Overview

Hospital readmission within 30 days of discharge after acute myocardial infarction (AMI) is one of the most scrutinized quality metrics in cardiovascular medicine. It affects approximately 15 to 22% of AMI survivors, generates substantial healthcare costs, and is increasingly used by payers and regulators as a surrogate for care quality. The READMITS score is a validated, bedside-applicable clinical decision tool that quantifies individual patient risk for 30-day all-cause readmission following AMI hospitalization. Derived from the landmark Enhanced Feedback for Effective Cardiac Treatment (EFFECT) study cohort in Ontario, Canada, and externally validated in large provincial administrative databases, READMITS synthesizes eight independently predictive variables into a single risk score that stratifies patients into low, intermediate, and high readmission risk categories.

The eight variables are encoded in the acronym READMITS: Renal failure, Ejection fraction, Age, Diabetes, MI type, Income, Transfer status, and Sex. Each variable is assigned a weighted integer point value derived from its adjusted hazard ratio in multivariable Cox regression modeling. The composite score enables clinicians, care coordinators, and hospital quality programs to identify patients at elevated readmission risk before discharge, allowing targeted deployment of post-discharge support resources such as cardiac rehabilitation referral, early follow-up appointments, medication reconciliation, and transitional care nurse navigator programs.

Epidemiology and Significance of Post-AMI Readmission

Incidence and Burden

Despite advances in reperfusion therapy, dual antiplatelet therapy, high-intensity statin use, renin-angiotensin-aldosterone system (RAAS) blockade, and evidence-based secondary prevention, the 30-day readmission rate following AMI remains persistently high across health systems worldwide. Population-based registry data from the United States (Get With The Guidelines-Coronary Artery Disease, PREMIER database), Canada (EFFECT, GRACE Canada), and Europe (FAST-MI, MINAP) consistently report 30-day readmission rates of 14 to 22%, with significant variation by patient subgroup, institution, and region.

In the United States, AMI was among the three conditions originally targeted by the Centers for Medicare and Medicaid Services (CMS) Hospital Readmissions Reduction Program (HRRP), enacted under the Affordable Care Act in 2012. Hospitals with excess 30-day readmission rates for AMI face financial penalties of up to 3% of total Medicare payments, creating strong institutional incentives for readmission reduction. However, risk adjustment for patient complexity is inherently imperfect in administrative models, underscoring the clinical value of individualized bedside risk scores such as READMITS.

Causes of Early Readmission

Post-AMI readmissions are heterogeneous in etiology. Approximately 30 to 40% are directly related to the index cardiovascular event or its complications, including:

• Recurrent ACS (unstable angina, NSTEMI, or reinfarction)
• Heart failure exacerbation secondary to reduced left ventricular ejection fraction (LVEF) or ischemic mitral regurgitation
• Arrhythmias (atrial fibrillation, ventricular tachycardia, complete heart block)
• Mechanical complications of MI: ventricular septal defect, free wall rupture, papillary muscle rupture
• Pericarditis or Dressler syndrome
• Contrast-induced or ischemia-related acute kidney injury progression
• Bleeding complications from antiplatelet and anticoagulant therapy

The remaining 60 to 70% of early readmissions are attributable to non-cardiac causes, including infections (pneumonia, urinary tract infections), exacerbations of comorbid conditions (COPD, diabetes, renal failure), and socioeconomic barriers to outpatient care. This mixed etiology underscores that post-AMI readmission risk is not solely a cardiovascular risk measure; it is an integrated index of patient vulnerability, comorbidity burden, and care system support.

Readmission as a Quality Metric: Controversies

The use of 30-day all-cause readmission as a hospital quality indicator has been debated extensively. Key concerns include:

• Inadequate risk adjustment: Administrative risk models used by CMS and equivalent bodies capture coded diagnoses but not granular clinical variables (LVEF, renal function, functional status), potentially penalizing hospitals serving higher-complexity or lower-income populations.
• Perverse incentives: Pressure to reduce readmissions may lead to avoidance of appropriate readmissions for genuinely ill patients, or to premature discharge of high-risk patients to reduce the denominator.
• Observation status gaming: Hospitals may reclassify readmitted patients as "observation" status rather than inpatient admissions to avoid counting them as readmissions, an approach that shifts cost burden to patients without improving outcomes.
• Some readmissions are unavoidable: For the highest-risk patients with severe heart failure, advanced renal disease, or multiple serious comorbidities, a readmission may represent appropriate escalation of care rather than a quality failure.

