, MS,‡ Kevin Chagin, MS,‡ Michael Kattan, PhD,†‡ Shannon C. Phillips, MD,† Eugene H. Blackstone, MD,‡|| and J. Michael Henderson, MD†¶ Background: Anemia during hospitalization is associated with poor health outcomes. Does anemia at discharge place patients at risk for hospital readmission within 30 days of discharge? Our objectives were to examine the prevalence and magnitude of anemia at hospital discharge and determine whether anemia at discharge was associated with 30-day readmissions among a cohort of hospitalizations in a single health care system. Methods: From January 1, 2009, to August 31, 2011, there were 152,757 eligible hospitalizations within a single health care system. The endpoint was any hospitalization within 30 days of discharge. The University HealthSystem Consortium's clinical database was used for demographics and comorbidities; hemoglobin values are from the hospitals' electronic medical records, and readmission status was obtained from the University HealthSystem Consortium administrative data systems. Mild anemia was defined as hemoglobin of greater than 11 to less than 12 g/dl in women and greater than 11 to less than 13 g/dl in men; moderate, greater than 9 to less than or equal to 11 g/dl; and severe, less than or equal to 9 g/dl. Logistic regression was used to assess the association of anemia and 30-day readmissions adjusted for demographics, comorbidity, and hospitalization type. Results: Among 152,757 hospitalizations, 72% of patients were discharged with anemia: 31,903 (21%), mild; 52,971 (35%), moderate; and 25,522 (17%), severe. Discharge anemia was associated with severity-dependent increased odds for 30-day hospital readmission compared with those without anemia: for mild anemia, 1.74 (1.65–1.82); moderate anemia, 2.76 (2.64–2.89); and severe anemia, 3.47 (3.30–3.65), P < 0.001. Conclusions: Anemia at discharge is associated with a severitydependent increased risk for 30-day readmission. A strategy focusing on anemia treatment care paths during index hospitalization offers an opportunity to influence subsequent readmissions. Key Words: 30-day hospital readmissions, anemia, resource use, mortality (J Patient Saf 2014;00: 00–00)
A
pproximately 19% of Medicare beneficiaries discharged from the hospital return within 30 days of discharge, at an annual estimated cost in the billions of dollars.1 Recently, a hospital readmission prediction score reported that a number of factors including lower hemoglobin values at discharge were associated with increased 30-day readmissions.2 We have previously addressed the role of present-on-admission anemia and the development of hospital-acquired anemia on health outcomes and resource use in our health system. We found a risk-adjusted increase in hospital From the *Department of Cardiothoracic Anesthesia, Heart and Vascular Institute, †Quality and Patient Safety Institute, ‡Department of Quantitative Health Sciences, Research Institute, §Business Intelligence, Medical Operations, ||Department of Thoracic and Cardiovascular Surgery, Heart and Vascular Institute, and ¶Department of General Surgery, Digestive Disease Institute, Cleveland Clinic, Cleveland, Ohio. Correspondence: Colleen G. Koch, MD, MS, MBA, Department of Cardiothoracic Anesthesia, Heart and Vascular Institute, Cleveland Clinic, 9500 Euclid Ave/Mail Stop J-4, Cleveland, OH 44195 (e‐mail: [email protected]). The authors disclose no conflict of interest. Copyright © 2014 by Lippincott Williams & Wilkins
J Patient Saf • Volume 00, Number 00, Month 2014
mortality, length of hospital stay, and total charges among anemic patients.3 It was unclear whether this heightened risk continued to the postdischarge setting by placing our patients at risk for hospital readmission within 30 days of discharge. Hence, our objectives were to (1) examine the prevalence and magnitude of anemia at the time of hospital discharge as well as (2) determine whether anemia at discharge was associated with 30-day readmissions among a large cohort of hospitalizations in a quaternary referral health care system.
METHODS Patient Population From January 1, 2009, to August 31, 2011, there were 417,301 hospitalizations within a single health care system consisting of 1 quaternary academic medical center (Cleveland Clinic) and 9 community hospitals. Exclusion criteria included hospitals without available electronic laboratory data, patients 18 years or younger, and hospitalizations without hemoglobin values recorded. We retained only the first hospitalization with readmission or the first hospitalization for patients who did not have a 30-day readmission. The final data set consisted of 152,757 hospitalizations (Fig. 1).
