FormalPara Key Summary Points

Inadequate adherence to diabetes care and to follow-up guidelines by physicians may contribute to suboptimal glycaemic, lipid and blood pressure target attainment rates in subjects with type 2 diabetes (T2DM).

The aim of the GLANCE study was to assess the level of physician adherence to the patient follow-up protocol (PFP) of the Hellenic Diabetes Association (HAD) 2017 T2DM guidelines, the achievement rates for glycaemic, blood pressure and low-density lipoprotein-cholesterol (LDL-C) targets and physician adherence to the HDA-recommended pharmacological treatment algorithm.

Overall, physician adherence to the PFP of the national guidelines for T2DM was found to be suboptimal, but the glycaemic target was achieved by eight of ten subjects.

It is important to identify the barriers to an adequate adherence by physicians to the full PFP provided in the 2017 HDA guidelines, with the aim to optimally support subjects with T2DM and improve outcomes.

Digital Features

This article is published with digital features to facilitate understanding of the article. You can access the digital features on the article’s associated Figshare page. To view digital features for this article go to https://doi.org/10.6084/m9.figshare.12967595.

Introduction

Diabetes is an emerging global epidemic of the twenty-first century, with an ever-increasing prevalence [1]. The International Diabetes Federation (IDF) 2019 estimate of diabetes prevalence in Greece was 7.4% of the national population [1]. Similarly, a study based on 2014–2015 Greek national electronic prescription database data estimated the prevalence of medication-prescribed diabetes at 8.2% for adults in general and at 30.3% for those aged ≥ 75 years [2].

Type 2 diabetes (T2DM) accounts for about 90% of all cases of diabetes [3]. Persons with T2DM carry a high risk of developing micro- and macrovascular complications, with atherosclerotic cardiovascular disease (CVD) being the primary source of morbidity and mortality [4,5,6]. Current management goals include achieving adequate glycaemic control and addressing vascular risk factors, notably blood pressure (BP), lipids, among others [7,8,9,10]. In general, the achievement of therapeutic targets in T2DM is not optimal [11, 12]; this also holds true for Greece [13,14,15,16], where the achievement rates of glycated haemoglobin (HbA1c) and BP targets have been reported to be 53 and 27%, respectively [15]. Achievement of therapeutic targets may be linked to adherence to diabetes standards of medical care recommendations [17,18,19]. In Greece, the Hellenic Diabetes Association (HDA) issues and regularly updates practice guidelines for the management and care of persons with diabetes [20], which are aligned with the guidelines of the American Diabetes Association (ADA) and European Association for the Study of Diabetes (EASD) [21,22,23]. The 2017 HDA guidelines include a patient follow-up protocol (PFP) that lists 62 items that should be part of a patient’s medical history and clinical and laboratory evaluations, with the aim to achieve optimal diabetes management [20].

To date, the adherence of Greek physicians to the PFP has not been systematically explored. The aims of this study were to assess (1) the adherence level of physicians to the PFP of the HDA 2017 T2DM guidelines [20] and (2) the achievement rates of glycaemic, BP and low-density lipoprotein-cholesterol (LDL-C) control; and (3) to examine physician adherence to the HDA-recommended pharmacological treatment algorithm.

Methods

Study Design

The present study (GLANCE) was a nationwide, multi-centre, cross-sectional study conducted in 15 hospital-based outpatient clinics and 38 private practices located across Greece. A total of 610 persons with T2DM were consecutively enrolled in the study (referred to as subjects) by 14 endocrinologists, 35 internists and four general practitioners with a special interest in diabetes between 27 June and 20 December 2018. The study was conducted in accordance with the International Society for Pharmacoepidemiology guidelines for Good Pharmacoepidemiology Practice, the ethical principles of the Declaration of Helsinki of 1964 and its later amendments and all standing regulations. The study was approved by the ethics committees of all participating hospitals (see Electronic Supplementary Material Table S1). Signed written informed consent was obtained from all participants.

The study included non-insulin-dependent persons with T2DM who were aged ≥ 19 years at enrolment, had ≥ 2 HbA1c measurements during the past year, had a HbA1c target < 7% and were receiving oral hypoglycaemic agents (OHAs) for ≥ 1 year. Exclusion criteria were patients who were insulin dependent; had type 1 DM or gestational diabetes; were hospitalised; were participating in any clinical trial at time of enrolment; had in the current or past (preceding) year a history of alcohol or drug abuse; were pregnant/lactating (at enrolment or within the previous year); subjects who did not have HbA1c < 7% as a therapeutic target were also excluded.

