Enhancing tuberculosis care in lower and middle-income countries through digital adherence technologies

Abstract

Tuberculosis (TB) remains a major public health challenge, particularly in low- and middle-income countries (LMICs), with treatment adherence being a critical issue. In recent years, digital adherence technologies (DATs) have emerged as a promising approach to enhance TB care and treatment adherence in these settings. Despite their potential to enhance treatment outcomes, their implementation faces challenges, including infrastructure demands, costs, and variability in effectiveness depending on the context. A critical aspect of DATs is their cost-effectiveness, which has shown mixed results across different LMICs. Some technologies, such as the medication sleeves/labels (e.g., 99DOTS), demonstrate low-cost alternatives to traditional directly observed therapy (DOT), while others report inconsistent outcomes. Throughout its implementation, DATs can be a source of intervention-generated inequality, where they may disproportionately benefit low-risk populations with better access to technology, potentially widening gaps in healthcare equity. However, when targeted effectively at high-risk groups, DATs can promote equitable health outcomes and enhance TB care delivery. The successful implementation of DATs requires context-specific strategies that address the unique challenges of LMICs, such as technology fatigue, data privacy, and the need for tailored interventions that consider socioeconomic and cultural factors. Leveraging the support of global health initiatives and collaborative funding mechanisms could help scale up DAT adoption and contribute to the global goal of TB elimination.

Shareable abstract

By utilising digital adherence technologies and data-driven insights we can make progress towards achieving the goals of TB elimination and ensuring care for people with TB in resource-limited settings https://bit.ly/47KGtLW

Introduction

Tuberculosis (TB) remains one of the leading infectious diseases worldwide, responsible for approximately 1.25 million deaths in 2023, [1] remaining a significant public health challenge in lower and middle-income countries (LMICs) where the burden of the disease is the highest [2] and healthcare services and resources may be scarce [3]. Though effective treatments are available worldwide, completing TB treatment remains a challenge [4], which not only worsens the TB epidemic by accelerating drug resistance and transmission but also highlights the urgent need for interventions that encourage treatment adherence [5]. Unfortunately, non-adherence to TB treatment is a pervasive problem in many resource-limited settings, often due to factors such as the lengthy treatment duration, lack of access to healthcare facilities, and socioeconomic barriers [6].

One promising approach to address this challenge is the implementation of digital adherence technologies (DATs) in TB care and support programmes. DATs have emerged as valuable tools to support people with TB in their adherence to treatment, fostering an improvement in person-centred care [7]. These technologies, which include video-supported treatment (VST) and smart pillboxes (see table 1), have the potential to enhance TB treatment outcomes by remotely monitoring patient adherence, providing real-time feedback, and enabling timely interventions to address non-adherence [7]. This in-depth perspective will explore the role of DATs in TB care and support within lower-middle-income countries, highlighting their advantages, obstacles, and future prospects.

TABLE 1.

Overview of digital adherence technologies (DATs) currently being pilot-tested or implemented in clinical settings for tuberculosis care

DAT	Description	Status in LMICs and effectiveness
Medication sleeve/label	Medication sleeves are wrapped around fixed dose combination blister packs. The person on treatment reports their medication intake daily, through the code found in the packaging, by calling a toll-free phone number, or sending a toll-free SMS message.	In use (India) [9]. The study showed no significant improvement in treatment completion.
VST	Asynchronous: through a secured mobile application, the person on TB treatment video records their medication intake. The video is then uploaded to a secure server accessible to the healthcare provider for later review. If the video is not sent or if the medication intake is unclear, the healthcare provider can contact to provide support and ensure medication adherence.	No significant effect in three randomised controlled trials; dependent on patient access to phones and data [10].
	Synchronous: the healthcare provider and the person on TB treatment schedule a daily video meeting through a secure interface, through which the healthcare provider can see the person proceed with their medication intake at home in real-time.	Synchronous VSTs are used in countries like Uganda and Vietnam [11, 12].
Smart pillbox	A medication container, equipped with a battery-powered sensor and a mobile data connection, automatically logs medication intake by sending a signal each time the box is opened.	In Ethiopia, the Adherence Support Coalition to End TB (ASCENT) project implemented a smart pillbox intervention across 78 health facilities [13]. The smart pillbox DAT was effective in identifying patterns of non-adherence among TB patients.
Ingestible sensors [8]	The system consists of an ingestible sensor, activated by gastric fluids, and an on-body wearable sensor or wearable patch. The ingestible sensors communicate unique identifying signatures to the body surface. The sensor counts the number of times each unique signature is received, thus registering medication intake. The ingestible sensors are designed to communicate for approximately 7 min, after which they are inactive and get eliminated in the faeces.	A study has been conducted at two sites in the USA. The study involved 30 participants who completed 10 directly observed therapy visits, resulting in 1080 co-ingestion events. The system demonstrated a 95.0% positive detection accuracy and 100% identification accuracy, with a specificity of 99.7%. Ingestible sensor-based systems are not yet widely implemented in LMICs, but ongoing research and stakeholder engagement are paving the way for their potential adoption in enhancing TB treatment adherence.