Anemia Definitions Presence and magnitude of anemia at hospital discharge were determined by the last available hemoglobin value before discharge. The World Health Organization defines anemia as a hemoglobin value of less than 12 g/dl in women and less than 13 g/dl in men.4 We further grouped hemoglobin by degree of severity into mild anemia, defined as hemoglobin of greater than 11 to less than 12 g/dl in women and greater than 11 to less than 13 g/dl in men; moderate anemia, defined as hemoglobin of greater than 9 to less than or equal to 11 g/dl; and severe anemia, defined as hemoglobin of less than or equal to 9 g/dl.
Endpoint The endpoint was hospital readmission to the health care system within 30 days of hospital discharge.
Data Patient demographics and baseline comorbidities were extracted from the clinical database/resource manager of the University HealthSystem Consortium (UHC). The UHC is composed of 119 U.S. academic medical centers and their 272 affiliated hospitals. It represents more than 90% of nonprofit academic medical centers in the country. Data were originally retrieved from the hospitals' administrative data systems, normalized according to UHC standardized data specifications, and submitted for inclusion in the national repository. Data quality was assessed using standardized error checking and data completeness algorithms and was required to meet established minimum thresholds to be www.journalpatientsafety.com
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
1
J Patient Saf • Volume 00, Number 00, Month 2014
Koch et al
FIGURE 1. Consolidated Standards of Reporting Trials (CONSORT)-style diagram of eligible hospitalizations within a single health care system.
included in the national repository. Measured hemoglobin values were taken from the hospitals' electronic medical records from complete blood count testing. The data source for hospital readmission status was obtained from Cleveland Clinic Health System administrative data systems. The Cleveland Clinic Institutional Review Board approved this investigation, with individual patient consent waived.
Statistical Analysis The primary analysis examined the association between the presence and magnitude (mild, moderate, and severe) of anemia at discharge and hospital readmissions within 30 days of discharge. Baseline risk factors and readmission status were summarized and compared between groups using frequencies, percentages, and χ2 tests for categorical variables as well as median, 25th and 75th quartiles, and the Kruskal-Wallis test for continuous variables. Logistic regression was used to assess the association between demographics, comorbidity, presence and magnitude of anemia, and hospital readmission. We used a comprehensive set of comorbidity measurements to control for the patients' underlying conditions in addition to discharge anemia groupings, age, race,
2
www.journalpatientsafety.com
sex, and Medicare severity diagnosis–related group (MS-DRG) status (surgery and medical status).5 Unless otherwise specified, a P value of less than 0.05 was considered statistically significant. All analyses were performed using SAS version 9.2 (SAS, Inc, Cary, NC).
RESULTS Among 152,757 hospitalizations, 72% of the patients were discharged with anemia: 31,903 (21%) with mild anemia, 52,971 (35%) with moderate anemia, and 25,522 (17%) with TABLE 1. Magnitude of Anemia at Hospital Discharge Determined by the Last Available Hemoglobin Value Discharge Anemia Groupings
n (%)
No discharge anemia Mild discharge anemia Moderate discharge anemia Severe discharge anemia Total
42,361 (28) 31,903 (21) 52,971 (35) 25,522 (17) 152,757 (100)
© 2014 Lippincott Williams & Wilkins
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
J Patient Saf • Volume 00, Number 00, Month 2014
Anemia at Discharge Increases Readmissions
TABLE 2. Patient Characteristics and Comorbidities Stratified by 4 Anemia Groups, Based on the Last Available Hemoglobin Value at Hospital Discharge No Discharge Mild Discharge Anemia (n = 42,361) Anemia (n = 31,903) Variable
n (%) *
