Patients’ Perspectives of Oral and Injectable Type 2 Diabetes Medicines, Their Body Weight and Medicine-Taking Behavior in the UK: A Systematic Review and Meta-Ethnography

Electronic databases (Cochrane Library, DARE/CRD/NHS (via http://​www.​crd.​york.​ac.​uk), PROSPERO, Web of Science, SCOPUS, ScienceDirect, PsychINFO, CINAHL Plus, MEDLINE, EMBASE, AMED, EBSCO) were searched from inception to February 2016 (full strategy is shown in the Electronic Supplementary Material [ESM]). A broad and specific approach was used to identify studies based on a combination of the following search terms: “type 2 diabetes,” “medication-taking,” “adherence,” “compliance,” “treatment,” “medication,” “oral hypoglycemic drugs/agents,” “anti-obesity drugs,” “insulin,” “beliefs,” “attitudes” and “weight.” The references of all retrieved articles were checked for relevant studies.

Published studies in English in which the methodology included a qualitative data collection and analysis with a primary focus on medication-taking behavior or beliefs/attitudes towards GLDs and/or AODs were included in the review. Mixed-method studies reporting qualitative data were also included. Studies were selected if the participants were (1) adults (> 18 years) with T2DM, (2) using pharmacotherapy to manage diabetes and/or weight and (3) managed in the UK (primary care or an outpatient clinic in secondary care). Studies were excluded if participants had gestational diabetes or were hospitalized, or they included all types of diabetes and types of therapies but did not focus on the above criteria in a sub-group analysis. The lead author (AP) performed the systematic literature search, of which a sample was verified by the second author (HC). This process involved screening papers by title, abstract and full text. Discrepancies were resolved through consensus. A further three studies were identified following screening of individual studies from two systematic reviews [20, 24] as they were eligible for this review. The review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) checklist for the selection of studies (Fig. 2).

Included studies were assessed for quality using the Critical Appraisal Skills Program (CASP) criteria for qualitative research [29], with an additional question regarding each study’s theoretical perspective, as illustrated elsewhere [30]. To ensure rigor (see ESM), articles were included only if they scored ≥ 7 (maximum score 11), resulting in one article being excluded at this stage (Fig. 2). One author (AP) critically appraised all studies, and a sub-sample was checked by a second author (HC). Both AP and HC examined papers with low CASP score to reach a consensus.

Papers reporting data from the same research study were collated as one. None of the eligible studies reported views on AODs and non-insulin injectable medicines. Papers were separated into four research-focus groups, namely, oral GLDs (OGLDs), insulin, diabetes medicines (OGLDs and insulin) and polypharmacy/multimorbidity, using the chronological approach to form the synthesis [28]. For example, articles which only described participants’ experiences with OGLDs were part of the first group and were synthesized chronologically. First- and second-order constructs (i.e. direct quotes from participants and interpretations of authors) were extracted by AP using a spreadsheet to manage the synthesis process. Second-order constructs retained the terminology of the authors who published the study to preserve the meaning and context of their findings. Themes were separated into two categories of views and experiences, namely, prior to and after initiation of a medication treatment, to identify changes in participants’ views about their medicines and their medication-taking behavior over time. Themes also included passages related to quality of care and weight to identify whether these aspects influenced medicine-taking behavior, although these did not represent third-order constructs. The studies were then translated into each other, one by one, to form a reciprocal (i.e. similar findings) or refutational (i.e. contradictory findings) synthesis [26, 30]. Briefly, the synthesis of the first two papers was compared with the third paper, and this process continued until all studies were translated into each other. This process was done separately for the “OGLD” and “insulin” groups. The findings from the “diabetes medicines” and “polypharmacy/multimorbidity” groups were then compared with each of the first two groups where relevant. In the process of comparing the studies against each other, explicit differences among the studies were noted (Table 1). Finally, a “line-of-argument” synthesis [26] was constructed by developing an overarching model that linked the translations and authors’ interpretations to address the aims of the review.

A total of 16 studies were reviewed (reported in 19 papers) (Table 1) using data from 360 individuals, mostly of Caucasian (n = 133) and South Asian origin (n = 122), all of whom were recruited from different UK regions, both in primary (n = 9) and secondary (n = 7) care. Participants (aged 21–89 years) had T2DM for a maximum of 35 years although their glycemic control and body mass index (or body weight) were rarely reported, except where details were extracted from a clinical trial [31] associated with one of the studies [40, 41].

