Methods
Study design
We under take a two-arm, multicentre RCT at selected health facilities in Northwest Ethiopia. The study population includes all adult (>18 years) patients with TB enrolled in continuation phase of TB treatment. Patients enrolled in intensive phase of TB treatment; aged >18 years; own a cellphone; decided not to transfer to another facility and able and willing to provide informed consent are eligible for enrolment. The study excludes patients who started continuation treatment phase; no/limited mobile networks at participant’s home; with multidrug-resistant TB; enrolled to ART and already enrolled in another study that affect outcome of interest.
Intervention package
Participants in the first arm, in addition to usual care, will receive daily medication and weekly pill refilling reminders using graphic and short text messages with local language (Amharic). The graphic reminder intended to engage illiterate is intended to engage illiterate patients who cannot read text messages. Graphic message on figure 1A represents pills for daily medication. The number of daily pills to ranges from two to four based on their body weight. A strip of pills represented on figure 1B aimed to remind clinic appointment for pill refilling as shown in figure 1.
Figure 1Graphic and Amharic text message to remind weekly pill refilling and daily medication, Northwest Ethiopia, 2019.
The Amharic message for daily pill reminder ‘እባክዎ የቲቢ መድሀኒትዎን ይዋጡ’ means ‘Please take your pills’. The weekly based Amharic message ‘የክሊኒክ ቀጠሮዎ ነገ ስለሆነ እባክዎ በሰዓቱ ይገኙ!’ means ‘your clinic appointment is on tomorrow, please come on time’. The combined graphic and text messages are displayed and compatible for both smart and simple cellphones, as shown in figure 2.
Figure 2Cellphone compatibility for graphic and text messaging with Amharic language, Northwest Ethiopia, 2019.
Daily medication reminder is sent on 07:30–08:00 every day to remind the conventional time (08:30) for taking daily TB pills, whereas pill refilling reminder is sent a day before the due date (appointment date) at 18:30–19:00. Before initiating the intervention, each participant on intervention arm will receive a 5 min orientation on use of graphic and text messages. We will pilot the system involving participants out of the main study area.
Usual care
The control arm will receive only usual care. In this study, patients taking their daily medication at home with support of TB treatment supporters (TTS) during continuation phase is considered as usual care. TTS is either health extension worker, family member, neighbour, workmates or community figures and trained to observe the optimal administration TB treatment outside the health facility.21 This study compares treatment adherence between intervention and control groups (see figure 3).
Figure 3Weekly pill refilling and daily TB medication reminder system, Northwest Ethiopia, 2019.
The study will be conducted from 15 April to 15 September 2019, an average of 2 months for open cohort enrolment of participants and 4 months (122 daily doses) for follow-up period.
Outcomes
Adherence to TB treatment as primary outcome is assessed with self-report using Adherence to Refills and Medications Scale.22 As secondary outcome, pill refill visits measured using at least once for a medication refill visit were classified as non-adherent.23 The study will also measure treatment success rate, defined as the sum of patients cured and/or completed their treatment.24
Sample size determination and sampling procedure
We calculated the sample size using STATA V.14 considering superiority design and the proportion of patient adherence to treatment during continuation phase (π1=66%) in non-intervention group and π2=81% in intervention group.8 We considered 15% minimum detectable effect size, with α=5%, power (β) of 80% and attrition rate of 10%, the total sample size was 262 (131 in each study arms). It is a multicentred study conducted in 13 districts and 1 town administration. Overall, 19 health centres, one health centre from each district and subcity, were included in the trial.
Eligibility criteria
The study recruits participants who fulfil the criteria: age 18+ years, own a mobile phone, planned to continue in current facility and willing to participate in the study, whereas those who are physically impaired (unable to see) will be excluded.
Recruitment
In communication with focal person of the clinic, the research assistant will take the list of all patients with TB enrolled in intensive phase. The research assistant will keep a record for all patients enrolled in intensive phase irrespective of eligibility criteria. After checking eligibility criteria, the research assistants will explain details of the study and fill consent form including reasons for rejection to participate in this trial. The procedure continues in open cohort until required sample size is captured (see figure 4).
Figure 4Flow of RCT implementation on cellphone reminder intervention to support TB treatment, Northwest Ethiopia, 2019. MDR, multidrug resistant; RCT, randomised controlled trial; TB, tuberculosis.
Randomisation and allocation
The research assistants will be responsible for randomisation and enrolment of study participants. The study will use a covariate adaptive randomisation technique to control and balance the influence of covariates on the study outcome.25 Participants’ educational status (literate and illiterate) and type of facility (health centre and hospital) were considered during randomisation to account individual and service related effect on the outcome variable. We randomly include one patient if more than one patient attending TB treatment from one household.
Masking
We cannot mask study participants due to the nature of the intervention that requires an overt participation in the study. However, outcome assessors and investigators will be masked.
Follow-up
The follow-up period will be 4 months beginning from the time when the patient enrolled in continuation treatment phase. Participants will receive 138 messages (122 daily medication reminders and 16 weekly pill refilling reminders). Data collectors will follow and capture end line data on patient adherence through patient self-report on third month of continuation phase and attendance to appointments and treatment outcomes using register review on fourth month of continuation phase.
Statistical methods
Data analyst will use summary measures like mean, median, SD and percentages. The study will apply intension-to-treat analysis technique. We will use t-tests or Kruskal-Wallis test based on statistical distribution to compare the mean adherence score between intervention and control groups. A generalised linear mixed model to take account the lack of independence of the outcomes of repeated measures of adherence and appointments for the same patient. A χ2 test will be used to compare the proportion of participants attended clinic appointments as well as the proportion of participants with successful treatment.
