Comments
(1)
Abstract
Rationale: Various causes can contribute to the high rates of readmission among patients hospitalized with chronic obstructive pulmonary disease (COPD).
Objectives: To determine the frequency and predictors of early readmission among patients aged 40–64 years, hospitalized with COPD.
Methods: In a retrospective cohort study, using a large national commercial insurance database, we obtained the clinical information within 12 months of the index hospitalization and 30 days after discharge.
Measurements and Main Results: Primary outcome was early readmission, defined as hospitalization within 30 days of discharge. We categorized predictor variables as patient, provider, and system factors, and compared these variables between patients readmitted and those not readmitted. Logistic regression was used for multivariable analysis. Of 8,263 patients who met the inclusion criteria, 741 (8.9%) had early readmission. Multivariable analysis showed patient factors (male, history of heart failure, lung cancer, osteoporosis, and depression), provider factors (no prior prescription of statin within 12 mo of the index hospitalization and no prescription of short-acting bronchodilator, oral steroid and antibiotic on discharge), and system factors (length of stay, <2 or >5 d and lack of follow-up visit after discharge) were associated with early readmission among patients hospitalized with COPD. The C-statistic of the model including patient characteristics was 0.677 (95% confidence interval, 0.656–0.697), which was improved to 0.717 (95% confidence interval, 0.702–0.732) after addition of provider- and system-based factors.
Conclusions: One of 11 patients hospitalized with COPD is readmitted within 30 days of discharge. Provider and system factors are important modifiable risk factors of early readmission.
In 2009, the Centers for Medicare and Medicaid Services began to publicly report 30-day risk-standardized readmission rates for pneumonia, congestive heart failure (CHF) and acute myocardial infarction, as a quality performance measure (1–3). Although some readmissions may be unavoidable, reducing hospital readmissions has been an objective of the Affordable Care Act and Accountable Care Organizations. The readmission reduction program seeks to reduce unnecessary readmissions of patients. One in five Medicare beneficiaries is hospitalized within 30 days of hospital discharge (4), at a cost of more than $15 billion annually (4, 5).
High rates of hospital readmissions and unexplained variation in those rates may indicate problems in delivery of care, transitions of care, or outpatient management after discharge. Improving the quality of care during care transitions at the time of hospital discharge may result in fewer hospital readmissions and associated costs (6, 7). In 2012, the Centers for Medicare and Medicaid Services started to reduce Medicare payments to hospitals with high readmission rates for patients admitted with CHF, myocardial infarction, and pneumonia (8). Starting from fiscal year 2015, this list will also include chronic obstructive pulmonary disease (COPD) (9). COPD is the third most common cause of hospital readmission among Medicare beneficiaries (4). Almost one in four (22.6%) Medicare beneficiaries (4) and 1 in 12 younger adults (40–64 yr old) (10) hospitalized for COPD were readmitted to the hospital within 30 days of discharge.
COPD affects more than 6.2% of adults in the United States, is the third most common cause of death (11), and is the only leading cause of mortality with rising morbidity and mortality (12, 13). The typical disease course of the patient with COPD is punctuated with exacerbations (14–16). Most are treated in outpatient settings; more severe cases, however, require hospitalization. Patients are at increased risk of reexacerbations within the first few weeks of the index exacerbation (17), which can result in hospital readmissions. High rates of readmission among patients with COPD may result from the complexity of the index hospitalization or problems in transition of care or outpatient follow-up after discharge (18).
Claim-based models to predict 30-day readmission exist for CHF, acute myocardial infarction, and pneumonia (19–22). However, similar models are lacking for patients hospitalized for COPD. We hypothesized that an array of patient, provider, and system factors contribute to early readmission among patients hospitalized with COPD. In the current work, we aimed to evaluate the frequency of early hospital readmission rate (within 30 d of discharge) for patients 40 to 64 years old and hospitalized for COPD, and build a risk prediction model based on patient, provider, and system factors that were associated with early readmission.
The study population included commercially ensured patients 40 to 64 years of age and hospitalized with a primary discharge diagnosis of COPD between January 2009 and November 2011 in a large private health insurance plan, which covers approximately 70 million Americans. The Clinformatics Data Mart, managed by OPTUMInsight (Minneapolis, MN), is composed of medical and pharmacy claims data for members in this health plan throughout the United States. The Research Database contains claims data from May 2000 to December 2011 for approximately 51,000,000 current and past members, with a mean continuous enrollment period of approximately 2 years. All files contain encrypted patient and provider identifiers, which permit patient-specific longitudinal tracking of patient, hospital, and medication histories. However, race and socioeconomic status are not captured in the database.
Using the International Classification of Diseases, Ninth Revision (ICD-9), we included patients hospitalized between January 2009 and November 2011 with primary diagnosis codes for COPD. We excluded patients 65 years of age or older as well as those younger than 40 years, those with incomplete data for the 12 months before the index hospitalization, those transferred to a long-term facility, and those with discharge diagnosis ICD-9 codes of 490 (nonspecific bronchitis) and 493 (asthma). Overall, 8,263 patients 40–64 years of age, with ICD-9 codes of 491.xx, 492.xx, and 496 as their primary diagnoses were included in the study (Figure 1). The university institutional review board approved the study protocol.
Figure 1. Establishment of a cohort of patients 40–64 years of age and hospitalized between 2009 and 2011 with a primary diagnosis of chronic obstructive pulmonary disease (COPD). dx = diagnosis; ICD-9 = International Classification of Diseases, Ninth Revision.
[More] [Minimize]The Research Database contains the following components: the member enrollment file, consisting of demographic information on all health plan enrollees (year of birth, sex, state, and benefits); medical claims records, including diagnosis and procedure codes from all health care sites (inpatient hospital, emergency room, physician’s office, or surgery center) for services provided to enrollees; pharmacy claims (drug code, drug names, drug strength, date prescription filled, and days of supply).
