Mortality prediction models for heart failure work well on whole populations but won't reliably tell an individual patient their risk, researchers found.
Across more than 10,000 ambulatory heart failure patients in three integrated health systems, the C statistic for predicting probability of death within 1 year was 0.66 for the Seattle Heart Failure Model (SHFM) and 0.69 for the Meta-Analysis Global Group in Chronic Heart Failure (MAGGIC) risk calculators.
However, at the individual level, neither predictor was better than a coin toss, Larry Allen, MD, MHS, of the University of Colorado in Aurora, and colleagues reported online in JAMA Cardiology.
Among the individual level findings, they found:
"The SHFM predicted a more than 50% probability of dying in the next year for eight of the 1661 patients who died (sensitivity for 1-year death was 0.5%) and for five patients who lived at least a year (positive predictive value, 61.5%).
"The MAGGIC risk calculator predicted a more than 50% probability of dying in the next year for 52 of the 1661 patients who died (sensitivity, 3.1%) and for 63 patients who lived at least a year (positive predictive value, 45.2%)."
And while the SHFM estimated 77.8% of patients had a less than 15% chance of dying at 1 year, nearly two-thirds of deaths were in that group. The MAGGIC risk calculator also estimated less than 25% chance of 1-year mortality for most patients who died by that point.
"Risk models cannot prospectively identify the vast majority of individual ambulatory patients who will die in the near future," they concluded. "For undulating diseases such as heart failure, improvements in risk modeling are only a partial solution. Meanwhile, we must embrace the certainty of uncertainty and better help patients plan for the worst while hoping for the best."