We conducted a convergent parallel mixed methods study, which was embedded in our cluster randomized trial (REMIND-TB), among patients with TB and TB care providers.

This study was approved by the Kilimanjaro Christian Medical College research ethics and review committee (approval 1157, dated December 10, 2018) and the National Health Research ethics subcommittee (ref NIMR/HQ/R.8a/VolIX/2992, dated January 14, 2019). We registered the trial at the Pan African Clinical Trials Registry under PACTR201811755733759 on November 8, 2018.

From 2019 to 2021, we conducted a 2-arm cluster randomized trial to investigate whether the evriMED pillbox with reminder cues and adherence feedback effectively improved adherence to anti-TB treatment among patients with TB in Kilimanjaro region in Tanzania. Study sites were randomized into 12 clusters: 6 intervention arms and 6 control arms. The inclusion criteria for the trial were patients’ diagnosis with presumptive sensitive TB, aged 18-65 years, attending care at any of the TB centers in the Kilimanjaro region of Tanzania, willing to use the evriMED pillbox, able to read and understand SMS text messages, and able to understand and willing to sign informed consent. Exclusion criteria were hospitalized patients and those who previously participated in similar studies. We provided each participant in the intervention arm with an evriMED1000 pillbox for their medication storage and intake. In the control arm, participants followed the standard of care procedures. In both arms, we followed participants for 6 months of treatment. In the evriMED arm, participants received a reminder SMS text message every day 30 minutes before their set time of taking medication. Detailed information about the REMIND-TB trial and the DAT can be found elsewhere [20].

The evriMED1000 is a type of tablet dispenser with a SIM card produced by Wisepill, based in South Africa. The pillbox records the opening of the box and stores the so-called medication events on a chip, along with the date and time whenever it is opened. This information is transmitted to a centralized server when one opens the device. Additionally, the evriMED1000 delivers a daily heartbeat event that includes information about the device’s identification, battery life, and signal quality. If the pillbox is not opened on a particular day, any pending events will be transmitted during the next heartbeat.

The evriMED1000 sends a reminder SMS text message to an individual’s mobile phone 30 minutes before intake time. If the individual does not take the medication within 1 hour of the intake period, a second reminder SMS text message is generated and sent. The patient does not require internet connectivity to receive the reminder SMS text message. The Wisepill pillbox uses a global roaming SIM chip that will connect to the best available mobile network in the area. These devices are designed to work in low-network resource settings (see Multimedia Appendix 1).

As the trial was an implementation study, TB care providers took full responsibility for participant care within their regular duties. They had to explain and demonstrate the use of the evriMED device to patients, provide medication through the device, and discuss adherence reports generated by the device during follow-up clinics. Participants were trained on device usage upon enrollment and were required to stay with it for the entire 6-month treatment period. Any challenges related to device functionality or misunderstandings were addressed through ongoing discussions between patients and care providers during follow-up visits.

To answer the objective regarding the acceptability of evriMED among patients with TB and TB care providers, we conducted a descriptive analysis of the exit survey responses using STATA (version 15; Stata Corp LLC). The results of the exit survey provided an overview of the frequency and percentages of each element of acceptability. In addition, we analyzed the qualitative responses of participants and TB care providers inductively and deductively by using thematic framework analyses. Three researchers independently read the transcripts (AEM, MSdB, and RAM). We developed memos and subthemes inductively based on the narratives and deductively by adopting preidentified themes from the theoretical framework constructs. We uploaded transcripts and memos in NVivo software (Lumivero) for coding and data organization. Narratives from the transcripts were then coded based on the predefined subthemes.

We enrolled 280 patients in the intervention arm; of these, 21 (7.5%) died before study completion, and 14 (5%) were excluded due to either being transferred to other regions or lost to follow-up. Of the 245 (87.5%) patients who completed treatment, 145 (59.2%) were not interviewed because their phone numbers were unreachable. We interviewed 100 (40.8%) patients. The details of the demographic characteristics of the patients with TB are shown below in Table 2.

In addition, we in-depth interviewed 18 patients and 15 TB care providers or directly observed therapy, and we reached data saturation in these interviews. Among the 18 patients, 12 (67%) were males and 6 (33%) were females. Detailed demographic characteristic are shown in Table 3. Of the 15 TB care providers who were interviewed, 3 (20%) were males and 12 (80%) were females, of whom 4 (27%) were clinicians, 1 (6%) was a pharmacist, 4 (27%) were medical attendants, and 6 (40%) were registered nurses.

The table summarizing the quantitative survey findings can be found in Multimedia Appendix 2.