In order to ensure country-wide representation of clinical practices, the sample of patients was distributed according to the percentage of the Greek population residing in the various regions. Consecutive patients from those attending the selected centres during a pre-specified period were invited to participate in the study.

Data were collected using a web-based data system during the single study visit, which occurred at enrolment, and through patient self-report and medical chart review. For the evaluation of adherence by physicians to the HDA guidelines, we used information available in participants’ medical records within the last 365 days and up tol the most recent visit prior to enrolment. For the assessment of metabolic target attainment rates, we used laboratory assessments available at enrolment or within 1 month prior to enrolment. Compliance of participants to their current antidiabetic treatment regimen and their adherence to physician-provided nutritional recommendations were rated by the physicians using a five-level Likert-type scale with responses of “very good”, “good”, “moderate”, “poor” and “very poor”.

Outcomes

The primary outcome was overall physican adherence to the PFP of the HDA guidelines, at a participant level. Other outcomes included: (1) glycaemic control, defined as the proportion of subjects attaining the HbA1c target < 7%; (2) adherence to the HDA-recommended therapeutic algorithm; (3) physician adherence to each domain and each item of the PFP; (4) adherence to the PFP according to each physician’s specialty (endocrinologist, internist, general practitioner); and (5) target achievement for LDL-C, BP and composite metabolic control, in accordance with the HDA T2DM guidelines [20], as follows: (1) systolic BP (SBP) < 140 mmHg (< 130 mmHg for subjects < 65 years old) and diastolic BP (DBP) < 85 mmHg (< 80 mmHg for subjects < 65 years old); (2) composite SBP/DBP target of SBP < 140 and DBP < 85 mmHg for subjects  aged ≥  65 years old, and SBP < 130 and DBP < 80 mmHg for those aged < 65 years; (3) LDL-C < 100 mg/dL or < 70 mg/dL in subjects with pre-existing CVD (i.e. coronary artery disease, peripheral arterial disease, stroke, transient ischaemic attack, carotid artery stenosis (> 50%) and abdominal aortic aneurysm); (4) composite metabolic control, i.e. simultaneous attainment of HbA1c, LDL-C, SBP and DBP targets as defined above.

Follow-Up Protocol

The PFP of the HDA 2017 guidelines [20] comprises 62 items (Table 1), organised in three distinct domains: complete medical history (items 1–30), physical examination (items 31–42) and laboratory evaluation (items 43–62). For the following items to be considered to be completed, the respective information should have been available in the patient’s medical record: items 1, 3–10, 13, 21–29, 31–37, 40, 43 on every visit, and items 44–58 and 62 at least once in the year prior to enrolment. The remaining items were considered fulfilled if the physician confirmed they had been evaluated.

Table 1 Items and domains based on the patient’s follow-up protocol of the Hellenic Diabetes Association guidelines and physicians’ adherence to each item
Full size table

Statistical Analyses

Statistical analyses were performed using SAS® software version 9.4 (SAS Institute Inc., Cary, NC, USA). Categorical variables are presented as frequencies (n, %), while continuous variables are presented as the mean and standard deviation (SD) or as the median and interquartile range (IQR). Normality of distribution was assessed using the Kolmogorov–Smirnov test. For each subject, a crude score of overall adherence was estimated as the sum of fulfilled items, divided by the total number of PFP items (n = 62), multiplied by 100. Each item was scored 1 point if fulfilled or 0 if not. The mean overall adherence 95% confidence interval (CI) was calculated. The 95% Wald CI for binomial proportions is presented for the glycaemic control rate.

To identify patient- and physician-related factors associated with physicians’ adherence to the HDA guidelines PFP, we performed univariable and multivariable linear regression analyses. The initial multivariable model included the following factors: participant’s age at enrolment (≥ 65 vs. < 65 years), sex, T2DM duration (> 5 vs. ≤ 5 years), obesity, number of past or ongoing medical conditions/comorbidities (> 3 vs. ≤ 3), presence of diabetic complications and physicians’ specialty (endocrinologist vs. general practitioner, internist vs. general practitioner). The final multivariable model was derived through a stepwise procedure based on the minimisation of the Akaike information criterion.

