This study explored the feasibility and acceptability of using MEMS-bottles to monitor adherence to ART in Tanzania's Kilimanjaro Region. Twenty-three patients on ART used a MEMS-bottle during three consecutive months. After this period, in-depth interviews were conducted with the participants. Overall, patients perceived MEMS-bottles as simple and acceptable for use. The absence of a label on the bottle was perceived as an advantage to the ordinary pill-bottles and, unexpectedly, reduced fear for HIV-status disclosure. The majority of patients stated that the use of MEMS-caps had improved their adherence because of increased family support and the knowledge that the MEMS-bottle was monitoring their behavior. This is in line with findings by Wendel et al. .
Whereas we found that the MEMS-bottle was readily accepted by patients on ART, other studies found that some patients, especially those who prefer not to disclose their HIV-status, do not like the white, bulky appearance of the MEMS for privacy concerns [18–20]. The high acceptability of MEMS-use by patients in the current study is particularly remarkable because fears of disclosure of one's HIV-status and consecutive stigmatization are widely present in Tanzania and patients prefer to carry their medication as discretely as possible [21, 22]. It turned out that the MEMS-design facilitated just that, with the absence of a medication label and the perception that the MEMS-bottle looks like a package of common products in this region, such as lotion or shampoo bottles.
The main barrier to using MEMS was on travel occasions. Three patients indicated to have left the MEMS bottle at home during travelling because of fears that it might be lost or stolen. This fear of losing the MEMS bottle on travel occasions was because patients perceived the MEMS-cap as a valuable device that could be stolen. This finding indicates the need of stressing patients that using the medication from the MEMS-caps when travelling is important and that the risk of it being stolen is acceptable and the responsibility of the researcher. This is especially important because adhering to the medication may be more challenging when travelling.
It has been suggested that MEMS monitoring can be perceived as intrusive or patronizing . Patients in the present study perceived the monitoring of their adherence behavior as an advantage; they reported that their adherence had improved due to MEMS-monitoring. Most participants explicitly said that they paid more attention to taking their medication on the correct time because they were aware that their adherence behavior was monitored. Some patients admitted that they had been skipping medication intake for weeks in a row prior to starting to use the MEMS-bottle, but not anymore since MEMS-use. Adherence rates were indeed high, especially during the first month of monitoring. However, the levels of adherence tended to decrease after the first month of the study and then stabilized, suggesting that patients returned to their usual adherence behavior. This pattern has also been observed in other studies [23, 24].
Examining the MEMS-data in order to determine periods of non-adherence or non-MEM-use, patterns of missed doses were found prior to clinic visits, because too few pills were being given to patients to bridge time between consecutive visits. Upon discussing the potential health consequences of this malpractice, the healthcare providers and dispensing nurse, acknowledging the necessity of a sufficient supply of medication, now provide patients with enough medication (including a buffer supply in case a patient is not able to visit the clinic on the appointed day).
This study showed that MEMS-devices have potential application in adherence research and monitoring of adherence in clinical practice in resource-limited settings. Although attention has to be paid to accurate MEMS-use, both at the time of instruction as well as at the moment of data interpretation, we did not observe some of the typical hurdles related to MEMS-use in resource-rich settings. For example, patients did not experience problems with the MEMS-design. In the absence of financial resources for regular viral load or CD4 testing, (intermittent) monitoring of adherence with the MEMS-cap may be possible and thus provide clinicians with valuable information to support the patients' medical treatment and medication adherence.
This study had some limitations. First, the sample size was small and the study was conducted in a single clinic, which limits external validity of the findings. However, the aim of the study was to explore the feasibility and acceptability of the MEMS-cap through qualitative procedures. Second, patients may not have understood well enough that their adherence data would not be shared with their doctor. The expectation that the doctor would act upon detection of non-adherence, might explain part of the intervention effect of MEMS-use, as described above. Third, the average adherence of patients with long periods of non-MEMS use should be interpreted with caution, since it cannot be verified whether medication was taken as prescribed in that period.