A Multifaceted Prospective Memory Intervention to Improve Medication Adherence: Design of a Randomized Control Trial

Introduction

Fifty percent of older adults do not adhere to antihypertensive agents. Nonadherence is associated with increased risk of major cardiovascular events and death [1] and is the single most modifiable risk reduction factor for stroke [2]. The need to control hypertension is clear and the problem is significant in numbers alone because 60 – 80% of older adults have hypertension [3]. Unfortunately, people do not fully adhere to medication regimens despite extensive interventions to educate, motivate, and support adherence behaviors [for reviews see 4, 5–8].

There is good evidence associating cognitive function and medication adherence [9, 10, for a review see 11]. According to Stone, the most common reasons for non-adherence are forgetfulness, changing medication schedules or busy lifestyles. Each of these reasons is potentially cognitive in nature. Specifically, forgetfulness is failing to remember to take medications at the designated times, changing schedules is likely to interfere with memory strategies that are based on routine and on associated environmental cues in one’s routine context, and busy lifestyles contribute to nonadherence by adding cognitive distraction that challenge limited working memory capacity. Adherence may especially depend on executive function and working memory [9] because past research demonstrates only small effects of interventions that address education [e.g. 12] or the design of medication instructions (and therefore patient understanding) [see for example 13]. Therefore, adherence may be improved by interventions that target executive function and working memory.

Assuming good motivation to faithfully follow one’s medication regimen, successful adherence to a medication schedule requires a complex set of cognitive processes. Over the past 20–25 years, there has been rapid growth in research on prospective memory resulting in better understanding of the component processes involved in remembering to perform actions in the future [14, 15]. Specifically, successful prospective remembering involves good encoding of the task demands, retaining these demands in memory, and retrieving them at the appropriate moment in time. It also involves initiating the execution of the action, often in the face of distraction, and then monitoring one’s action so that a dose is not mistakenly repeated. Given that successful medication adherence involves many processes and that each is vulnerable to failure, it is not surprising that adherence is often less than optimal and that proper adherence is related to levels of cognitive functioning [10, 16].

As we age, we are much more likely to develop chronic conditions that are treated with medications and yet we also experience changes in cognitive functioning that can undermine successful adherence to prescribed medication regimens. Research on aging and memory shows that aging causes general declines on a variety of cognitive functions, including processing speed, attention, working memory capacity, the ability to learn new information, and the ability to retrieve information [17–19]. These age related cognitive declines are robust [e.g., 20, 21]. Given the multiple cognitive processes described above that are involved in remembering to follow a medication regimen, it is not surprising that compromised cognitive functioning significantly impairs medication adherence [10]. The purpose of this article is to describe an in-home, tailored, multifaceted intervention that specifically targets remembering strategies for improving adherence to antihypertensive medication.

Overview of the Intervention

There are two results in the prospective memory and aging literatures that suggest that an intervention strategy that targets memory processes will be successful. One is that there is remarkable variability in age effects on laboratory tests of prospective memory, with some studies showing minimal or no age differences [22–25] and others showing striking age differences [e.g., 26, 27, 28]. Another is that there is evidence that aging has more pronounced disruptive effects on cognitive functioning mediated by the frontal lobes (e.g., working memory and executive attention [29–31] than on cognitive functioning mediated by the medial temporal areas (e.g., relatively automatic associative retrieval processes). Consistent with recent reviews of the aging and prospective memory literature [18, 25, 31], the emerging pattern is that older adults show substantial deficits when they rely on working memory and executive resources for prospective remembering but minimal deficits when they rely on mostly preserved and relatively automatic associative retrieval processes. For example, age differences are robust on prospective memory tasks that require strategic monitoring for the correct opportunity for performing an intended action or when the retrieved action must be delayed (thus placing a demand on working memory and attentional resources for keeping the retrieved intention activated over the delay). In contrast, age differences are minimal when there are good external or environmental cues that support relatively automatic retrieval of the intended action (cues that have been associated with intended actions) and when the action can be performed immediately. Thus, the general focus of our intervention is to encourage older adults’ use of environmentally supported associative retrieval processes (thought to be relatively spared with age) in place of prospective memory processes that heavily rely on working memory and executive resources (thought to be compromised with age).

The intervention targets all of the memory and attentional processes thought to be critical for successful prospective memory performance in everyday tasks such as medication taking [15, 32]. These include (1) forming a good encoding of the intended action and the condition(s) that is appropriate for initiating the action, (2) remembering the intention over the retention interval, (3) retrieving the intention at the appropriate point in time, (4) inhibiting other ongoing activities at the critical time and actually executing the action, and (5) monitoring performance of the action so that the person remembers performing it (and therefore does not repeat it). Failure at any one of these tasks could compromise medication adherence. For example, a person may form only a general intention to take her or his medicine in the morning and may fail to think of it at the appropriate time. Or, that person may retrieve the intention to take the medication at the right time but because of distractions or interruptions at that time may fail to follow through and execute the action. Or, after establishing a routine, the person may automatically take a dose while deeply engaged in other thoughts and a few minutes later forget that he or she had already taken the medication and take it again resulting in taking too much medication. Consequently, the intervention directly targets each of these tasks.