Sample Size

Due to the descriptive nature of the study, sample size calculation was based on the study’s primary endpoint, and no hypothesis-specific calculation was performed. Assuming that the estimated SD of the overall adherence score could range from 0.5 to approximately 1, a sample size of 600 participants was considered appropriate to estimate the adherence level with a 95% CI half-width < 0.1, with at least 80% power.

Results

Baseline Characteristics at Enrolment

A total of 610 subjects were enrolled. Of these, 601 [mean (SD); age 65.2 (10.3) years; 54.6% men] were included in the analysis (Fig. 1). Participants’ demographic and anthropometric characteristics are presented in Table 2. Mean (SD) age at T2DM diagnosis was 57.3 (10.4) years, with a median disease duration of 5.9 years (Table 2). Median (IQR) subject follow-up duration was 2.6 (1.5–4.7) years, with 29.5% of subjects having been diagnosed with T2DM by the study physicians.

Fig. 1
figure 1

Subject disposition and flow chart. Flow chart displaying the number of subjects enrolled and analysed, and the number of enrolled subjects analysed per physician specialty and healthcare setting

Full size image
Table 2 Demographic and anthropometric characteristics of persons enrolled in the study
Full size table

Overall, 96.5% of participants had ≥ 1 [median (IQR): 3.0 (2.0–4.0)] clinically relevant medical condition/comorbidity. Conditions/comorbidities reported in ≥ 10% of the participants are shown in Table 2. At enrolment, 93.8% (563/600) of subjects were receiving at least one concomitant medication other than antidiabetics. Medication classes received by > 10% of the participants included lipid-modifying agents (73.5%), cardiac therapy/anti-hypertensives (66.5%) and antithrombotic or antiplatelet agents (30.5%).

Frequency of diabetic complications was 16.1%: 6.5% of the subjects had macrovascular, 6.0% microvascular and 7.0% other diabetic complications (Table 3). Additionally, 12 months prior to enrolment, 4.4% of the 571 evaluable subjects had experienced a total of 134 (median 3) hypoglycaemic events. The latter included 75 of unknown severity and 59 mild ones. Prior to enrolment, most subjects had received training on diabetes self-management and personalised nutrition counselling, and most were using a self-monitoring glucose device (Table 3).

Table 3 Type 2 diabetes history, training on diabetes self-management, nutrition counselling and healthcare resource utilisation in the year prior to enrolment
Full size table

Treatment for T2DM and Subjects’ Compliance to Medication

Antidiabetic therapy since T2DM diagnosis was known for 509 (84.7%) of subjects. Among these, 30.5% had received one, 31.2% two and 38.3% ≥ 3 antidiabetic agents (Table 4). A median (IQR) of 4.0 (2.1–9.0) years had elapsed from first receipt of an OHA to enrolment.

Table 4 Medications used by study subjects for management of type 2 diabetes mellitus prior to study enrolment and at enrolment
Full size table

At enrolment, all subjects were receiving OHAs with a median (IQR) treatment duration of 3.8 (2.0–8.9) years. Specifically, 27.6% of subjects were receiving metformin monotherapy, 63.4% were receiving metformin + other OHAs and 9.0% were receiving therapy that did not include metformin (Table 4). Participants’ compliance with antidiabetic medication was rated by the physicians on a five-level scale: 67.5% (400/593) of the subjects were rated as having very good compliance, 26.0% (154/593) as good, 3.9% (23/593) as moderate and 2.7% (16/593) as poor or very poor. In addition, compliance was rated very good/good in 95.4% (188/197) of those receiving monotherapy, in 95.4% (208/218) of those on dual therapy, in 88.8% (135/152) of those receiving triple therapy and in 92.0% (23/25) of those on quadruple therapy.

Guidelines Followed by Physicians

Physicians followed the HDA therapeutic guidelines for managing 77.2% of the subjects, the ADA/EASD guidelines for 43.4% and the American Association of Clinical Endocrinologists/American College of Endocrinology (AACE/ACE) guidelines for 8.8%, while different guidelines were followed according to physicians’ specialty (Fig. 2). Hospital-based and office-based physicians followed the HDA guidelines for managing 85.6% (173/202) and 72.9% (291/399) of the subjects, the ADA/EASD guidelines for 40.1% (81/202) and 45.1% (180/399), and the AACE/ACE guidelines for 12.4% (25/202) and 7.0% (28/399), respectively.