Two studies employed a longitudinal study design to identify changes in participants’ views [34, 36]; however only one analyzed data longitudinally [34]. While the behavioral and social theories identified were linked to individuals’ health-related behavior, the studies did not provide much detail beyond stating the theory applied.

The studies examined in general various issues around the use of OGLDs and insulin, but none specifically addressed other injectable or OGLDs (i.e. glucogen-like peptide-1 [GLP-1] agonists, dipeptidyl peptidase 4 [DPP-4] inhibitors, or sodium/glucose cotransporter 2 [SGLT-2] inhibitors) or AODs (i.e. orlistat). One participant was taking insulin along with exenatide, a GLP-1 analog; however, the focus of that study was solely on insulin [43]. Medication name or regime was rarely reported, except where participants’ quotes indicated multiple use of diabetes and non-diabetes medicines or specified generic drug names.

Studies related to participants’ views following the initiation of insulin therapy included those involving participants who took insulin for at least 1 month [36] as well as those in which the participants were taking insulin for more than 2 years [37, 40, 41]. Studies related to barriers and refusal to initiate insulin therapy included participants who had either refused insulin in the last 3–6 months [38, 42] or had started insulin within 6 months [42]. Two studies [39, 43], related to perceptions about insulin, included individuals who either were already taking insulin (0.4–22 years) or were insulin naïve, while only two studies specific to OGLD treatment specified treatment duration (0.3–10 years) [33‐35, 39].

The reciprocal synthesis (Fig. 3) shows a timeline illustrating patients’ emotions arising from the prospect of initiating T2DM medication, including their perceptions of T2DM and respective medicines and their expectations of these medicines, as well as how these related to outcomes such as patterns of medication-taking and lifestyle behavior. The line-of-argument synthesis (Fig. 4) demonstrates the key processes to understanding patients’ journeys in medication management.

Devastation, shock, anger and disappointment were the major emotions which appeared with the prospect of taking medicines [32, 34, 36, 37, 40, 48]. Starting OGLD therapy was linked to becoming a “sick person” [32, 34] and entering “a slippery slope” [34] leading to taking insulin, often associated as “the last resort” [38, 39, 43]. Patients referred to being “defeated” and letting themselves down by failing to self-manage their diabetes and diabetes treatment [34, 37, 39, 42]. Personal failure also emerged when patients were advised to start insulin therapy, mainly when initiation was imminent [37], and because they would be judged by others as a “bad diabetic” [48].

Over time patients, particularly if symptomatic [48], rationalized that diabetes is a progressive disease [32, 34, 49], believing that lifestyle can cause it but cannot cure it [35], “…no matter how careful you are” [32]. This group of patients, many of whom were aware of T2DM and its treatment from family members [45], placed importance on the role of medicines in managing diabetes [36, 37, 39, 48, 49]. Once they had initiated medicines, they expressed the view that they would have to take them for the rest of their life [32, 49, 50], otherwise, they would “risk dying” [32]. This view was reinforced by self-regulation strategies which supported the acceptance of new medication prescriptions (top box, Fig. 3). The combination of results from blood glucose testing and their own observations over time of an increase in the number of diabetes tablets and doses were good indicators of deteriorating diabetes despite adherence to treatment [40, 43], justifying for some that eventual insulin treatment was an integral part of diabetes. Nonetheless, for others, blood glucose testing [34, 45] (bottom box, Fig. 3), over time, proved that medicines had helped with blood glucose control [34, 36, 41, 45] if they were asymptomatic or “unconvinced” of the medicine’s effectiveness, reinforcing their perceptions and positive experiences. In contrast, some on insulin therapy felt frustrated when they could not see the results of their efforts to achieve good diabetes control [37].

In our review this group of patients consists of those prescribed OGLDs but who had either no experience of taking insulin or were advised to take insulin but refused to do so [33, 35, 38, 39, 42‐44]. These patients did not want to actively think about their diagnosis [34, 45], denying the possibility of future complications [46] or believing these are not imminent and believing that their diabetes is “insignificant” [33] if symptoms ceased after being prescribed OGLDs or if T2DM was managed with tablets [37, 38, 42, 43, 48]. Insulin was an indicator of the worse type of diabetes [37‐39] and associated with more side effects than the tablets [39], and the reluctance to initiating insulin therapy was associated with negative stigma (Section Perceptions and Experiences of Medicines) and past medication experiences, particularly if OGLDs were perceived unsuccessful in diabetes management [38]. Additionally, Lawton et al. [33] suggested that the type and location of care can change an individual’s perception of his/her diabetes, where care provided by the general practitioner is perceived as less serious than hospital care.

