Abstract
Objectives:
Thirty-day readmission following hospitalization for acute coronary syndrome (ACS), atrial fibrillation (AF), or congestive heart failure (CHF) is common, and many occur within one week of discharge. Using a cohort of patients hospitalized for ACS, AF, or CHF, we sought to identify predictors of 30-day, early (0–7 day), and late (8–30 day) all-cause readmission.
Methods:
We identified 3531 hospitalizations for ACS, AF, or CHF at a large academic medical center between 2008 and 2018. Multivariable logistic regression models were created to identify predictors of 30-day, early, and late unplanned, all-cause readmission, adjusting for discharge diagnosis and other demographics and comorbidities.
Results:
Of 3531 patients hospitalized for ACS, AF, or CHF, 700 (19.8%) were readmitted within 30 days, and 205 (29.3%) readmissions were early. Of all 30-day readmissions, 34.8% of ACS, 16.8% of AF, and 26.0% of the CHF cohorts’ readmissions occurred early. Higher hemoglobin was associated with lower 30-day readmission [adjusted (adj) OR 0.92, 95% CI 0.88–0.97] while patients requiring intensive care unit (ICU) admission were more likely readmitted within 30 days (adj OR 1.31, 95% CI 1.03–1.67). Among patients with a 30-day readmission, females (adj OR 1.73, 95% CI 1.22, 2.47) and patients requiring ICU admission (adj OR 2.03, 95% CI 1.27, 3.26) were more likely readmitted early than late. Readmission predictors did not vary substantively by discharge diagnosis.
Conclusion:
Patients admitted to the ICU were more likely readmitted in the early and 30-day periods. Other predictors varied between readmission groups. Since outpatient follow-up often occurs beyond 1 week of discharge, early readmission predictors can help healthcare providers identify patients who may benefit from particular post-discharge services.
Keywords: acute coronary syndrome, heart failure, atrial fibrillation, readmission, intensive care unit
Introduction
Acute coronary syndrome (ACS), atrial fibrillation (AF), and congestive heart failure (CHF) are the three cardiac-related hospitalization diagnoses that most commonly result in hospital readmission within 30 days of discharge [1–9]. Not only do hospital readmissions following hospitalization for these conditions place significant financial burden on the health care system, but hospitals with risk-adjusted readmission rates greater than average readmission rates following ACS or CHF hospitalizations are penalized through the Center for Medicare and Medicaid Services’ Hospital Readmission Reduction Program [1,5,8,10–12]. Many post-discharge transitional care initiatives have been developed to aid patients’ home-to-hospital transition and minimize readmissions, but such programs may have differential benefit for patients depending on when readmission occurs. For example, a readmission within the first 7 days post-discharge may not allow sufficient time for outpatient follow-up. Understanding why patients are readmitted at varying points during the 30-day period post-discharge is crucial for minimizing 30-day readmission rates, particularly for health systems seeking to reduce overall cardiovascular re-hospitalizations through unified processes instead of implementing multiple concurrent programs for individual hospital diagnoses.
While other work has investigated predictors of early (0–7 day) readmission in the context of CHF and general medicine patients, no study has investigated early readmissions with a predictor model inclusive of patients hospitalized with either ACS, AF, or CHF [13,14]. For cardiologists managing a variety of cardiac patients, early readmission predictors for common cardiac conditions could help inform discharge planning and offer insight on how to minimize overall readmissions, as 7-day readmissions may be more amenable to prevention than late (8–30 day) readmissions [15,16]. Using a cohort of patients discharged following hospitalization for ACS, AF, or CHF, we aimed to (1) identify predictors of 30-day readmission while adjusting for discharge diagnosis and (2) identify predictors of early (0–7 day) and late (8–30 day) readmissions while also adjusting for discharge diagnosis. We hypothesized that certain characteristics would be predictive of early or late readmissions, regardless of principal discharge diagnosis (ACS, AF, or CHF). We further hypothesized that patients with CHF would be more likely to be readmitted within the first week after discharge than patients with other diagnoses.
Methods
BRIDGE Registry
The BRIDGE registry is a retrospective dataset of cardiac patients discharged from Michigan Medicine and referred to the BRIDGE clinic within 14 days of discharge [17]. The details of this registry have previously been described [18]. Briefly, the BRIDGE registry is a de-identified clinical database consisting of consecutive data that are manually abstracted from an electronic medical record for all patients referred to the BRIDGE program [17]. Data is abstracted manually by trained data abstractors, and 10% of all abstractions are audited for accuracy. The Human Subjects Internal Review Board of Michigan Medicine approved this study (HUM00035421) with a waiver of informed consent.
