Table 1.
Selected Heart Failure and General Readmission Risk Models Focusing on Patient, System, and Environmental Level Covariates
| Year | Model | Patient | System | Environment | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Demographic Covariates | Indicator of Frailty or Functional Status | Comorbidities | Markers of Illness Severity | Use Patterns | Hospital Characteristics and Postdischarge Services | Readiness for Discharge or Inpatient Quality | Finances, Education, Stability, and Support | Patient Behavior | ||
| 1985 | Predicting hospital readmissions in the Medicare population15 | Age, sex, race | Disability status | None | None | No. of discharges in previous 60 days, no. of discharges with same dx in past 60 days, LOS, hospital reimbursement, admission for chronic vs acute diagnosis | Hospital based characteristics, region of the country | None | Disability status, supplemental Medicaid coverage | None |
| 1988 | Identifying factors associated with health care use: A hospital‐based risk screening index16 | Age > 75 | Dependent ambulation, incontinence, poor mental status, terminal illness | 2+ chronic condition, terminal illness, psychiatric disease | None | Emergency admission, prior hospitalization within the past 2 months | None | None | Unmarried, less than subsistence level income, lives alone or in SNF, dependent self care (requires help with ADLs), unemployment or receiving disability, poor social support | History of alcoholism |
| 1988 | Postdischarge care and readmission17 | Age, sex, race | None | None | BUN, paO2, WBC, hemoglobin | ER visits in previous 6 months | Post discharge care including RN calls, mailings, and appointment reminder vs usual care | None | None | None |
| 1990 | Risk Factors for early readmission among veterans18 | Age, sex, race, period of military service, county of residence | None | Spinal cord injury, number of surgeries performed, risk category for admitting diagnosis | None | LOS, unit type (medical, intermediate, neurological, surgical), discharged against medical advice | VA auspices, place and type of disposition | None | Compensation/pension status, distance from hospital services, marital status, former POW | None |
| 1991 | Factors predicting readmission of older general medicine patients19 | Age, race | Cognition (MMSE) | Depression, diagnoses at admission (CHF or COPD) | Illness severity (Computerized Severity of Illness Index) | Emergent hospitalization, no. of hospitalizations in the last year, no. of days hospitalized in the last year, LOS, admitted from home | None | None | Income, level of education, marital status, lives alone, meeting ADLs | Novne |
| 1992 | Contribution of a measure of disease complexity (COMPLEX) to prediction of outcome and charges among hospitalized patients20 | Age, sex | None | Used a metric comprised of a count of significantly effected body systems as well as a comorbidity severity score | None | None | None | None | None | None |
| 1996 | Does risk‐adjusted readmission rate provide valid information on hospital quality19 | Age, sex | None | “complexity” measured as number of PMCs present | PMC Relative Intensity Score | LOS | None | None | None | None |
| 1997 | Correlates of early hospital readmission or death in patients with Congestive Heart Failure21 | Age, sex, race | None | History of MI, HF, VT/VF, DM, Charlson Comorbidity Index | EF, systolic BP, respiratory rate, serum sodium, serum creatinine, cardiomegaly on admission CXR, NSR on admission EKG, absence of new ST‐T changes on admission EKG | None | Has a PCP | Symptoms at discharge, laboratory abnormalities at discharge | Income, education, single, person at home to help with medical care | None |
| 1999 | Prediction of hospital readmission for heart failure: development of a simple risk score on administrative data22 | Age, sex, race | None | Charlson Comorbidity Index, specific comorbid conditions | See use pattern | LOS, total hospital discharge dollars, use of an ICU, procedural complication, discharge to SNF, transfer to acute care hospital, home health services after discharge, discharged AMA, Cardiology service, PT/OT, specific noninvasive cardiology procedures (echo, telemetry monitoring, EST, etc.), invasive cardiac procedures (PCI, CABG, etc.), critical care procedures (pulmonary artery catheterization, inotropic agents, mechanic ventilator support, HD, etc.) | Hospital location, hospital type | None | Insurance type | History of drug or alcohol abuse |
