This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR mHealth and uHealth, is properly cited. The complete bibliographic information, a link to the original publication on http://mhealth.jmir.org/, as well as this copyright and license information must be included.
Problems with prospective memory, which refers to the ability to remember future intentions, cause deficits in basic and instrumental activities of daily living, such as taking medications. Older adults show minimal deficits when they rely on mostly preserved and relatively automatic associative retrieval processes. On the basis of this, we propose to provide external cues to support the automatic retrieval of an intended action, that is, to take medicines. To reach this end, we developed the Medication Ambient Display (MAD), a system that unobtrusively presents relevant information (unless it requires the users’ attention) and uses different abstract modalities to provide external cues that enable older adults to easily take their medications on time and be aware of their medication adherence.
This study aimed to assess the adoption and effect of external cues provided through ambient displays on medication adherence in older adults.
A total of 16 older adults, who took at least three medications and had mild cognitive impairment, participated in the study. We conducted a 12-week feasibility study in which we used a mixed methods approach to collect qualitative and quantitative evidence. The study included baseline, intervention, and postintervention phases. Half of the participants were randomly allocated to the treatment group (n=8), and the other half was assigned to the control group (n=8). During the study phases, research assistants measured medication adherence weekly through the pill counting technique.
The treatment group improved their adherence behavior from 80.9% at baseline to 95.97% using the MAD in the intervention phase. This decreased to 76.71% in the postintervention phase when the MAD was no longer being used. Using a one-way repeated measures analysis of variance and a post hoc analysis using the Tukey honestly significant difference test, we identified a significant statistical difference between the preintervention and intervention phases (
The findings of this study demonstrate that using ambient modalities for implementing external cues is useful for drawing the attention of older adults to remind them to take medications and to provide immediate awareness on adherence behavior.
ClinicalTrials.gov NCT04289246; https://tinyurl.com/ufjcz97
One of the most common reasons for medication nonadherence among older adults is forgetfulness [
An ambient display unobtrusively presents relevant information unless it requires the users’ attention [
In the last decade, different technological-based interventions for supporting the medication adherence of older adults have been studied. Owing to the rapid penetration of mobile phones, one of the approaches that has been widely studied is text message (SMS) reminders [
Some research has been conducted to explore new computing approaches, such as ambient computing technologies, mobile games, and conversational agents specifically designed for older adults [
Our study aimed to address the following research questions (RQ):
RQ1: What is the effect of the external cues provided by the MAD on older adults’ medication adherence?
RQ2: How do the MAD design features promote its adoption?
We used a mixed methods approach to obtain quantitative evidence regarding how several variables associated with adherence to medication improved and qualitative evidence regarding the adoption of the system by the participants. To describe our study, we used the suggestions by the Consolidated Standards of Reporting Trials to report pilot investigations [
The evaluation of the MAD was designed as a small trial study. This section presents the study timeline, the activities conducted, and the instruments used to collect data (see
Study activities and instruments administered to participants during each study phase. CG: control group; MAQ-8: 8-item Medication Adherence Questionnaire; MedMaIDE: Medication Management Instrument for Deficiencies in the Elderly; SPMSQ: Short Portable Mental Status Questionnaire; TG: treatment group.
To be eligible, older adults had to meet the following criteria: be older than 60 years, take at least three medications prescribed by a physician (ie, polypharmacy), have mild cognitive impairment, report medication-forgetting events, and live with a relative who could provide us with information on the assistance required by the study participant to take their medications. The exclusion criteria were as follows: being unable to self-administer medications due to a functionality problem or severe cognitive impairment, and not taking pill-based medications (it may be difficult to assess adherence otherwise). To participate in the study, it was not a requirement that older adults have experience in the use of the internet or mobile devices.