Key motivators for accepting diabetes medicines included instant success, specifically with an injectable treatment, and benefits of both weight loss and diabetes control [35, 43]. Generally, individuals wished to consult a healthcare profession who they considered to be an expert or trustworthy, an individual who they trusted would prescribe medications appropriately [32, 33, 39, 45], and they wished for continuity of care between healthcare services and health professionals (HPs) [37].

Patients’ information needs were found to be high [34, 35, 49, 50], despite expecting an ongoing prescription [33, 34] and not questioning medication changes [34]. They preferred non-judgmental guidance [34] on alleviating negative medication experiences and managing missed doses and information that uses “layman’s” language [45, 50]. Specific to insulin, they wanted to know the advantages and disadvantages of taking insulin [37, 43, 48], what type and how many doses are required to control diabetes and if any changes are required in these [36, 39, 41], how to inject insulin correctly and guidance on dietary changes [39]. They also wanted to be better informed about the need for their medicines, particularly if they hold negative views about them [49].

Patients’ positive medication experiences reinforced positive perceptions that medications are beneficial, effective, relieve symptoms and preserve health [32, 34, 36, 37, 39, 41, 49]. When medication effects were seen over time [36, 41], some patients indicated that medicines were more effective than lifestyle measures [34, 35, 49], prioritizing certain T2DM medicines over others [32, 49, 50] and reporting insulin being better than tablets [37, 39].

However, medications were unnatural [32, 34, 36, 46]. Therefore, those taking multiple medicines (whether on complex diabetes and non-diabetes medication regimes) were worried these could counteract their individual effects [32] or “that’s three lots attacking the kidney” [34], resulting in long-term health damage [32, 34, 46]. They wished to minimize the daily medication numbers or simplify the regime to help overcome their ongoing concerns and cope with the demands of self-managing diabetes [50].

Moreover, insulin was perceived negatively by all patients whether they refused it or not, with many expressing they would rather take more OGLDs than inject insulin [43]. Common worries were needle anxiety [36, 38‐41, 48], fear of consequences/complications of permanently taking injections [37, 43], fear of weight gain or inability to lose weight [38, 42] and fear of hypoglycemia [37‐39, 42]. Insulin was perceived as restricting and interfering with their lives [38], including daily and social activities and traveling [37‐39, 42], and with increased dependence on others [38, 39]. Those participants who were experienced with insulin therapy reported that they had lost their identity and normality [36, 48], found it harder to self-manage diabetes [37] and worried about injecting in public, particularly if they needed to inject with meals. Insulin was also perceived as a threat [36, 39, 42], often used by HPs to encourage patients to better control their diabetes. However, insulin perception transformed into “punishment” [37] for those starting insulin due to its association with negative stigma [37, 43] and the belief it was unjust [36], perhaps originated from the way it was introduced.

Overall, five adherence patterns were identified (Fig. 3). There were those who fully adhered to diabetes medicines and who strongly believed in the positive aspects of these medicines. However, they felt guilty when they unintentionally non-adhered to their medicines [33, 34, 41, 49] due to forgetfulness, broken routines, being asymptomatic and confusion when taking multiple medicines [34, 41]. Adherent and unintentional non-adherents adopted strategies to promote adherence, such as establishing routines that included identifying logistics of where and when to inject, using visual and family reminders, using medication boxes or carrying medicines with them or receiving group support [34, 37, 41].

Conversely, there were some patients who unintentionally non-adhered to medicines due to forgetfulness, who did not appear to feel guilty about that behavior, perhaps due to lack of symptoms and their belief that diabetes is not a serious condition [33]. In those who exemplified intentional non-adherence, guilt was also rarely present [34]. Reasons for non-adherence included strong negative perceptions of their medication, such as insulin stigma [41], adverse effects [32, 33] and the medication being detrimental to health [32, 50]. Strategies of intentional non-adherents included adjusting doses, delaying or omitting taking the medicines altogether [32, 34, 35, 41, 50], often manipulated by using self-monitoring of blood glucose and/or diet, because they believed that medicines work instantly by alleviating symptoms [32, 35]. HPs also indirectly influenced medication-taking behavior if they had in the past advised patients to adjust medicines; so in similar situations, patients were adjusting doses without seeking medical advice [32]. For those taking insulin, strategies included talking to friends and family about T2DM and treatment of hypoglycemia [37, 39] by regularly referring to written material and actively seeking further information on diabetes treatments [37], often encouraged by their family [36, 37]. They manipulated their diet by eating more often to avoid feeling weak, which according to the authors was unlikely to be linked with low blood glucose [46], or restricted social activities such as eating out to avoid injecting in public [41].

