The authors have declared that no competing interests exist.
Conceived and designed the experiments: XL JJL HZ JL XLi SH SC LW SJ DPC KLF. Performed the experiments: XL HZ WL SZ GZ LB JL XLi HC JLu. Analyzed the data: XL JJL HZ JL XLi HC ML RC JC SH SC LW SJ DPC KLF. Wrote the first draft of the manuscript: XL JJL DPC KLF. Contributed to the writing of the manuscript: XL JJL HZ WL SZ GZ LB JL XLi HC ML RC JC JLu SH SC LW SJ DPC KLF. Enrolled patients: WL SZ GZ LB JLu. Agree with the manuscript’s results and conclusions: XL JJL HZ WL SZ GZ LB JL XLi HC ML RC JC JLu SH SC LW SJ DPC KLF. All authors have read, and confirm that they meet, ICMJE criteria for authorship. SZ was unable to confirm for himself and all other authors have confirmed on his behalf that he meets ICMJE authorship criteria.
‡ Author Shun Zhang was unable to confirm his authorship contributions. On his behalf, all other authors have reported his contributions to the best of their knowledge.
Mobile text messaging and medication monitors (medication monitor boxes) have the potential to improve adherence to tuberculosis (TB) treatment and reduce the need for directly observed treatment (DOT), but to our knowledge they have not been properly evaluated in TB patients. We assessed the effectiveness of text messaging and medication monitors to improve medication adherence in TB patients.
In a pragmatic cluster-randomised trial, 36 districts/counties (each with at least 300 active pulmonary TB patients registered in 2009) within the provinces of Heilongjiang, Jiangsu, Hunan, and Chongqing, China, were randomised using stratification and restriction to one of four case-management approaches in which patients received reminders via text messages, a medication monitor, combined, or neither (control). Patients in the intervention arms received reminders to take their drugs and reminders for monthly follow-up visits, and the managing doctor was recommended to switch patients with adherence problems to more intensive management or DOT. In all arms, patients took medications out of a medication monitor box, which recorded when the box was opened, but the box gave reminders only in the medication monitor and combined arms. Patients were followed up for 6 mo. The primary endpoint was the percentage of patient-months on TB treatment where at least 20% of doses were missed as measured by pill count and failure to open the medication monitor box. Secondary endpoints included additional adherence and standard treatment outcome measures. Interventions were not masked to study staff and patients. From 1 June 2011 to 7 March 2012, 4,292 new pulmonary TB patients were enrolled across the 36 clusters. A total of 119 patients (by arm: 33 control, 33 text messaging, 23 medication monitor, 30 combined) withdrew from the study in the first month because they were reassessed as not having TB by their managing doctor (61 patients) or were switched to a different treatment model because of hospitalisation or travel (58 patients), leaving 4,173 TB patients (by arm: 1,104 control, 1,008 text messaging, 997 medication monitor, 1,064 combined). The cluster geometric mean of the percentage of patient-months on TB treatment where at least 20% of doses were missed was 29.9% in the control arm; in comparison, this percentage was 27.3% in the text messaging arm (adjusted mean ratio [aMR] 0.94, 95% CI 0.71, 1.24), 17.0% in the medication monitor arm (aMR 0.58, 95% CI 0.42, 0.79), and 13.9% in the combined arm (aMR 0.49, 95% CI 0.27, 0.88). Patient loss to follow-up was lower in the text messaging arm than the control arm (aMR 0.42, 95% CI 0.18–0.98). Equipment malfunction or operation error was reported in all study arms. Analyses separating patients with and without medication monitor problems did not change the results. Initiation of intensive management was underutilised.
This study is the first to our knowledge to utilise a randomised trial design to demonstrate the effectiveness of a medication monitor to improve medication adherence in TB patients. Reminders from medication monitors improved medication adherence in TB patients, but text messaging reminders did not. In a setting such as China where universal use of DOT is not feasible, innovative approaches to support patients in adhering to TB treatment, such as this, are needed.
Current Controlled Trials,
In a cluster-randomized controlled trial, Katherine Fielding and colleagues examine the effectiveness of electronic reminders for improving tuberculosis medication adherence.
