Using electronic medication monitoring to guide differential management of tuberculosis patients at the community level in China

According to the World Health Organization’s (WHO’s) 2018 Global Tuberculosis Report, China has the world’s second highest tuberculosis (TB) burden, with an estimated 889,000 cases (new and existing) in 2017, accounting for 8.9% of the global total [1]. Although the use of directly observed therapy (DOT) by health care workers is the most widely known approach to ensure treatment adherence, recent evidence indicates that the program is not being implemented systematically, resulting in lower patient adherence rates [2‐4]. Poor adherence is cited as the primary reason for suboptimal clinical benefits of TB treatment, and it leads to worse clinical outcomes, including development of drug resistance, increased duration of infectivity, and transmission to others [5]. A systematic review in China found that only 50% of TB patients received DOT (20% from a health care worker and 30% from a family member), and 50% self-administered treatment (SAT) without any adherence support [6]. Meanwhile, evidence showed that SAT patients had poor adherence compared to those on DOT [7]. According to the 2010 national TB prevalence survey in China, 20% of TB patients treated in the public health system were lost to follow-up or were not taking their medication regularly [8].

The use of digital health technologies may help to reach the new global End TB targets [9]. According to the updated WHO treatment guidelines for drug-susceptible TB, three digital technologies are recommended: short message service (SMS, or text messaging), the electronic medication monitor (EMM), and video-observed treatment (VOT) [10]. The three digital technologies have individual strengths and weaknesses that determine their suitability for distinct circumstances. The EMM, which can operate without mobile broadband Internet coverage, is currently the most accessible, affordable, and scalable solution in resource-limited settings [11]. Hence, between 2009 and 2015, the Chinese Center for Disease Control and Prevention (China CDC) explored the use of an indigenously developed EMM for improving treatment adherence. Under trial conditions, EMM use decreased the frequency of missed doses by 40 to 50% compared to SAT [7]. A usability study conducted in rural China with the EMM also showed a high level of user performance, acceptability, and satisfaction among both TB patients and health care workers [12].To strengthen community-based case management, China’s National 13th Five Year Plan on TB Control (2016–2020) included adoption of the EMM as part of TB case management approaches [13].

Feasibility of EMM implementation, user acceptability, and qualitative assessments at scale are important considerations, but evidence from implementation studies remains sparse [14, 15]. To extend the use of the EMM, the China CDC developed an EMM scale-up plan for three provinces involving 138 counties. Since this is the first large-scale use of the EMM in China, on behalf of the China CDC, we conducted this operational research to evaluate the acceptance of the EMM among health care workers and patients while implementing the device in actual routine practice by the local TB program.

This study found high patient acceptability with the EMM; among the eligible patients, only 11.3% of patients (26/231) refused to use the device initially, and another 8.2% (19/231) subsequently refused to use the device. The total acceptance rate was 80.5% (186/231). Positive bacteriological patients were more likely to refuse to use the device in this study, and further research is required to identify the reasons for this correlation. Moreover, similar to other studies [4, 17, 18], our study noted that retreated TB patients were more likely to have poor adherence compared to new TB patients, migrant TB patients were more likely to have poor adherence compared to patients from the local county, and farmers or migrant workers were more likely to have stronger treatment adherence results compared to those in other occupations, but since there were only 15 patients who switched to DOT due to poor adherence, the results need to be confirmed by further studies. Another important finding was that most patients who stopped using the EMM did so in the last 2 months of treatment; a systematic review likewise found that the effectiveness of interventions to improve adherence decreased by 1.1% each month [19], suggesting that patients need more interventions to enhance medication adherence in the middle and late stages of treatment.

Additionally, results from data collection with the structured questionnaire suggested good acceptance of the EMM among health care workers. Because the EMM did not provide real-time data transmission, which limited physicians’ timely access to patient dosing data, the physicians were unable to offer timely intervention to correct dosing behaviors. However, the LED screen has been associated with improved adherence [20], and EMMIMS, the Internet-based system for uploading data, allowed users at different levels to query and review patient medication adherence data.

