Comparison of Medication Adherence Assessment Tools in Patients: A Review

Medication adherence is the degree to which a person’s actual behavior of taking medication matches the recommendations agreed upon with a healthcare provider ^[2]. It is a multidimensional process. A patient might fail to start treatment, might not carry out the prescribed regimen properly, or might stop the treatment before the intended time. As a result, assessing adherence is more complicated than just finding out whether or not a patient takes their medicines ^[4,5].

Non-adherence to medication is common and can lead to poor control of the disease, to complications that are preventable and to greater use of healthcare services. It is therefore important to carry out accurate assessments in order to identify those patients who need counselling, having their regimen simplified, receiving adherence–support interventions or requiring further clinical assessment ^[1,2].

A key difficulty is that there is no single adherence measure that works well in all situations. Although objective methods can offer better evidence of actual medication-taking behavior, subjective methods

can give information on the purpose of the assessment and, where possible, complementary methods should be combined ^[1,3].

The systematic review by Nguyen et al. identified 43 self-report adherence scales across 60 articles and showed that these instruments do not all measure the same construct. Some primarily measure medication-taking behavior, whereas others assess barriers, beliefs, self-efficacy, or combinations of these domains ^[2]. This distinction is central to interpreting adherence scores and is the main focus of the present review.

2. WHY COMPARISON OF ADHERENCE TOOLS IS IMPORTANT

 For the same patient, different adherence tools may lead to different classifications since these tools examine different aspect of adherence, employ different response formats, and have been validated on different groups of people. A brief behavioral scale might be appropriate for use in screening, whereas a more multidimensional instrument would be more useful if the aim is to find out why a patient is non-adherent. Also, a belief instrument should not be automatically regarded as a direct measure of medication-taking behavior ^[2,4].

Measure the following:

• Medication-taking behavior, barriers, beliefs, self-efficacy, refill behavior or any combination thereof.
• The length of the study and the time it takes to administer it are important in busy clinical settings and in large studies.
• Some of the tools are broad in scope while others have had their effectiveness validated specifically for diseases.
• The way in which literacy and language are expressed and administered affects usability.
• With regard to scoring and interpretation, certain instruments include categories or cut-off points, whereas others are mainly used for describing levels or domains.
• The evidence from psychometrics indicates that reliability, validity, and responsiveness should be taken into account rather than assuming that a tool which is widely used is necessarily the best one ^[3,4,5].

3. CLASSIFICATION OF MEDICATION ADHERENCE MEASUREMENT METHODS

TABLE 1: CLASSIFICATION OF MEDICATION ADHERENCE MEASUREMENT METHODS

Lam and Fresco categorised adherence measurement into subjective and objective approaches and pointed out that each has advantages and disadvantages. They concluded that there is no perfect measure and a multi-measure approach is often preferable ^[1].

4. CENTRAL COMPARISON OF MAJOR MEDICATION ADHERENCE TOOLS

The following comparison is the heart of this review. The tools differ not only in number of items but, more importantly, in what they actually attempt to measure. When choosing and interpreting an instrument, the distinction between direct adherence behaviour and explanatory constructs such as beliefs or self-efficacy should be maintained ^[2,3].

TABLE 2: COMPARISON OF MAJOR MEDICATION ADHERENCE TOOLS

5. DETAILED REVIEW OF INDIVIDUAL TOOLS

5.1 MEDICATION ADHERENCE QUESTIONNAIRE (MAQ)

The Medication Adherence Questionnaire (Morisky-Green Version) is one of the shortest and most popular self-report measures. The comparison studies found that it was quick to administer and score and had good applicability to a variety of diseases ^[3,6]. Its major advantage was the ability to quickly screen patients for adherence behavior, but it was found to provide limited insight into the spectrum of adherence issues ^[3]. As a brief screening tool, the MAQ is convenient in an outpatient setting, but a low score should not be the sole guide to patient management. Further assessment of adherence determinants, such as access to drugs, adverse effects, beliefs about medication, regimen complexity, and self-confidence in regimen management, is needed to fully understand the reason for non-adherence ^[2,3].

5.2 MMAS-4

The four-item Morisky Medication Adherence Scale is a brief measure intended for screening purposes that asks the patients about behaviors indicative of non-adherence such as forgetting to take the medication, being careless and stopping the drugs under specified conditions ^[6]. The tool’s primary advantage is its brevity which makes it convenient to use in practice and research.

The main shortcoming is the amount of explanation that is required for each item which might be tiring for some patients. In addition, a patient who positively screens for this tool might need further assessment to identify the reason for medication non-adherence. This reason can be various ranging from forgetfulness, intentional non-use of drugs, adverse effects, lack of access to drugs, concern about medications or other reasons.