Tuberculosis—a contagious bacterial disease that usually infects the lungs—is a major global public health problem. Every year, about 9 million people develop tuberculosis and at least 1.3 million people die as a result.
Because tuberculosis treatment is long and unpleasant, patients often fail to take all their drugs. To improve medication adherence, the World Health Organization recommends that health care workers supervise patients while they take their medication (directly observed treatment, DOT). However, DOT can be hard to implement. In China, for example, where 11% of tuberculosis cases occur, DOT cannot be implemented in many parts of the country, and the national tuberculosis control policy permits self-administered treatment and treatment monitored by family members. It is estimated that over half of individuals with tuberculosis in China self-administer their treatment, but, in 2010, 20% of patients treated using nationally recommended case-management approaches were lost to follow-up or failed to take their medications regularly. In this pragmatic cluster-randomized trial, the researchers investigate whether reminders delivered by mobile phone or by medication monitor boxes (which provide audio reminders to patients and record when the box is opened) might improve tuberculosis medication adherence in China. A pragmatic trial asks whether an intervention works under real-life conditions; a cluster-randomized trial randomly assigns groups of people (here, people living in different counties/districts) to receive alternative interventions and compares outcomes in the differently treated “clusters.”
The researchers assigned people newly diagnosed with tuberculosis in counties/districts to receive reminders about taking their antibiotics and about monthly follow-up visits via text messaging, a medication monitor, or both text messaging and a medication monitor (the intervention arms), or to receive standard nationally recommended care without electronic reminders (the control arm). All the trial participants (about 1,000 per arm) took their drugs out of a medication monitor box, but the box’s audio reminder function was switched off in the text messaging only and control arms. In the intervention arms, doctors were advised to switch participants with poor medication adherence (evaluated at follow-up visits) to either more intensive management or DOT, depending on the level of missed treatment doses. Compared to the control arm, the percentage of patient-months with at least 20% of the drug doses missed (called “poor adherence” and measured by pill counts and data from the medication monitor) was not significantly reduced in the text messaging arm, whereas poor adherence was significantly reduced by 42% and 51% in the medication monitor and the combined arms, respectively (a significant reduction is unlikely to have happened by chance). Notably, fewer patients were switched to intensive management or DOT than expected based on medication adherence evaluations.
These findings show that, in China, the use of an electronic medication monitor box to remind patients to take their anti-tuberculosis drugs improved medication adherence. Interestingly, text messaging alone, which has been shown to improve adherence to antiretroviral therapy among HIV-positive individuals, did not improve medication adherence among patients with tuberculosis, possibly because the messages were too frequent or too impersonal, although this intervention (but none of the others) did reduce patient loss to follow-up. Battery problems with the medication monitor may have resulted in over-estimation of poor adherence to treatment. Moreover, the researchers’ assumption that opening the medication monitor box is synonymous with taking the medication may have introduced some inaccuracies into these findings. Despite these limitations and the underuse of more intensive case management in patients with poor adherence, these findings suggest that using medication monitors to deliver electronic drug reminders to patients might improve medication adherence among patients with tuberculosis in China and in other settings.
This list of resources contains links that can be accessed when viewing the PDF on a device or via the online version of the article at This study is further discussed in a The World Health Organization provides information (in several languages) on The The US Centers for Disease Control and Prevention provides information about The US National Institute of Allergy and Infectious Diseases also has detailed information on all aspects of MedlinePlus has links to further information about More
In 2013, China ranked second in the world in number of tuberculosis (TB) cases, accounting for 11% of the estimated 9 million global cases [
Electronic reminders and monitoring have been used in several disease conditions to improve medication adherence. The potential of mobile phone technology to improve the quality and delivery of health care, including diagnosis, treatment adherence, and data collection, has been recognised [
Electronic medication packaging (EMP) devices can remind patients to take medications on time, monitor time of drug intake, and alert health care workers to patients who have missed doses [
To evaluate the use of electronic reminders to improve medication adherence in TB patients, we conducted a cluster-randomised controlled trial to assess the effectiveness of three case-management approaches—using reminders via text messaging, a medication monitor, or both—compared to the standard of care in China.