Predictor variables were classified as patient, provider, and system factors. The patient factors included age, sex, and type and number of comorbid conditions before the index admission. We examined comorbidities prevalent in this population based on existing literature and also in our cohort (CHF, lung cancer, alcohol abuse, obesity, depression, osteoporosis, chronic kidney disease (23), diabetes mellitus, hyperlipidemia, and obstructive sleep apnea). We examined the effect of both individual comorbidities, the number of unique comorbidities, and certain combinations.
The provider factors focused on the quality of care within the 12 months before hospitalization and the prescriptions given on discharge. The type of inhaler medications prescribed within a year of the index hospitalization and any prescription of in-home oxygen were recorded for each subject. COPD medications included any prescription of short-acting β agonist, short-acting muscarinic antagonist, long-acting β agonist, long-acting muscarinic antagonist, or inhaled corticosteroid or oral steroid from the 12 months before the index hospitalization to 30 days after discharge. In addition, we recorded any prescription of antibiotics within 30 days of discharge. To examine the protective effect of statins and/or angiotensin-converting enzyme inhibitors (ACE-Is) (24) on COPD readmission we recorded any prescription of statins and ACE-I within the 12 months before the index hospitalization. Claims for oxygen therapy over the 12 months before the index hospitalization and within 30 days of discharge were also recorded.
System factors included number of outpatient visits over 12 months before the index hospitalization, type of COPD care provider seen (primary care physician [PCP], pulmonary specialist, both PCP and specialist), number of inpatient hospitalizations over the 12 months before the index hospitalization, length of stay in the hospital during index hospitalization, discharge follow-up visit, and type of provider seen after discharge. Length of stay during the index hospitalization was categorized as 1–2, 3–4, 5–6, or at least 7 days or more.
The primary outcome was defined as all-cause readmission within 30 days of the discharge day of index hospitalization. The secondary outcome was the reason for and the factors associated with readmission.
In this retrospective cohort study, we used descriptive statistics to report the characteristics of all cohorts and the rates of rehospitalization within 30 days by patient, provider, and system characteristics. The readmission rates across different levels of each categorical and continuous variable were compared by chi-squared (χ2) and Student t test, respectively.
Two multivariable logistic regression models were built to determine the independent predictors of early readmission. The first model included baseline patient demographic and clinical characteristics, provider factors (prescriptions over the 12 months before index hospitalization), and system factors (type of provider and number of hospitalization within the year before index hospitalization). Then, provider (prescriptions) and system factors (length of stay, discharge follow-up, and type of provider in the discharge follow-up) during hospitalization and within 30 days of discharge were added into the second model. C-statistics with 95% confidence interval (CI) from the receiving operating curves were calculated. We cross-validated the model by splitting the data into derivation and validation cohorts. We used SAS version 9.2 (SAS Institute Inc., Cary, NC) for all statistical analyses. All hypothesis testing was two-sided with significance set at P less than or equal to 0.05.
Between January 2009 and November 2011, 8,263 patients 40–64 years of age were hospitalized for COPD. All patients were followed up for 30 days after discharge. Of these, 741 (8.9%) were readmitted within the follow-up period. Table 1 presents the baseline characteristics of the entire cohort and those readmitted within 30 days.
| Total (n = 8,263) | 30-Day Readmission | P Value | ||
|---|---|---|---|---|
| No (n = 7,522): n (%) | Yes (n = 741): n (%) | |||
| Patient Factors | ||||
| Age (yr), mean (SD) | 56.55 (5.73) | N/A | 57.02 (5.53) | 0.257 |
| Age group, yr | ||||
| 40–54 | 2,729 | 2,507 (91.9) | 222 (8.1) | 0.063 |
| 55–64 | 5,534 | 5,015 (90.6) | 519 (9.4) | |
| Sex | ||||
| Female | 4,862 | 4,423 (90.9) | 439 (9.1) | 0.82 |
| Male | 3,401 | 3,099 (91.1) | 302 (8.9) | |
| Congestive heart failure | ||||
| Yes | 2,066 | 1,797 (86.9) | 269 (13.1) | <0.001 |
| No | 6,197 | 5,725 (92.4) | 472 (7.6) | |
| Lung cancer | ||||
| Yes | 347 | 298 (85.9) | 49 (14.1) | 0.001 |
| No | 7,916 | 7,224 (91.3) | 692 (8.7) | |
| Anxiety | ||||
| Yes | 2,206 | 1,947 (88.3) | 259 (11.7) | <0.001 |
| No | 6,057 | 5,575 (92.0) | 482 (8.0) | |
| Alcohol abuse | ||||
| Yes | 4,372 | 3,981 (91.1) | 391 (8.9) | 0.93 |
| No | 3,891 | 3,541 (91.0) | 350 (9.0) | |
| Obesity | ||||
| Yes | 1,488 | 1,319 (88.6) | 169 (11.4) | 0.001 |
| No | 6,775 | 6,203 (91.6) | 572 (8.4) | |
| Depression | ||||
| Yes | 811 | 713 (87.9) | 98 (12.1) | 0.001 |
| No | 7,452 | 6,809 (91.4) | 643 (8.6) | |
