Medication Adherence and the Role of Pictograms in Medication Counselling of Chronic Patients: a Review

Pharmaceutical care requires a patient-centered approach, focusing on the ability of patients to understand drug-related information and follow the instructions delivered by pharmacists as well as other health-care providers included in the circle of care. With the goal of ensuring the prescribed use of medications, called medication adherence, health-care providers have to consider many risk factors such as geography (culture), social economic status, age, and low literacy that may predispose patients to non-adherence, and considerations have to be made for chronic patients living with life-long disease states. The aim of this review is to provide a balanced and comprehensive review outlining a number of different medication counselling and education approaches that have been used to try to improve medication adherence and health outcomes with the use of clear and concise graphic illustrations—called pictograms. By highlighting the current landscape of the general use and efficacy of pharmaceutical pictograms to aid in the knowledge and recall of drug-related information, as well as outlining specific medication adherence outcomes with pharmaceutical pictograms in chronic patients, the current review describes the need for health-care providers to move beyond the traditional didactic methods of oral and verbal communication with patients regarding medication-taking behavior.


INTRODUCTION
Pharmaceutical care requires a patient-centered approach, focusing on the ability of patients to understand drug-related information and follow instructions delivered by pharmacists as well as other health-care providers included in the circle of care. Pharmaceutical care was defined by Hepler and Strand as the responsible provision of drug therapy for the purpose of achieving specific drugrelated outcomes (i.e., cure a disease, eliminate or reduce symptomatology, arrest or slow disease progression, or prevent disease onset) as well as improve overall wellbeing and quality of life of the patient (Hepler and Strand, 1990). Pharmaceutical care involves a multidirectional communication process whereby the pharmacist cooperates with the patient, physician, and other healthcare professionals in designing, implementing, and monitoring a therapeutic plan that will result in specific therapeutic outcomes with the goal of providing direct benefits for the patient. At the center of this balancing of treatment and outcomes is the patient, where medication adherence is the fulcrum that ultimately dictates success and literacy is the tool that tips the balance towards successful health outcomes.

The Role of Literacy in Health: A Focus on Chronic Patients
Poor literacy and health literacy are serious problems with worldwide health burden and are major contributors to negative patient outcomes. In the United States, it is estimated that 47.0% of adults have difficulty understanding health-related information and lack the necessary skills to manage their health adequately (Institute of Medicine US Committee on Health Literacy, 2004); in Canada, it is an estimated 60.0% people (Canadian Council on Learning, 2008); and a recent survey of European countries suggests that 47.6% of all adults have limited health literacy skills with major differences among countries, ranging from 28.7% in the Netherlands to 62.1% in Bulgaria (Sørensen et al., 2015). For Africa, data on health literacy is scarce but it has been noted that the average literacy rate was 63% in 2015, where approximately one third of the population could not read and write (UNESCO Institute for Statistics, 2013). Clearly there is a global health literacy crisis that affects both developing and developed nations. It is also clear that improving the patient's ability to understand and make appropriate decisions about medication-taking behavior depends on improving both literacy and health literacy by ensuring the information is both accessible and understood.
Low health literacy has also been associated with poor medication and treatment adherence in healthy and in chronic patients (Kalichman et al., 2000;Youmans and Schillinger, 2003;Ngoh, 2009).
Patients with chronic diseases and poor health literacy are generally at risk for poor health outcomes and low quality of life, and they generate a greater burden for healthcare systems. Specifically, poor health literacy is associated with higher allcause mortality rates among patients with heart failure (Peterson et al., 2011). Furthermore, in the setting of chronic diseases such as hypertension, poor adherence limits the effectiveness of therapies proven to improve cardiovascular outcomes (Roccella, 1997) and can result in hospitalization and higher health care costs (Sokol et al., 2005). In patients living with diabetes, low health literacy has been associated with poor glycaemic control and therefore exposure to diabetes complications (retinopathy) (Bailey et al., 2014). In patients living with chronic-obstructive pulmonary disease a positive association exists with disease severity and utility of emergency healthcare services, whereas a negative association emerges with disease severity and quality of life specific to respiratory disorders, as well as helplessness in the face of illness (Omachi et al., 2013). It should be noted that some of the data on lower health literacy and nonadherence are clear (e.g. medication dosing), however the relationship between literacy and adherence is not equivocal as there are studies that have failed to find any relationships (Davis et al., 2006;Pignone and DeWalt, 2006).