The study was approved by the ethics committees of the Chinese Center for Disease Control and Prevention (201008) and the London School of Hygiene & Tropical Medicine (5704). All patients provided written consent prior to inclusion in the study.
This study was a pragmatic cluster-randomised trial with one control and three intervention arms. New pulmonary TB patients, starting on standard 6-mo short-course chemotherapy and managed as outpatients, were recruited into the study. Those in the control arm were managed according to the standard of care of the National Tuberculosis Control Program. Those in the three intervention arms also received reminders to take their medications from text messages via short message service (SMS), a medication monitor, or both. If adherence problems were subsequently detected, more intensive management was recommended. For logistical simplicity, randomisation was conducted at the cluster level.
Clusters were defined as rural counties or urban districts within the provinces of Heilongjiang, Jiangsu, Hunan, and Chongqing—located in northern, eastern, central, and western China, respectively. Each cluster had at least 300 active pulmonary TB patients registered in 2009 (
In each cluster, consecutive pulmonary TB patients newly registered at the public health TB clinic were screened for study eligibility. Inclusion criteria included the following: no communication impairment (mental, visual, auditory, or speech), patient at least 18 y old, and patient or family member able to use mobile phone to read SMS text messages and use the medication monitor after training. Because of the nature of the study, interventions were not masked to study staff and patients.
The 36 clusters were randomised to the four arms by rural/urban stratum and restricted such that each province had at least two clusters in each arm. From 5,000 randomly generated acceptable allocations, one was chosen at random as the final allocation using Stata version 12.0. See
All patients were treated according to National Tuberculosis Control Program guidelines including the use of isoniazid, rifampin, ethambutol, and pyrazinamide for 2 mo, followed by isoniazid and rifampin for 4 mo; the programme used every other day dosing for the entire treatment course. Patients received their blister-pack medications in a medication monitor box that electronically collected the date and time of each opening. In the control and text messaging arms, the medication monitor box was in silent mode and was not used as a reminder tool for patients. At each monthly visit, patients were dispensed enough medications for a 1-mo period.
At the start of treatment, the doctor and the patient selected one of three treatment monitoring approaches as per National Tuberculosis Control Program protocol: self-administered treatment, treatment supervised by family members, or treatment supervised by health care workers. The local doctor monitoring treatment at the township or village/community level was given 60 renminbi (RMB; equivalent to US$10) for each patient.
As in the control arm, patients and their doctors in the intervention arms selected one of the three treatment monitoring approaches as per National Tuberculosis Control Program protocol. The interventions had three common components: reminders for timely drug intake, reminders for monthly follow-up visits, and a recommendation for doctors to switch patients from self-administered treatment to a more intensive treatment monitoring approach when patients were found to have adherence problems based on data available to the managing doctor (
Intervention Arm | Component | ||
---|---|---|---|
Reminding Patient to Take Medication | Reminding Patient of the Monthly Dispensing Visit |
Assessment of Adherence by Doctor at the Monthly Dispensing Visit | |
There is an agreed time (based on patient preference) for the medication to be taken. Up to three SMS reminders are sent to the patient on the day of medication, depending on whether the patient replies or not. These reminders are sent at the agreed time medication is to be taken and subsequently at 12 noon and 6 |
SMS reminder sent 4, 3, 2, and 1 d before the scheduled monthly follow-up visit. | Adherence patterns based on patient interview, pill count from medication monitor box, and SMS feedback. |
|
There is an agreed time (based on patient preference) for the medication to be taken. If the box is not opened at that time, there are up to eight further reminders (bleep), taking place at 5 min, 20 min, 30 min, 1 h, 2 h, 4 h, 6 h, and 8 h after the agreed time. Once the box has been opened, the reminders stop for that day. | Medication box reminder (human voice) 4, 3, 2, and 1 d before the scheduled monthly follow-up visit. | Adherence patterns based on patient interview, pill count from medication monitor box, and electronic data on dates and times of opening of the medication monitor box. Intensive management/DOT initiation and incentives as above. | |
A combination of the SMS and medication monitor reminders, as described above. | A combination of the SMS and medication monitor reminders, as described above. | Adherence patterns based on patient interview, pill count from medication monitor box, and electronic data on dates and times of opening of the medication monitor box. Intensive management/DOT initiation and incentives as above. |
1In all three intervention arms and in the control arm there is a National Tuberculosis Control Program requirement for the managing doctor to contact the patient after 3 d following a missed visit, using all available contact methods.