The study was conducted in Mexicali, Mexico. Ten students from the Faculty of Nursing at the Universidad Autónoma de Baja California participated as research assistants. These students were enrolled in a social service program at the Community Center of the University (known as the UNICOM), which aims to provide seniors with occupational therapy and provide some health care assistance. The UNICOM is strategically located in a neighborhood where aging inhabitants predominate. For recruiting participants, research assistants contacted older adults in the vicinity of the UNICOM and administered a set of instruments as summarized in
Instruments used to assess the eligibility criteria.
Eligibility criteria | Instrument | Score to be eligible |
Mild cognitive impairment | Short Portable Mental State Questionnaire [ |
3 or 4 points |
Medication deficiency | Medication Management Instrument for Deficiencies in the Elderly [ |
<13 points |
Adherence for medicating | 8-item Medication Adherence Questionnaire, also known as Morisky scale [ |
1-2 points=low and 3-8 points=medium |
Caregiver involvement | Semistructured interview to find out how caregivers assisted older adults | —a |
aNot applicable.
Once the participants were recruited, we realized that they were primarily of low socioeconomic status and affiliated with the Mexican Institute of Social Security (IMSS), the largest medical institution in Mexico. Periodically (monthly or bimonthly), they attended an IMSS clinic for follow-up consultation and to retrieve an updated prescription to get their medications from the clinic’s pharmacy. The lack of adequate health care and pharmaceutical policies to rationally manage medications and monitor the treatment of patients increases the vulnerability of Mexican seniors to medication errors [
Baseline data were collected during weeks 6 to 10 on medication adherence by using the pill counting technique. We noticed that participants accumulated containers with the same medications. Under those circumstances, we provided seniors with a basket to arrange the medications that should be taken each week (see
A research assistant arranging medications in a basket.
Outcome variables and methods used to collect data to address research question 1.
Variable | Description | Collection method |
|
The number of pills taken by participants in a period divided by the number of pills expected to be taken for that perioda | Pill counting |
|
The number of medication episodes reported as taken by participants in a period divided by the number of episodes expected to be recorded for that period | MAD’sc log |
Timelyd | Indicates whether the medication was taken 30 min before or after the time expected to take the medication. This is the number of medication episodes registered in the time window during a period divided by the number of episodes registered as taken for that period | MAD’s log |
Self-reported medication adherencee | A score estimated based on reported nonadherent behaviors; for example, drug omissions, medication forgetting, carelessness, or stopping a medication when feeling worse [ |
8-item Medication Adherence Questionnaire |
aIt measures whether the medication is not being taken as prescribed, which may affect the clinical outcome. We estimated it for both groups (treatment group and control group).
bIt enabled us to understand how much older adults used the MAD reminders. It was estimated for the treatment group.
cMAD: Medication Ambient Display.
dIt measures whether doses were taken during the prescribed interval. It was estimated for the treatment group.
eIt is an assessment instrument to identify individuals’ perception about their medication adherence. The 8-item Medication Adherence Questionnaire was administered during the recruitment phase and at the end of the intervention phase.
For medical information systems to be more specific to the needs of users, an iterative approach must be followed, which consists of different usability studies [
We implemented the MAD for Android tablets to be placed as portrait frames in the older adults’ homes and to provide the following cues:
Abstract and stylized representations of their medication adherence
Auditory and visual reminders to call older adults’ attention
Events that may enhance older adults’ awareness about whether the medication was taken
The MAD shows a virtual birdcage, which has the aim of raising elders’ consciousness about how they have to take responsibility for caring for their health, in a way similar to willingly caring for a pet. As presented in
Abstract representations of medication adherence based on parakeet growth.
Detailed information about the medication adherence corresponding to the current day and notation used to represent if medicines were taken on time.
The parakeet provides auditory reminders (ie, parakeet whistle) and pictograms that inform how to take medications. For instance,
Medication Ambient Display reminding to medicate.
These cues refer to actions performed by older adults to make taking medications more memorable [
We also implemented an administration component (see
Registering the medication: After a senior medicates, she/he should move the corresponding medication container closer to the tablet in order for the attached NFC tag can be recognized by the tablet NFC reader (left); then, MAD acknowledges that the medicine was registered as taken on time (right).