5.3 MMAS-8

The eight-item Morisky scale was constructed from the former MAQ. The current review points out that the first seven items are to be answered yes/no, while the last item uses a five-point Likert response. Those additional items give more insight into medication-taking behaviors, and why patients may not take their medicines ^[1]. 

The MMAS-8 gives more information than the brief four-item version but is still relatively quick to complete. It has seen its widespread use in research on patients with chronic disease. However, systematic reviews of medication adherence measures suggest that self-report tools should not be presumed to be equivalent to more objective measures of adherence. In the light of the validation evidence present for the MMAS-8 itself, and the legal and practical context in which it is applied, the use of the MMAS-8 should be carefully considered. ^[1,4,7]

5.4 BRIEF MEDICATION QUESTIONNAIRE (BRIEF BMQ)

The Brief Medication Questionnaire is most appropriate when the researcher’s aim is to evaluate medication-taking behavior in combination with barriers to adherence. The provided literature revealed that the questionnaire comprises 3 primary screens: a 5-item Regimen Screen, a 2-item Belief Screen, and a 2-item Recall Screen ^[1,8]. The Regimen Screen assess the manner in which the participant took each drug during the past week; the Belief Screen was created to identify troublesome or challenging aspects; and the Recall Screen discovers whether it was difficult to remember the regimen.

The Access screen, which refers to the inability to obtain or refill the medication, is also available in few sources. Notably, the structure of the Brief Medication Questionnaire may vary depending on the author’s preferences, which means that it is essential to specify the exact version of the instrument rather than referring to generic versions ^[8].

The key advantage of the Brief BMQ is that it is appropriate for assessing multidrug regimens, allowing the researcher to identify adherence issues on an individual drug basis. This ability is particularly valuable in cases when the objective is to differentiate between various forms of non-adherence, thus providing more information than a simple measure of behavior would. However, a limitation of the questionnaire is that a detailed review of the regimen and the completion of multiple screens can take more time than a very brief assessment ^[1,8].

It is also important to note that the Brief Medication Questionnaire should not be confused with another patient-reported outcome measure, that is the Beliefs about Medicines Questionnaire (BMQ-Specific). Although the two questionnaires share some similarities and may be mentioned in the same literature pieces, they are clearly different, and the researcher’s goal will determine the choice of the instrument.

5.5 HILL-BONE COMPLIANCE SCALE

The Hill-Bone Compliance Scale was developed mainly for hypertensive patients. The 14-item form assesses three behavioral domains: medication-taking behavior, appointment keeping and sodium intake, with items rated on a four-point response scale ^[1,8].

Its multidimensional structure is useful when adherence to antihypertensive therapy is being considered together with appointment attendance and dietary behavior. The main limitation is generalizability: because the instrument is strongly linked to hypertension care, it is less appropriate as a universal adherence tool for unrelated conditions ^[1,3].

5.6 MEDICATION ADHERENCE RATING SCALE (MARS-5)

The 5-item version of the Medication Adherence Report Scale (MARS-5) is useful for evaluating medication-taking behaviors and attitudes toward medicines and has been evaluated among patient-reported outcome measures used to measure medication adherence ^[4]. In previous comparative work, strong associations were found between MARS-5 responses and psychiatric populations, although recent work has also examined applications for other disease groups ^[3,4].

MARS-5 is useful when a short behavioral measure is required, but investigators should confirm that the version has been validated for their population and condition. It should not be assumed that psychometric performance in one disease group will automatically transfer to another.

5.7 ADHERENCE TO REFILLS AND MEDICATIONS SCALE (ARMS)

ARMS was designed to evaluate both taking medications as directed and obtaining/refilling medications. This is an important distinction because patients can be adherent when they possess medicines but fail to refill them on time, or can obtain medicines without taking them correctly. The supplied Nguyen review reports validation evidence including association with clinical outcomes, pharmacy records and other adherence measures ^[2].

ARMS is consequently attractive for chronic disease and primary-care research when refill behavior is relevant. Its longer format, compared with MAQ or MMAS-4, increases the information obtained but also increases respondent burden.

5.8 SELF-EFFICACY FOR APPROPRIATE MEDICATION USE SCALE (SEAMS)

SEAMS comprises 13 questions that explore self-efficacy which is the patient’s confidence in his or her ability to take medicines appropriately in spite of challenges. Comparative studies established SEAMS as a clinically significant adherence-related patient-reported outcome measure ^[3,4].

The critical interpretive issue is that SEAMS evaluates self-efficacy rather than the medication-taking behavior. Thus, a patient may report high self-efficacy but be nonadherent or have low self-efficacy but currently adhere to the therapy. For this reason, SEAMS is particularly informative as an adjunct tool when the research question addresses potential barriers to adherence that may be surmounted by educational, counselling or behavioral interventions.