2No incentives were paid to township doctors (urban).
In the text messaging and combined arms, a text message reminded patients to take their medication at the time previously agreed on with the patient. If patient did not reply to the text message, another two text messages would be sent later in the day. Once the patient replied to the SMS reminder, with or without text, the reminders were stopped for that day. Similarly, in the medication monitor and combined arms, an audio reminder from the medication monitor box reminded patients to take their medication. If the patient did not open the medication monitor by a pre-specified time, up to eight additional reminders sounded. Once the box was opened, the reminders were stopped for that day. In all three intervention arms, patients received four reminders to attend the monthly dispensing visit (
At each monthly follow-up visit, the managing doctor evaluated adherence patterns. Missed doses were defined as the larger of (1) missed doses based on pill count or (2) missed doses from missing SMS reply (in the text messaging only arm) or from failure to open the medication monitor box (in the other two intervention arms). If the patient reported any equipment malfunction or operation error during the previous month, the number of missed doses was based on pill count only.
If 1–2 doses were missed in the previous month, we recommended the doctor counsel the patient on the importance of adherence to medication but allowed self-administered treatment to continue. If 3–6 doses were missed, we recommended the doctor switch the patient to “intensive management”, in which township or village/community doctors visited the patient twice a month or once a week, respectively, for the rest of treatment. If seven or more doses were missed the previous month or if 3–6 doses were missed in two prior months, we recommended the doctor switch the patient to DOT, with each dose of treatment supervised by the township or village/community doctor. The local doctors monitoring treatment at the township or village/community level were given 5 RMB (US$0.8) every time they made a visit to a patient as part of the intensive management or DOT, in addition to the 60 RMB (US$10) they received for every patient.
All study endpoints were measured at the individual level. The primary study endpoint for treatment adherence was the percentage of patient-months where at least 20% of doses (equivalent to missing three of 15 doses) were missed (“poor adherence”). The secondary treatment adherence endpoints were (1) percentage of patient-months where at least 47% of doses (equivalent to seven of 15 doses) were missed, (2) percentage of total doses missed over the 6 mo of treatment, (3) binary categorisation of secondary endpoint 2 as <10% versus ≥10% of total doses missed (National Tuberculosis Control Program definition of non-adherent), and (4) percentage of patient-months on TB treatment where at least 20% of doses were missed based on pill count only.
Measurement of the adherence endpoints utilised the same data for all four study arms and included data from the medication monitor box, downloaded into a database when patients returned for their monthly medication refill. All adherence endpoints, except secondary endpoint 4, measured the number of missed doses per month as the larger of the number of missed doses from pill count or the number of failures to open the medication monitor box. A month was defined as the number of days between successive appointments, typically 30 d, during which 15 doses should have been taken, but this was adjusted for early/late or missed visits and reduced by the number of days a patient was hospitalised or temporarily discontinued treatment on doctor’s recommendation. Data were censored when a patient died, moved, or permanently discontinued treatment based on a doctor’s decision. For those who were lost to follow-up during treatment, we assumed no drug intake (100% non-adherence) for the period from the date of being lost to follow-up to the date when they should have completed treatment. We also conducted a post hoc sensitivity analysis censoring adherence measurement at the time of loss to follow-up.
The secondary TB treatment outcome endpoints, following standard WHO definitions, were (1) poor treatment outcome, defined as death, treatment failure, or patient loss to follow-up and (2) patient loss to follow-up (
Sample size calculations were based on a binary endpoint of non-adherence and took into account the study design [
Analysis of all endpoints used standard methods for a small number of clusters [
There were problems with loose batteries in some of the medication monitors, resulting in a power outage during which data on box openings were not captured. The problem could be easily fixed by the patient or the doctor when they became aware of the problem We conducted a post hoc stratified analysis separating patient-months into those that had a record of a medication monitor problem and those that did not (
Analysis was conducted using Stata version 13.