These controversies reinforce the value of patient-level risk stratification tools like READMITS, which allow differentiation of avoidable from high-risk-but-inevitable readmissions and more precise targeting of preventive interventions.

Derivation: The EFFECT Study

The Enhanced Feedback for Effective Cardiac Treatment (EFFECT) study was a cluster-randomized trial conducted in Ontario, Canada, that enrolled AMI and heart failure patients discharged from 86 acute care hospitals between 1999 and 2001 (Phase 1) and 2004 and 2005 (Phase 2). The primary aim of EFFECT was to evaluate the effect of performance feedback to hospitals on adherence to evidence-based care processes and patient outcomes.

The READMITS score was derived from the EFFECT AMI cohort by Southern and colleagues (Circulation: Cardiovascular Quality and Outcomes, 2014). The derivation cohort comprised approximately 9,000 AMI patients across both EFFECT phases. Patients who died during the index hospitalization were excluded. The outcome was 30-day all-cause readmission to any acute care hospital.

Candidate variables were selected based on clinical plausibility, availability at the time of hospital discharge, and prior literature. Multivariable Cox proportional hazards regression with backward selection identified the eight variables comprising the READMITS score. Integer point values were assigned proportional to the natural logarithm of each variable's adjusted hazard ratio, following established methods for converting regression coefficients to integer score weights (analogous to the Framingham risk score methodology).

The score was externally validated in a separate province-wide Ontario administrative database cohort of over 40,000 AMI patients, demonstrating consistent discriminatory performance (c-statistic approximately 0.64 to 0.66) and good calibration across risk deciles.

READMITS Score: Variables and Weighting

The eight READMITS variables, their definitions, and their assigned point values are as follows:

Variable	Definition	Points
R — Renal failure	Serum creatinine ≥200 µmol/L (approximately ≥2.3 mg/dL) or chronic dialysis	7
E — Ejection fraction	Left ventricular ejection fraction <40%	3
A — Age	Age ≥75 years	3
D — Diabetes mellitus	Known diagnosis of diabetes mellitus (type 1 or type 2)	2
M — MI type	Non-ST-elevation MI (NSTEMI) vs. STEMI	2
I — Income	Lowest neighborhood income quintile	2
T — Transfer	Transferred from another acute care facility for the index AMI admission	2
S — Sex	Female biological sex	1

Total score range: 0 to 22 points.

Risk Stratification

READMITS Score	Risk Category	Approximate 30-Day Readmission Risk
0–3	Low	~10%
4–6	Low-Intermediate	~14%
7–9	Intermediate	~18%
10–13	Intermediate-High	~22–25%
≥14	High	>30%

Detailed Variable Analysis

R: Renal Failure (7 points)

Renal failure is by far the most heavily weighted READMITS variable, reflecting the profound impact of impaired renal function on virtually every adverse outcome following AMI. The threshold used is serum creatinine greater than or equal to 200 micromol/L (approximately 2.3 mg/dL) or dependence on chronic renal replacement therapy (hemodialysis or peritoneal dialysis). At this level of renal impairment, the estimated glomerular filtration rate (eGFR) by CKD-EPI or MDRD equation is typically below 25 to 30 mL/min/1.73m².

