Franklin Jessica M, Shrank William H, Lii Joyce, Krumme Alexis K, Matlin Olga S, Brennan Troyen A, Choudhry Niteesh K
Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women's Hospital and Harvard Medical School, Boston, MA.
CVS Caremark, Woonsocket, RI.
Health Serv Res. 2016 Feb;51(1):220-39. doi: 10.1111/1475-6773.12310. Epub 2015 Apr 16.
Despite the proliferation of databases with increasingly rich patient data, prediction of medication adherence remains poor. We proposed and evaluated approaches for improved adherence prediction.
We identified Medicare beneficiaries who received prescription drug coverage through CVS Caremark and initiated a statin.
A total of 643 variables were identified at baseline from prior claims and linked Census data. In addition, we identified three postbaseline predictors, indicators of adherence to statins during each of the first 3 months of follow-up. We estimated 10 models predicting subsequent adherence, using logistic regression and boosted logistic regression, a nonparametric data-mining technique. Models were also estimated within strata defined by the index days supply.
In 77,703 statin initiators, prediction using baseline variables only was poor with maximum cross-validated C-statistics of 0.606 and 0.577 among patients with index supply ≤30 days and >30 days, respectively. Using only indicators of initial statin adherence improved prediction accuracy substantially among patients with shorter initial dispensings (C = 0.827/0.518), and, when combined with investigator-specified variables, prediction accuracy was further improved (C = 0.842/0.596).
Observed adherence immediately after initiation predicted future adherence for patients whose initial dispensings were relatively short.
尽管拥有越来越丰富患者数据的数据库不断增多，但药物依从性的预测仍然很差。我们提出并评估了改善依从性预测的方法。
我们确定了通过CVS Caremark获得处方药保险并开始服用他汀类药物的医疗保险受益人。
从先前的理赔记录和相关人口普查数据中，在基线时共识别出643个变量。此外，我们确定了三个基线后预测指标，即随访前3个月中每个月他汀类药物依从性的指标。我们使用逻辑回归和增强逻辑回归（一种非参数数据挖掘技术）估计了10个预测后续依从性的模型。模型也在由索引日供应量定义的分层内进行估计。
在77,703名开始服用他汀类药物的患者中，仅使用基线变量进行预测的效果很差，索引供应量≤30天和>30天的患者中，最大交叉验证C统计量分别为0.606和0.577。仅使用初始他汀类药物依从性指标可大幅提高初始配药时间较短患者的预测准确性（C = 0.827/0.518），并且与研究者指定的变量相结合时，预测准确性进一步提高（C = 0.842/0.596）。
开始用药后立即观察到的依从性可预测初始配药时间相对较短患者未来的依从性。