The EMM in this study not only provided patient medication reminders but also generated dosing data that helped health care providers identify patients who were non-adherent to therapy. We successfully implemented the differential patient management program, as only 8.9% of patients (15/169) were found non-adherent by EMM and shifted to DOT. Among the other 91.1% of patients (154/169), who used the EMM and completed treatment, the median adherence rate was 99.3%. Moreover, EMM use significantly reduced the workload of community physicians compared to DOT. The resources saved by not giving DOT to the adherent patients could help to strengthen management of non-adherent patients—for example, by implementing DOT, tracking the interrupted patients, increasing counselling, and providing more home visits. Several studies [7, 14, 19, 21, 22] have demonstrated the effectiveness of an EMM in improving medication adherence among TB patients. Furthermore, two studies [23, 24] have suggested that implementing differential patient management programs through the use of an EMM, using a package of interventions tailored to each patient’s needs and values, is more likely to improve TB outcomes. While the most appropriate thresholds for switching to DOT need to be discussed further, according to a recent study, less than 90% medication adherence among TB patients is a significant risk factor for unfavorable outcomes [25]. If we had followed this criterion, 10.1% of patients (17/169) in this study would have been switched to DOT, an increase of 1.3 percentage points.

Development and implementation of digital health products at a large scale requires continued engagement of both TB patients and health care workers, in addition to innovators, funders, policymakers, advocacy groups, and affected communities [9]. Although use of the EMM greatly reduced the workload of community physicians, county clinics expended additional resources on training, EMM configuration, data downloading and uploading, and data-driven counseling. Even though most health care workers reported that use of the EMM only moderately increased their workload, it is important to note that the patient management workload was partially transferred from the community to the county, and it is important to think about optimizing resource allocations accordingly. Another study from China found that physicians from county designated hospitals largely ignored the adherence data when deciding whether to switch patients to a more intensive case management approach because of insufficient financial incentives to do so [12]. Moreover, it was noted that health care workers spent nearly twice as much time with patients during their first visit than during follow-up visits, suggesting that health care workers could benefit from training materials or videos on the EMM.

There were several limitations of this study. First, the intervention was implemented in only one prefecture, located in eastern China, where the level of economic development is higher than the national average and the health care community generally has sufficient human resources. Therefore, the results obtained cannot predict the success of national scale-up but support efforts to further study and expand use of the EMM. Second, the study lacked cost-effectiveness and intervention impact analyses, which would yield essential information for considering large-scale use of the EMM [26]. A cluster-randomized, controlled trial currently being undertaken in China by the China CDC will assess whether an EMM-based treatment strategy can improve clinical outcomes for TB patients [27]. Third, the study did not collect data from patients about their perceptions of EMM use. Based on this study, the authors recommend more operational research, including studies on coverage of EMM use, adherence using the EMM, and TB treatment outcomes under programmatic conditions in more areas. Additional research should also assess patients’ perceptions of EMM use and include cost minimization analysis as well as cost-effectiveness analysis, to provide public health planners in low-resource settings with adequate data to evaluate the return on investment of treatment adherence technology for broader implementation [28].

Despite these limitations, the authors believe this is the first operational research to assess the acceptability of an EMM that does not provide real-time data transmission in actual routine practice by the local TB program. Based on the findings from this research, the same EMM device has been used in three selected provinces since Jun 2018, and since January 2019, all 138 counties from the three provinces have started using the EMM.

In settings such as China, where universal use of DOT is not feasible, innovative approaches are needed to support patient adherence to TB treatment. Our study successfully used an EMM to guide differential patient management at the community level in China, and the results support further expansion of this locally developed, low-cost, quality-assured device under the national TB control program. Evidence from implementation studies is sparse, however. To advance this intervention, more operational research needs to be conducted on the coverage of EMM, adherence using the EMM under programmatic conditions in more areas, cost-effectiveness, and the impact of the intervention. The results are needed to consider introduction and scale-up of the technology countrywide.