Fig. 2
figure 2

Diabetes guidelines followed by physicians overall and according to physicians’ specialty. Bars and labels indicate percentage of subjects for which each guideline was followed. More than one guideline was followed for some participants. AACE/ACE American Association of Clinical Endocrinologists/American College of Endocrinology, ADA/EASD American Diabetes Association/European Association for the Study of Diabetes, HDA Hellenic Diabetes Association

Full size image

Physicians’ Adherence to the PFP of the HDA Guidelines

Mean overall physician adherence to the HDA guidelines PFP was 43.6% [95% CI 42.3–45.0, median (IQR) 41.9% (32.3–54.8)] (Fig. 3). The highest adherence was noted for the laboratory evaluation domain (mean 48.3%, 95% CI 46.2–50.3), followed by the complete medical history domain (mean 45.5%, 95% CI 43.9–47.1) and then by the physical examination domain (mean 31.2%, 95% CI 29.3–33.2) (Fig. 2).

Fig. 3
figure 3

Adherence to the patient follow-up protocol (PFP) of the HDA guidelines overall and according to physicians’ specialty. Bars represent the mean and standard deviation, while the labels indicate means

Full size image

Mean overall adherence and the results for adherence to each PFP domain according to physicians’ specialty are presented in Fig. 2. The percentage of subjects for whom each of the 62 PFP items was completed are shown in Table 1. Items completed in > 90% of subjects included the items “current antidiabetic treatment”, “age or date of birth”, “response to treatment”, “year of T2DM diagnosis” and “HbA1c measurement”. Items fulfilled in < 10% of the subjects were the items “alcohol consumption”, “psychosocial problems”, “surgical procedures” and “electrocardiogram at rest’.

The overall adherence of physicians to the PFP was higher for female subjects (p = 0.026), subjects with > 3 medical conditions/comorbidities (p = 0.043), those with diabetic complications (p < 0.001) and those seen by endocrinologists versus general practitioners (p < 0.001) and by internists versus general practitioners (p = 0.001), when adjusted for other factors in the final multivariable model (Table 5). Significant predictors of physicians’ adherence overall and to each of the PFP domains are shown in Table 5.

Table 5 Predictors of physicians’ adherence to the type 2 diabetes mellitus patient follow-up protocol (PFP) of the Hellenic Diabetes Association guidelines (multivariable linear regression), overall and for each PFP domain
Full size table

Achievement of Glycaemic Control at Enrolment

Based on available HbA1c measurements, adequate glycaemic control (HbA1c < 7%) was achieved by 82.1% of all participants (363/442, 95% CI 78.6–85.7; Fig. 4a). Among the 79 (17.9%) subjects without adequate glycemic control, 71 had HbA1c 7–8.5%, one had HbA1c > 8.5 and ≤ 9%, and seven had HbA1c > 9%. The glycaemic control achievement rate according to the physicians’ specialty is shown in Fig. 4a. The glycaemic control rate decreased as the number of OHAs increased, ranging from 91.6% in those receiving antidiabetic monotherapy to 58.8% in those receiving four OHAs (Fig. 4b).

Fig. 4
figure 4

Achievement of glycaemic control and action taken with current antidiabetic medication by physicians at enrolment. a Glycaemic control achievement rate, overall and according to physicians’ specialty. Labels indicate the percentages of participants achieving control, and N is the number of patients for whom data were available on haemoglobin A1c (HbA1c). b Glycaemic control achievement rate by number of antidiabetic treatment classes received. Labels indicate the percentages of participants achieving control, and N is the number of patients for whom data were available on HbA1c. c Distribution of participants [in absolute numbers (N)] attaining the target HbA1c < 7% according to the action to be taken with their antidiabetic medication at enrolment. d Distribution of participants [in absolute numbers (N)] failing to attain the target HbA1c < 7% according to the action to be taken with their antidiabetic medication at enrolment

Full size image

LDL-C, Blood Pressure, and Composite Metabolic Control at Enrolment

Overall, the percentages of subjects achieving the LDL-C, SBP, DBP, composite SBP/DBP and composite metabolic targets were 57.0% (174/305) [with established CVD: 26.7% (12/45), without CVD: 62.3% (162/260)], 55.2% (289/524), 60.7% (318/524), 42.6% (223/524) [≥ 65 years: 59.6% (159/267), < 65 years: 24.9% (64/257)] and 21.6% (58/268), respectively. The corresponding rates for SBP/DBP target achievement per physicians’ specialty were 44.4% (59/133), 42.2% (152/360, and 38.7% (12/31) for endocrinologists, internists and general practitioners respectively; achievement of the composite metabolic target was 21.6% (21/97), 21.5% (34/158) and 23.1% (3/13) for the aforementioned specialties.