Study Participants
For this retrospective study, patients enrolled in the BRIDGE registry between March 2008 and October 2018 were identified if they were admitted to the hospital with a principal discharge diagnosis of ACS (includes unstable angina, ST-segment elevation myocardial infarction [STEMI], and non-ST segment elevation myocardial infarction [NSTEMI]), AF, or CHF (included systolic and diastolic heart failure). Of note, all patients enrolled in the registry who met the preceding criteria were included in the study, regardless of whether or not they actually attended a BRIDGE clinic visit.
Outcomes
The primary outcomes of this study were all-cause readmission in the following time periods: 30 days post-discharge, early period (0–7 days post-discharge), and late period (8–30 days post-discharge). Only unplanned readmissions were considered in the analysis.
Statistical Analysis
Patients with a principal discharge diagnosis of ACS, AF, or CHF from March 2008 to October 2018 were considered for inclusion. After excluding patients lost to follow-up, patients were first dichotomized into those readmitted within 30 days of discharge and those who were not readmitted within 30 days of discharge (Figure 1). Patients readmitted within 30 days were further dichotomized into those readmitted early (0–7 days) and late (8–30 days). Baseline characteristics of each readmission group were then determined. For the Charlson Comorbidity Index (CCI), a modified score was calculated based on similar assumptions utilized by Chang et al [19]. Admission to the intensive care unit (ICU) was based on the clinical needs of the patient as assessed both by the ICU physicians and nursing staff. Typically, this includes patients who required advanced respiratory or hemodynamic support, patients undergoing pulmonary artery catheter-directed diuresis, patients requiring temporary pacing, and patients within the first 24 hours after a STEMI. While time to all-cause readmission was the primary outcome, a secondary analysis was performed to characterize the readmission diagnoses for each readmission group. Readmission diagnoses were categorized as ACS, CHF, bleeding, trauma, ‘other cardiac diagnosis’, or ‘other non-cardiac diagnosis.’ For ACS hospitalizations, a readmission diagnosis of ‘ACS’ means recurrent ACS. For CHF hospitalizations, a readmission diagnosis of CHF means recurrent CHF.
Figure 1.
Patient Flow Diagram: Patients who were hospitalized with either acute coronary syndrome (ACS), atrial fibrillation (AF), or congestive heart failure (CHF) were dichotomized into those readmitted or not readmitted within 30 days of discharge. Patients readmitted within 30 days were further dichotomized into early (0–7 days) and late (8–30 days) readmissions. Patients lost to follow-up were excluded.
Bivariate analyses were conducted to compare 30-day readmissions to those not readmitted within 30 days as well as early to late readmissions. Chi-square tests were used for categorical variables while parametric and Wilcoxon-rank sum tests were used to for continuous variables with normal or non-normal distributions, respectively. All reported p-values are two-sided, and those ≤ 0.05 were considered significant. Multivariable logistic regression models were then created to identify predictors of 30-day, early, and late all-cause readmission. Based on review of key cardiovascular readmissions literature, we identified 12 variables to include in our models a priori: sex, age, anemia, diabetes, acute and chronic renal disease, CCI, ICU admission, length of hospital stay, hemoglobin upon admission, creatinine upon admission, and blood urea nitrogen upon admission (Supplemental Table 1). Discharge diagnosis was also included a priori to address our research question. Reported odds ratios and 95% confidence intervals are adjusted for all other variables in the model (Supplemental Table 1).
Results
Of 3531 patients referred for a BRIDGE clinic visit following hospital discharge for ACS, AF, or CHF, 700 (19.8%) resulted in an unplanned, all-cause readmission within 30 days. Of these 30-day readmissions, 205 (29.3%) were early. Demographics, principal discharge diagnoses, comorbidities, hospitalization characteristics, discharge medications, and 6-month all-cause mortality rates are summarized in Table 1 for 30-day, early, and late readmissions as well as for patients not readmitted within 30 days of discharge. The percentage of 30-day, early, and late readmissions involving female patients were 44.9%, 53.2%, and 41.4%, respectively. On average, patients readmitted within 30 days were older than those not readmitted within 30 days (p<0.0001), while the average age for early and late readmissions did not significantly differ (p=0.30). Racial composition was similar between 30-day readmissions and non-30-day readmission (p=0.51) as well as for early and late readmissions (p=0.64).