| Osteoporosis | ||||
| Yes | 1,277 | 1,112 (87.1) | 165 (12.9) | <0.001 |
| No | 6,986 | 6,410 (91.7) | 576 (8.3) | |
| Chronic kidney disease | ||||
| Yes | 468 | 401 (85.7) | 67 (14.3) | <0.001 |
| No | 7,795 | 7,121 (91.3) | 674 (8.7) | |
| Diabetes mellitus | ||||
| Yes | 2,549 | 2,272 (89.1) | 277 (10.9) | <0.001 |
| No | 5,714 | 5,528 (96.7) | 186 (7.3) | |
| Hypertension | ||||
| Yes | 5,713 | 5,158 (90.3) | 555 (9.7) | <0.001 |
| No | 2,550 | 2,364 (92.7) | 186 (7.3) | |
| Hyperlipidemia | ||||
| Yes | 4,242 | 3,840 (90.5) | 402 (9.5) | 0.096 |
| No | 4,021 | 3,682 (91.6) | 339 (8.4) | |
| Obstructive sleep apnea | ||||
| Yes | 1,435 | 1,276 (88.9) | 159 (11.1) | 0.002 |
| No | 6,828 | 6,246 (91.5) | 582 (8.5) | |
| Number of comorbidities | ||||
| No comorbid conditions | 1,261 | 1,184 (93.9) | 77 (6.1) | <0.001 |
| One comorbidities | 3,128 | 2,912 (93.1) | 216 (6.9) | |
| Two comorbidities | 2,328 | 2,110 (90.6) | 218 (9.4) | |
| Three or more comorbidities | 1,546 | 1,316 (85.1) | 230 (14.9) | |
| Provider Factors | ||||
| Any COPD medication* 1 yr before hospitalization | ||||
| Yes | 3,892 | 3,704 (95.2) | 188 (4.8) | <0.001 |
| No | 4,371 | 3,818 (87.3) | 553 (12.7) | |
| ACE inhibitors | ||||
| Yes | 1,566 | 1,466 (93.6) | 100 (6.4) | <0.001 |
| No | 6,697 | 6,056 (90.4) | 641 (9.6) | |
| Statins | ||||
| Yes | 1,715 | 1,614 (94.1) | 101 (5.9) | <0.001 |
| No | 6,548 | 5,908 (90.2) | 640 (9.8) | |
| Long-term oxygen therapy 1 yr before index admittance | ||||
| Yes | 2,032 | 1,769 (87.1) | 263 (12.9) | <0.001 |
| No | 6,231 | 5,753 (92.3) | 478 (7.7) | |
| System Factors | ||||
| Outpatient care in previous 12 mo | ||||
| Visit to PCP | 3,215 | 2,921 (90.9) | 294 (9.1) | <0.001 |
| Visit to pulmonary specialist | 659 | 576 (87.4) | 83 (12.6) | |
| Visit to both PCP and pulmonary specialist | 1,510 | 1,338 (88.6) | 172 (11.4) | |
| No visit to PCP or specialist | 2,879 | 2,687 (93.3) | 192 (6.7) | |
| Total outpatient visits in previous 12 mo, mean (SD), d | 3.23 (4.0) | N/A | 4.18 (4.5) | <0.001 |
| PCP visits only, mean (SD) | 2.05 (2.9) | N/A | 2.56 (3.7) | <0.001 |
| Pulmonary specialist only, mean (SD) | 0.83 (2.0) | N/A | 1.17 (2.3) | 0.008 |
| Number of inpatient hospital visits in 12 mo before index hospitalization, mean (SD) | 1.00 (1.1) | N/A | 2.01 (2.7) | <0.001 |
As illustrated in Table 1, patients with a history of CHF, lung cancer, anxiety, obesity, depression, osteoporosis, chronic kidney disease, diabetes mellitus, hypertension, and obstructive sleep apnea were more likely to be readmitted within 30 days, compared with those who did not have these comorbidities. There was a stepwise increase in rates of 30-day readmission with increasing number of comorbid medical conditions. For example, patients with none, one, two, and three or more comorbidities had 6.1, 6.9, 9.4, and 14.9% rates of readmissions, respectively (P < 0.001). In addition, patients with COPD with certain combination of comorbidities had much higher risk of readmission than others, suggesting a differential role for coexisting comorbidities. For example, patients with COPD with CHF and osteoporosis (n = 74) had the highest risk for readmission (21.0%), followed by those with CHF and anxiety (n = 100; 18.2%), CHF and depression (n = 83; 15.9%), and CHF and alcohol abuse (n = 129; 14.4%).
Patients who were not receiving any COPD medications, statins, or ACE-Is in the year before hospitalization had higher rates of 30-day readmissions. As expected, patients with severe disease (as suggested by long-term oxygen therapy use or being seen by a pulmonary specialist) had higher rates of 30-day readmissions (Table 1). We tested for differential effects of statins on 30-day readmission in patients with and without hyperlipidemia. Patients receiving statins for hyperlipidemia had lower odds of 30-day readmission (odds ratio [OR], 0.6; 95% CI, 0.4–0.8) compared with those patients without hyperlipidemia (OR, 1.1; 95% CI, 0.6–2.0). However, the effect of ACE inhibitors on 30-day readmission was not different in patients with or without history of CHF (P = 0.06).
Table 2 shows the process of care measures during index hospitalization. As shown, patients who did not receive a prescription for a short-acting β agonist/short-acting muscarinic antagonist, long-acting β agonist, long-acting muscarinic antagonist, or inhaled glucocorticoids within 30 days of discharge had higher rates of readmission. Patients who received any prescription for COPD medications and/or antibiotics had lower rates of 30-day readmission. For example, rates of 30-day readmission were 16.8% for patients who received no prescriptions for COPD medications within 30 days of discharge, compared with 5.9% in those who received prescriptions for any COPD medications (P < 0.001). Patients who had a follow-up visit within 30 days of discharge had a readmission rate of 7.6%, compared with 10.9% in those who did not have follow-up visit postdischarge (P < 0.001). There was no statistically significant difference in rate of readmission in terms of time to follow-up visit since discharge or type of provider seen after discharge.