Changes in OHAs at Enrolment

For 89.4% (537/601) of subjects, OHAs were not changed. Conversely, antidiabetic treatment was changed for 64 subjects (10.6%), including 34 with HbA1c < 7%, 21 with HbA1c ≥ 7% (Fig. 4c, d), and nine for whom HbA1c data were unavailable. Changes in OHAs adhered to the HDA-recommended therapeutic algorithm in 83.5% (369/442) of subjects with available HbA1c [in 96.7% (351/363) of those with HbA1c < 7% and in 22.8% (18/79) of those with HbA1c ≥ 7% at enrolment].

Discussion

The GLANCE study provides novel evidence on the adherence of Greek physicians to the PFP of the HDA guidelines in the routine care setting. It also provides data on the achievement rates for glycemic, blood pressure and lipid and composite metabolic control targets of non-insulin dependent subjects with T2DM in routine care settings.

The mean overall adherence of physicians to the PFP of the HDA guidelines was below 50%. Physican adherence was highest for the laboratory evaluation domain (48%), followed by the complete medical history domain (46%) and then the physical examination domain (31%). Slightly higher physician adherence rates were reported in the ADMIRE (Adherence of physicians to guidelines for the management of type 2 diabetes) study [18, 24], which had a similar design as the GLANCE study and evaluated physician adherence to the Turkish diabetes guidelines PFP in 1790 patients [18, 25]. The Turkish and Greek PFP have several items in common, and the former separates the items into the same three domains as examined in the GLANCE study [20]. In ADMIRE, the mean overall adherence score for the 12 months prior to study enrolment was 70%, with the scores (on 10-point scales) being 8.8 for the 5-item medical history domain, 6.3 for the 3-item laboratory evaluation domain, and 5.9 for the 10-item physical examination domain [18]. The difference in the adherence scores between the two studies may be partly due to all 62 items of the HDA PFP being considered in the GLANCE study, whereas only 18 of the 39 items of the Turkish guidelines were evaluated in the ADMIRE study [18, 24, 25]. Additionally, the higher physician adherence rate in ADMIRE could, at least partly, be attributed some items of the PFP examined possibly being among those more commonly performed by physicians. For example, mean adherence score for the laboratory domain in the present study would be about 60% (and thus similar to the mean of 6.3 on a 10-point scale reported in ADMIRE), if it was based on the laboratory domain items included in ADMIRE [18, 24]. However, caution is needed when comparing the results of these two studies given that they differ in terms of patient characteristics. Importantly, in ADMIRE [24], 35% of subjects received insulin, which was an exclusion criterion in the present study. Moreover, female sex was more frequent in ADMIRE (62%) [24] than in the present study (45%), and this has been shown to affect adherence, both in the present study and in a Luxembourg study [26]. Finally, in both the present study and in ADMIRE, diabetic complications and physicians’ specialty were associated with adherence [18].

The results from the multivariable model on the factors affecting the overall PFP adherence require further investigation. For example, there are no specific cultural characteristics that may explain the increased adherence related to female subjects; and mixed findings have been reported in published data on the sex differences affecting the management of people with diabetes [27]. Furthermore, subjects with multiple comorbidities may have more regular access to the healthcare system and they also may receive more attention from their treating physician related to diabetes management and regular follow-up. In addition, the specialised and continuous education of endocrinologists and internal medicine physicians in diabetes could further provide a potential explanation of these results.

The general reasons reported for the poor adherence of physicians to guidelines include, among others, a lack of awareness or familiarity with the guidelines and/or the absence of any education on the use of the guidelines [17, 28, 29]. Indeed, in the ADMIRE study, overall physican adherence increased following the introduction of an education programme on the guidelines and diabetes complications [18]. Additional barriers to adherence may include poor healthcare finances and resources, inadequate healthcare system infrastructure and logistic support, excessive physician workload and the fact that physicians often doubt the utility of guidelines [29]. During the financial crisis in Greece, it was reported that prescriptions for OHAs were not affected [30], but decreased income and unemployment were identified as the major factors affecting access to healthcare services [31]. Specifically, the concerns of subjects with T2DM about treatment costs may affect their adherence to treatment and frequency of prescription fulfilment or even influence physicians’ prescription behaviour towards choosing less expensive treatment options [32]. Moreover, budget cutbacks related to biochemical testing for and poor follow-up of diabetic complications that may prevent regular patient monitoring [32] could also play an important role in physician adherence to PFP. However, an inquiry into these factors was beyond the scope of GLANCE.