Table 1.
Baseline characteristics, discharge diagnoses, comorbid conditions, discharge medications, and post-hospitalization characteristics of 30-day readmissions versus no 30-day readmission and early (0–7 day) versus late (8–30 day) readmission.
| 30 day readmissions (n=700) | No 30-day readmissions (n=2831) | p-value | 0–7 day readmissions (n=205) | 8–30 day readmissions (n=495) | p-value | |
|---|---|---|---|---|---|---|
|
| ||||||
| Demographics | ||||||
|
| ||||||
| Female, n (%) | 314 (44.9) | 1156 (40.8) | 0.05 | 109 (53.2) | 205 (41.4) | 0.004 |
| Age (years), (mean ± SD) |
69.48 ± 13.12 | 66.98 ± 14.04 | <0.0001 | 70.28 ± 13.9 | 69.15 ± 12.79 | 0.30 |
| Non-white, n (%) | 95 (13.7) | 412 (14.7) | 0.51 | 30 (14.6) | 65 (13.3) | 0.64 |
|
| ||||||
| Principal Discharge Diagnosis, n (%) | ||||||
|
| ||||||
| ACS | 224 (32.0) | 1065 (37.6) | <0.0001 | 78 (38.1) | 146 (29.5) | 0.08 |
| AF | 161 (23.0) | 758 (26.8) | 45 (22.0) | 116 (23.4) | ||
| CHF | 315 (45.0) | 1008 (35.6) | 82 (40.0) | 233 (47.1) | ||
|
| ||||||
| Prevalence of Comorbid Conditions, n (%) | ||||||
|
| ||||||
| AF/Atrial Flutter | 365 (52.1) | 1428 (50.4) | 0.42 | 108 (52.7) | 257 (51.9) | 0.85 |
| Anemia | 231 (33.0) | 665 (23.5) | <0.0001 | 68 (33.2) | 163 (32.9) | 0.95 |
| Aortic Stenosis | 122 (17.4) | 328 (11.6) | <0.0001 | 32 (15.6) | 90 (18.2) | 0.41 |
| Any Cardiomyopathy | 193 (27.6) | 642 (22.7) | 0.006 | 54 (26.3) | 139 (28.1) | 0.64 |
| Cerebrovascular disease | 143 (20.4) | 447 (15.8) | 0.003 | 46 (22.4) | 97 (19.6) | 0.40 |
| Current smoker | 68 (9.7) | 349 (12.3) | 0.06 | 15 (7.32) | 53 (10.71) | 0.17 |
| Diabetes mellitus | 279 (39.9) | 992 (35.0) | 0.02 | 77 (37.6) | 202 (40.8) | 0.42 |
| Dyslipidemia | 391 (55.9) | 1684 (59.5) | 0.08 | 119 (58.1) | 272 (55.0) | 0.45 |
| GERD | 216 (30.9) | 790 (27.9) | 0.12 | 67 (32.7) | 149 (30.1) | 0.50 |
| Hypertension | 530 (75.7) | 2086 (73.7) | 0.27 | 149 (72.7) | 381 (77.0) | 0.23 |
| ICD/Pacemaker | 110 (15.7) | 400 (14.1) | 0.29 | 33 (16.1) | 77 (15.6) | 0.86 |
| Liver Disease | 35 (5.0) | 139 (4.9) | 0.92 | 13 (6.3) | 22 (4.4) | 0.29 |
| Malignancy | 156 (22.3) | 570 (20.1) | 0.21 | 47 (22.9) | 109 (22.0) | 0.80 |
| Obesity (BMI ≥ 30) | 159 (22.7) | 721 (25.5) | 0.13 | 44 (21.5) | 115 (23.2) | 0.61 |
| Major Psychiatric Disorder | 248 (35.4) | 958 (33.8) | 0.43 | 77 (37.6) | 171 (34.6) | 0.45 |
| Acute renal diease | 186 (26.6) | 483 (17.1) | <0.0001 | 55 (26.8) | 131 (26.5) | 0.92 |
| Chronic renal disease | 253 (36.1) | 749 (26.5) | <0.0001 | 75 (36.6) | 178 (36.0) | 0.88 |
| Pulmonary disease | 362 (51.7) | 1265 (44.7) | 0.001 | 112 (54.6) | 250 (50.5) | 0.32 |