55.1 (18.4) Age at admission, y Female 22,019 (52) Race White 34,043 (80) Black 6524 (15) Other 1794 (4.2) 9995 (24) MS-DRG type (surgery)† Hypertension 21,597 (51) Heart failure 2061 (4.9) Valvular disease 994 (2.3) Pulmonary circulation disease 585 (1.4) Peripheral arterial disease 1958 (4.6) Paralysis 1002 (2.4) Other neurologic disorders 2428 (5.7) Chronic pulmonary disease 7376 (17) Diabetes without chronic complications 6733 (16) Diabetes with chronic complications 1140 (2.7) Hypothyroidism 4269 (10) Renal failure 1680 (4.0) Liver disease 1202 (2.8) Peptic ulcer disease 8 (0.019) AIDS 44 (0.1) Lymphoma 186 (0.44) Metastatic cancer 419 (0.99) Solid tumor without metastasis 424 (1.0) Rheumatoid arthritis/collagen 1050 (2.5) vascular disease Coagulopathy 1120 (2.6) Obesity 5999 (14) Weight loss 993 (2.3) Fluid and electrolyte disorders 6690 (16) Alcohol abuse 2619 (6.2) Drug abuse 2620 (6.2) Psychoses 2669 (6.3) Depression 4828 (11)
Moderate Discharge Anemia (n = 52,971)
Severe Discharge Anemia (n = 25,522)
n (%)
n (%)
n (%)
58.2 (18.7) 14,791 (46)
59.2 (19.3) 33,246 (63)
59.3 (18) 14,852 (58)
25,743 (81) 4829 (15) 1331 (4.2) 13,172 (41) 16,860 (53) 2145 (6.7) 970 (3.0) 556 (1.7) 2168 (6.8) 815 (2.6) 2000 (6.3) 5626 (18) 5620 (18) 1315 (4.1) 3302 (10) 2505 (7.9) 970 (3.0) 6 (0.019) 46 (0.14) 255 (0.8) 710 (2.2) 590 (1.8) 939 (2.9)
42,709 (81) 8043 (15) 2219 (4.2) 27,166 (51) 28,515 (54) 4684 (8.8) 2142 (4.0) 1300 (2.5) 4420 (8.3) 1236 (2.3) 3223 (6.1) 9646 (18) 9292 (18) 2937 (5.5) 6674 (13) 6333 (12) 1732 (3.3) 25 (0.047) 53 (0.1) 576 (1.1) 1812 (3.4) 1124 (2.1) 2076 (3.9)
19,847 (78) 4500 (18) 1175 (4.6) 14,514 (57) 14,360 (56) 2576 (10) 1199 (4.7) 836 (3.3) 2466 (9.7) 662 (2.6) 1663 (6.5) 4858 (19) 4648 (18) 1736 (6.8) 3302 (13) 4162 (16) 1092 (4.3) 22 (0.086) 44 (0.17) 309 (1.2) 967 (3.8) 591 (2.3) 1016 (4.0)
Magnitude of Anemia at Discharge Increases 30-Day Hospital Readmissions Colleen G. Koch, MD, MS, MBA,*† Liang Li, PhD,‡ Zhiyuan Sun, MS,‡ Eric D. Hixson, PhD,§ Anne Tang, MS,‡ Kevin Chagin, MS,‡ Michael Kattan, PhD,†‡ Shannon C. Phillips, MD,† Eugene H. Blackstone, MD,‡|| and J. Michael Henderson, MD†¶ Background: Anemia during hospitalization is associated with poor health outcomes. Does anemia at discharge place patients at risk for hospital readmission within 30 days of discharge? Our objectives were to examine the prevalence and magnitude of anemia at hospital discharge and determine whether anemia at discharge was associated with 30-day readmissions among a cohort of hospitalizations in a single health care system. Methods: From January 1, 2009, to August 31, 2011, there were 152,757 eligible hospitalizations within a single health care system. The endpoint was any hospitalization within 30 days of discharge. The University HealthSystem Consortium's clinical database was used for demographics and comorbidities; hemoglobin values are from the hospitals' electronic medical records, and readmission status was obtained from the University HealthSystem Consortium administrative data systems. Mild anemia was defined as hemoglobin of greater than 11 to less than 12 g/dl in women and greater than 11 to less than 13 g/dl in men; moderate, greater than 9 to less than or equal to 11 g/dl; and severe, less than or equal to 9 g/dl. Logistic regression was used to assess the association of anemia and 30-day readmissions adjusted for demographics, comorbidity, and hospitalization type. Results: Among 152,757 hospitalizations, 72% of patients were discharged with anemia: 31,903 (21%), mild; 52,971 (35%), moderate; and 25,522 (17%), severe. Discharge anemia was associated with severity-dependent increased odds for 30-day hospital readmission compared with those without anemia: for mild anemia, 1.74 (1.65–1.82); moderate anemia, 2.76 (2.64–2.89); and severe anemia, 3.47 (3.30–3.65), P < 0.001. Conclusions: Anemia at discharge is associated with a severitydependent increased risk for 30-day readmission. A strategy focusing on anemia treatment care paths during index hospitalization offers an opportunity to influence subsequent readmissions. Key Words: 30-day hospital readmissions, anemia, resource use, mortality (J Patient Saf 2014;00: 00–00)
A