Of note, some studies have reported high adherence to guideline recommendations in terms of laboratory tests but inadequate glycemic control [33,34,35]. We found the reverse discrepancy, i.e. inadequate PFP adherence, but high glycaemic control (82.1%). Impressively, this glycaemic control rate is the highest reported over the last decade in multi-centre studies in routine care Greek settings. In the recent AGREEMENT registry of 1191 adult patients with T2DM, adequate glycaemic control was achieved by 53% of subjects overall and by 65% of non-insulin-treated subjects [36]. Other studies have reported glycaemic control rates of 53–67% [15, 30, 37]. In the GLANCE study, the LDL-C target was achieved by 57% of subjects, compared to 31% [15] and 60% [36] in previous Greek studies. In addition, adequate SBP control was achieved in the GLANCE study by 55% of subjects compared to a previously reported rate of 27% [15], while the composite metabolic control rate was 22% in this work compared to about 5% in a previous study [15]. Arguably, our results point to some improvement—or at least some degree of stability—in the observed rates compared to those reported previously. Nonetheless, it should be mentioned that the present work was of a different design to those of the aforementioned Greek studies and had the primary aim to assess physician adherence to the HDA diabetes care guideline, with a secondary aim to capture the rates of achievement of glycemic, blood pressure and lipid targets.

Patient adherence to medication is known to affect the achievement of therapeutic targets [38, 39]. In the present study, physicians reported that compliance to OHAs was very good/good in the vast majority (93%) of subjects. Importantly, therapeutic decisions for changes in OHAs adhered to the HDA therapeutic algorithm in nearly nine of ten subjects. This high percentage probably reflects the wide acceptability of the HDA guidelines and the quality of the HDA-initiated physician education initiative on antidiabetic therapy.

Overall, the findings of GLANCE indicate that further efforts should focus on identifying potential barriers to the suboptimal adherence of physicians to the 62 PFP items and towards increasing physician awareness of the importance of comprehensive medical evaluation as outlined by HDA and international diabetes guidelines [20,21,22,23]. Moreover, given the available resources of physicians and taking into consideration potential financial and time constraints in routine care, it may be useful to take a closer look at the 62 items in the national PFP that were most met and least commonly met and potentially prioritise some items in the PFP. This could possibly optimise care for subjects with T2DM under the realistic situations encountered by physicians in daily practice [40]. On the other hand, the rapid evolution in digital solutions for diabetes management could further support patient follow-up protocols and improve outcomes [41]. However, as highlighted by the EASD and the ADA Diabetes Technology Working Group, several issues regarding the standardisation of diabetes digital applications should be addressed so that they can be validated in terms of their clinical significance and ensure patient safety and data privacy [42].

The strengths of the study include the large sample size and its multi-centric and nation-wide nature. The recruitment of subjects from 10 of the 13 administrative regions of Greece (residence to 93.4% of the country’s population) and the inclusion of endocrinologists, internists and general practitioners reflect variations in clinical practice. However, the study also has a number of limitations. First, bias in subject selection and information cannot be entirely ruled out. The former was minimised by consecutive enrolment. Regarding the latter, the rate of missing data on HbA1c, LDL-C and SBP/DBP measurements was between 13% (SBP/DBP) and 49% (LDL-C). Moreover, data on diabetic complications were only collected from medical records; therefore, underestimation of their actual prevalence cannot be ruled out. A further limitation is that we did not use a central laboratory for measurements. Moreover, the inclusion of subjects with ≥ 2 HbA1c measurements may have contributed to overestimation of adherence. Also, the Likert scale used by physicians to evaluate patient compliance to antidiabetic treatment as well as patient adherence to physicians’ recommendations represents per se a subjective rating method, which does not necessarily reflect the real situation. Lastly, results should be interpreted with caution, and a prospective study using all subjects with T2DM from the national diabetes registry could shed more light on current T2DM management in Greece.

Conclusions

The adherence of Greek physicians to the full PFP of the national diabetes guidelines is suboptimal. Of note, this adherence is affected by physicians’ specialty as well as patient characteristics, such as sex, comorbidities and diabetic complications. Hence, future efforts should strive to identify potential barriers to guideline adherence in order to improve it and potentially prioritise the items included in the PFP, with the ultimate goal to optimally support subjects with T2DM.