User interfaces of the administration component of Medication Ambient Display, which shows how it enabled the information registration of each medication.
We conducted a session with the research assistants, who made a random and blind allocation of the participants to the treatment group (TG) and the control group (CG). On the first day of this phase, research assistants visited older adults in the TG to introduce the MAD in the presence of caregivers by using the
After the intervention was completed, we removed the MAD from the participants’ homes. Research assistants then carried out weekly visits (weeks 16-17) to older adults from the TG to collect data that enabled us to understand how the withdrawal of the MAD affected their medication routine and adherence.
Medication compliance (known as adherence as well) refers to “the act of conforming to the recommendations made by the provider concerning timing, dosage, and frequency of medication-taking” [
During the intervention phase, we also collected qualitative evidence about the system’s adoption, which enabled us to address RQ2. We interviewed the older adults in the TG regarding the system’s functionalities that they perceived as most useful, less useful, and the difficulties faced while using it. At the end of the postintervention stage, we interviewed participants to obtain their perceptions of how withdrawal from the MAD impacted their medication adherence. In addition, those caregivers who were at home during our visits were interviewed to obtain information on their involvement in the seniors’ medication activities. Our questions centered on the specific activities associated with the older adult’s medication regimen that caregivers were involved in and how they knew if the older adult took his or her pills in a given week. These semistructured interviews were administered by the first 3 authors of this paper.
We used Student
For the qualitative analysis, we transcribed the collected data from their original Spanish version, that is, audio, handwritten notes, and photographs taken during the interviews. Individual quotes were translated into English for use in this paper. We followed the thematic analysis approach, which consists of generating initial codes from the data, searching for potential themes, contrasting the identified themes with the data, and iteratively refining them [
The ethics review board of Faculty of Nursing approved the study protocol once we proposed how to address their suggestions on how to handle the withdrawal of the technology at the study end. We agreed to provide the participants of the TG with an adequate financial incentive that would allow them (if desired) to obtain a PC tablet similar to the one used during the study. Every week, participants received an economic incentive, approximately US $7 if they were in the CG and US $14 if they were in the TG. We obtained informed written consent from all individual participants.
The research assistants contacted approximately 100 older adults to participate in the study (see
Flow diagram of the participants’ progress through the study phases. CG: control group; MCI: mild cognitive impairment; TG: treatment group.
Characteristics of the participants.
Characteristic | Group | Statistics | ||||
|
Control | Treatment | Chi-square test ( |
|||
Age (years), mean (SD) | 73.5 (8.3) | 68.62 (6.2) | 1.32 (15) | —a | .21 | |
Education (years), mean (SD) | 5.25 (3.8) | 6.75 (2.1) | 0.97 (15) | — | .35 | |
Number of medications, mean (SD) | 5.75 (1.8) | 4.88 (1.6) | 1.01 (15) | — | .33 | |
Cognition | Mildb | Mildb | — | — | — | |
|
1.3 (1) | .25 | ||||
|
Female | 5 | 7 | — |
|
|
|
Male | 3 | 1 | — |
|
|
|
1.1 (1) | .57 | ||||
|
Spouse | 3 | 3 | — |
|
|
|
Child | 4 | 5 | — |
|
|
|
Other | 1 | 0 | — |
|
|
|
0.2 (1) | .59 | ||||
|
Low | 2 | 3 | — |
|
|
|
Middle | 6 | 5 | — |
|
|
|
0.30 (15) | .77 | ||||
|
Baseline | 79.87 (17.9) | 80.9 (16) | — |
|
|
aNot applicable.