| 2000 | Predicting nonelective hospital readmissions: A multi‐site study23 | Age, gender, race | SF‐36 score physical component summary | SF‐36 mental component summary | Disease specific severity markers (eg, insulin dependence, home O2 use, NYHA class), discharge lab values including BUN, Hb, WBC | No. of ED visits in previous 6 months, no. of admissions in previous 6 months, LOS, patient satisfaction scores from survey data | None | None | Marital status, highest grade completed, distance from VAMC, employment status, service connection | None |
| 2000 | Predictors of readmission among elderly survivors of admission with heart failure24 | Age, sex, race | Discharge mobility | Specific comorbid conditions | Presences of PND, orthopnea, chest pain, systolic/diastolic blood pressure, respiratory rate, pulmonary edema on CXR, LVEF, occurrence during hospitalization of a major complication (cardiac arrest, shock, MI, stroke), major procedure during hospitalization (CABG, cardiac catheterization), labs at discharge including: sodium, BUN, creatinine, BUN/CR ratio, ACE inhibitor prescription, digoxin prescription | Previous admission within 1 year, LOS | None | None | None | None |
| 2004 | Posthospital care transitions: patterns, complications, and risk identification25 | Age, sex | Premorbid functional status score, self‐rated general health, visual impairment, need for assistance with ADLs | Charlson Comorbidity Index, specific comorbid conditions, Alzheimer's disease | None | Previous admission and average LOS in the previous 6 months, number of prior SNF stays and average LOS in previous 6 months | None | None | Medicaid status, unmarried | None |
| 2004 | Risk stratification after hospitalization for decompensated Heart Failure26 | Age, gender, race | None | Specific comorbid conditions | Duration of HF diagnosis, HF etiology, history of PCI, presence of peripheral edema, S3 murmur, EF, NYHA class, JVD, HJR, rales, heart rate, systolic BP, diastolic BP, respiratory rate, K, BUN, Cr, Na, platelets, Hb | Number of prior HF hospitalizations in the previous 12 months | None | None | None | None |
| 2006 | Identifying patients at high risk of emergency hospital admissions: A logistic regression analysis27 | Age, sex, ethnicity, Std admission ratio | None | Charleson Comorbidity Index, presence of an “ambulatory care sensitive condition” | None | No. of ED visits in past 365 days, no. of ED visits in the past 366 days to 36 months, number of consultant episodes in “index spell” | None | None | Area‐level lifestyle group, area‐level deprivation, source of admission | None |
| 2006 | Validation of the potentially avoidable hospital readmission rate as a routine indicator of the quality of hospital care28 | Age, sex | None | Charlson Comorbidity Index, number of comorbidities, diagnoses at admission | None | Previous admission in the past 6 months | None | None | None | None |
| 2007 | Improving the management of care for high‐cost Medicaid patients29 | Age, sex, race/ethnicity | None | Number and type of comorbidities, history of mental illness, history of alcohol or substance abuse | None | Frequency and interval of hospitalizations, ED visits, primary care visits, and specialist care visits in previous 3 years, Use of home health care, personal care, rehab services, substance abuse services, prescription medications, inpatient spending | None | None | Socioeconomic status of the zip code of residence | None |
| 2007 | Prediction of Rehospitalization and Death in Severe Heart Failure by Physicians and Nurses of the ESCAPE Trial30 | Age, sex, race | 6 minute walk distance | None | NYHA class, need for “high dose” loop diuretic, ischemic vs nonischemic HF, systolic BP, diastolic BP, HR, NA, Cr, BUN, EF, required CPR, required mechanical ventilation | None | None | Beta blocker prescribed at discharge, ACE inhibitor prescribed at discharge | None | None |
| 2008 | Hospital 30‐day Heart Failure readmission measure: methodology31 | Age, sex | Protein calorie malnutrition | Specific comorbid conditions | None | None | None | None | None | Drug or alcohol abuse |
| 2008 | Risk factors for 30‐day hospital readmission in patients ≥ 65 years of age32 | Age, sex, race/ethnicity | None | Specific comorbid conditions | None | Service type (medical vs surgical) | Discharge destination | None | Insurance status, distance from hospital, median income of zip code of residence | None |