pproximately 19% of Medicare beneficiaries discharged from the hospital return within 30 days of discharge, at an annual estimated cost in the billions of dollars.1 Recently, a hospital readmission prediction score reported that a number of factors including lower hemoglobin values at discharge were associated with increased 30-day readmissions.2 We have previously addressed the role of present-on-admission anemia and the development of hospital-acquired anemia on health outcomes and resource use in our health system. We found a risk-adjusted increase in hospital From the *Department of Cardiothoracic Anesthesia, Heart and Vascular Institute, †Quality and Patient Safety Institute, ‡Department of Quantitative Health Sciences, Research Institute, §Business Intelligence, Medical Operations, ||Department of Thoracic and Cardiovascular Surgery, Heart and Vascular Institute, and ¶Department of General Surgery, Digestive Disease Institute, Cleveland Clinic, Cleveland, Ohio. Correspondence: Colleen G. Koch, MD, MS, MBA, Department of Cardiothoracic Anesthesia, Heart and Vascular Institute, Cleveland Clinic, 9500 Euclid Ave/Mail Stop J-4, Cleveland, OH 44195 (e‐mail: [email protected]). The authors disclose no conflict of interest. Copyright © 2014 by Lippincott Williams & Wilkins
J Patient Saf • Volume 00, Number 00, Month 2014
mortality, length of hospital stay, and total charges among anemic patients.3 It was unclear whether this heightened risk continued to the postdischarge setting by placing our patients at risk for hospital readmission within 30 days of discharge. Hence, our objectives were to (1) examine the prevalence and magnitude of anemia at the time of hospital discharge as well as (2) determine whether anemia at discharge was associated with 30-day readmissions among a large cohort of hospitalizations in a quaternary referral health care system.
METHODS Patient Population From January 1, 2009, to August 31, 2011, there were 417,301 hospitalizations within a single health care system consisting of 1 quaternary academic medical center (Cleveland Clinic) and 9 community hospitals. Exclusion criteria included hospitals without available electronic laboratory data, patients 18 years or younger, and hospitalizations without hemoglobin values recorded. We retained only the first hospitalization with readmission or the first hospitalization for patients who did not have a 30-day readmission. The final data set consisted of 152,757 hospitalizations (Fig. 1).
Anemia Definitions Presence and magnitude of anemia at hospital discharge were determined by the last available hemoglobin value before discharge. The World Health Organization defines anemia as a hemoglobin value of less than 12 g/dl in women and less than 13 g/dl in men.4 We further grouped hemoglobin by degree of severity into mild anemia, defined as hemoglobin of greater than 11 to less than 12 g/dl in women and greater than 11 to less than 13 g/dl in men; moderate anemia, defined as hemoglobin of greater than 9 to less than or equal to 11 g/dl; and severe anemia, defined as hemoglobin of less than or equal to 9 g/dl.
Endpoint The endpoint was hospital readmission to the health care system within 30 days of hospital discharge.
Data Patient demographics and baseline comorbidities were extracted from the clinical database/resource manager of the University HealthSystem Consortium (UHC). The UHC is composed of 119 U.S. academic medical centers and their 272 affiliated hospitals. It represents more than 90% of nonprofit academic medical centers in the country. Data were originally retrieved from the hospitals' administrative data systems, normalized according to UHC standardized data specifications, and submitted for inclusion in the national repository. Data quality was assessed using standardized error checking and data completeness algorithms and was required to meet established minimum thresholds to be www.journalpatientsafety.com
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
1
J Patient Saf • Volume 00, Number 00, Month 2014
Koch et al
FIGURE 1. Consolidated Standards of Reporting Trials (CONSORT)-style diagram of eligible hospitalizations within a single health care system.
included in the national repository. Measured hemoglobin values were taken from the hospitals' electronic medical records from complete blood count testing. The data source for hospital readmission status was obtained from Cleveland Clinic Health System administrative data systems. The Cleveland Clinic Institutional Review Board approved this investigation, with individual patient consent waived.