bCognitive impairment was assessed with the Short Portable Mental Status Questionnaire because it is appropriate for low-literacy persons [
Although underadherence was the most predominant behavior, overadherence was also found during the stages of different studies, for both the TG and the CG (see
The TG improved their adherence behavior (dosage), increasing from 80.9% in the preintervention phase to 95.97% in the intervention phase. However, it decreased to 76.71% in the postintervention phase. Using a one-way repeated measures ANOVA, we compared the effect of the MAD in the TG during the 3 phases. It showed a significant statistical difference between at least two of the phases (
The CG increased their medication adherence between the preintervention (mean 79.87% [SD 20.49%]) and intervention phases (mean 88.18% [SD 20.49%]). However, according to a paired-samples
The medication adherence rate of the TG (mean 95.97% [SD 6.08%]) was higher than that of the CG (mean 88.18% [SD 13.06%]) during the intervention phase. However, according to a independent samples
A one-way ANOVA was conducted to compare the effect of
We obtained 2224 medication episodes registered in the MAD’s log, of which 93.17% (2072/2224) were registered as taken on time, and 88.35% (1830/2224) of these timely episodes were taken after the MAD reminders. As illustrated in
Timely and reminder dependency rates estimated by participants.
Participants in the treatment group | Medication episodes rate (%) | Timely episodes rate (%) | Reminder dependency rate (%) |
P1 | 86.46 | 87.37 | 93.64 |
P2 | 100.67 | 92 | 94.2 |
P3 | 93.83 | 94.92 | 100 |
P4 | 100 | 92.44 | 32.19 |
P5 | 94.71 | 97.21 | 98.56 |
P6 | 99.23 | 89.92 | 93.97 |
P7 | 98.56 | 98.06 | 99.5 |
P8 | 100.43 | 93.16 | 98.62 |
We found that MAQ-8 scores and
The collected qualitative evidence helps us to address RQ2 and to complement the quantitative results. We analyzed the data collected from the interviews administered to participants P1 to P8 of the TG, and we discovered the findings described in the following subsections.
All the older adults provided answers that lead us to conclude that the most useful cues were the auditory reminders, followed by the stylized representation of medication adherence (eg, parakeet growth). Some participants explained how these external cues caught their attention so that they can medicate themselves. Participant P6 said:
Sometimes I am busy or just thinking about something else, and I forgot what I have to do; but now when the parakeet sings and sings, it reminds me to medicate.
This finding is supported by the high rates in timely and reminder dependency measures. For instance, participant P1 improved her medication-taking behavior because she developed a high dependency on the ambient reminders (see
When I heard the parakeet whistle, I came to the kitchen to take my medicines...It was better when I had the system.
All older adults, except participant P4, reported that they did not consult the detailed information on their medication adherence, but they verified if the parakeet grew after registering the medication as taken. Thus, participant P4 was the only older adult who did not develop a high reminder dependency rate (see
Although adverse drug events associated with overmedication and undermedication were identified, older adults hardly recognized that before using the MAD; there were times when they might have forgotten to take a medication. For instance, although participant P4 explicitly stated that she did not forget to take her pills, her husband contradicted her. Participant P6 was the only participant who, during the interviews, admitted forgetting to take her medication because as a consequence, she had symptoms of her disease:
[Before using the MAD] I realized that I had forgotten to take the [night doses] pill for my blood pressure, until the next morning my head hurt and I felt dizzy.
On the other hand, we have no evidence that older adults were aware that they sometimes overmedicate because they forgot that they had taken their medication.
Older adults intend to take their medications when carrying out some of their daily routines. We identified that some older adults fail to associate medication with their daily activities effectively. For instance, participant P3 indicated:
I plan to take my night medication before going to sleep, but before using MAD, sometimes I realized that I had forgotten to medicate until I was in bed, and murmured: “Ay! I have not taken the pill.” However, now, it [MAD] sounds, I take my medication, and then, I go to sleep.
On the other hand, we found that leaving home or traveling affected the use of the MAD and probably the medication adherence of participants. For instance, several participants (P2, P3, and P6) reported situations in which they could not register a medication intake timely because they had to leave home. For instance, participant P6 reported:
I did not take my medication this Wednesday since I had to leave the house urgently, but I took it later.