| Total COPD Admission | Patients with 30-Day Readmission | P Value | ||
|---|---|---|---|---|
| No | Yes | |||
| Provider Factors (within 30 d of Discharge) | ||||
| SABA and/or SAMA | ||||
| Yes | 3,153 | 2,973 (94.3) | 180 (5.7) | <0.001 |
| No | 5,110 | 4,549 (89.0) | 561 (11.0) | |
| Long-acting β agonist | ||||
| Yes | 172 | 164 (95.3) | 8 (4.7) | <0.001 |
| No | 8,091 | 7,358 (90.9) | 733 (9.1) | |
| Long-acting muscarinic antagonist | ||||
| Yes | 1,709 | 1,615 (94.5) | 94 (5.5) | <0.001 |
| No | 6,554 | 5,907 (90.1) | 647 (9.9) | |
| Oral glucocorticoids | ||||
| Yes | 4,894 | 4,610 (94.2) | 284 (5.8) | <0.001 |
| No | 3,369 | 2,912 (86.4) | 457 (13.6) | |
| ICS | ||||
| Yes | 480 | 450 (93.7) | 30 (6.3) | 0.03 |
| No | 7,783 | 7,072 (90.9) | 711 (9.1) | |
| ICS plus LABA | ||||
| Yes | 2,078 | 1,965 (94.6) | 113 (5.4) | <0.001 |
| No | 6,185 | 5,557 (89.8) | 628 (10.2) | |
| Theophylline | ||||
| Yes | 308 | 286 (92.9) | 22 (7.1) | 0.25 |
| No | 7,955 | 7,236 (91.0) | 719 (9.0) | |
| Any COPD medication during 30 d postdischarge | ||||
| Yes | 5,807 | 5,479 (94.3) | 328 (5.7) | <0.001 |
| No | 2,456 | 2,043 (83.2) | 413 (16.8) | |
| Any antibiotics during 30 d after index admission | ||||
| Yes | 4,616 | 4,328 (93.8) | 288 (6.2) | <0.001 |
| No | 3,647 | 3,194 (87.6) | 453 (12.4) | |
| LT/OT (30 d after index admittance) | ||||
| Yes | 2,266 | 2,054 (90.6) | 212 (9.4) | 0.45 |
| No | 5,997 | 5,468 (91.2) | 529 (8.8) | |
| System Factors | ||||
| Length of stay during index hospitalization (d), mean (SD) | 5.1 (14.9) | N/A | 5.7 (14.0) | <0.001 |
| Had a follow-up visit within the 30 d | 4,732 | 4,374 (92.4) | 358 (7.6) | <0.001 |
| No follow-up within 30 d of discharge | 3,531 | 3,148 (89.1) | 383 (10.9) | |
Table 3 presents the primary reason for each readmission. The most common reason for readmission, within the first 30 days after discharge, was COPD (27.4%), followed by respiratory failure (14.3%) and symptoms involving the respiratory system (8.5%). Other reasons for readmission were related to cardiovascular comorbidities. Half of the patients were readmitted within the first 15 days and 19% within the first 7 days (Figure 2).
| Reason for Readmission* | n (%) |
|---|---|
| COPD (491, 492, 496) | 203 (27.4) |
| Respiratory failure (518) | 106 (14.3) |
| Symptoms involving respiratory system (786) | 63 (8.5) |
| Pneumonia (486) | 31 (4.2) |
| Heart failure (428) | 17 (2.3) |
| Cardiac dysrhythmia (427) | 16 (2.2) |
| Coronary atherosclerosis (414) | 15 (2.0) |
| Lung cancer (162) | 13 (1.8) |
| Septicemia (038) | 12 (1.6) |
| Acute pulmonary heart disease (415) | 7 (0.9) |
| Other | 258 (34.8) |
| Total | 741 |
Figure 2. Distribution of early readmission among patients 40–64 years of age and hospitalized with chronic obstructive pulmonary disease between 2009 and 2011.
[More] [Minimize]Table 4 presents the multivariable analysis of factors associated with 30-day readmission. As shown, female patients had lower odds of readmission (OR, 0.9; 95% CI, 0.7–0.9). The highest odds of readmission were associated with patients with lung cancer comorbidities (OR, 1.6; 95% CI, 1.3–2.1). Patients who had a prescription for statins in the 12 months before hospitalization had lower odds of readmission (OR, 0.6; 95% CI, 0.5–0.7). Prescriptions for antibiotics within 30 days of discharge were associated with a lower likelihood of a 30-day readmission (OR, 0.8; 95% CI, 0.6–0.9).