The mechanisms underlying the strong association between advanced renal failure and post-AMI readmission are multifactorial:

• Volume overload and heart failure: Impaired urinary sodium and water excretion predisposes to fluid retention, pulmonary congestion, and decompensated heart failure, the most common cardiovascular cause of early readmission in AMI patients with reduced LVEF.
• Accelerated coronary artery disease: CKD is an independent risk factor for accelerated atherosclerosis and plaque instability, driven by uremic dyslipidemia, oxidative stress, endothelial dysfunction, and systemic inflammation. Patients with advanced CKD have higher rates of recurrent ACS.
• Electrolyte dysregulation: Hyperkalemia (from RAAS blockade in the setting of impaired potassium excretion) and hyponatremia are common, may limit guideline-directed medical therapy, and predispose to life-threatening arrhythmias necessitating readmission.
• Contrast-induced nephropathy progression: Patients undergoing coronary angiography or PCI during the index hospitalization may develop contrast-associated acute kidney injury superimposed on CKD, further reducing renal reserve and increasing post-discharge morbidity.
• Polypharmacy and medication complexity: AMI patients with CKD require dose adjustments for multiple guideline-directed medications (antiplatelet agents, anticoagulants, RAAS inhibitors, diuretics), increasing the risk of medication errors, under-treatment, and adverse drug events that precipitate readmission.
• Dialysis scheduling: Patients on chronic hemodialysis have multiple scheduled and unscheduled healthcare encounters per week that increase the probability of being admitted for any of the above complications within 30 days of AMI discharge.

E: Ejection Fraction <40% (3 points)

Left ventricular systolic dysfunction (LVEF below 40%), also known as heart failure with reduced ejection fraction (HFrEF) when symptomatic, is a direct consequence of myocardial necrosis in AMI. Anterior STEMI with large infarct territory (LAD occlusion) produces the most severe LVEF reductions, though significant dysfunction can follow any AMI type depending on infarct size, pre-existing myocardial fibrosis, and collateral circulation.

Post-AMI LVEF below 40% is associated with readmission through multiple pathways:

• Decompensated heart failure: Reduced systolic function leads to elevated filling pressures, neurohormonal activation, and progressive fluid retention. Despite in-hospital diuresis, patients may be discharged in a partially compensated state, with symptoms recurring rapidly after discharge as neurohormonal systems are re-activated.
• Ventricular arrhythmias: Post-MI scar tissue creates re-entrant circuits predisposing to ventricular tachycardia and fibrillation, particularly in the first weeks after AMI when the infarcted tissue is undergoing remodeling and the peri-infarct zone is electrically unstable.
• Functional mitral regurgitation: LV dilation secondary to large infarcts distorts the mitral valve apparatus, causing functional (secondary) mitral regurgitation that worsens heart failure symptoms and may require urgent readmission.
• ICD eligibility evaluation: Patients with LVEF below 35% after AMI are potential candidates for implantable cardioverter-defibrillator (ICD) placement after a 40-day waiting period (per current guidelines). This evaluation pathway itself generates scheduled and unscheduled care encounters.
• Neurohormonal optimization complexity: Titrating evidence-based therapies (ACE inhibitors/ARBs/ARNIs, beta-blockers, mineralocorticoid receptor antagonists) in the post-MI period requires close monitoring for hypotension, bradycardia, hyperkalemia, and worsening renal function, and patients may require readmission when dose titration is not manageable in the outpatient setting.

A: Age ≥75 Years (3 points)

Advanced age is a universally recognized predictor of adverse outcomes across virtually all acute illness episodes. For post-AMI readmission specifically, age at or above 75 years captures a constellation of factors associated with high short-term readmission risk:

• Greater comorbidity burden: Older AMI patients carry higher prevalences of COPD, renal insufficiency, atrial fibrillation, peripheral vascular disease, prior stroke, and cognitive impairment, all of which independently increase readmission risk.
• Frailty and functional decline: Physiological reserve is reduced in older adults; even a brief ICU stay or period of bed rest during AMI hospitalization can cause significant functional decline, deconditioning, and sarcopenia that impairs post-discharge recovery and increases dependency.
• Polypharmacy and adherence barriers: Older patients are prescribed a greater number of medications and are more susceptible to drug-drug interactions, adverse drug reactions, and cognitive or functional barriers to medication adherence.
• Social support deficits: Older adults are more likely to live alone, have limited social networks, and lack the functional capacity for self-care, increasing the probability of symptom under-recognition and delayed care-seeking.
• Atypical presentations and diagnostic delays: Older adults with recurrent ACS or decompensated heart failure may present with atypical symptoms (dyspnea, fatigue, confusion) rather than classic chest pain, delaying recognition and treatment.