Chronic Patients, Health Literacy, and Medication Adherence
In 2005, 35 million people were estimated to have died from chronic diseases worldwide, representing more than 60% of all deaths globally (World Health Organization, 2005). Nearly half of all adults and approximately 8% of children (aged 5-17 years) worldwide have a chronic disease, where chronic diseases are defined broadly as conditions that last 1 year or more and require ongoing medical attention or limit activities of daily living or both (Centres for Disease Control and Prevention). Furthermore, four in ten adults have two or more chronic diseases, leaving millions requiring multiple lifelong medications for control (Buttorff et al., 2017). Poor medication adherence is a key hindrance in combating the challenges of public health related to the management of chronic conditions in both developed and developing countries (Ingersoll and Cohen, 2008). The World Health Organization defines adherence as "the extent to which a person's behavior -taking medication, following a diet, and/or executing lifestyle changes, corresponds with agreed recommendations from a health care provider" (World Health Organization, 2003). Medication adherence is defined as the degree to which a patient adheres to the prescribed dose and interval of their medication regimen and it includes the initiation of the treatment, the implementation of the prescribed regime, and the discontinuation of the pharmacotherapy (Vrijens et al., 2012). Nonadherence to medication ranges between 25 and 50%, with higher numbers of nonadherence found in patients living in less developed countries and in patients living with chronic disease (DiMatteo, 2004). The impact of nonadherence on the patient's health can be severe due to the reduced benefit of their medication, resulting in further disease progression or delayed recovery. This can lead to an increased risk of hospitalization and possibly death. Notably, the risk of hospitalization can be increased by over 100% in patients suffering from certain chronic conditions such as hypertension, diabetes, congestive heart failure and hypercholesterolaemia (Sullivan et al., 1990).
Indeed, in reports of medication adherence in chronically ill patients, it has consistently been reported that approximately 50% of these patients do not take their medications as prescribed (Dunbar-Jacob and Mortimer-Stephens, 2001) and in certain diseases such as hypertension, where the patient may present as asymptomatic, the incidence of non-adherence may approach 80% (Brown et al., 2016). Such findings are not shocking when considering the complexity of taking multiple medications, where patients are required to read their medication labels and associated medical information, comprehend instructions, perform numeric tasks (e.g., calculating the number of tablets to take in a day and in a single dose), and must decide what actions are required in the case of a missed dose or to monitor themselves for side effects (Youmans and Schillinger, 2003). Taken together, in order to ensure overall adherence and patient safety, a proactive approach of the health care provider is integral in identifying and preventing potential drug-related problems and resolving actual ones (Hepler and Strand, 1990), as well as communicating this complex information in the context of each patient's own health literacy level.

Review Aims: Medication Adherence in Chronic Patients and the Role of Pictograms
The problem of low health literacy should be considered in the broad spectrum of the current challenges of healthcare systems. Both physicians and pharmacists tend to have less time to spend on education for the patient (Gazmararian et al., 2003) so the modern healthcare systems must focus on the inter-relationship between self-management or self-efficacy from the patient, and an open dialogue with the health care team. People with lowliteracy indicate that they experience a high cognitive load when required to read written drug information, which is reflected in their comments about the time and effort it takes to read and process the information (Mayer, 2002); indeed, similar findings exist in older patients living with chronic conditions (Park et al., 1994). Taken together, the aim of this review is to provide a balanced and comprehensive review outlining a number of different medication counselling and education approaches that have been used to try to improve medication adherence and health outcomes, with a specific focus on the use of clear and concise graphic illustrations-called pictograms, to help improve patient outcomes. By highlighting the current landscape of the general use and efficacy of pharmaceutical pictograms to aid in the knowledge and recall of drug-related information, as well as outlining specific medication adherence outcomes with pharmaceutical pictograms in chronic patients, the objective of this review is to describe the need for health-care providers from low-and middle-income countries to high-income countries to move beyond the traditional didactic methods of oral and verbal communication with patients in order to improve medicationtaking behavior.