Also, only 1 participant (P4) felt confident enough to take the system with her when she left home.
The family caregivers commonly served as the main support actor and assisted with managing the seniors’ medications. Before using the MAD, caregivers tended to be aware of the medication timetables to remind older adults or ask older adults if they had medicated. We observed that the external cues of the MAD acted as triggers that facilitated caregivers’ assistance. That is, the cues did not overwhelm family members but were an appropriate mediation strategy to support seniors’ medication routines. For instance, the husband of participant P4 perceived that the MAD enabled him to be aware when she took the correct medications. For an adolescent caregiver, the system enabled him to feel less worried about having to remind her grandmother (participant P1) to medicate:
If I have to do my homework, I can focus on doing it.
The MAD also helped to assure caregivers that older adults would not forget to take their medications; for example, participant P6 said:
My children used to forget reminding me to medicate [before using the MAD]. Currently, they hear the parakeet, and then they make sure if I took them.
Our quantitative results show that providing the external cues supported by the MAD resulted in significant improvements in the average rates of dosage outcomes for older adults. This is because the ambient modalities used for implementing these external cues were useful for drawing the attention of older adults. We found that external cues (1) reminded them to take medications, (2) enabled them to recognize if a medication was recorded as taken, and (3) provided immediate awareness about how they followed their medication regimens.
We learned that providing older adults with an abstract and stylized representation of their medication adherence, which could be peripherally perceived, was better accepted than medication adherence reports that need to be explicitly evoked. However, this stylized and abstract modality of representation was not enough to make participants aware of their medication problems related to undermedication and overmedication. Previous research has demonstrated that feedback-based systems that are consulted explicitly and daily help seniors identify their medication errors and then self-regulate their medication behavior [
Similar to our results, other studies have shown that when older adults stop using medication aiding systems, their medication adherence is affected [
Furthermore, we identified that external cues of the MAD provided caregivers with better awareness of older adults’ medication adherence. This awareness was 2-fold: (1) when auditory reminders were perceived by caregivers, they made sure that the reminders reached the target recipients, and (2) medication adherence representations enabled caregivers to be aware of medications taken. Therefore, providing external cues through ambient displays helps family caregivers to better support seniors to follow their medication regimens.
The use of financial incentives has been questioned because they may provide inducements to participate in a study for financial purposes only, and vulnerable populations are prone to be enticed by the financial reward and be more willing to accept any study risks [
We observed that some participants faced problems in managing their medications, such as accumulating medications, confusing medications because they look alike, and tending to give medications to others, which may not be an appropriate practice. The design of our study was limited in that the MAD system was personalized according to the prescribed medication regimens and timetables that the participants followed. Although using our technology did not introduce any risk, it might have supported inappropriate medication routines adopted by older adults to overcome some of the barriers imposed by the setting. We recognize the importance of conducting a contextual study before conducting a technology evaluation. The contextual study should be designed in collaboration with clinical or nursing specialists to reduce the complexity of older adults’ medication regimens and the risks associated with the way they manage medications. Finally, we are not able to state that our results are generalizable to the whole Mexican elderly population. This is because participants were primarily from a low-income socioeconomic stratum; therefore, exploring this technology among high-income elders who have access to private health care services might produce different findings.
The pill counting technique can be prone to human error. So, one limitation of this study is that we were not able to identify which overmedication and undermedication events were registered by the research assistants erroneously. Using electronic monitoring devices (EMDs), such as the Medication Event Monitoring System, could have overcome this limitation to some extent, although using EMDs does not guarantee that a person has taken their medication [
Another limitation of the study is its duration time. Some research works have identified that older adults should have an adaptation period to an intervention, and after a specific period of using it, reliable data can be collected to measure its effectiveness for improving medication compliance [
Overview of some studies to assess different technological approaches to support older adults’ medication adherence.