| Valvular heart disease | 197 (28.1) | 614 (21.7) | 0.0003 | 60 (29.3) | 137 (27.7) | 0.67 |
| Vascular disease | 201 (28.7) | 570 (20.1) | <0.0001 | 56 (27.3) | 145 (29.3) | 0.60 |
| CCI, median, (25th,75th) | 6.10 (4.5, 7.7) | 5.30 (3.5, 7.0) | <0.0001 | 6.30 (4.8, 7.7) | 6.0 (4.3, 7.6) | 0.17 |
|
| ||||||
| Hospitalization Characteristics | ||||||
|
| ||||||
| ICU admission, n (%) | 124 (17.8) | 432 (15.3) | 0.11 | 47 (23.04) | 77 (15.7) | 0.02 |
| Total length of stay in days, (mean ± SD) | 6.59 ± 6.01 | 5.76 ± 6.55 | <0.0001 | 6.22 ± 5.14 | 6.74 ± 6.34 | 0.54 |
| Hemoglobin on arrival, (mean ± SD) | 11.95 ± 2.25 | 12.72 ± 2.32 | <0.0001 | 11.98 ± 2.16 | 11.93 ± 2.29 | 0.79 |
| Creatinine on arrival, (mean ± SD) | 1.55 ± 1.30 | 1.31 ± 1.14 | <0.0001 | 1.56 ± 1.52 | 1.54 ± 1.20 | 0.41 |
| BUN on arrival, (mean ± SD) |
32.52 ± 21.70 | 26.66 ± 17.85 | <0.0001 | 32.03 ± 20.32 | 32.73 ± 22.26 | 0.92 |
|
| ||||||
| Discharge Medications, n (%) | ||||||
|
| ||||||
| ACE/ARB | 389 (55.6) | 1670 (59.0) | 0.10 | 116 (56.6) | 273 (55.2) | 0.73 |
| Amiodarone | 63 (9.0) | 255 (9.1) | 0.98 | 18 (8.8) | 45 (9.1) | 0.89 |
| P2Y12 Inhibitor | 226 (32.3) | 915 (32.4) | 0.97 | 65 (31.7) | 161 (32.5) | 0.83 |
| Aspirin | 508 (72.6) | 2053 (72.6) | 0.99 | 152 (74.2) | 356 (71.9) | 0.55 |
| Beta Blocker | 556 (79.4) | 2260 (79.9) | 0.79 | 162 (79.0) | 394 (79.6) | 0.86 |
| CCB | 153 (21.9) | 636 (22.6) | 0.69 | 47 (22.9) | 106 (21.4) | 0.66 |
| Digoxin | 62 (8.9) | 230 (8.2) | 0.56 | 16 (7.8) | 46 (9.3) | 0.53 |
| Nitrate | 196 (36.0) | 764 (33.6) | 0.29 | 63 (40.1) | 133 (34.4) | 0.21 |
| Statin | 516 (73.7) | 2061 (72.9) | 0.65 | 154 (75.1) | 362 (73.1) | 0.59 |
| Loop Diuretic | 389 (55.7) | 1247 (44.2) | <0.0001 | 98 (47.8) | 291 (58.9) | 0.007 |
| Thiazide Diuretic | 81 (11.6) | 265 (9.4) | 0.08 | 26 (12.7) | 55 (11.1) | 0.55 |
| MRA | 71 (13.1) | 280 (12.4) | 0.65 | 16 (10.3) | 55 (14.2) | 0.21 |
| *Oral Anticoagulant | 282 (40.3) | 1136 (40.1) | 0.94 | 78 (38.1) | 204 (41.2) | 0.44 |
|
| ||||||
| Post-Hospitalization Characteristics | ||||||
|
| ||||||
| All-Cause Mortality at 180 days post-discharge, n (%) | 137 (19.8) | 189 (6.7) | <0.0001 | 40 (19.7) | 97 (19.8) | 0.94 |
Abbreviations:
ACS = acute coronary syndrome; ACE = angiotensin-converting enzyme inhibitor; AF = atrial fibrillation; ARB = angiotensin receptor blocker; BMI = body mass index; BUN = blood urea nitrogen; CCB = calcium channel blocker; CCI = Charlson Comorbidity Index; CHF = congestive heart failure; GERD = gastroesophageal reflux disease; ICD = implantable cardioverter-defibrillator; ICU = intensive care unit; MRA = mineralocorticoid receptor antagonist
Oral Anticoagulants included were warfarin or Xa inhibitor.