| Variable | Odds Ratio (95% CI) | P Value |
|---|---|---|
| Patient Factors | ||
| Age | ||
| 40–54 yr | 1.0 | 0.386 |
| 55–64 yr | 1.1 (0.9, 12.9) | |
| Sex | ||
| Male | 1.0 | 0.025 |
| Female | 0.9 (0.7, 0.9) | |
| Congestive heart failure | ||
| No | 1.0 | 0.021 |
| Yes | 1.2 (1.0, 1.5) | |
| Lung cancer | ||
| No | 1.0 | 0.035 |
| Yes | 1.6 (1.3, 2.1) | |
| Alcohol abuse | ||
| No | 1.0 | 0.457 |
| Yes | 1.1 (0.9, 1.3) | |
| Obesity | ||
| No | 1.0 | 0.407 |
| Yes | 1.1 (0.9 1.4) | |
| Anxiety | ||
| No | 1.0 | 0.003 |
| Yes | 1.5 (1.2, 1.8) | |
| Depression | ||
| No | 1.0 | 0.016 |
| Yes | 1.3 (1.1, 1.8) | |
| Osteoporosis | ||
| No | 1.0 | 0.008 |
| Yes | 1.3 (1.1, 1.6) | |
| Hypertension | ||
| No | 1.0 | 0.507 |
| Yes | 1.1 (0.9, 1.3) | |
| Chronic kidney disease | ||
| No | 1.0 | 0.395 |
| Yes | 1.1 (0.8, 1.5) | |
| Hyperlipidemia | ||
| No | 1.0 | 0.064 |
| Yes | 1.2 (0.9, 1.4) | |
| Diabetes mellitus | ||
| No | 1.0 | 0.816 |
| Yes | 1.0 (0.9, 1.2) | |
| Obstructive sleep apnea | ||
| No | 1.0 | 0.811 |
| Yes | 0.9 (0.8, 1.2) | |
| Provider Factors | ||
| Prescription of ACE inhibitors 12 mo before index admission | ||
| No | 1.0 | 0.029 |
| Yes | 0.8 (0.6, 0.9) | |
| Prescription for statins 12 mo before index admission | ||
| No | 1.0 | <0.001 |
| Yes | 0.6 (0.5, 0.7) | |
| Being on long-term oxygen treatment 12 mo before hospitalization | ||
| No | 1.0 | 0.344 |
| Yes | 1.1 (0.9, 1.3) | |
| Prescription of SABA/SAMA within 30 d of discharge | ||
| No | 1.0 | 0.021 |
| Yes | 0.8 (0.6, 0.9) | |
| Prescription of LABA within 30 d of discharge | ||
| No | 1.0 | 0.132 |
| Yes | 0.6 (0.3, 1.2) | |
| Prescription of LAMA within 30 d of discharge | ||
| No | 1.0 | 0.159 |
| Yes | 0.8 (0.7, 1.1) | |
| Prescription of oral steroid within 30 d of discharge | ||
| No | 1.0 | <0.001 |
| Yes | 0.7 (0.5, 0.8) | |
| Prescription of ICS within 30 d of discharge | ||
| No | 1.0 | 0.456 |
| Yes | 1.0 (0.7, 1.5) | |
| Prescription of LABA/ICS within 30 d of discharge | ||
| No | 1.0 | 0.194 |
| Yes | 0.9 (0.7, 1.1) | |
| Prescription of any antibiotic within 30 d of discharge | ||
| No | 1.0 | 0.001 |
| Yes | 0.8 (0.6, 0.9) | |
| System Factors | ||
| Length of stay | ||
| 1–2 vs. 3–4 d | 1.3 (1.0, 1.5) | 0.019 |
| 5–6 vs. 3–4 d | 1.1 (0.8, 1.4) | 0.626 |
| >7 vs. 3–4 d | 1.2 (1.1, 1.5) | 0.010 |
| Hospitalization | ||
| None vs. ≥3 admissions | 0.4 (0.3, 0.6) | <0.001 |
| One vs. ≥3 admissions | 0.6 (0.4, 0.7) | <0.001 |
| Two vs. ≥3 admissions | 0.7 (0.6, 0.9) | 0.021 |
| Follow-up visit within 30 d | ||
| No | 1.0 | <0.001 |
| Yes | 0.7 (0.6, 0.9) | |
Patients with shorter (≤2 d) and longer (≥7 d) stays had increased odds of a 30-day readmission. Patients who had a follow-up within 30 days of discharge had lower odds of readmission (OR, 0.7; 95% CI, 0.6–0.9) compared with those who did not have a follow-up visit to an outpatient provider postdischarge.
Finally, we examined the C-statistics of the model in predicting 30-day readmission. A model including only the patient characteristics had a C-statistic of 0.677 (95% CI, 0.656–0.697). The addition of provider- and system-level factors postdischarge significantly enhanced the model prediction to 0.717 (95% CI, 0.702–0.732). We cross-validated the final model by randomly splitting the cohort into a deviation set and a validation set. The C-statistic of the validation sample was 0.73, with a 95% CI of 0.70–0.76. In addition, the assessment of the predictive accuracy of our final model indicated that the rate of readmissions was 4.3% for patients with the predicted probability of readmission in the bottom quartile compared with 20.3% for the patients with the predicted probability of readmission in the top quartile. There was a 4.7-fold increase in readmission rates between these two groups of patients.
One in 11 patients 40–64 years of age and hospitalized with COPD was readmitted within 30 days of the index hospitalization. Patient (sex, comorbid conditions), provider (specific medications prescribed), and system (process of care) factors predicted early readmission. To our knowledge, this is the first study to examine potentially modifiable provider- and system-level factors associated with readmission.
The early readmission rate of 8.9% in our cohort is similar to another study on a similar patient population (10). Prior reports have shown the 30-day readmission rate for patients with COPD to be 14.3–24% (4, 25). Among Medicare beneficiaries, the national readmission rates are much higher (4, 5). Low early readmission rates in our study can be explained by the relatively younger age group and fewer comorbidities.
In concordance with previous studies the number of comorbidities directly correlates with the frequency of 30-day readmission after hospital discharge (26). Those with conditions such as CHF, lung cancer, anxiety, depression, and osteoporosis had a higher likelihood of readmission. Among patients with chronic debilitating medical conditions, anxiety and depression are common ailments (27). The prevalence of anxiety and depression is high in the COPD population (28–30), with the risk of depression in those with severe disease nearly 2.5 times higher than normal (30). Whether the worsening of COPD induces anxiety and depression or a history of anxiety and depression causes lack of adherence to COPD management is unknown. The presence of these psychosocial factors has been associated with increased readmission in patients with COPD (31–33). Patients with anxiety and depression have poor compliance with medications (34, 35). This association further underscores the need to tailor efforts toward proper identification and treatment of anxiety and depression. However, the benefit of treating these psychological conditions in reducing readmission is unknown.