The following general principles guided the development of the intervention. First, we used techniques that have been empirically demonstrated to improve medication adherence and/or prospective memory and to be effective with older adults. Second, consistent with the view outlined above that medication adherence is a complex process [8], we developed a comprehensive approach to adherence. Specifically, we developed a multifaceted intervention that targets all of the cognitive processes thought to be involved in remembering to take medication. As illustrated above, the idea is that problems in any one component process can undermine successful performance and thus it is critical to support all of the component processes. Third, because successful performance of any one process is probabilistic in nature, the intervention included redundancy so that there were backup routes for accomplishing each process. The intervention especially focuses on retrieval of the intention, which according to some researchers is particularly vulnerable in older adults [27], and each process of prospective remembering (encoding, retention, retrieval, execution, and monitoring) is targeted by at least two strategies in the intervention. Fourth, as developed earlier, the intervention capitalizes on abilities that are relatively preserved with age and reduces reliance on age-vulnerable abilities. In particular, it relies extensively on relatively well-preserved associative processes that respond more or less automatically to stimuli that have been associated with the intention (e.g., on cuing from environmental stimuli), while reducing reliance on working memory and executive function (e.g., self-initiated monitoring), which prior research has shown are susceptible to the effects of aging [30, 33]. Consistent with this, Henry and colleagues’ [34] meta-analysis showed significantly smaller effects of age on prospective memory tasks supported by relatively automatic processes, r=−.14, compared to prospective memory tasks demanding active monitoring processes, r=−.40; [18, 25, 31]. For example, if people think that they should take their medication at 8:00 am, then they need to actively monitor the clock in order to determine whether the time is appropriate for taking their pill. According to current cognitive theories, actively monitoring the clock and inhibiting ongoing activities in order to actually check the clock require extensive working memory and executive resources [35, 36]. By getting people to reinterpret their task as one of taking medication with breakfast and then getting them to create an associative link between the external signal of breakfast (the cue or stimulus) and taking medication (the intended action), it is likely that the occurrence of breakfast will trigger retrieval of the intended action [see 37]. Thus, through this simple reinterpretation of the task, prospective memory retrieval is accomplished through associative processes that are thought to be spared with age as opposed to working memory and executive processes that have been shown to decline with age.

Our final principle in designing our intervention was based on Bargh and Chartrand’s [38] theory that humans have a limited capacity for conscious control over behavior, that conscious self-regulation is a limited capacity that is quickly depleted [39], and that many, if not most, of our behaviors are unconsciously initiated by stimuli in our environment. For example Holland, Hendriks and Aarts have shown that introducing the weak scent of a citrus cleaner into a room leads subjects to unconsciously engage in more cleaning behaviors [40]. Translated into a medication adherence intervention, this means that the average person is unlikely to consistently use cumbersome capacity-consuming processes like monitoring. Evidence indicates that even memory psychologists rarely use burdensome mnemonic devices in retrospective memory situations, despite their clearly demonstrated effectiveness [41]. Thus, whenever possible, the fifth principle was to reduce reliance on consciously demanding processes. As examples, we encouraged intervention participants to use relatively automatic processes that are set in motion in response to stimuli and also to perform an intended action as soon as they thought about it (thereby obviating the need to use executive and working memory resources to monitor for critical events over delays).

The strategies to be implemented, as well as the processes that they were designed to affect, are outlined in Table 1. We ask participants to link medication taking to their daily routines and we work with them to do this. Emphasizing routine and linking medication taking to this routine facilitates prospective retrieval because the routine promotes the association of medication taking with existing patterns and bypasses the more vulnerable cognitive processes of executive function and working memory. We work with participants to help them identify good cues within their daily practices in their homes that function as triggers for performing the intended action. A good cue is one that is easily noticed and will trigger the retrieval of the intention to take medications, for example, linking medication taking with brushing teeth or eating a meal. Medication taking can also be linked to visual reminders. Placing the medicines in an area where they will be visible, for example, in the center of the breakfast table rather than in a drawer, is also a cue.

We ask participants to use events rather than time as a cue or trigger to take the medication as intended. Changing a time-based task (take medication at 7:00 a.m.), to an event based task (take medication when you set the breakfast plate on the table, or make coffee, or turn on a particular television program) provides associative cues and facilitates retrieval [35].