The binary threshold of 75 years in READMITS reflects the point at which age-related risk acceleration becomes most clinically meaningful in the EFFECT cohort. Continuous age models generally show an exponential increase in readmission hazard beyond 70 to 75 years.

D: Diabetes Mellitus (2 points)

Diabetes mellitus is present in 25 to 35% of hospitalized AMI patients and is one of the most important modifiable determinants of post-AMI outcomes. Its contribution to readmission risk operates through multiple mechanisms:

• Accelerated coronary artery disease and higher rates of multivessel disease: Diabetic patients have diffuse, distal coronary atherosclerosis, higher residual ischemic burden after revascularization, and higher rates of in-stent restenosis, all increasing the likelihood of recurrent ACS.
• Microvascular disease and impaired healing: Diabetic microangiopathy impairs coronary microvascular reperfusion after PCI, resulting in larger effective infarct sizes and worse LVEF recovery compared with matched non-diabetic patients.
• Glycemic instability during hospitalization and post-discharge: AMI-associated catecholamine surge and corticosteroid release (in patients receiving steroids for contrast allergy prophylaxis) cause hyperglycemia that exacerbates myocardial injury. Post-discharge glycemic management requires careful drug selection (some antidiabetic agents require dose adjustment in the setting of reduced LVEF or renal impairment) and increases care complexity.
• Renal disease synergy: Diabetic nephropathy is the most common cause of chronic kidney disease globally, and the combination of diabetes and CKD in post-AMI patients creates compounding readmission risk exceeding the sum of either factor alone.
• Heart failure risk: Diabetes independently increases heart failure risk through diabetic cardiomyopathy (myocardial fibrosis and diastolic dysfunction independent of coronary disease) as well as through worsening of post-infarction LV remodeling.

M: MI Type — NSTEMI (2 points)

Counterintuitively, NSTEMI (rather than the typically more dramatic STEMI) carries higher 30-day readmission risk in the READMITS model. This seemingly paradoxical finding reflects several important clinical realities:

• Greater comorbidity burden in NSTEMI patients: NSTEMI patients are on average older, have higher rates of diabetes, renal insufficiency, prior heart failure, and multivessel coronary disease than STEMI patients. The index NSTEMI event may represent decompensation of a complex, long-standing cardiovascular risk profile rather than an isolated acute coronary event.
• Less complete revascularization: STEMI management protocols prioritize rapid culprit vessel reperfusion (primary PCI), resulting in high rates of single-vessel revascularization. NSTEMI management involves risk-stratified invasive evaluation, and a substantial proportion of NSTEMI patients are managed conservatively or undergo incomplete revascularization of multivessel disease, leaving residual ischemic substrate for recurrent events.
• Shorter length of stay in STEMI: Modern STEMI pathways prioritize rapid discharge after successful primary PCI in hemodynamically stable patients, resulting in shorter inpatient stays. Brief hospitalizations may actually reduce the opportunity for in-hospital complication development (which would lead to in-hospital death rather than post-discharge readmission), while NSTEMI patients with longer stays and more comorbidities may be more vulnerable post-discharge.
• Residual ischemic burden: NSTEMI frequently represents a non-occlusive coronary event with subendocardial ischemia and preserved or only mildly impaired LVEF. While acute mortality may be lower than STEMI, the non-culprit disease and ongoing plaque instability drive early recurrent events.

I: Income — Lowest Quintile (2 points)

The inclusion of socioeconomic status (SES) as an independent predictor of post-AMI readmission in the READMITS score is both clinically important and policy-relevant. Patients residing in neighborhoods in the lowest income quintile (as determined by census-derived area-level income data in the EFFECT Ontario cohort) face structural barriers to effective post-discharge recovery:

• Medication affordability: Post-AMI guideline-directed therapy includes aspirin, P2Y12 inhibitors, statins, beta-blockers, and RAAS inhibitors, representing a substantial out-of-pocket cost burden for patients without comprehensive drug coverage. Rates of medication non-adherence after AMI are significantly higher in lower-income patients, directly increasing recurrent event and readmission risk.
• Transportation and geography barriers: Access to follow-up cardiac care requires reliable transportation to outpatient clinics and cardiac rehabilitation programs. Lower-income patients are less likely to own vehicles, more likely to live in areas with poor public transit, and more likely to be unable to take time off work for medical appointments.
• Food insecurity and dietary compliance: Heart-healthy dietary patterns (low sodium for heart failure management, Mediterranean-style diets for secondary prevention) require access to affordable fresh produce and the time and knowledge to prepare appropriate meals, which are disproportionately unavailable to patients in poverty.
• Housing instability: Unstable housing is associated with missed follow-up, inability to manage medications, and higher rates of non-cardiac illnesses (infections, trauma, substance use) that drive non-cardiac readmissions.
• Health literacy and self-monitoring capacity: Low-income patients have on average lower health literacy, making it harder to recognize early warning signs of decompensating heart failure (weight gain, ankle edema, dyspnea on exertion) that should prompt outpatient contact before hospital-level care is required.

The incorporation of SES into READMITS reflects the recognition that readmission is not purely a biological outcome — it is a social determinant-laden event where structural inequity substantially amplifies physiological risk.

T: Transfer from Another Hospital (2 points)

Patients transferred from a non-PCI-capable community hospital to a PCI-capable tertiary or quaternary center for their index AMI are at elevated readmission risk compared with patients presenting directly to the receiving facility. Several mechanisms contribute:

• Greater infarct severity and higher-risk presentations: Transfers are typically driven by clinical necessity — patients with high-risk STEMI, cardiogenic shock, complex anatomy, or failed thrombolysis who require advanced interventional capabilities unavailable at the presenting facility. These patients tend to have larger infarcts, lower post-PCI LVEF, and more hemodynamic compromise than self-presenting direct admission patients.
• Care fragmentation and communication failures: A patient transferred from one institution to another may have incomplete medical record transfer, missed medications, unresolved pending test results, and follow-up appointments arranged at the transferring hospital rather than the receiving facility. These communication gaps create gaps in post-discharge care continuity.
• Geographic dislocation: Transferred patients may live far from the receiving tertiary center, making follow-up at that institution impractical. Yet the transferring community hospital may not have the necessary cardiology expertise for optimal post-AMI follow-up. This creates a care vacuum in the post-discharge period.
• Family and social support disruption: Transfer separates patients from their local support networks during hospitalization and may complicate discharge planning when family members are geographically distant from the receiving facility.

S: Female Sex (1 point)

Female sex contributes the smallest independent weight to the READMITS score (+1 point), reflecting a modest but statistically significant elevation in 30-day readmission risk among women after AMI. While female sex carries less total point weight than other variables, its inclusion is clinically meaningful given the well-documented sex disparities in post-AMI care and outcomes:

• Later presentation and higher baseline risk: Women presenting with AMI are on average older than men (reflecting the protective effect of estrogen on coronary artery disease until menopause), have higher rates of hypertension, diabetes, and heart failure, and are more likely to present with atypical symptoms, all contributing to delayed diagnosis and treatment.
• Higher prevalence of MINOCA: Myocardial infarction with non-obstructive coronary arteries (MINOCA) is 2 to 4 times more common in women than men. MINOCA has distinct etiologies (coronary vasospasm, spontaneous coronary artery dissection, plaque erosion, microvascular dysfunction) and may be managed differently, potentially with less complete secondary prevention, contributing to recurrent events.
• Differential response to RAAS blockade: Women may experience higher rates of ACE inhibitor-induced cough and lower rates of RAAS inhibitor prescription at discharge, reducing the completeness of evidence-based post-MI neurohormonal blockade.
• Psychosocial factors: Depression and anxiety, both independent predictors of post-AMI readmission and mortality, are significantly more prevalent in women after AMI. The impact of post-AMI depression on medication adherence, return to physical activity, and engagement with cardiac rehabilitation is well established.
• Caregiver role: Women are disproportionately primary caregivers for dependent family members, which may interfere with cardiac rehabilitation participation, follow-up appointment attendance, and adequate post-discharge rest and recovery.

