National Center for Tuberculosis Control and Prevention, Chinese Center for Disease Control and Prevention, Beijing, China.
Duke Global Health Institute, Duke University, Durham, NC, USA.
Infect Dis Poverty. 2021 Mar 17;10(1):29. doi: 10.1186/s40249-021-00818-3.
In China, an indigenously developed electronic medication monitor (EMM) was designed and used in 138 counties from three provinces. Previous studies showed positive results on accuracy, effectiveness, acceptability, and feasibility, but also found some ineffective implementations. In this paper, we assessed the effect of implementation of EMMs on treatment outcomes.
The longitudinal ecological method was used at the county level with aggregate secondary programmatic data. All the notified TB cases in 138 counties were involved in this study from April 2017 to June 2019, and rifampicin-resistant cases were excluded. We fitted a multilevel model to assess the relative change in the quarterly treatment success rate with increasing quarterly EMM coverage rate, in which a mixed effects maximum likelihood regression using random intercept model was applied, by adjusting for seasonal trends, population size, sociodemographic and clinical characteristics, and clustering within counties.
Among all 69 678 notified TB cases, the treatment success rate was slightly increased from 93.5% [95% confidence interval (CI): 93.0-94.0] in second quarter of 2018 to 94.9% (95% CI: 94.4-95.4) in second quarter of 2019 after implementing EMMs. There was a statistically significant effect between quarterly EMM coverage and treatment success rate after adjusting for potential confounders (P = 0.0036), increasing 10% of EMM coverage rate will lead to 0.2% treatment success rate augment. Besides, an increase of 10% of elderly or bacteriologically confirmed TB will lead to a decrease of 0.4% and 0.9% of the treatment success rate.
Under programmatic settings, we found a statistically significant effect between increasing coverage of EMM and treatment success rate at the county level. More prospective studies are needed to confirm the effect of using EMM on TB treatment outcomes. We suggest performing operational research on EMMs that provides real-time data under programmatic conditions in the future.
在中国，一种自主研发的电子药物监测器（EMM）已在三省的 138 个县设计并使用。先前的研究表明其在准确性、有效性、可接受性和可行性方面均取得了积极成果，但也发现了一些执行不力的情况。本文评估了 EMM 的实施对治疗结果的影响。
采用县级纵向生态学方法，使用汇总的二级规划数据。本研究纳入了 2017 年 4 月至 2019 年 6 月期间所有在 138 个县报告的结核病例，且排除了利福平耐药病例。我们采用多水平模型评估随着每季度 EMM 覆盖率的增加，季度治疗成功率的相对变化，其中应用混合效应最大似然回归模型，通过调整季节性趋势、人口规模、社会人口学和临床特征以及县内聚类，对结果进行校正。
在所有 69678 例报告的结核病例中，实施 EMM 后，治疗成功率从 2018 年第二季度的 93.5%（95%置信区间：93.0-94.0）略有上升至 2019 年第二季度的 94.9%（95%置信区间：94.4-95.4）。在调整潜在混杂因素后，每季度 EMM 覆盖率与治疗成功率之间存在统计学显著关联（P=0.0036），EMM 覆盖率增加 10%，治疗成功率将提高 0.2%。此外，每增加 10%的老年或细菌学确诊结核病例，治疗成功率将分别下降 0.4%和 0.9%。
在规划环境下，我们发现县级 EMM 覆盖率的增加与治疗成功率之间存在统计学显著关联。需要更多的前瞻性研究来证实 EMM 使用对结核治疗结果的影响。我们建议未来在规划条件下对 EMM 进行操作研究，以提供实时数据。