METHODOLOGY
The methodology was as follows: the search was conducted with the use of Medline, EmBase, Cochrane Library and Google Scholar. We used key words like: pictogram, graphic, picture, pictorial aid, label. The search was conducted from January to March 2020. No limitation regarding the date of publication was set. No documentation regarding the collection of publications was made.
In the first step of the review, the titles of the articles were carefully checked. If the title obviously indicated that the article was not suitable for the scope of our review, the reference was rejected. If the title suggested the presence of data consistent with our interests, the publication was further analyzed. In the next step, abstract was assessed, and if still applicable, full text was analyzed. We wanted to include original studies assessing the utility, efficacy and patient feedback regarding pictograms in the context of chronic diseases among teenagers and adults, and so the criteria were defined. In our review, we aimed to present the broad spectrum of the concepts of pharmaceutical care and how pictograms can be implemented into this philosophy.

MEDICATION ADHERENCE: THE CURRENT LANDSCAPE
The topic of medication adherence has been widely studied as evidenced by the fact that there are over 200 systematic reviews on the topic. According to the WHO, there are several reasons that have been identified for non-adherence, including: low health literacy, cost to patients, fear of adverse drug effects, lack of social support, etc.). Regarding the study design of the interventions that have assessed adherence, a Cochrane systematic review by Nieuwlaat et al. identified RCTs at lowest risk of bias, where they found that studies assessing adherence ". . .generally involved complex interventions with multiple components, trying to overcome barriers to adherence by means of tailored ongoing support from allied health professionals such as pharmacists, who often delivered intense education, counseling (including motivational interviewing or cognitive behavioral therapy by professionals) or daily treatment support (or both), and sometimes additional support from family or peers (Nieuwlaat et al., 2014). However, as noted but the authors, these interventions did not lead to large improvements in adherence or clinical outcomes.
Perhaps most important in any study of adherence is to first properly operationalize how this outcome is measured. In a recent review by Anderson et al., the authors identified 25 high quality systematic reviews (with a mean of 36 primary studies) assessing interventions to improve medication adherence (Anderson et al., 2020). Overall it was found that 96% of these reviews did not restrict the method for measuring medication adherence, where the main outcomes of adherence were the following: 84% of reviews had at least one primary study that employed electronic monitoring to measure adherence (e.g., mobile text messaging, pill bottles with alarm features, etc.), 84% of reviews had at least one primary study that used fill count or had at least one primary study that used patient self-report, whereas, 68% had at least one primary study that used pharmacy refill data, and 8% had at least one primary study that used patient blood levels. Overall, it was determined that the top three measures that led to improvements in adherence were: dose simplification (5 systematic reviews), electronic reminders (4 systematic reviews), and patient education (4 systematic review).
Taken together, and with the understanding that medication adherence is best promoted with the aid of health professionals such as pharmacists, findings highlighted in this section demonstrate that interventions that focus on dose simplification-with tailored ongoing support-are the most successful in improving patient outcomes for adherence (Wilhelmsen and Eriksson, 2019). For these reasons, and for the fact that pictograms have been used for many years to improve patient information concerning their medications, the following sections will discuss the potential role of pictograms in Frontiers in Pharmacology | www.frontiersin.org August 2021 | Volume 12 | Article 582200 improving medication adherence, with a focus on chronic patients.