From 1 June 2011 to 7 March 2012, 6,203 pulmonary TB patients were screened in the 36 clusters, and 5,057 (81.5%) met enrolment criteria, of whom 4,292 (84.9%) gave informed consent. Of these, 61 (1.4%) were reassessed as not having TB by their managing doctor, and 58 (1.4%) were withdrawn from the study as they had switched to a different treatment model within the first month (due to hospitalisation or travel) and were therefore excluded from all analyses (
Reasons for non-eligibility: SMS req = unable to use mobile phone after training; <18y = less than 18 y of age; comm dis = communication disability. *Withdrew from the study but continued treatment in the local Center for Disease Control and Prevention.
Overall, 71.0% of participants were male, median age was 43 y (inter-quartile range [IQR] 29 to 56 y), 56.0% were farmers, 7.9% were illiterate, median household income was 20,000 RMB (IQR 10,000 to 30,000 RMB), and 36.3% were smear positive (
Characteristic | Subcategory | Control Arm ( |
Text Messaging Arm ( |
Medication Monitor Arm ( |
Combined Arm ( |
||||
---|---|---|---|---|---|---|---|---|---|
Percent | Percent | Percent | Percent | ||||||
70.1% | 774 | 71.3% | 719 | 71.1% | 709 | 71.6% | 762 | ||
<30 | 30.2% | 333 | 23.3% | 235 | 23.1% | 230 | 24.2% | 258 | |
30–39 | 16.0% | 177 | 19.0% | 192 | 11.5% | 115 | 16.8% | 179 | |
40–59 | 39.1% | 432 | 41.1% | 414 | 39.1% | 390 | 41.2% | 438 | |
60+ | 14.7% | 162 | 16.6% | 167 | 26.3% | 262 | 17.8% | 189 | |
48.9% | 540 | 60.7% | 612 | 66.0% | 658 | 49.5% | 527 | ||
Illiterate | 7.3% | 81 | 5.3% | 53 | 11.2% | 112 | 8.0% | 85 | |
Lower middle | 62.8% | 693 | 75.8% | 764 | 66.7% | 665 | 63.3% | 674 | |
Upper middle | 17.6% | 194 | 12.9% | 130 | 13.2% | 132 | 19.1% | 203 | |
University | 12.3% | 136 | 6.1% | 61 | 8.8% | 88 | 9.6% | 102 | |
Not married | 23.9% | 264 | 15.8% | 159 | 18.3% | 182 | 19.0% | 202 | |
First marriage | 69.7% | 770 | 77.8% | 784 | 76.0% | 758 | 73.0% | 777 | |
Other | 6.3% | 70 | 6.4% | 65 | 5.7% | 57 | 8.0% | 85 | |
84.3% | 931 | 92.4% | 931 | 97.8% | 975 | 91.5% | 974 | ||
41.1% | 454 | 27.9% | 281 | 28.2% | 281 | 26.0% | 277 | ||
<10 | 23.3% | 257 | 24.9% | 251 | 17.5% | 174 | 35.9% | 382 | |
10–29 | 38.6% | 426 | 42.0% | 423 | 37.0% | 369 | 32.1% | 342 | |
20–39 | 18.1% | 200 | 12.1% | 122 | 15.6% | 156 | 13.3% | 141 | |
≥40 | 20.0% | 221 | 21.0% | 212 | 29.9% | 298 | 18.7% | 199 | |
≤1 | 66.6% | 735 | 49.5% | 499 | 63.5% | 633 | 66.1% | 703 | |
2 | 21.8% | 241 | 32.0% | 323 | 21.1% | 210 | 20.1% | 214 | |
>2 | 11.6% | 128 | 18.5% | 186 | 15.4% | 154 | 13.8% | 147 | |
33.8% | 373 | 38.0% | 383 | 39.0% | 389 | 34.6% | 368 |
Table excludes 61 patients who were reassessed as not having TB by their managing doctor and 58 patients who were withdrawn from the study as they switched to a different treatment model within the first month (due to hospitalisation or travel).