Clinical Application of the READMITS Score

Timing and Workflow

The READMITS score is optimally calculated on the day of planned discharge, once echocardiographic data (or catheterization-based LVEF assessment) is available. All eight variables should be determinable from the index hospitalization record: admission serum creatinine, formal LVEF assessment (echocardiogram or left ventriculogram), patient age, documented diabetes diagnosis, MI classification (STEMI vs. NSTEMI) on the discharge summary, neighborhood income level (available in electronic health records linked to census data, or approximated by zip/postal code in systems with income mapping), transfer status (documented in admission note), and biological sex.

In health systems where neighborhood income quintile is not readily available in the EHR, clinical teams may use proxy indicators such as insurance type, documented housing instability, social work referral flags, or pharmacy benefit status to approximate income-related risk.

Targeted Interventions by Risk Category

The clinical value of the READMITS score lies in its ability to guide the allocation of resource-intensive post-discharge support to patients most likely to benefit. Proposed interventions stratified by READMITS risk category include:

Low Risk (Score 0–3)

• Standard discharge education regarding AMI warning symptoms, medication adherence, diet, and activity restrictions.
• Scheduled follow-up with cardiologist or primary care physician within 7 to 14 days.
• Referral to cardiac rehabilitation (strong class I recommendation for all post-MI patients regardless of risk score).
• Prescription reconciliation with written medication list provided to patient and primary care physician.

Intermediate Risk (Score 4–9)

• All low-risk interventions, plus:
• Pharmacist-led comprehensive medication reconciliation and adherence counseling before discharge.
• Transitional care nurse navigator contact within 48 to 72 hours of discharge (telephone or home visit).
• Expedited follow-up within 5 to 7 days with cardiologist.
• Remote monitoring for patients with LVEF below 40% (weight monitoring with daily weight diary; heart failure action plan for symptom escalation).
• Social work evaluation for patients scoring the income variable, with assessment of medication cost support, transportation assistance, and community resource linkage.

High Risk (Score ≥10)

• All intermediate-risk interventions, plus:
• Dedicated heart failure disease management program enrollment for patients with LVEF below 40%.
• Home health nursing visits for the first 2 to 4 weeks post-discharge.
• Structured telephone follow-up protocol with defined escalation pathways for symptom changes.
• Nephrologist co-management or expedited nephrology outpatient referral for patients scoring the renal failure variable.
• Potential extension of inpatient stay for further stabilization in patients with scores of 14 or above, when clinically appropriate.
• Palliative care consultation for patients with very high scores in the setting of advanced age, severe renal failure, and reduced LVEF, where prognosis may support goals-of-care discussion.

Secondary Prevention After AMI: Context for Readmission Reduction

Regardless of READMITS risk category, all AMI survivors should receive the full spectrum of guideline-directed secondary prevention therapy unless contraindicated. These therapies directly reduce the cardiovascular causes of readmission:

Antiplatelet Therapy

Dual antiplatelet therapy (DAPT) with aspirin plus a P2Y12 inhibitor (ticagrelor or prasugrel preferred over clopidogrel for most ACS patients, per PLATO and TRITON-TIMI 38 trial data) is the cornerstone of post-MI antithrombotic therapy. Duration is typically 6 to 12 months after PCI with drug-eluting stent placement, then aspirin indefinitely. Premature DAPT discontinuation is associated with stent thrombosis and recurrent MI, a potentially preventable readmission cause that is entirely avoidable with patient education and adherence support.

High-Intensity Statin Therapy

Atorvastatin 40 to 80 mg or rosuvastatin 20 to 40 mg daily is recommended for all post-ACS patients regardless of baseline LDL, targeting LDL-C below 1.4 mmol/L (55 mg/dL) with a secondary goal of at least 50% LDL reduction from baseline. Statins reduce recurrent ACS and cardiovascular death, both of which generate readmissions. Statin non-adherence (common in patients with side effects, low health literacy, or financial barriers) is a modifiable readmission risk factor directly addressable at discharge.