WHAT IS A PICTOGRAM?
In general, pictograms are symbols or drawings representing a concept or idea (Peregrin, 2010). A pictogram should be considered as a two-part construct: a graphical symbol and its intended meaning. The "theory of semiotics" by Saussure introduces the distinction between the signifier (physical description of the object) and signified (its mental concept created by cultural convention) (Yang and Hsu, 2015), and importantly, the signified can have different meanings based on different cultures (Cloutier et al., 2014;Kheir et al., 2014). According to Dowse and Ehlers, "Well-designed pictograms should be simple and clear and able to convey their intended meaning to all patients, including those who are illiterate, elderly, or visually impaired." (Dowse et al., 1998). The guiding theory of pictogram use in any field of study is that when exposed to an image, the verbal memory may be triggered by reinforcing memory traces and subsequent recall; in order to do so, the message needs to be clear, appropriate for the intended audience, and must focus on actions rather than information (see Figure 1 for examples of validated pictograms for medication counselling). Furthermore, it is important to note that pictograms should be used in combination with traditional counselling/education. Specifically, prior patient counseling by pharmacists on the intended meaning and use of pictograms has been shown to improve the pictogram effectiveness (Ngoh, 2009), a fact that has been reiterated by (Montagne, 2013) in their description of best practices for the development and assessment of pictograms (Montagne, 2013).

Pictograms in Health-Related Research: A Focus on Chronic Patients
Why should healthcare providers consider using pictograms to convey health-related information during patient counselling? It has been reported that more than two-thirds of physicians provide written patient education materials for chronic patients (Carrier and Reschovsky, 2009), and between 40 and 80% of verbal information communicated during a health care consultation can be forgotten almost immediately (Kessels, 2003).
Introduction of this review has already painted a picture of an epidemic of poor literacy, and despite abysmal worldwide levels of health literacy, the reading level of most health literature is above eighth grade (Cotugna et al., 2005), signifying a disparity between knowledge transfer. There is a sufficient body of evidence to show that pictograms not only make complicated patient information more attractive (Houts et al., 2006), but these images can also improve comprehension and recall of proper medication-taking behavior (Chan et al., 2015;Sletvold et al., 2020). Although the data on literacy and pictogram effectiveness is scant, results remain equivocal, where some groups have shown no impact where caregivers had adequate levels of literacy (Yin et al., 2011), whereas others have shown low and high literacy caregivers can both benefit from pictogram-based counselling (Hu et al., 2013;Tork, 2013).
There is a paucity of information on the use of pictograms in chronic patients to aid in medication adherence; however, one can see in Table 1 that the majority of studies in chronic patients are in patients living with HIV/AIDS. Although each study showed a positive outcome for the effectiveness of pictograms to improve recall and comprehension of dosing instructions, it was noted in a systematic review by Chan et al. that these are all studies with high risk of bias (Chan et al., 2015).
As seen in Table 1, there were several underlying themes that emerged in these interventions that deployed pictograms: the impact on pharmacotherapy safety, patient awareness and improvement of treatment results, patient involvement in the therapeutic process, and patient communication and trust in the delivery of their healthcare. It is worthwhile to note that a recent systematic review by Nguyen et al. identified at least 43 validated self-report adherence scales (Nguyen et al., 2014), indicating that it is crucial to consider the tool being used to assess the outcome of adherence, as well as the quality of the pictograms being tested. Overall, there is no single measure that can assess all the behaviors involved in being adherent to medication(s) and how this can impact patient outcomes. Selecting two (or more) medication adherence measures such as can be seen in Table 1 can allow strengths of one method (indirect methods of verbally explaining medication instructions) to help compensate putative weakness and to more accurately capture the information needed to determine adherence levels when incorporating another method (such as pictograms).
In patients living with chronic illness, data suggests that individuals with low health literacy can benefit more from  • A significant improvement in interpretation of the individual pictograms at the second interview was noted for two of the pictograms "pain medication" and "slow digestion".
• Among 13 participants who assessed their satisfaction with the pictograms, 11 (Continued on following page) Frontiers in Pharmacology | www.frontiersin.org August 2021 | Volume 12 | Article 582200    interventions employing pictograms (Machtinger et al., 2007;Negarandeh et al., 2013;Mohan et al., 2014;Phimarn et al., 2019), whereas, individuals with high health literacy do not seem to benefit to the same degree (Kripalani et al., 2012). Table 1 summarizes the results of studies on the use of pictograms among chronically ill patients in distinction for the following diseases: asthma (n 2), diabetes (n 4), AIDS (n 7), chronic kidney disease (n 1), and cardiac disorders (n 4). Studies focusing on the elderly were analyzed separately (n 3). Studies focusing on self-efficacy were analyzed separately (n 2). The data obtained in the selected studies indicate that the use of pictograms significantly increased various aspects of disease self-management of the patients with chronic conditions. However, not all pictogram interventions were found to be effective, where several groups have failed to find a significant effect of pictograms in helping with medication adherence in patients living with chronic illness (Kripalani et al., 2012;Advani et al., 2013;Chan and Hassali, 2014). In our work, apart from comprehensive review on the impact of pictograms on adherence in chronically ill patients, we also would like to point out some methodological issues, which may be an obstacle to a reliable assessment of the impact of pictograms on adherence.
Publications summarize in Table 1 describes the role of pictograms in chronically ill patients, but from methodological point of view, they are different in many respects. First of all, the results were obtained in many different countries and among subjects with different level of literacy. As we stated previously these factors can have a considerable impact on the usefulness and efficacy of pictograms. Additionally to this, interventions were carried out in patients with different chronic diseases. Moreover, the number of patients participating in assessing studies varied a lot (n 17-446). Apart from that, these studies lacked information of the clarity and appropriateness of pictograms chosen. There was also a lack of explanations on the validity and reliability of instruments used. One should also pay attention to the differences in the methods of patient education. Interventions described in cited publications were carried out both in pharmacies and clinics, so we can anticipate many differences in approach to education as well as the amount of time devoted to the patient.
These all above described factors impact on heterogenicity of published trials and consequently have an impact on the possibility of comparing obtained results, drawing conclusions and preparing more credible publications like metanalysis. Future research should also devote more attention to aforementioned limitations of the included studies. Maybe good tool, which can contribute to overcome these methodological problems is developing specific protocols that will be used in further studies. That protocols could allow collect more coherent results and precise assessment of the impact of pictograms on adherence.