LMICs: low- and middle-income countries; VST: video-supported treatment; TB: tuberculosis.

Implementation of DATs in resource-limited settings

DATs operate through various mechanisms designed to support TB treatment in resource-limited settings. Wirelessly observed therapy (WOT) uses ingestible sensors paired with wearable devices to automatically log medication intake. VST enables real-time (synchronous) or recorded (asynchronous) observation of patients taking their medication via mobile devices. SMS reminders are sent daily to prompt adherence and often require a confirmation reply. Smart pillboxes record the date and time the container is opened and transmit this data to healthcare providers for remote monitoring.

The use of DAT has yielded inconsistent findings, with some studies showing improved treatment outcomes and others demonstrating no significant benefits [14]. A recent systematic review analysed 16 randomised controlled trials (RCTs), primarily conducted in LMICs, found that participants using WOT (i.e. ingestible sensors) were 7.69 times more likely to complete treatment when compared to those receiving directly observed therapy (DOT) (95% CI 4.51–14.48) [10]. However, three RCTs on VST and three on SMS reminders did not find a significant impact on treatment completion [10].

Medication sleeves (99DOTS) have also emerged as a potential, cost-effective, and low-infrastructure alternative to conventional DOT [15]. Its core component is an innovative sleeve folded around the medication blister pack, which reveals a unique toll-free phone number as patients remove their pills. Patients confirm their dosage by calling this number, triggering an automated system to log their adherence in real time via an online dashboard. The effectiveness of medication sleeves like 99DOTS has varied across trials, largely due to site-specific factors such as mobile network reliability, patient literacy, cultural acceptability, health worker engagement and system-level follow-up protocols [9]. In some settings, poor connectivity limited real-time adherence logging, while in others, low literacy hindered patients’ ability to use the system correctly. Health worker responsiveness also influenced outcomes, as a lack of timely follow-up diminished the tool's effectiveness. These context-dependent challenges suggest that results from 99DOTS trials may not be directly generalisable, and successful implementation likely depends on careful adaptation to local infrastructure and health system capacities [9].

The accessibility and adaptability of DATs seem to be heavily dependent on the specific technology used and the setting in which they are employed [16]. RCTs in Pakistan [17] and Uganda [18] demonstrated that most participants did not achieve favourable treatment outcomes with the use of DATs. In the trials conducted in Pakistan and Uganda, participants were generally not provided with mobile phones, and inclusion was often limited to individuals who already owned a functioning phone with network access. Similarly, in Peru, people with TB with limited access to cell phones were more likely to experience unsatisfactory treatment outcomes [19]. As a result, a substantial number of eligible patients were precluded from participation, though exact exclusion figures were not consistently reported. This selective inclusion based on device ownership may introduce bias and limit the generalisability of these findings to the broader TB population in resource-limited settings, particularly women and rural residents who face systemic barriers to digital access. In contrast, a smart pillbox proved effective in identifying a subset of patients struggling with adherence to their HIV and drug-resistant TB treatments in a South African trial [20].

Cost-effectiveness of DATs in TB care

One of the key advantages of DATs in TB care is their potential cost-effectiveness. In resource-constrained settings, DATs may present themselves as a solution to reduce the burden of healthcare costs associated with unfavourable outcomes, such as drug resistance and hospitalisation [16].

A recent systematic review found that the median provider cost for DOT ranged between USD 200 and USD 700 per patient in LMICs, primarily due to intensive labour and travel requirements [21]. Meanwhile, SMS-based DATs cost around USD 115 per patient, and medication sleeves (e.g. 99DOTS) are similarly low-cost, requiring minimal infrastructure [21]. VST, while more expensive (up to USD 1364 per patient), can still be less costly than DOT in urban or remote settings where staffing and logistics pose challenges.