Renin-Angiotensin-Aldosterone System Blockade

ACE inhibitors (or ARBs if ACE inhibitor-intolerant) are a class I recommendation for all AMI patients with LVEF below 40%, anterior MI, or heart failure, and a class IIa recommendation for all other AMI patients. In patients with LVEF below 35% and NYHA class II or III heart failure on optimal medical therapy, sacubitril/valsartan (ARNI) is preferred over ACE inhibitor based on the PARADIGM-HF and PIONEER-HF trials. Mineralocorticoid receptor antagonists (eplerenone or spironolactone) are additionally indicated for patients with LVEF below 40% and either diabetes or symptomatic heart failure, based on the EPHESUS trial.

Beta-Blockers

Oral beta-blocker therapy (metoprolol succinate, carvedilol, or bisoprolol) reduces all-cause mortality, reinfarction, and sudden cardiac death after AMI and is a class I recommendation for patients with LVEF below 40% or ongoing ischemia. In the absence of these indications, current evidence for routine long-term beta-blockade in AMI patients with preserved LVEF is more limited (the REDUCE-AMI trial, 2024, found no significant mortality benefit of routine beta-blocker prescription after AMI with preserved EF), and guidelines are evolving in this area.

Cardiac Rehabilitation

Participation in a structured cardiac rehabilitation program is a class IA recommendation for all post-MI patients. Cardiac rehabilitation delivers a comprehensive package of supervised exercise training, risk factor modification, medication optimization, psychological support, and patient education. Meta-analyses consistently demonstrate 20 to 30% reductions in cardiovascular mortality and hospital readmissions in post-MI patients who complete cardiac rehabilitation compared with those who do not. Despite this evidence, uptake remains below 30% in most health systems due to access, logistical, and awareness barriers, particularly in lower-income and older patient groups that carry the highest READMITS scores.

Comparison with Other Post-AMI Readmission Prediction Tools

LACE Index

The LACE index (Length of stay, Acuity of admission, Comorbidity burden by Charlson index, and Emergency department visits in the prior 6 months) is a generic 30-day readmission tool validated across multiple medical conditions, not specific to AMI. While widely implemented due to its simplicity and use of routinely captured administrative variables, LACE lacks the AMI-specific variables (LVEF, HBV DNA equivalent variables in other contexts) that give disease-specific tools like READMITS greater predictive precision for this patient population. Direct comparisons show READMITS outperforms LACE specifically for post-AMI readmission prediction.

HOSPITAL Score

The HOSPITAL score (Hemoglobin, discharge from Oncology, Sodium, Procedure during hospitalization, Index admission type, number of Admissions in the prior year, and Length of stay) is another generic readmission tool validated in large multicenter studies. Like LACE, it is not specific to AMI and does not include LVEF, MI type, or the SES components that characterize READMITS. In general medical readmission contexts, HOSPITAL has demonstrated robust performance (c-statistics ~0.71 to 0.76) but has not been systematically compared with READMITS in AMI-specific cohorts.

GRACE Score (Repurposed)

The GRACE risk score was originally developed to predict in-hospital and 6-month mortality in ACS, not 30-day readmission. However, several studies have explored its utility as a readmission predictor given its widespread clinical adoption. GRACE incorporates Killip class, cardiac arrest at admission, ST-segment changes, elevated biomarkers, age, heart rate, systolic blood pressure, and creatinine. While it captures cardiac severity well, it does not incorporate SES, transfer status, or diabetes in the same format as READMITS, and its readmission-specific predictive performance is generally inferior to READMITS in head-to-head comparisons.

CMS Administrative Readmission Models

CMS uses hierarchical logistic regression models built on administrative claims data (ICD diagnosis and procedure codes, age, sex, and selected comorbidities) to calculate the Expected Readmission Rate (ERR) for hospital-level performance reporting. These models have several well-documented limitations: they cannot capture granular clinical variables (LVEF, creatinine) that READMITS incorporates, they rely on coded diagnoses that vary in accuracy across hospitals, and they produce hospital-level (not patient-level) risk estimates. READMITS operates at the individual patient level, making it more actionable for clinical decision support and discharge planning.