1Over last calendar year.
The cluster geometric mean of the percentage of patient-months on TB treatment where at least 20% of doses were missed was 29.9% in the control arm (range 16.0%–48.1%;
Solid bars represent geometric means of cluster-level proportions.
Endpoint and Study Arm | Number of Patients | Geometric Mean of Cluster-Level Endpoint | Unadjusted Analysis | Adjusted Analysis |
||
---|---|---|---|---|---|---|
MR (95% CI) | MR (95% CI) | |||||
Control | 1,091 | 29.9% | 1 | 1 | ||
Text messaging | 996 | 27.3% | 0.91 (0.66, 1.25) | 0.536 | 0.94 (0.71, 1.24) | 0.622 |
Medication monitor | 992 | 17.0% | 0.57 (0.40, 0.81) | 0.004 | 0.58 (0.42, 0.79) | 0.002 |
Combined | 1,059 | 13.9% | 0.46 (0.25, 0.86) | 0.018 | 0.49 (0.27, 0.88) | 0.020 |
Control | 1,091 | 18.9% | 1 | 1 | ||
Text messaging | 996 | 17.8% | 0.94 (0.63, 1.41) | 0.744 | 0.96 (0.67, 1.38) | 0.808 |
Medication monitor | 992 | 11.1% | 0.59 (0.38, 0.91) | 0.022 | 0.60 (0.40, 0.89) | 0.015 |
Combined | 1,059 | 9.4% | 0.50 (0.26, 0.94) | 0.034 | 0.52 (0.28, 0.97) | 0.042 |
Control | 1,091 | 22.6% | 1 | 1 | ||
Text messaging | 996 | 20.7% | 0.92 (0.66, 1.28) | 0.584 | 0.94 (0.70, 1.26) | 0.649 |
Medication monitor | 992 | 13.9% | 0.61 (0.44, 0.86) | 0.008 | 0.62 (0.46, 0.84) | 0.004 |
Combined | 1,059 | 11.4% | 0.51 (0.28, 0.92) | 0.029 | 0.53 (0.29, 0.95) | 0.034 |
Control | 1,091 | 57.4% | 1 | 1 | ||
Text messaging | 996 | 54.7% | 0.95 (0.74, 1.23) | 0.690 | 0.97 (0.77, 1.23) | 0.807 |
Medication monitor | 992 | 38.7% | 0.67 (0.50, 0.90) | 0.011 | 0.68 (0.52, 0.89) | 0.008 |
Combined | 1,059 | 31.0% | 0.54 (0.30, 0.96) | 0.037 | 0.56 (0.33, 0.97) | 0.041 |
Control | 1,091 | 9.2% | 1 | 1 | ||
Text messaging | 996 | 3.8% | 0.41 (0.20, 0.87) | 0.023 | 0.39 (0.18, 0.83) | 0.018 |
Medication monitor | 992 | 5.5% | 0.60 (0.33, 1.08) | 0.084 | 0.58 (0.35, 0.96) | 0.037 |
Combined | 1,059 | 6.4% | 0.70 (0.34, 1.45) | 0.307 | 0.67 (0.31, 1.47) | 0.294 |
Control | 1,066 | 8.6% | 1 | 1 | ||
Text messaging | 966 | 3.9% | 0.45 (0.18, 1.16) | 0.092 | 0.44 (0.17, 1.13) | 0.084 |
Medication monitor | 955 | 6.1% | 0.70 (0.32, 1.53) | 0.264 | 0.71 (0.33, 1.51) | 0.346 |
Combined | 992 | 8.8% | 1.01 (0.46, 2.22) | 0.973 | 1.00 (0.45, 2.20) | 0.991 |
Control | 1,057 | 8.5% | 1 | 1 | ||
Text messaging | 954 | 3.6% | 0.42 (0.18, 1.00) | 0.050 | 0.42 (0.18, 0.98) | 0.046 |
Medication monitor | 946 | 5.0% | 0.58 (0.23, 1.51) | 0.243 | 0.61 (0.25, 1.51) | 0.264 |
Combined | 982 | 7.6% | 0.90 (0.38, 2.08) | 0.783 | 0.90 (0.38, 2.09) | 0.784 |
1Adjusted for individual-level variables of gender, age group, occupation (farmer or not), local resident or not, distance to nearest TB clinic, education level, income category, and smear result at start of treatment, and for the cluster-level variable of pre-randomisation stratum (rural/urban).