Of all 30-day readmissions, we observed that 34.8% of ACS, 16.8% of AF, and 26.0% of the CHF cohorts’ readmissions occurred during days 0 through 7 post-discharge (Figure 2). For days 0 through 14 following discharge, 65.6%, 65.8%, and 57.5% of all-cause 30-day readmissions occurred for the ACS, AF, and CHF patient cohorts, respectively. Throughout the entire 30-day post-discharge period, the cumulative timing of 30-day all-cause readmissions by day was similar across all three discharge diagnoses.
Figure 2.
Thirty-Day readmissions by day (0–30) following discharge for acute coronary syndrome (ACS), atrial fibrillation (AF), or congestive heart failure (CHF). Bars represent the percentage of all-cause 30-day readmissions (primary vertical axis) per day by discharge diagnosis while lines represent the cumulative percentage of all-cause 30-day readmissions (secondary vertical axis) per day by discharge diagnosis.
Reasons for readmission in the 30-day, early, and late time periods are described in Supplemental Table 2. In summary, the most common reason for 30-day readmission or late readmission among the ACS and AF cohorts was ‘other cardiac diagnosis’ while recurrent CHF was the most common reason for 30-day readmission in the CHF cohort. ‘Other non-cardiac diagnosis,’ ‘other cardiac diagnosis,’ and recurrent CHF were the most common reasons for early readmission in the ACS, AF, and CHF cohorts, respectively.
The results of multivariable logistic regression models identifying predictors of 30-day, early, and late all-cause readmission are included in Figure 3. Briefly, each unit increase in hemoglobin (grams per deciliter) was associated with 8% lower odds of 30-day readmission (OR 0.92, 95% CI 0.88–0.97), adjusting for all variables in the model. In contrast, requiring ICU admission during index hospitalization was associated with greater odds of 30-day readmission (OR 1.31, 95% CI 1.03–1.67), after adjustment. Relative to late readmissions, females (OR 1.73, 95% CI 1.22, 2.47) and patients requiring ICU admission during index hospitalization (OR 2.03, 95% CI 1.27, 3.26), both had greater adjusted odds of early readmission. Discharge diagnosis was not a significant independent predictor in either the 30-day readmission or early versus late readmission models.
Figure 3.
Multivariable logistic regression models comparing 30-day readmissions versus no readmission in 30 days (reference) as well as early (0–7 day) versus late (8–30 day) readmissions (reference). Reported odds ratio estimates and 95% confidence intervals are adjusted for every variable listed in this figure. Specific odds ratio values are reported in Supplemental Table 3.
Discussion
For this study analyzing a cohort of patients discharged from the hospital with ACS, AF, or CHF, we identified three key findings. Relative to the AF and CHF cohorts, the ACS cohort had a greater proportion of 30-day readmissions occur within the first week of discharge. Female patients were more likely than males to be readmitted within the first week of discharge. Additionally, ICU admission during index hospitalization was associated with 30-day readmission and greater odds of early readmission relative to patients readmitted late, regardless of initial discharge diagnosis.
In this study, the proportion of 30-day readmissions occurring within the first week of discharge was greater among patients discharged following an ACS than for patients discharged following an AF or CHF hospitalization. This finding is important, as it highlights an opportunity for health systems seeking to minimize early readmissions to consider diagnosis-specific readmission-reduction interventions. For example, there is an abundance of literature describing how revascularization by percutaneous coronary intervention is associated with lower 30-day readmissions among ACS patients [9,20–22]. Considering interventions specific to ACS may be beneficial for health systems seeking to optimally minimize early readmissions.
The association between female sex and readmission in our study is concordant with a growing body of literature [2,8,9,21,23,24]. In our analyses, the association between female sex and 30-day readmission became statistically insignificant after adjusting for potential confounding variables. However, among patients readmitted within 30 days, women had roughly 70% greater odds of readmission within the first 7 days of discharge relative to men, regardless of initial discharge diagnosis (ACS, AF, or CHF). This contrasts with other studies showing that readmission timing did not substantively vary by sex following hospitalization for acute myocardial infarction or heart failure [4,24]. Further research is needed to understand the implications of this finding in order to better inform discharge planning, especially since early readmissions are potentially more preventable and often occur before outpatient follow-up can occur [15,16,18].