Healthcare Policy and System Implications

Risk-Adjusted Benchmarking

The READMITS score enables hospital quality programs to perform more granular risk adjustment of readmission rates than administrative models allow. By stratifying the post-AMI patient population into READMITS risk categories, quality improvement teams can:

• Calculate expected versus observed readmission rates within each risk stratum, identifying whether excess readmissions are concentrated in high-risk or low-risk patients.
• Evaluate the effectiveness of specific readmission reduction interventions (e.g., transitional care programs) in the population most likely to benefit.
• Advocate to payers and regulators for more sophisticated risk adjustment that accounts for patient SES, comorbidity, and clinical complexity beyond what ICD coding captures.

Social Determinants and Health Equity

The explicit inclusion of neighborhood income quintile in the READMITS score is a methodological acknowledgment that post-discharge outcomes are shaped by social determinants of health that fall outside the purview of traditional biomedical risk models. This has important implications for health equity:

• Hospitals serving predominantly low-income communities will systematically produce higher READMITS scores in their AMI population, reflecting their patient mix rather than quality failures. Any readmission benchmark that does not risk-adjust for SES effectively penalizes safety-net hospitals for the communities they serve.
• The income variable identifies a specific, actionable intervention target: connecting patients with financial assistance programs for medications, transportation vouchers, community health workers, and social service resources that directly address SES-related readmission risk.
• Health systems using READMITS in clinical practice can use the income variable to trigger automatic social work referrals at the time of READMITS score calculation, institutionalizing an equity-oriented response to identified social risk.

Limitations

• Moderate discriminatory performance: The c-statistic of READMITS (approximately 0.64 to 0.66) reflects the inherently complex and partially unpredictable nature of 30-day readmission. No clinical readmission prediction tool has achieved a c-statistic above 0.70 in large external validation cohorts, reflecting the substantial contribution of unmeasured factors (patient preferences, social network strength, primary care access, post-discharge illness course) to readmission risk. READMITS should be viewed as a risk stratification tool rather than an absolute readmission predictor.
• Ontario, Canada derivation cohort: The EFFECT cohort is drawn from a universal healthcare system with comprehensive drug coverage for patients over 65 (ODB program) and universal hospital and physician coverage. The readmission risk associated with low income and medication costs may differ in health systems with greater insurance-related care access barriers, potentially underweighting the income variable in US or other contexts with higher rates of underinsurance.
• Temporal validity: The EFFECT cohort was enrolled between 1999 and 2005. Contemporary AMI management includes more widespread use of drug-eluting stents, potent P2Y12 inhibitors (ticagrelor, prasugrel), PCSK9 inhibitors, SGLT2 inhibitors, and GLP-1 receptor agonists, all of which reduce recurrent cardiovascular events. Readmission rates and their predictors may have shifted since the derivation period, and periodic recalibration of the score against contemporary populations is warranted.
• LVEF assessment availability: Formal echocardiographic or angiographic LVEF assessment is required for the ejection fraction variable. In centers without routine pre-discharge echocardiography, or in patients who undergo only electrocardiographic evaluation, LVEF may be unavailable for scoring. Estimated or absent LVEF data limits score completeness.
• Neighborhood income vs. individual income: The income variable in READMITS uses area-level (census-derived neighborhood) income as a proxy for individual SES, which introduces ecological bias. A patient of high individual income may reside in a low-income neighborhood, or vice versa. Individual-level income, insurance status, or social vulnerability index scores may perform differently as SES proxies in non-Canadian health system contexts.
• Does not capture all relevant comorbidities: Important post-AMI readmission predictors not incorporated into READMITS include COPD/asthma, prior stroke, peripheral vascular disease, atrial fibrillation, cancer, functional status, frailty index, and depression/anxiety. The parsimony of an 8-variable score comes at the cost of capturing these additional risk dimensions.
• Not validated for specific readmission causes: READMITS predicts all-cause readmission and does not distinguish between cardiovascular readmissions (recurrent ACS, heart failure, arrhythmia) and non-cardiovascular readmissions (infections, bleeding, non-cardiac illness). Cause-specific readmission prediction would require distinct models.
• This calculator is for clinical decision support and educational use only. Discharge planning, post-acute care allocation, and readmission reduction intervention decisions require comprehensive clinical assessment by the treating cardiology and multidisciplinary team and should not be based solely on the READMITS score.