The physician plays a vital role in preventing hospital readmission for COPD. Patients who were prescribed statins in the 12 months before admission had lower rates of 30-day readmission. Current interest has focused on the benefit of statins in reducing inflammatory markers that contribute to COPD inflammation (36), specifically regarding mortality. In the Rotterdam Study of statin use in 363 patients with COPD, long-term (>2 yr) statin use was associated with a nearly 40% decreased risk of death compared with no prior use (37). In a cohort study it was shown that statin use in patients with COPD led to a 30% reduction in all-cause mortality at 3–4 years after first hospital admission with adjustment for prior history of diabetes and cardiovascular disease (38). The benefit of statins in reducing all-cause mortality and respiratory-related mortality has been shown in several studies (38, 39). The National Heart, Lung, and Blood Institute sponsored randomized clinical trial entitled Simvastatin Therapy for Moderate and Severe COPD (STATCOPE) is currently investigating the effects of statin therapy in patients with COPD. Further studies will be needed to confirm the benefit of statin therapy in reducing readmissions.
Patients who were prescribed steroids and antibiotics on discharge were less likely to be readmitted within 30 days. The positive effects of steroids in the management of COPD have been well demonstrated in randomized clinical trials (40) and administrative data (41). Patients who are prescribed oral steroids on admission will have faster improvement in FEV1, shorter hospital stays, and improvements in patient-assessed dyspnea and quality of life (40, 41). In a retrospective cohort study of patients 40 years of age or older, Rothberg and colleagues found that, in patients with COPD, early antibiotic administration was associated with lower rates of readmission and lower rates of inpatient mortality (42). In a study of more than 53,000 patients, Stefan and colleagues found that, in patients admitted with COPD and who were treated with steroids, the addition of antibiotics was associated with a 40% reduction in risk of in-hospital mortality and a 13% improvement in rate of readmission (43). Despite promising results from both randomized clinical trials and numerous administrative data on the impact of steroids and antibiotics on COPD exacerbation, not all exacerbations are bacterial and routine use of antibiotics remains controversial.
Although modifiable patient and physician factors may reduce the likelihood of readmission, system factors that can be modified should not be overlooked. In our study, we found that those patients with a follow-up visit with either a PCP or pulmonologist had lower rates of readmission. A study of the Medicare population found similar results (44). Efforts to improve follow-up visits through automated scheduling on admission or reminder phone calls should be targeted. The efficacy of this intervention was shown in conditions such as CHF (45). Our results of higher readmission rates for patients with low (≤2 d) or higher (≥7 d) lengths of stay are consistent with a prior study (46). The observed difference in readmission rates depending on the type of provider seen in the 12 months before index hospitalization could be attributable to selection bias, as sicker patients are more likely to be seen by a pulmonary physician. Patients with too short or too long a hospital stay have higher rates of readmission, consistent with prior studies (46, 47). Our results may indicate that patients discharged after 2 days or less are undertreated, causing earlier readmission. Patients who are admitted for 5 days or more may have more severe hospitalizations, refractory disease, or suffer from nosocomial infections causing earlier readmission.
The 30-day readmission risk prediction models for CHF, acute myocardial infarction, and pneumonia are based on administrative claims and have moderately discriminatory C-statistics and are mostly driven by patient factors (21, 22). Provider and system factors are potentially modifiable and, in our study, adding these factors to patient factors significantly improved the C-statistics of the model to predict 30-day readmission. Future studies should include provider and system factors in risk prediction models.
Our study has several limitations. We used ICD-9 to identify patients hospitalized with COPD. Previous studies suggested that this approach had low sensitivity (48). We repeated the analysis by addition of ICD-9 codes 490 and 493 to the baseline cohort. The 30-day readmission rates were 7.73 and 7.68%, respectively. Using administrative claim data, we do not have information on the severity of patients’ COPD or the complexity of the index hospitalization. These factors can confound the rate of readmission, as sicker patients are more likely to be readmitted. Similarly, because our study cohort was limited to ages 40–64 years these results cannot be generalized to fee-for-service Medicare beneficiaries. Socioeconomic status is an important determinant of readmission and was not captured in the database. Although the prescription of certain medications is associated with lower readmission rate, the analysis does not take into account adherence to medications. This limitation may underinflate the benefit of steroids, antibiotics, and long-acting β agonists/long-acting muscarinic antagonists as patients who were prescribed these medications, but were nonadherent, and then were readmitted within 30 days should be excluded. Finally, using patient, provider, and system factors, we were able to build a risk prediction model that was cross-validated in the current data set but needs validation in an independent sample.
In summary, patient, provider, and system factors predict COPD readmission. Modifiable provider and system factors should be targeted to decrease the rate of readmission within 30 days.