Best Practices in Pictogram Design
Pictograms, when properly designed and validated, have been implemented to help convey medication information in many health-related settings and humanitarian missions (e.g., hospitals, pharmacies, temporary aid stations, etc.) with varying degrees of successful outcomes (Houts et al., 2006;Sorfleet et al., 2009;Dowse et al., 2014). But in order to make any conclusions on the effectiveness of pictograms on medication adherence, and prior to having to consider how to answer a research question with a large and expensive randomized trial, one first needs to be sure the image meets specific standards of best practice in pictogram design.
Researchers in the field of health-related pictograms should consider the following. First, and most important to ensuring best practices are being followed, is the rigorous and time consuming process of designing new pictograms. Pictograms must go through a systematic design process, whereby images are piloted and tested on scales of transparency (or "guessability of meaning") and translucency ("agreeability of intended meaning"), each of which contribute to independent elements that come together to determine the overall comprehensibility of the pictogram (Vaillancourt et al., 2017), which can also be evaluated with open choice face-to-face methods (Mansoor and Dowse, 2003). It is important to note that there are differences in patient populations to consider (e.g., education, beliefs, attitudes, etc.) in this initial process of testing and design (Dowse and Ehlers, 2003). In particular, many studies have now demonstrated the importance of considering differences in pictorial interpretation between countries, and several studies have also demonstrated cultural differences within countries, and these differences can determine whether the information relayed in the pictogram is correctly understood (Doak et al., 1996;Ehlers et al., 2004;Grenier et al., 2011). As such, if the researcher is to ensure an effective evaluation process for pictogram design that can provide data that is both valid and reliable, it is integral to have the target population involved at all stages of the design process.
Furthermore, in order to be able to systematically evaluate differences in study outcomes in the field of health-related pictogram research there are several other important elements to consider to ensure that best practices in pictogram design are followed (for extensive details see: Dowse, 2020;Abdullah et al., 2006;Mansoor and Dowse, 2004). These are guiding documents that should serve to help groups in health-related research to produce high quality pictograms with optimal legibility and visibility. Sletvold et al. (2020) concluded that studies assessing the value of pharmaceutical pictograms must establish best practices in the design and use of pictograms to ensure we can isolate the independent effects of pictograms on medication adherence (Sletvold et al., 2020).