To date, there are mixed results in the literature regarding whether ICU admission confers increased, negligible, or decreased risk of all-cause hospital readmission [25–30]. However, it is widely recognized that patients surviving ICU admission represent a vulnerable group at risk for the post-intensive care syndrome, which makes returning to baseline quality of life difficult [25,31]. In a recent study by Chesley et al [25], ICU survivors had comparable 30-day readmission risk relative to those not requiring ICU admission, but the former were more likely to be readmitted within 7 days. We similarly observed that ICU survivors were more likely readmitted in 7 days compared to those not requiring ICU care during index hospitalization. However, we also observed an increased risk of readmission across the entire 30-day post-discharge period among ICU survivors. Our findings suggest that ICU survivors may have a differential risk of all-cause readmission within the first 7 days compared to later in the 30-day time period relative to patients not admitted to the ICU, irrespective of initial cardiovascular hospital discharge diagnosis (ACS, AF, or CHF). This finding is important, as the average time for a patient to see a provider in outpatient follow-up is 2 weeks post-discharge [18]; therefore, a substantial proportion of patients may not be receiving outpatient care prior to requiring readmission. Prior work has demonstrated that hospitals with greater rates of 1-week follow-up for patients following CHF hospitalization have lower 30-day readmission rates [32]. Considering prompt referral to post-ICU clinics or encouraging patients to utilize telehealth services within the first week of discharge may be beneficial for health systems seeking to minimize early readmissions [31,33,34].
This study has important strengths. Since the BRIDGE registry has been maintained for more than 10 years, we had a large sample of ACS, AF, and CHF patients, and a considerable number of these patients were readmitted within 30 days of discharge. Additionally, the data in our study were manually abstracted by trained data abstractors and were not limited to claims-based data sources. Furthermore, our cohort was not limited by age or insurance status.
This study also has several limitations. As this is an observational study based on retrospective data, causality cannot be determined. Furthermore, because the BRIDGE registry contains data from a single academic center with limited racial/ethnic diversity, the results should be applied cautiously to other patient populations. Information bias via misclassification must also be considered, since patients with multiple active conditions had to be categorized into a single principal discharge diagnosis. Lastly, because this registry is maintained within one institution, the registry may underreport readmission to outside sites.
Conclusion
In summary, the proportion of 30-day readmissions occurring within the first week of discharge was greater among patients discharged following an ACS than for patients discharged following an AF or CHF hospitalization. Additionally, female patients were more likely to be readmitted early versus late within the first 30-days following hospital discharge, regardless of discharge diagnosis. ICU survivors were more likely readmitted in the early and 30-day periods compared to patients not requiring ICU care during index hospitalization. Since outpatient follow-up often occurs after the first week of discharge, early readmission predictors can help healthcare providers identify patients who may benefit from greater medical attention prior to discharge or referral to particular post-discharge services.
Supplementary Material
Supplemental Table 3. Adjusted odds ratios and 95% confidence intervals reported in figure 3 for 30-day readmission and early versus late multivariable logistic readmission models.
Supplemental Table 1. Studies supporting use of particular variables in 30-day vs. no 30-day readmission and early (0–7 day) vs. late (8–30 day) readmission multivariable logistic regression models.
Supplemental Table 2. Reasons for readmission stratified by time to readmission and principal discharge diagnosis from index hospitalization.
Acknowledgements
The authors would like to acknowledge the BRIDGE clinic nurse practitioners for their hard work in the clinic and their contribution to the registry, as well as the MCORRP student interns, who have abstracted the majority of the clinical data since the BRIDGE registry’s inception.
Declaration of funding
George Cholack is funded by grant 5TL1R002242 from the National Center for Advancing Translational Sciences (NCATS), which is awarded to Michigan Institute for Clinical and Health Research at the University of Michigan, Ann Arbor, MI, USA. The funding source had no role in creating the study design, data collection, analysis, or interpretation of results.
Declaration of financial/other relationships
GDB: discloses consulting fees from Pfizer/Bristol-Myers Squibb, Janssen, AMAG Pharmaceuticals and Acelis Connected Health.
EKR: discloses consulting services to Janssen as well as being a member of the Board of Directors for Anticoagulation Forum.