2Doses missed based on the larger of missed doses from (1) pill count or (2) the number of failures to open the medication monitor.
3Excludes 35 patients who had no adherence data (by arm: 13 in control, 12 in text messaging, five in medication monitor, and five in combined).
4Data collected monthly, then aggregated at the patient level as a proportion. The arithmetic means of these proportions were used to produce a cluster-level summary. Finally, the geometric mean (as a log transformation of the cluster-level summaries; see
5Excludes 188 patients with outcome of side effect on treatment, resulting in an extension on TB treatment and the final outcome not being documented (by arm: 38 in control, 42 in text messaging, 41 in medication monitor, and 67 in combined), five patients who transferred to another clinic (all in combined arm; unknown outcome in new clinic), and one patient with missing outcome (in medication monitor arm). The numbers of patients with a poor treatment outcome by arm, ignoring cluster, are as follows: control arm—121/1,066; text messaging arm—53/966; medication monitor arm—68/955; combined arm—99/992.
6Excludes 188 patients with outcome of side effect on treatment, resulting in an extension on TB treatment and the final outcome not being documented (by arm: 38 in control, 42 in text messaging, 41 in medication monitor, 67 in combined), 13 patients with treatment failure (by arm: three in control, six in text messaging, one in medication monitor, three in combined), 27 deaths (by arm: six in control, six in text messaging, eight in medication monitor, seven in combined), five patients who transferred to another clinic (all in combined arm; unknown outcome in new clinic), and one patient with missing outcome (medication monitor arm). The numbers of patients lost to follow-up by arm, ignoring cluster, are as follows: control arm—112/1,057; text messaging arm—41/954; medication monitor arm—59/946; combined arm—89/982.
There were similar reductions in the intervention arms versus the control arm in the percentage of months with at least 47% of doses missed (equivalent to 7/15 doses), the percentage of doses missed over the whole treatment period, and the percentage of patients who missed at least 10% of their doses, in both unadjusted and adjusted analyses (
The text messaging arm had a lower patient loss to follow-up and occurrence of poor treatment outcome than the control arm. Modest reductions in patient loss to follow-up were also seen for the medication monitor and combined arms, though confidence intervals for the effect estimates included one. A post hoc sensitivity analysis that censored adherence measurement at the time of loss to follow-up showed a strengthening of the evidence for a reduction in poor adherence as measured by pill count in the three intervention arms, but otherwise very similar results (
Problems with the medication monitor box, recorded either by the doctor at the monthly visit or by the medication monitor as power interruption, were more common in the medication monitor (49.4% of patients;
Process Measure | Control Arm ( |
Text Messaging Arm ( |
Medication Monitor Arm ( |
Combined Arm ( |
||||
---|---|---|---|---|---|---|---|---|
Percent |
Percent |
Percent |
Percent |
|||||
Reported by doctor | 7.4% | 82 | 9.0% | 91 | 44.2% | 441 | 41.1% | 437 |
Recorded by medication monitor |
12.0% | 132 | 9.8% | 99 | 20.9% | 208 | 21.4% | 228 |
Any problem |
17.8% | 196 | 16.7% | 168 | 49.4% | 492 | 48.0% | 511 |
Reported by doctor | 56.5% | 569 | 27.3% | 290 | ||||
Should start |
4.1% | 41 | 4.3% | 43 | 4.4% | 47 | ||
Started | 4.0% | 40 | 3.2% | 32 | 4.1% | 44 | ||
Should start |
0.8% | 8 | 1.3% | 13 | 1.4% | 15 | ||
Started | 0.8% | 8 | 0.9% | 9 | 0.9% | 10 |
1Percentage denominator is total number of patients in arm.