| 1 . | Krumholz HM, Merrill AR, Schone EM, Schreiner GC, Chen J, Bradley EH, Wang Y, Wang Y, Lin Z, Straube BM, et al. Patterns of hospital performance in acute myocardial infarction and heart failure 30-day mortality and readmission. Circ Cardiovasc Qual Outcomes 2009;2:407–413. |
| 2 . | Krumholz HM, Normand SL, Spertus JA, Shahian DM, Bradley EH. Measuring performance for treating heart attacks and heart failure: the case for outcomes measurement. Health Aff (Millwood) 2007;26:75–85. |
| 3 . | Department of Health and Human Services, Centers for Medicare and Medicaid Services. Hospital Compare, 2014. Available from: http://www.medicare.gov/hospitalcompare/search.html |
| 4 . | Jencks SF, Williams MV, Coleman EA. Rehospitalizations among patients in the Medicare fee-for-service program. N Engl J Med 2009;360:1418–1428. |
| 5 . | Camberg LC, Smith NE, Beaudet M, Daley J, Cagan M, Thibault G. Discharge destination and repeat hospitalizations. Med Care 1997;35:756–767. |
| 6 . | Westert GP, Lagoe RJ, Keskimäki I, Leyland A, Murphy M. An international study of hospital readmissions and related utilization in Europe and the USA. Health Policy 2002;61:269–278. |
| 7 . | Garcia-Aymerich J, Farrero E, Félez MA, Izquierdo J, Marrades RM, Antó JM; Estudi del Factors de Risc d’Agudització de la MPOC Investigators. Risk factors of readmission to hospital for a COPD exacerbation: a prospective study. Thorax 2003;58:100–105. |
| 8 . | Proposed rule clarifies admission policies. Hosp Case Manag 2013;21:95–96. |
| 9 . | Readmission reduction has begun and the penalties will escalate. Hosp Case Manag 2013;21:45–47. |
| 10 . | Baker CL, Zou KH, Su J. Risk assessment of readmissions following an initial COPD-related hospitalization. Int J Chron Obstruct Pulmon Dis 2013;8:551–559. |
| 11 . | Centers for Disease Control and Prevention (CDC). Deaths from chronic obstructive pulmonary disease—United States, 2000–2005. MMWR Morb Mortal Wkly Rep 2008;57:1229–1232. |
| 12 . | Murray CJ, Lopez AD. Alternative projections of mortality and disability by cause 1990–2020: Global Burden of Disease Study. Lancet 1997;349:1498–1504. |
| 13 . | Gershon AS, Wang C, Wilton AS, Raut R, To T. Trends in chronic obstructive pulmonary disease prevalence, incidence, and mortality in Ontario, Canada, 1996 to 2007: a population-based study. Arch Intern Med 2010;170:560–565. |
| 14 . | Donaldson GC, Müllerova H, Locantore N, Hurst JR, Calverley PM, Vestbo J, Anzueto A, Wedzicha JA. Factors associated with change in exacerbation frequency in COPD. Respir Res 2013;14:79. |
| 15 . | Leidy NK, Wilcox TK, Jones PW, Roberts L, Powers JH, Sethi S; EXACT-PRO Study Group. Standardizing measurement of chronic obstructive pulmonary disease exacerbations: reliability and validity of a patient-reported diary. Am J Respir Crit Care Med 2011;183:323–329. |
| 16 . | Lange P, Marott JL, Vestbo J, Olsen KR, Ingebrigtsen TS, Dahl M, Nordestgaard BG. Prediction of the clinical course of chronic obstructive pulmonary disease, using the new GOLD classification: a study of the general population. Am J Respir Crit Care Med 2012;186:975–981. |
| 17 . | Eriksen N, Vestbo J. Management and survival of patients admitted with an exacerbation of COPD: comparison of two Danish patient cohorts. Clin Respir J 2010;4:208–214. |
| 18 . | Elixhauser A, Au DH, Podulka J. Statistical Brief #121: readmissions for chronic obstructive pulmonary disease, 2008. Rockville, MD: Agency for Healthcare Research and Quality; 2011. |
| 19 . | Krumholz HM, Lin Z, Keenan PS, Chen J, Ross JS, Drye EE, Bernheim SM, Wang Y, Bradley EH, Han LF, et al. Relationship between hospital readmission and mortality rates for patients hospitalized with acute myocardial infarction, heart failure, or pneumonia. JAMA 2013;309:587–593. |
| 20 . | Krumholz HM, Lin Z, Drye EE, Desai MM, Han LF, Rapp MT, Mattera JA, Normand SL. An administrative claims measure suitable for profiling hospital performance based on 30-day all-cause readmission rates among patients with acute myocardial infarction. Circ Cardiovasc Qual Outcomes 2011;4:243–252. |
| 21 . | Keenan PS, Normand SL, Lin Z, Drye EE, Bhat KR, Ross JS, Schuur JD, Stauffer BD, Bernheim SM, Epstein AJ, et al. An administrative claims measure suitable for profiling hospital performance on the basis of 30-day all-cause readmission rates among patients with heart failure. Circ Cardiovasc Qual Outcomes 2008;1:29–37. |
| 22 . | Lindenauer PK, Normand SL, Drye EE, Lin Z, Goodrich K, Desai MM, Bratzler DW, O’Donnell WJ, Metersky ML, Krumholz HM. Development, validation, and results of a measure of 30-day readmission following hospitalization for pneumonia. J Hosp Med 2011;6:142–150. |
| 23 . | Divo M, Cote C, de Torres JP, Casanova C, Marin JM, Pinto-Plata V, Zulueta J, Cabrera C, Zagaceta J, Hunninghake G, et al.; BODE Collaborative Group. Comorbidities and risk of mortality in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2012;186:155–161. |
| 24 . | Mortensen EM, Copeland LA, Pugh MJ, Restrepo MI, de Molina RM, Nakashima B, Anzueto A. Impact of statins and ACE inhibitors on mortality after COPD exacerbations. Respir Res 2009;10:45. |
| 25 . | Kawasumi Y, Paterson MJ, Morrow RL, Miller TA, Bassett K, Wright JM, Dormuth CR. Comparative effectiveness of tiotropium and ipratropium in prevention of hospital readmission for COPD: a population-based cohort study. Clin Ther 2013;35:523–531, e1. |
| 26 . | McGhan R, Radcliff T, Fish R, Sutherland ER, Welsh C, Make B. Predictors of rehospitalization and death after a severe exacerbation of COPD. Chest 2007;132:1748–1755. |
| 27 . | PrattLA, Brody DJ. NCHS Data Brief: depression in the United States household population, 2005–2006. Hyattsville, MD: National Center for Health Statistics; 2008. |