GC: Funding from National Center for Advancing Translational Sciences (as above).
The remaining authors declare no conflicts of interest.
Footnotes
Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.
References
- 1.Amin AN, Jhaveri M, Lin J. Hospital readmissions in us atrial fibrillation patients: occurrence and costs. Am J Ther. 2013;20(2):143–150. [DOI] [PubMed] [Google Scholar]
- 2.Bilal Munir M, Sharbaugh MS, Ahmad S, et al. Causes and predictors of 30-day readmissions in atrial fibrillation (from the nationwide readmissions database). Am J Cardiol. 2017;120(3):399–403. [DOI] [PubMed] [Google Scholar]
- 3.Cowie MR, Anker SD, Cleland JGF, et al. Improving care for patients with acute heart failure: before, during and after hospitalization. ESC Hear Fail. 2014;1(2):110–145. [DOI] [PubMed] [Google Scholar]
- 4.Dharmarajan K, Hsieh AF, Lin Z, et al. Diagnoses and timing of 30-day readmissions after hospitalization for heart failure, acute myocardial infarction, or pneumonia. JAMA - J Am Med Assoc. 2013;309(4):355–363. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Jencks SF, Williams M V, Coleman EA. Rehospitalizations among patients in the medicare fee-for-service program. N Engl J Med. 2009;360(14):1418–1428. [DOI] [PubMed] [Google Scholar]
- 6.Wang H, Zhao T, Wei X, et al. The prevalence of 30-day readmission after acute myocardial infarction: a systematic review and meta-analysis. Clin Cardiol. 2019;42(10):889–898. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Blecker S, Herrin J, Li L, et al. Trends in hospital readmission of medicare-covered patients with heart failure. J Am Coll Cardiol. 2019;73(9):1004–1012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Kim LK, Yeo I, Cheung JW, et al. Thirty-day readmission rates, timing, causes, and costs after st-segment-elevation myocardial infarction in the united states: a national readmission database analysis 2010–2014. J Am Heart Assoc. 2018;7(18):e009863. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Lemor A, Hernandez GA, Patel N, et al. Predictors and etiologies of 30-day readmissions in patients with non-st-elevation acute coronary syndrome. Catheter Cardiovasc Interv. 2018;93(3):373–379. [DOI] [PubMed] [Google Scholar]
- 10.Centers for Medicare & Medicaid Services. Readmissions Reduction Program (HRPP). Available from:https://www.cms.gov/medicare/medicare-fee-for-service-payment/acuteinpatientpps/readmissions-reduction-program.html. Last updated August 6th, 2021; cited May 25th, 2021.
- 11.Lahewala S, Arora S, Tripathi B, et al. Heart failure: same-hospital vs. different-hospital readmission outcomes. Int J Cardiol. 2019;278:186–191. [DOI] [PubMed] [Google Scholar]
- 12.Lu C, Clark CM, Tober R, et al. Readmissions and costs among younger and older adults for targeted conditions during the enactment of the hospital readmission reduction program. BMC Heal Serv Res. 2021;21(1):386. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Fudim M, O’Connor CM, Dunning A, et al. Aetiology, timing and clinical predictors of early vs. late readmission following index hospitalization for acute heart failure: insights from ascend-hf. Eur J Heart Fail. 2018;20(2):304–314. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Graham KL, Wilker EH, Howell MD, et al. Differences between early and late readmissions among patients: a cohort study. Ann Intern Med. 2015;162(11):741–749. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Graham KL, Dike O, Doctoroff L, et al. Preventability of early vs. late readmissions in an academic medical center. PLoS One. 2017;12(6):e0178718. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Graham KL, Auerbach AD, Schnipper JL, et al. Preventability of early versus late hospital readmissions in a national cohort of general medicine patients. Ann Intern Med. 2018;168(11):766–774. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Bumpus SM, Brush BL, Pressler SJ, et al. A transitional care model for patients with acute coronary syndrome. Am J Accountable Care. 2014;2(2):39–52. [Google Scholar]
- 18.Bumpus SM, Krallman R, Kline-Rogers E, et al. Transitional care to reduce cardiac readmissions: 5-year results from the bridge clinic. J Fam Med Dis Prev. 2017;3(3):3–9. [Google Scholar]
