CHARM prognostic model (primary endpoint: cardiovascular death or heart-failure hospitalization — 2-year risk)
The CHARM prognostic models were developed using 7,599 patients enrolled in the CHARM programme and identify 21 independent predictors of outcome across the spectrum of left ventricular systolic function. The model forms a linear risk score by summing each predictor multiplied by its coefficient (log hazard). For presentation the authors multiply the linear predictor by 10 to produce a convenient “risk score.” Higher scores are associated with higher 2-year risk; for example, the publication examples show a low-risk patient with score ≈3.4 and ~5.3% 2-year probability, and a high-risk patient with score ≈31 and ~58% 2-year probability. This calculator implements the published predictor hazard ratios (converted to coefficients = ln(HR)) and reproduces the paper’s example probabilities by mapping the score to 2-year probability using a logistic mapping calibrated to the published examples.
Clinical interpretation: use the estimated probability as an adjunct for risk-stratification and shared decision-making; the CHARM model was derived in patients enrolled 1999–2003, so absolute risks may differ in contemporary cohorts receiving more recent guideline-directed therapies. Always combine model output with clinical judgement and individual patient context.
Reference:
Pocock SJ, et al. Predicting death and hospitalisation in patients with chronic heart failure: results from the CHARM programme. European Heart Journal. 2006;27:65–75.