| 28 . | Light RW, Merrill EJ, Despars JA, Gordon GH, Mutalipassi LR. Prevalence of depression and anxiety in patients with COPD: relationship to functional capacity. Chest 1985;87:35–38. |
| 29 . | van Ede L, Yzermans CJ, Brouwer HJ. Prevalence of depression in patients with chronic obstructive pulmonary disease: a systematic review. Thorax 1999;54:688–692. |
| 30 . | Felker B, Bush KR, Harel O, Shofer JB, Shores MM, Au DH. Added burden of mental disorders on health status among patients with chronic obstructive pulmonary disease. Prim Care Companion J Clin Psychiatry 2010;12:e1–e8. |
| 31 . | Abrams TE, Vaughan-Sarrazin M, Van der Weg MW. Acute exacerbations of chronic obstructive pulmonary disease and the effect of existing psychiatric comorbidity on subsequent mortality. Psychosomatics 2011;52:441–449. |
| 32 . | Coventry PA, Gemmell I, Todd CJ. Psychosocial risk factors for hospital readmission in COPD patients on early discharge services: a cohort study. BMC Pulm Med 2011;11:49. |
| 33 . | Carneiro R, Sousa C, Pinto A, Almeida F, Oliveira JR, Rocha N. Risk factors for readmission after hospital discharge in chronic obstructive pulmonary disease: the role of quality of life indicators. Rev Port Pneumol 2010;16:759–777. |
| 34 . | DiMatteo MR, Lepper HS, Croghan TW. Depression is a risk factor for noncompliance with medical treatment: meta-analysis of the effects of anxiety and depression on patient adherence. Arch Intern Med 2000;160:2101–2107. |
| 35 . | Sundbom LT, Bingefors K. The influence of symptoms of anxiety and depression on medication nonadherence and its causes: a population based survey of prescription drug users in Sweden. Patient Prefer Adherence 2013;7:805–811. |
| 36 . | Gross NJ. Novel antiinflammatory therapies for COPD. Chest 2012;142:1300–1307. |
| 37 . | Lahousse L, Loth DW, Joos GF, Hofman A, Leufkens HG, Brusselle GG, Stricker BH. Statins, systemic inflammation and risk of death in COPD: the Rotterdam Study. Pulm Pharmacol Ther 2013;26:212–217. |
| 38 . | Young RP, Hopkins R, Eaton TE. Potential benefits of statins on morbidity and mortality in chronic obstructive pulmonary disease: a review of the evidence. Postgrad Med J 2009;85:414–421. |
| 39 . | Young RP, Hopkins R, Chan WC, Thornley S, Gamble GD. Effect of statin therapy on mortality in COPD: analysis of cause specific death in a national cohort study [abstract]. Am J Respir Crit Care Med 2013;187:A6017. |
| 40 . | Davies L, Angus RM, Calverley PM. Oral corticosteroids in patients admitted to hospital with exacerbations of chronic obstructive pulmonary disease: a prospective randomised controlled trial. Lancet 1999;354:456–460. |
| 41 . | Niewoehner DE, Erbland ML, Deupree RH, Collins D, Gross NJ, Light RW, Anderson P, Morgan NA; Department of Veterans Affairs Cooperative Study Group. Effect of systemic glucocorticoids on exacerbations of chronic obstructive pulmonary disease. N Engl J Med 1999;340:1941–1947. |
| 42 . | Rothberg MB, Pekow PS, Lahti M, Brody O, Skiest DJ, Lindenauer PK. Antibiotic therapy and treatment failure in patients hospitalized for acute exacerbations of chronic obstructive pulmonary disease. JAMA 2010;303:2035–2042. |
| 43 . | Stefan MS, Rothberg MB, Shieh MS, Pekow PS, Lindenauer PK. Association between antibiotic treatment and outcomes in patients hospitalized with acute exacerbation of COPD treated with systemic steroids. Chest 2013;143:82–90. |
| 44 . | Sharma G, Kuo YF, Freeman JL, Zhang DD, Goodwin JS. Outpatient follow-up visit and 30-day emergency department visit and readmission in patients hospitalized for chronic obstructive pulmonary disease. Arch Intern Med 2010;170:1664–1670. |
| 45 . | Phillips CO, Wright SM, Kern DE, Singa RM, Shepperd S, Rubin HR. Comprehensive discharge planning with postdischarge support for older patients with congestive heart failure: a meta-analysis. JAMA 2004;291:1358–1367. |
| 46 . | Sin DD, Tu JV. Are elderly patients with obstructive airway disease being prematurely discharged? Am J Respir Crit Care Med 2000;161:1513–1517. |
| 47 . | Prieto-Centurion V, Markos MA, Ramey NI, Gussin HA, Nyenhuis SM, Joo MJ, Prasad B, Bracken N, Didomenico R, Godwin PO, et al. Interventions to reduce rehospitalizations after chronic obstructive pulmonary disease exacerbations: a systematic review. Ann Am Thorac Soc 2014;11:417–424. |
| 48 . | Stein BD, Bautista A, Schumock GT, Lee TA, Charbeneau JT, Lauderdale DS, Naureckas ET, Meltzer DO, Krishnan JA. The validity of International Classification of Diseases, Ninth Revision, Clinical Modification diagnosis codes for identifying patients hospitalized for COPD exacerbations. Chest 2012;141:87–93. |
Supported by the National Institutes of Health (K08-AG31583), the Agency for Healthcare Research and Quality (R01-HS020642 and 1R24HS02213401), and by the Patient-Centered Outcomes Research Institute (PCORI). This study was also funded by a University of Texas System Health Information Technology grant.
Author Contributions: All of the authors contributed to the study design, extraction, analysis, and interpretation of results, as well as manuscript preparation. The corresponding author is responsible for the content of the manuscript, including the data and analysis.
Author disclosures are available with the text of this article at www.atsjournals.org.
Predictors of COPD re-admissions
I noticed that the article does not mention whether these patients continued to smoke post hospitalization, I find in my experience that a major factor in the re-admission of these patients to the hospital is that they continue to smoke after discharge, I think it is time for the system to begin tasking these patients for personal responsibility in the management of their healthcare and the way to begin accomplishing this is to make them financially responsible for some of the cost of their healthcare.