Great importance to lifestyle measures was given by those who perceived their diabetes as serious and who had converted to insulin [35], by strictly following a healthy eating plan [37, 39] or using a moderate approach to eating “…a little bit of what you fancy” [48]. In contrast, others could not identify if any changes to their lifestyle had an impact on diabetes control and were uncertain of their diet’s effectiveness [45], struggling to keep up with the diet and physical activity [33, 35, 37, 44]. Yet others indicated that the lifestyle and medication are two alternative ways in which they could self-manage diabetes, suggesting that once you take medicines they could eat what they want [35, 39, 48], particularly when on insulin [39].

Concerns of weight gain were apparent, but these were not linked to diabetes medicines, except when fear of weight gain was associated to the refusal to initiate insulin therapy [38, 42, 46]. Hence, weight was related to lifestyle measures, even though many patients were “naughty” and “cheating” [33] or were unsure how to reduce calories [39]; they felt misunderstood by HPs in relation to the difficulties they faced [44, 47] with following the recommended diet. Conversely, others suggested that they would lose weight if they were to avoid taking medicines [35, 49], although not all believed that it would be diabetes medicines [49].

Patients’ journeys started either at the time of diagnosis and/or at the time of medication prescription, both creating grief reactions. As patients progressed from diet to tablets to insulin, their emotional reactions were progressively stronger and impacted upon the way they coped. However, perceptions of diabetes severity appeared to change over time and were influenced by their community and family experiences [36, 38, 39, 43, 45], the presence/lack of diabetes symptoms and complications [33, 35, 37, 43, 45‐48] and their own medication beliefs. Taken together, beliefs and emotions reinforced an evaluation of oneself, leading to either accepting or denying the “altered self-image” [36]. Generally, acceptance to change meant that medicines were accepted as part of the successful management of diabetes. However, resistance to change did not necessarily mean resistance to taking medicines.

Medication-taking and lifestyle behavior coupled with self-regulation strategies were all tactics patients engaged to regain self-control. These were often associated with patients’ treatment adherence or non-adherence [34, 35, 41, 49] and to being passive or active [34, 36, 43] in self-managing diabetes, terms commonly used in clinical research and practice. However, for patients these strategies are rational decisions/actions enabling them to live as normally as possible and to minimize the impact of their diabetes and medicines in everyday life.

Patients were committed to taking medicines [34, 37] with a “sense of growing confidence” [36] and were making conscious and deliberate “effort” [37, 41] to find out about diabetes and adapt their daily practices to fit medication, with some “requesting” or “negotiating” their new regime [34, 41]. Still others reported feeling “powerless” [36], conveying “worry” [39] because they had “tried hard” [37] but had been unable to make successful changes. Hence, the experience of successful and unsuccessful practices reinforces or undermines patients’ T2DM treatment beliefs and leads yet again to a re-evaluation of self.

The review’s strength is reflected by the systematic identification of papers using both broad and specific search strategies and multiple databases. Its rigor is enhanced by the inclusion of high-quality papers based on the CASP score following critical appraisal by two of the research team. The transparency and trustworthiness of the findings were ascertained by detailed discussions among the research team about the interpretation of the findings. However the transferability of findings and conceptual models to patients from other countries is limited due to our restriction on language and focus on medication experiences from a selected group of UK participants who are medically/pharmacologically treated and do not pay for their prescriptions.

The intention of the synthesis was to retain a rich context of data by looking systematically at the influences of various contextual factors on medication-taking behavior, such as socio-economic status, diabetes control, body weight, disease and treatment duration, treatment type and use of healthcare services, as seen elsewhere [19, 24, 70]. However, this was difficult as the authors of many studies did not provide adequate descriptions of context and the impact of context on their findings (Table 1). We believe such information would be useful to better describe the medication adherence phenomenon. Only a few papers provided rich and insightful data to fully appreciate patients’ experiences and practices with diabetes medicines. While there were only two longitudinal studies, most studies were based on peoples’ memories of past experiences and their views after medication changes had occurred. The synthesis related to OGLDs was based on two studies, which were mainly conducted by the same group of authors. Furthermore, the process of synthesizing is inherently interpretive, so other reviewers may produce difference conceptual frameworks. Nonetheless, 19 papers provided sufficient data to reach a “line-of-argument” resonating findings from other published reviews [14, 17, 20, 22, 24, 71]. This suggests that our findings are triangulated and credible.