5.9 BELIEFS ABOUT MEDICINES QUESTIONNAIRE (BMQ-SPECIFIC)

BMQ-Specific is primarily a beliefs measure, rather than a stand-alone adherence scale. The authors of the BMQ-Specific state its focus on “whether the patient regarded their medicines as necessary and whether they had any concerns about the medicines” ^[2]. Patients’ beliefs about the necessity and concerns related to the medicines they take can affect their adherence to treatment and provide clinicians with insight into patients’ intentional non-adherence to treatment.

Thus, the BMQ-Specific adherence scale is intended to work in combination with a direct measurement of adherence. The authors of the BMQ-Specific note that “a high concern score is not sufficient evidence of intentional non-adherence to treatment; however, it does provide an important therapeutic communication opportunity” ^[2]. In fact, Nguyen et al. insist that belief-focused measures should not be used as standalone adherence indicators ^[2].

6. COMPARISON ACCORDING TO WHAT THE TOOL MEASURES

TABLE 3: COMPARISON ACCORDING TO WHAT THE TOOL MEASURES

7. PSYCHOMETRIC CONSIDERATIONS: RELIABILITY AND VALIDITY

Reliability is a property of an instrument that determines its consistency, and validity is a property that determines if the instrument measures what it is supposed to measure. The provided literature review allows identifying that different instruments can demonstrate different psychometric properties and that different instruments can work differently with different populations. The conducted review by Culig and Leppée provides an example of such differences, which are demonstrated by different Cronbach’s alpha values for different scales and studies, which can inform decisions about the reliability of an instrument for different populations ^[9].

TABLE 4: RELIABILITY & VALIDITY OF TOOLS

A 2020 systematic review reported a total of 121 unique medication-adherence patient-reported outcome measures from 214 studies. Several of these instruments, including the MMAS-4, MMAS-8, MARS-5, Hill-Bone and SEAMS, had at least moderate evidence for five or more of the measurement properties, but there was no instrument that could be regarded as universally superior ^[4]. A later systematic review in cardiovascular disease and type 2 diabetes mellitus also noted substantial variability and did not classify any of the 27 assessed PROMs as having sufficiently strong evidence for any of the measurement properties to be placed in its highest category ^[7].

8. ADVANTAGES AND LIMITATIONS OF SELF-REPORT TOOLS

ADVANTAGES:

• Inexpensive, rapid, non-invasive and feasible for large studies.
• Are self-administered or interviewer administered depending on level of literacy and study design
• Are able to ascertain barriers, beliefs and behavioral patterns which are not revealed by objective measures
• Are a useful screening tool for patients who may benefit from a more intensive assessment of adherence

LIMITATIONS:

• Recall bias, social-desirability bias and intentional misreporting of non-adherence.
• Different wording, recall periods and scoring systems between instruments make direct comparison between them difficult.
• Performance may vary when the instrument is used in a different population than that in which it was validated ^[1,2,3,4].

9. OBJECTIVE AND INDIRECT MEASURES AS COMPLEMENTS

The potential objective and indirect measures of adherence can be used for self-reported adherence assessments. Firstly, pharmacy dispensing records can provide information on the number of dispensed doses surrendered by the patient whereas the pill count method determines the number of tablets or capsules by weight and electronic monitoring objectively measures the opening of the medication container. Secondly, the concentration of drugs or their metabolites in fluids is an objective measure of exposure, which typically requires significant resources to assess and can be invasive and variable due to inter-individual differences in pharmacokinetics ^[1,5].

However, most importantly, different medication adherence measures assess different constructs of adherence behavior. For example, pharmacy records determine the number of doses dispensed to the patient, electronic monitoring measures only register the number of times the medication container has been opened, self-administration evaluates individuals’ behavior and attitudes towards medication use, whereas clinical outcomes reflect the overall effect of adherence and other factors on health status. Therefore, the fact that different medication adherence measures yield different results is not surprising, as each evaluates a unique construct of medication adherence. Indeed, when measuring adherence, an individual behaves differently depending on which component of adherence is assessed by the selected method and therefore interacts with it differently.

10. MULTIMETHOD APPROACH

Because no single measure is a gold standard, a multimethod strategy can provide a more informative assessment ^[1,3]. For example, a study may combine a brief behavioral questionnaire with pharmacy refill data and a short barrier/belief assessment. This approach can answer both questions: 'Is the patient likely to be non-adherent?' and 'Why might the patient be non-adherent?'

11. SELECTION OF THE MOST APPROPRIATE TOOL

TABLE 5: SELECTION OF APPROPRIATE TOOLS