Data extracted from the studies | Morawski et al [ |
Mertens et al [ |
Robiner et al [ |
Grindrod et al [ |
Park et al [ |
Lee and Dey [ |
Perera et al [ |
Patel et al [ |
Reeder et al [ |
De Oliveira et al [ |
|
|
|||||||||||
|
Mobile phone apps | Xa | —b | — | — | — | — | X | X | — | X |
|
SMS | — | — | — | — | X | — | — | — | — | — |
|
Tablet | — | X | — | X | — | X | — | — | — | — |
|
Dispenser | — | — | — | — | — | — | — | — | X | — |
|
Monitoring | — | — | X | — | — | X | — | — | — | — |
|
|||||||||||
|
Remind | X | X | — | X | — | — | X | X | X | — |
|
Register taken doses | X | X | — | — | — | — | X | X | X | X |
|
Educate | — | — | — | — | X | — | X | X | — | — |
|
Feedback | X | X | X | — | — | X | — | — | — | — |
|
Caregiver participation | X | X | X | — | — | — | — | — | X | — |
|
Social game | — | — | — | — | — | — | — | — | — | X |
|
|||||||||||
|
Total participants (older adults) | 413 | 24 | 6 | 35 | 90 | 12 | 28 | 48 | 96 | 16 |
|
Medicines | 2Hipertension | +3cardvascular desease | 1renal | +1several | 2cardiovascular desease | N/Sc | 3HIV | 3Hipertension | +11several | +1serveral |
|
Methods to measure adherence: subjective and objective | Subjective and objective | Subjective and objective | Objective | — | Subjective and objective | Subjective and objective | Subjective and objective | Subjective and objective | — | — |
|
Acceptance | — | X | X | X | X | — | X | X | X | X |
aX: studies that assessed the acceptance of the system by the participants.
bNot applicable.
cN/S: nonsignificant.
The external cues provided by our ambient display not only improved the medication adherence of the elderly but also encouraged caregiver involvement. We found that the external cues perceived as most useful were those that reminded participants to take medications, helped seniors recognize if medications were recorded as taken, and provided immediate and abstract representations of their medication adherence. We identified that external cues did not overwhelm family members but were an appropriate mediation strategy to support older adults’ medication routines. We also recognized the potential of providing external cues to enable older adults to associate medication routines with their daily routines appropriately. For future work, we plan to conduct studies to assess the feasibility of external ambient cues to support the seamless integration of medication regimens into the daily routines of the elderly.
Individual dosage rates estimated through the pill counting technique (Dosagepill) and the medication episodes recorded in MAD’s log (DosageMAD). MAD: Medication Ambient Display.
CONSORT-EHEALTH checklist (V 1.6.1).
analysis of variance
control group
National Council of Science and Technology
electronic monitoring device
Mexican Institute of Social Security
Medication Ambient Display
8-item Medication Adherence Questionnaire
Medication Management Instrument for Deficiencies in the Elderly
Near Field Communication
research question
treatment group
Community Center of the University
The National Council of Science and Technology (CONACyT) in Mexico and the Autonomous University of Baja California supported this work under grant numbers 153863 and 1914, respectively. The authors want to thank CONACyT for the scholarships provided to the first and seventh authors. The authors thank the students from the Faculty of Nursing who helped with participant recruitment and data collection, and professors Betzabé Arizona, Rosa Esparza, and José Aguero for their support with conducting the study.
None declared.
This randomized study was only retrospectively registered, explained by authors as: “our paper presents a small study to provide evidence about the feasibility to be accepted and its efficacy to improve the medication adherence of older adults.” The editor granted an exception from ICMJE rules mandating prospective registration of randomized trials because the risk of bias appears low and the study was considered formative, guiding the development of the application. However, readers are advised to carefully assess the validity of any potential explicit or implicit claims related to primary outcomes or effectiveness, as retrospective registration does not prevent authors from changing their outcome measures retrospectively.