Evaluation Metric | No SDH | SDH |
---|

R^{2} (95% CI)^{a} |

Linear | 0.327 (0.300, 0.353) | 0.327 (0.300, 0.354) |

ML | 0.388 (0.357, 0.420) | 0.387 (0.357, 0.419) |

MAE (95% CI)^{b} |

Linear | 6992 (6889, 7094) | 6991 (6889, 7094) |

ML | 6637 (6539, 6735) | 6634 (6536, 6732) |

C-statistic (95% CI)^{c} |

Linear | 0.703 (0.701, 0.705) | 0.700 (0.699, 0.702) |

ML | 0.717 (0.715, 0.718) | 0.716 (0.714, 0.717) |

- Comparison of performance measures between linear regression and machine learning prospective risk adjustment models, predicting 2017 yearly top-coded spending from 2016 characteristics. The SDH model additionally includes SDH variables obtained from U.S. Census data (see Table 1)
^{a}Confidence intervals for R^{2} were constructed using the nonparametric bootstrap [21]^{b}Confidence intervals for MAE were constructed using a paired t-test^{c}Confidence intervals for C-statistic were constructed using a jackknife procedure [25]