We ask participants to elaborate the action of taking medicines to make the event more salient and distinct, so that they are more likely to remember having taken it [42]. We explain that doing something to make the action of taking the medicine more memorable, for example, shaking the bottle of pills before taking the pill, will prevent the repetition of a dose. One concern about having a routine is that it may make the monitoring of the completion of an activity more difficult. As the event becomes routine it can also become less salient which can lead to confusion such as the following: “did I take the medicine or do I just think I took it because I’ve been taking the same medication every day for the past two years?” Shaking the bottle makes the event more complex, involves more conscious awareness and creates more cues for monitoring whether the medication was taken as intended.

We also encourage participants to “Do it now” because even a short delay, as brief as 5 seconds, can cause forgetting to take the medication as intended [43]. This strategy is important for supporting execution of the intended action.

We ask participants to use the provided organizer to facilitate monitoring of whether the medication was taken as intended. We give participants a box with 7 sections. Each section labeled with the first letter of each day of the week that accommodates the Medication Electronic Monitoring System (MEMS) cap. We give one to everyone who was using a medication organizer prior to the study and to everyone in the intervention group. We provide two boxes: one labeled A.M. and one P.M. in the event the monitored medication is to be taken twice a day. After taking their medications, we ask participants to move the MEMS bottle to the next slot in order to help them remember that they have already taken their medication at the current time.

Prospective memory is also supported by having participants create implementation intentions, which involve imagining exactly when and where they will perform an intended action. The principle is “when X arises, I’m going to perform Y” and it needs to be tailored so that the participant can imagine doing it in the context of their daily routine [37, 44]. For example, instead of forming the general intention to take medication in the morning, participants might be encouraged to form the implementation, “When I have my orange juice with breakfast in the morning, I will take my medications.” As part of the intervention, we also encourage participants to imagine performing the action in that particular context, and we walk through (rehearse) the planned medication taking event with the participant in that context.

We also use spaced retrieval, (teach, ask, wait, ask again, wait ask again) in order to support encoding, retention, and retrieval processes involved in medication taking. We ask participant to recall the information we have just presented. Then we wait one minute and ask participant to recall the information again and importantly how they are going to enact the medication taking event, e.g. where, when, how and imagine doing it. Then we ask again in 15 minutes. When the session ends, we ask participants once more to recall the information. The benefit of having the participants actively retrieve the intention is that research has shown that spaced retrieval improves memory for the intention [45].

Finally, we ask participants to practice with us in their home setting. For example, we have participants practice walking to the kitchen to fill the orange juice glass and take the medication. This final step helps participants consolidate their strategies and provides another opportunity to address any remaining questions.

The Education Control Condition

We are using an education comparison group to control for the possible benefit of developing a relationship with a health care provider and education on adherence. Using the education comparison as the control group acknowledges evidence of the value of relationships with health care providers to improve adherence [46, 47] and the role of education to improve adherence (the necessary components of encoding and retention are involved in the education portion of the intervention), although education-only interventions provide only small effects [for a review of medication adherence interventions see 8, 48]. Nurses assigned to each group are trained in the principles of developing a relationship and encouraged to establish rapport with participants. Participants in both groups receive education on hypertension and antihypertensives. The education portion was developed from internet materials from the American Heart Association and the National Heart Lung and Blood Institute [49, 50] that were designed for the general public and addressed issues of hypertension and types of antihypertensives. Both groups are given information on the importance of controlling high blood pressure to prevent stroke and other end organ damage. Nurses for each group are in the home the same amount of time. Once the education comparison group receives education, the nurse assigned to this group discusses prevention of falls (for example, the benefit of eliminating scatter rugs, use of hand rails on stairs) and engages in social interaction. Therefore, the only difference between the two groups is the use of the prospective memory strategies in the intervention group. Using the education control comparison group provides a strong test of the effectiveness of the prospective memory intervention.

Research Design and Methods

The main hypothesis to be tested in this intervention study is whether individuals in the multifaceted prospective memory intervention group are more adherent to antihypertensive medications than those in the education comparison group both immediately following the intervention and over the six months of testing. By targeting memory and attentional processes, we anticipate that the intervention will improve self-management of medications. Moreover, because the intervention is designed to reduce reliance on executive and working memory resources and to increase reliance on associative processes spared by aging, we further hypothesize that individual differences in working memory capacity will be associated with medication adherence in the control condition, but not in the intervention condition.

Discussion

Self-management and specifically adhering to antihypertensive medications remains an important concern because adherence tends to be low for this asymptomatic disease, and uncontrolled or undercontrolled hypertension is associated with poor health outcomes including white matter hyperintensities and stroke [77, 78]. Based on research and theories in the areas of prospective memory and aging, the key and novel features of this described intervention are that it addressed all components of prospective remembering in multiple ways and in an integrated and comprehensive fashion. It also focused on strategies that rely on processes that are relatively well preserved in older adults. The intervention has the potential to address the significant problem of nonadherence to prescribed medication with the goal of improving health outcomes among older adults.