Peering into the future to accurately predict health outcomes is challenging and imprecise. Yet, this has not stopped the development of a broad array of models to predict future disease. The widespread availability of regression modeling approaches has made it seem easy to develop prediction models. A substantial body of literature documents the development, calibration, and validation of risk prediction models using large data sets. Typically, these models use individual patient risk factors to estimate the probability of a future health outcome. Patient-friendly Web sites provide tools that enable patients to calculate their own risk for such conditions as cardiovascular disease or cancer, in theory to motivate them to improve their risk profiles. However, the prediction modeling enterprise faces substantial challenges that seldom attract as much attention as they should.

展望未来以准确预测健康结果具有挑战性和不精确性。然而，这并没有阻止预测未来疾病的各种模型的发展。回归建模方法的广泛应用使开发预测模型变得容易。大量文献记录了使用大数据集的风险预测模型的开发，校准和验证。通常，这些模型使用个体患者风险因素来估计未来健康结果的可能性。对患者友好的网站提供了一些工具，使患者能够计算出自己对心血管疾病或癌症等疾病的风险，从理论上讲可以激励他们改善风险状况。然而，