Best Practices in Pictogram Evaluation
An important consideration for the successful evaluation of pictograms in the delivery of health information is to ensure that the healthcare professionals who will ultimately deliver the education and counselling are involved in the entire process of design and evaluation. For example, when implementing pharmaceutical pictograms that help with medication adherence, patients generally still need to be taught how to interpret the pictograms (Ngoh, 2009), and these increased task demands on the pharmacist must be perceived as a beneficial use of time and resources in order to get the most benefit for the patient. Much research in the study of pictograms in the field of health information-and in particular research on medication adherence-still does not explicitly specify if pharmacists were involved in the design or implementation of the specific pictogram intervention. Results described by Wilhelmsen & Eriksson (2019) in a recent systematic review possibly speak to this importance, where they noted: ". . .that interventions delivered by pharmacists and nurses showed a better result in improving adherence and outcomes than interventions delivered by general practitioners (Wilhelmsen and Eriksson, 2019). Although this is not direct evidence for the case, future research can benefit from a standardization process whereby these details are systematically documented, for example, when trials are registered. Finally, once best practices have been employed in the design and evaluation of the pharmaceutical pictograms, the research team is ready to design a randomized controlled trial that can isolate the independent effects of pictograms on, for example, medication adherence. As noted by (Anderson et al., 2020) systematic overview of systematic reviews, much of the research on medication adherence is highly heterogeneous in study designs, settings, and methods (Anderson et al., 2020), so attention must be focused on the fidelity of design to ensure reproducibility of the data. These summaries of fundamental problems were echoed by (Sletvod et al., 2020), where the authors concluded that heterogeneity in the design and conduct of the current landscape of RCTs examining pictograms in medication adherence preclude any meta-analysis of observed pictogram effects (Sletvold et al., 2020).

CONCLUSIONS AND FUTURE PERSPECTIVES
On the basis of the reviewed data, we conclude that effective spoken and written communication of information about medicine, in combination with health-related pictograms, can lead to improved treatment outcomes for the patient. Low health literacy levels negatively affect patients-including chronic patients-across the continuum of care. With this in mind, pharmacists and health care-providers are well-situated to ensure that patients adhere to their medication regimens correctly, especially so for lower health literacy populations. Chronically ill patients are a population that can benefit from interventions (e.g. "teach back" and Ask Me 3 ™ ) because their pharmacotherapy is often a life-long challenge, which often requires coping with complex daily multi-drug regimens that can require high levels of literacy and self-efficacy.
The current review highlighted adherence in chronic patients. Due to a paucity of research on pictograms and medication adherence in this population and due to the lack of best practice in much of the existing research, we could not conclude that these patients benefitted any more or less from pictogram interventions. Indeed, in the recent systematic review by Sletvold et al. (2020) the authors found a possible effect of pictograms on medication adherence, but in order to make stronger conclusions, researchers in this field need to follow best practices not only in the design and evaluation of healthrelated pictograms, but also in the design of strong RCTs that can isolate the independent effects of pictograms on study outcomes (Sletvold et al., 2020). Taken together, research in the field of health-related pictograms and medication adherence should focus on creating content that focuses on dose simplification and electronic reminders. By adhering to best practices in the design, testing, and evaluation of pictograms, we can strengthen future findings from this exciting and visually stimulating field of health-related research.

AUTHOR CONTRIBUTIONS
Conceptualization, PM, KB, and RV; methodology, RV; validation, MJ, and RV; formal analysis, TC, KB, PM, TH, KB, KS; investigation, PM; resources, DS; data curation, PM; writingoriginal draft preparation, PM, JC, and RV; writing-review and editing, TC; visualization, MJ; supervision, RV; project administration, PM. All authors have read and agreed to the published version of the article.