Statistical Analysis The primary analysis examined the association between the presence and magnitude (mild, moderate, and severe) of anemia at discharge and hospital readmissions within 30 days of discharge. Baseline risk factors and readmission status were summarized and compared between groups using frequencies, percentages, and χ2 tests for categorical variables as well as median, 25th and 75th quartiles, and the Kruskal-Wallis test for continuous variables. Logistic regression was used to assess the association between demographics, comorbidity, presence and magnitude of anemia, and hospital readmission. We used a comprehensive set of comorbidity measurements to control for the patients' underlying conditions in addition to discharge anemia groupings, age, race,
2
www.journalpatientsafety.com
sex, and Medicare severity diagnosis–related group (MS-DRG) status (surgery and medical status).5 Unless otherwise specified, a P value of less than 0.05 was considered statistically significant. All analyses were performed using SAS version 9.2 (SAS, Inc, Cary, NC).
RESULTS Among 152,757 hospitalizations, 72% of the patients were discharged with anemia: 31,903 (21%) with mild anemia, 52,971 (35%) with moderate anemia, and 25,522 (17%) with TABLE 1. Magnitude of Anemia at Hospital Discharge Determined by the Last Available Hemoglobin Value Discharge Anemia Groupings
n (%)
No discharge anemia Mild discharge anemia Moderate discharge anemia Severe discharge anemia Total
42,361 (28) 31,903 (21) 52,971 (35) 25,522 (17) 152,757 (100)
© 2014 Lippincott Williams & Wilkins
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
J Patient Saf • Volume 00, Number 00, Month 2014
Anemia at Discharge Increases Readmissions
TABLE 2. Patient Characteristics and Comorbidities Stratified by 4 Anemia Groups, Based on the Last Available Hemoglobin Value at Hospital Discharge No Discharge Mild Discharge Anemia (n = 42,361) Anemia (n = 31,903) Variable
n (%) *
55.1 (18.4) Age at admission, y Female 22,019 (52) Race White 34,043 (80) Black 6524 (15) Other 1794 (4.2) 9995 (24) MS-DRG type (surgery)† Hypertension 21,597 (51) Heart failure 2061 (4.9) Valvular disease 994 (2.3) Pulmonary circulation disease 585 (1.4) Peripheral arterial disease 1958 (4.6) Paralysis 1002 (2.4) Other neurologic disorders 2428 (5.7) Chronic pulmonary disease 7376 (17) Diabetes without chronic complications 6733 (16) Diabetes with chronic complications 1140 (2.7) Hypothyroidism 4269 (10) Renal failure 1680 (4.0) Liver disease 1202 (2.8) Peptic ulcer disease 8 (0.019) AIDS 44 (0.1) Lymphoma 186 (0.44) Metastatic cancer 419 (0.99) Solid tumor without metastasis 424 (1.0) Rheumatoid arthritis/collagen 1050 (2.5) vascular disease Coagulopathy 1120 (2.6) Obesity 5999 (14) Weight loss 993 (2.3) Fluid and electrolyte disorders 6690 (16) Alcohol abuse 2619 (6.2) Drug abuse 2620 (6.2) Psychoses 2669 (6.3) Depression 4828 (11)
Moderate Discharge Anemia (n = 52,971)
Severe Discharge Anemia (n = 25,522)
n (%)
n (%)
n (%)
58.2 (18.7) 14,791 (46)
59.2 (19.3) 33,246 (63)
59.3 (18) 14,852 (58)
25,743 (81) 4829 (15) 1331 (4.2) 13,172 (41) 16,860 (53) 2145 (6.7) 970 (3.0) 556 (1.7) 2168 (6.8) 815 (2.6) 2000 (6.3) 5626 (18) 5620 (18) 1315 (4.1) 3302 (10) 2505 (7.9) 970 (3.0) 6 (0.019) 46 (0.14) 255 (0.8) 710 (2.2) 590 (1.8) 939 (2.9)
42,709 (81) 8043 (15) 2219 (4.2) 27,166 (51) 28,515 (54) 4684 (8.8) 2142 (4.0) 1300 (2.5) 4420 (8.3) 1236 (2.3) 3223 (6.1) 9646 (18) 9292 (18) 2937 (5.5) 6674 (13) 6333 (12) 1732 (3.3) 25 (0.047) 53 (0.1) 576 (1.1) 1812 (3.4) 1124 (2.1) 2076 (3.9)
19,847 (78) 4500 (18) 1175 (4.6) 14,514 (57) 14,360 (56) 2576 (10) 1199 (4.7) 836 (3.3) 2466 (9.7) 662 (2.6) 1663 (6.5) 4858 (19) 4648 (18) 1736 (6.8) 3302 (13) 4162 (16) 1092 (4.3) 22 (0.086) 44 (0.17) 309 (1.2) 967 (3.8) 591 (2.3) 1016 (4.0)