This study aims to evaluate the acceptability of a DAT (evriMED1000 pillbox) among patients and TB care providers to improve adherence to anti-TB drugs in Kilimanjaro, Tanzania. The overall findings of this study indicate high acceptance of DAT among patients with TB and TB care providers. We found that the high acceptance of DATs was based on the positive attitude toward using the DAT (affective attitude), wherein 83% (83/100) of the participants were satisfied with the intervention, 98% (98/100) expressed good experiences, 78% (78/100) understood how the intervention works, and 92% (92/100) could use the pillbox, such as opening the device, refilling the pills, and recharging the box (intervention coherence). Of the 100 participants, 84 reported improved adherence (perceived effectiveness), and they preferred to continue receiving reminder SMS text messages (self-efficacy). A few participants reported experiencing difficulties while using DATs. Some participants reported experiencing TB-related stigma; 12% (12/100) were not happy being reminded daily, and 9% (9/100) reported experiencing difficulties keeping the device at home. Additionally, 30% (30/100) reported experiencing challenges with mobile network connectivity issues.

Our findings support similar studies reporting on the acceptability of DATs among patients and health care providers [12,13,15,24]. The potential benefit of DAT, such as its ability to monitor medication adherence or ease of use, was deemed valuable by patients and TB care providers. A study done in South Africa reported that the acceptability of DAT was highly associated with its ease of use among patients [12]. The real-time medication monitoring reports and feedback on adherence helped patients understand their health conditions and led to improved patient and health care provider relationships [12,24,25]. In addition, it led to improved care practice and behavior change among TB care providers. Many health care providers reported feeling more accountable for patient follow-up and motivating patients to adhere to the time of medication intake. Feedback on adherence counseling also generated a sense of care among patients, which had an impact on the psychosocial life of the patients.

However, participants reported several challenges with DATs, such as incorrect sending of SMS text messages due to network failure, large size of the pillbox, and existence of the reminder alarm, which led to fear of disclosure and, consequently, nonuse of the device during travelling. Similar challenges have been reported in other studies [14,26], which, if not well addressed, might contribute significantly to the nonuse of DATs and less uptake of DATs [24]. TB care providers reported increased workload during the use of DATs. Similar findings were reported in a study done in China, where health care providers reported a moderate workload increase during DAT implementation [16]. However, this contradicts a study in India, which reported a decreased health care workload [24]. We found that the increased workload by health care workers was reported as a major concern in settings with shortage of care staff—mainly dispensaries and health centers. Larger facilities such as hospitals reported a slight increase in the workload. In addition, 48% (48/100) of the participants stated that adherence reports from the device were not shown nor utilized in their conversations with health care workers. The health care provider was likely to have little knowledge of the value of adherence reports as a tool for counseling. Regular training should be conducted to reinforce their understanding of the intervention for effective scale-up. Fear of TB stigma and unwanted disclosure should be considered for effective intervention scale-up.

This study had certain limitations. One significant limitation was that because of the COVID-19 pandemic, all exit interviews were conducted via phone calls, which posed challenges related to network connectivity and potential interruptions during the questioning process. Mitigation strategies were employed, such as recapping participant responses to ensure accurate information capture. Another limitation was the small sample size in the exit survey compared to the total number of enrolled participants in the trial. The small sample size can be attributed to the public policy implemented in July 2021, wherein unregistered SIM cards were blocked, making it difficult to reach most participants. Nonetheless, we found that the demographic characteristics of the participants who were interviewed (100/245) did not differ from those of the participants who were not interviewed due to the change in the government policy, indicating that the interviewed participants were likely representative of those who were not interviewed. Additionally, 59.2% (145/245) of the participants who were not reached due to the change in the government policy were not affected during the medication period. The change in the government policy regarding SIM card registration occurred when many of our participants had already completed the treatment follow-up and were waiting for the exit survey. In this case, the change in the government policy impacted the exit survey process rather than the intervention itself.

This study has notable strengths that enhance its significance and scope. First, we enrolled participants from all TB-providing facilities in the Kilimanjaro region, thereby offering a comprehensive understanding of acceptability from a broader perspective. Notably, our research pioneers the investigation of the evriMED reminder pillbox’s acceptability among patients with TB in East Africa, providing valuable insights on the acceptance. Additionally, using the TFA facilitated a robust understanding of the acceptability of the evriMED reminder pillbox among patients with TB.

Our study demonstrates the positive acceptance of a DAT (evriMED) among patients and TB care providers for improving anti-TB drug adherence in Kilimanjaro, Tanzania. Although the potential acceptability of DATs is evident, addressing concerns related to mobile network connectivity and participants’ preferences regarding the number of reminder SMS text messages, and providing adequate training and technical support to health care providers are critical for successful implementation. Future research should explore the impact of evriMED on large-scale implementation in different settings.