Moreover, traditional DOT has been associated with catastrophic expenditures for patients in many LMICs. DATs reduce this burden by enabling remote monitoring, thus eliminating the need for daily facility visits [22]. Another study further emphasises the economic burden of traditional DOT, including catastrophic costs for patients and the need for intensive resources, suggesting that DATs could offer a more cost-effective approach [23]. Notably, a meta-analysis of implementation feedback found that DATs reduce indirect patient costs such as travel, lost wages, and clinic visits, thereby mitigating catastrophic health expenditures [23]. Moreover, improved adherence linked to DATs may prevent treatment failure and reduce multidrug-resistant TB incidence, leading to substantial long-term health system savings. While initial investment varies, these figures suggest that DATs can generate both direct provider savings and significant downstream economic benefits compared to conventional DOT. In summary, while not universally cheaper in all contexts, DATs can provide a more cost-effective and patient-centred alternative to DOT in many LMIC settings.

A closer look at intervention-generated inequality

Research on whether DATs are more effective for high-risk populations remains limited, raising questions about whether their primary benefits for low-risk groups might reduce their impact or inadvertently contribute to intervention-generated inequality. Technology-driven health interventions could foster intervention-generated inequality by favouring low-risk populations who have access to technologies or are more skilled at using them [24]. However, they can also support people who are at high risk, encouraging more equitable outcomes and improving the effectiveness of healthcare systems, particularly if targeted precisely at these populations [24].

Despite these possibilities, a recent study on the effects of DAT support in adherence to TB treatment confirmed that, although heterogeneous, the highest intervention effects were associated with people from higher-risk groups; i.e. the intervention appears to have disproportionately benefited people at an increased risk [25]. Person-centred care, with the involvement of trained support sponsors, and differentiated enrolment policies appeared as potential explanations for this phenomenon [25].

Gender disparities further compound digital inequality in LMICs, as women often face limited access to mobile phones, the internet and digital literacy opportunities. This digital divide may restrict their ability to engage with DATs, reinforcing existing barriers to adherence and care. Therefore, the design and deployment of DATs must explicitly consider gender-sensitive approaches such as shared access models, low-bandwidth tools, or community-based support, to avoid widening the treatment gap for women with TB.

Technology fatigue in DAT implementation

With the increasing usage of smartphones and the internet, DATs have gained greater acceptance and utility, making them useful resources for person-centred adherence support [26]. However, the accuracy might be impaired when the measures of adherence are exclusively dependent on self-reporting logs [16]. This issue takes particular relevance when focusing on SMS or phone reminders, where previous studies have identified an emerging challenge due to technology fatigue, as illustrated by Mukora et al. [20], (see figure 1). Despite these challenges, a recent time-trend analysis of 14 projects in 12 countries with high TB burden identified an increase in adherence patterns followed by a decrease. This trend was not attributable to technology fatigue but factors associated with common non-adherence patterns (e.g. suffering fewer symptoms being misinterpreted as being cured) [27].

FIGURE 1.

Response rates over time in treatment. RR: response rate. Reproduced and modified from [17] with permission.

Conclusion

The World Health Organization has acknowledged DATs as tools for enhancing adherence to TB treatment in less privileged countries [25]. The efficacy of DATs in LMICs can be increased by creating and executing customised treatments depending on contextual elements such as socioeconomic status, cultural norms and healthcare infrastructure. When implementing new digital solutions, operability and usability are crucial factors to consider. The effectiveness of a new system during operation is measured by its operability. On the other hand, acceptability describes how a user feels about a system or product, such as a website, software, gadget, or application. Therefore, before the digital intervention is applied widely, related workers should receive training. Apart from that, it is essential to guarantee data confidentiality, privacy, and adherence to standards of care by creating and implementing legislative frameworks and rules for the ethical use of DATs.

DATs are qualified for funding from the Global Fund and may now be routinely purchased by country programmes through the Stop TB Partnership's Global Drug Facility, which may help to scale up DATs in other high-burden locations [7]. This effective integration and implementation in TB care and support can be facilitated by these collaborative partnerships, capacity-building initiatives, and strategic interventions, though current hindrances like data security and access hurdles need to be addressed. We can improve the health and well-being of impacted populations globally and move closer to the goal of eliminating TB by utilising the potential of DATs.