- 19.Chang CM, Yin WY, Wei CK, et al. Adjusted age-adjusted charlson comorbidity index score as a risk measure of perioperative mortality before cancer surgery. PLoS One. 2016;11(2):e0148076. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Kwok CS, Wong CW, Shufflebotham H, et al. Early readmissions after acute myocardial infarction. Am J Cardiol. 2017;120(5):723–728. [DOI] [PubMed] [Google Scholar]
- 21.Nguyen OK, Makam AN, Clark C, et al. Predicting 30-day hospital readmissions in acute myocardial infarction: the ami “readmits” (renal function, elevated brain natriuretic peptide, age, diabetes mellitus, nonmale sex, intervention with timely percutaneous coronary intervention, and low systolic blood pressure) score. J Am Heart Assoc. 2018;7(8):e008882. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Sangu PV, Ranasinghe I, Aliprandi Costa B, et al. Trends and predictors of rehospitalisation following an acute coronary syndrome: report from the australian and new zealand population of the global registry of acute coronary events (grace). Heart. 2012;98(23):1728–1731. [DOI] [PubMed] [Google Scholar]
- 23.Dreyer RP, Dharmarajan K, Hsieh AF, et al. Sex differences in trajectories of risk after rehospitalization for heart failure, acute myocardial infarction, or pneumonia. Circ Cardiovasc Qual Outcomes. 2017;10(5):e003271. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Dreyer RP, Ranasinghe I, Wang Y, et al. Sex differences in the rate, timing, and principal diagnoses of 30-day readmissions in younger patients with acute myocardial infarction. Circulation. 2015;132(3):158–166. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Chesley CF, Harhay MO, Small DS, et al. Hospital readmission and post-acute care use after intensive care unit admissions: new icu quality metrics? J Intensive Care Med. Published online 2020:1–9. [DOI] [PubMed] [Google Scholar]
- 26.Wunsch H, Guerra C, Barnato AE, et al. Three-year outcomes for medicare beneficiaries. JAMA - J Am Med Assoc. 2010;303(9):849–856. [DOI] [PubMed] [Google Scholar]
- 27.Lone NI, Gillies MA, Haddow C, et al. Five-year mortality and hospital costs associated with surviving intensive care. Am J Respir Crit Care Med. 2016;194(2):198–208. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Hill AD, Fowler RA, Pinto R, et al. Long-term outcomes and healthcare utilization following critical illness - a population-based study. Crit Care. 2016;20(1):1–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Garland A, Olafson K, Ramsey CD, et al. A population-based observational study of intensive care unit-related outcomes with emphasis on post-hospital outcomes. Ann Am Thorac Soc. 2015;12(2):202–208. [DOI] [PubMed] [Google Scholar]
- 30.Keenan SP, Dodek P, Chan K, et al. Intensive care unit survivors have fewer hospital readmissions and readmission days than other hospitalized patients in british columbia. Crit Care Med. 2004;32(2):391–398. [DOI] [PubMed] [Google Scholar]
- 31.Kuehn BM. Clinics aim to improve post-icu recovery. JAMA - J Am Med Assoc. 2019;321(11):1036–1038. [DOI] [PubMed] [Google Scholar]
- 32.Hernandez AF, Greiner MA, Fonarow GC, et al. Relationship between early physician follow-up and 30-day readmission among medicare beneficiaries hospitalized for heart failure. JAMA - J Am Med Assoc. 2010;303(17):1716–1722. [DOI] [PubMed] [Google Scholar]
- 33.Jensen JF, Thomsen T, Overgaard D, et al. Impact of follow-up consultations for icu survivors on post-icu syndrome: a systematic review and meta-analysis. Intensive Care Med. 2015;41(5):763–775. [DOI] [PubMed] [Google Scholar]
- 34.Connor MO, Asdornwised U, Dempsey ML, et al. Using telehealth to reduce all-cause 30-day hospital readmissions among heart failure patients receiving skilled home health services. Appl Clin Inform. 2016;7(2):238–247. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental Table 3. Adjusted odds ratios and 95% confidence intervals reported in figure 3 for 30-day readmission and early versus late multivariable logistic readmission models.
Supplemental Table 1. Studies supporting use of particular variables in 30-day vs. no 30-day readmission and early (0–7 day) vs. late (8–30 day) readmission multivariable logistic regression models.
Supplemental Table 2. Reasons for readmission stratified by time to readmission and principal discharge diagnosis from index hospitalization.