| 2009 | Using routine inpatient data to identify patients at risk of hospital readmission33 | Age, sex, indigenous status, | None | Specific comorbid conditions | None | Previous admission in the preceding 90 days, 1 year, or 3 years, previous emergency admission in the preceding 90 days, 1 year, and 3 years | None | None | Marital status, socioeconomic status, rurality, geographic remoteness | None |
| 2010 | An automated model to identify heart failure patients at risk for 30‐day readmission or death using electronic medical record data13 | Age, sex, race | See markers of severity | Depression or anxiety | Tabak Mortality Risk Score (derived from albumin, total bilirubin, CK, creatinine, sodium, BUN, Pco2, WBC, troponin‐I, glucose, INR, BNP, ph, temperature, pulse, diastolic BP, systolic BP) | Number of prior admissions, ED visits, and outpatient visits, presentation to ED from 6 am to 6 pm | None | None | Socioeconomic status, single, payment method, use of a health system pharmacy | Cocaine use, history of leaving AMA, missed outpatient appointments, number of home address changes |
| 2010 | Derivation and validation on an index to predict early death or unplanned readmission after discharge from hospital to the community34 | Age, sex | Dependent for one or more ADL | Charlson Comorbidity Index | None | LOS, visit to the ED in the previous 6 months, hospital admissions within the previous 6 months, medical vs surgical admission, emergent admission, number of medications at discharge, number of new medications at discharge, season at discharge, consultation, number of complications while hospitalized | Has a PCP | None | Lives alone | None |
| 2010 | Hospital readmission in general medicine patients: A prediction model35 | Age, sex, race/ethnicity | SF‐12 physical component, MMSE, presence of functional limitation | Charlson Comorbidity Index, SF‐12 mental component | None | Number of admissions in the previous year, LOS, need for extra day stay during current admission | Has a PCP | None | Household income, education, primary insurance, marital status, lives alone, someone available to help with care | None |
| 2011 | Inability of providers to predict unplanned readmissions36 | Age, sex | Poor self‐rated general health | CAD, DM2 | None | Admission in prior year, more than 6 doctor visits in prior year | None | None | None | None |
| 2011 | Incremental value of clinical data beyond claims data in predicting 30‐day outcomes after heart failure hospitalization37 | Age, sex | None | Diagnoses at admission including psychiatric diagnoses | EF, heart rate, hemoglobin, serum creatinine, serum sodium, systolic blood pressure, weight | None | None | None | None | None |
| 2011 | Unplanned readmissions after hospital discharge among patients identified as being at high risk for readmission using a validated predictive algorithm38 | None | None | Charlson Comorbidity Index | None | LOS, number of ED in the previous 6 months, emergent admission | None | None | None | None |
Bolded covariates were included in the final model. Nonbolded covariates were proposed, but not included. LOS indicates length of stay; SNF, skilled nursing facility; ADL, activities of daily living; WBC, white blood cell; MMSE, mini‐mental state examination; CHF, congestive heart failure; HF, heart failure; VT, ventricular tachycardia; EF, ejection fraction; PCP, primary care provider; EST, exercise stress test; PCI, percutaneous coronary intervention; CABG, coronary artery bypass graft; HD, hemodialysis; Hb, hemoglobin; LVEF, left ventricular ejection fraction; ACE, angiotensin‐converting enzyme; CK, creatine kinase; INR, International Normalized Ratio; BNP, brain natriuretic peptide; CAD, coronary artery disease.BUN, blood urea nitrogen; RN, registered nurse, VA, Veteran's Administration; POW, prisoner of war; COPD, chronic obstructive pulmonary disease; PMC, patient management category; MI, myocardial infarction; VF, ventricular fibrillation; DM, diabetes mellitus; BP, blood pressure; CXR, chest x‐ray; NSR, normal sinus rhythm; EKG, electrocardiogram; ST‐T, ST or T segment; ICU, intensive care unit; PT/OT, physical therapy/occupational therapy; SF‐36, Short Form‐36; NYHA, New York Heart Association; ED, emergency department; VAMC, Veteran's Affairs Medical Center; PND, paroxysmal nocturnal dyspnea; CR, creatinine; JVD, jugular venous distention; HJR, hepatojugular reflux; ESCAPE, Evaluation Study of Congestive Heart Failure and Pulmonary Artery Catheterization Effectiveness; HR, heart rate; NA, sodium; CPR, cardiopulmonary resuscitation.