2An incorrect date was recorded by the medication monitor, indicating the power had failed and then been resolved without resetting the internal clock to the correct date.
3Reported by doctor or recorded by medication monitor.
4According to information available to the patient’s dispensing doctor.
Similar percentages of patients in the three intervention arms were switched to intensive management (3.2%–4.1%) and DOT (0.8%–0.9%) (
Minor problems with the mobile phones used to receive text messages were also common and were reported by 56.5% of those in the text messaging arm and 27.3% of those in the combined arm (
Our study found that the use of a medication monitor to remind TB patients to take their drugs reduced poor medication adherence by 40%–50% compared to the standard of care in China’s National Tuberculosis Control Program. This reduction was seen for all TB treatment adherence measures in this study. The use of text messaging did not reduce poor medication adherence but did reduce patient loss to follow-up by 58%. The use of a medication monitor alone resulted in a smaller, and not statistically significant, reduction in patient loss to follow-up compared to control; however, the study was not powered for this treatment outcome.
Even though many types of EMP devices exist and have been used for different disease conditions, a recent systematic review concluded that there were limited data supporting their use in improving adherence [
Our results demonstrate that text messaging did not reduce poor medication adherence among TB patients. This contrasts with available evidence supporting the use of text messaging among HIV patients on antiretroviral therapy [
In our intervention arms, we recommended that doctors switch patients to intensive patient management or DOT when adherence problems were documented. However, this rarely happened, despite data suggesting that a substantial percentage of patients should have switched. The trial was designed to be pragmatic, and so we did not enforce the initiation of more intensive management or DOT. Because problems with medication monitors, mobile phones, or their use were frequently reported, it is possible that doctors largely chose to ignore the electronic adherence data when deciding whether to switch patients to more intensive case-management approaches. In addition, doctors may not have had sufficient financial incentives to carry out more intensive case management.
Even though more intensive case-management approaches were underutilised in the presence of recorded treatment non-adherence, we still observed better adherence in the medication monitor and combined arms. This suggests that the use of a medication monitor to remind patients to take their medications can improve treatment adherence by itself. If information on poor treatment adherence had been used by providers to switch patients to more intensive case-management approaches, as intended, it is likely we would have seen an even greater reduction in poor treatment adherence with the use of medication monitors.
Interestingly, text messaging reduced the risk of patient loss to follow-up. Perhaps text messaging is an effective approach to remind patients of follow-up visits and resulted in better attendance at monthly visits. However, a recent meta-analysis suggests that the effectiveness of SMS reminders for appointments is modest at best and not more effective than other types of reminders [
The differences in the effects of the interventions in terms of adherence and treatment outcome endpoints suggest these do not correlate well. However, adherence is complex, and a recent taxonomy divides it into three constructs—initiation (patient takes the first dose), implementation (measure of how patient’s actual dosing history corresponds to the prescribed dosing regimen from initiation until the last dose is taken), and discontinuation (patient stops taking the prescribed medication) [
Our study had several limitations. First, the battery problems with the medication monitors in our study led to loss of data in some patients, potentially resulting in an over-estimation of poor adherence. However, when we performed a stratified analysis using patient-months with or without this problem, we found similar reductions in poor treatment adherence. Second, more intensive case-management approaches were underutilised, possibly because doctors disregarded information from the medication monitor or SMS feedback. In addition, the financial incentives given to the doctors to perform more intensive management may have been inadequate. Third, for the adherence endpoints, we assumed that opening the medication monitor box was synonymous with drug intake, which may not have been the case, though our measure of poor adherence using a combination of this and pill count is arguably more robust than pill count alone. Pill counts have often been shown to under-report poor adherence or non-adherence [
The use of a medication monitor as a reminder for drug intake in combination with the identification of patients requiring more intensive management has been suggested as an approach for improving TB treatment adherence [
(DOCX)
(DOCX)
(DOCX)
(DOCX)
(DOCX)
(DOCX)
(DOCX)
(DOCX)
(DOCX)
We thank the thousands of participants who consented to take part in this study.
adjusted mean ratio
directly observed treatment
electronic medication packaging
inter-quartile range
mean ratio
renminbi
short message service
tuberculosis