Factors affecting prescription of sodium-glucose co-transporter 2 inhibitors in patients with type 2 diabetes mellitus with established cardiovascular disease/ chronic kidney disease in Hong Kong: a qualitative study

Diabetes management aims to prevent complications and maintain quality of life [1]. Atherosclerotic cardiovascular disease (ASCVD) and chronic kidney disease (CKD) remain the two most essential complications for diabetes worldwide. The crude incidence of diabetes complications in Hong Kong (HK) were 33.7% for albuminuria, 16.4% for stage 3 CKD or above, 6.1% for the history of coronary heart disease, and 5.1% for the history of stroke according to a recent territory-wide study [2]. ASCVD is one of leading causes of death and disability in diabetes [3]. Meanwhile, patients with CKD have a greater risk of cardiovascular complications [4], and CKD due to diabetes mellitus has become the major cause of end-stage renal failure (ESRF) requiring renal replacement therapy in HK [5]. These complications have posed a significant financial burden on the health care system. A local analysis on direct medical expenditure revealed a 5.94-fold increase in costs for new stroke events, and a 2 to 3-fold rise in expenses for acute myocardial infarction (MI), heart failure (HF), stroke, and ESRF [6].

Sodium-glucose co-transporter 2 inhibitor (SGLT2 I) belongs to a novel class of oral glucose-lowering agents. Besides glycaemic control, they have been shown to have other beneficial effects for the cardiovascular and renal systems, which include diuresis, reduction in weight, lowering of blood pressure [1, 7‐9], as well as having low risks of hypoglycaemia [8]. They also have been proven to have substantial cardio-renal protective benefits, including a significant decrease in the composite outcome of MI, stroke, and cardiovascular death, reduced risk of cardiovascular mortality, reduction of all-cause mortality, reduction of hospitalisation for HF as well as progression of renal disease in four landmark clinical outcome trials [9‐12] and real-world studies [13, 14]. The trials also suggested that the shown cardiovascular risk reduction effects were not mediated by glycaemic control [9‐11]. The above findings have led to a significant shift from a sole emphasis on glycaemic control to improving cardio-renal outcomes when choosing glucose-lowering agents. SGLT2 I have been strongly recommended for patients with diabetes and established ASCVD /CKD/HF by the American Diabetes Association /European Association for the Study of Diabetes (ADA/EASD) guidelines [1], European Society of Cardiology (ESC) guidelines [15], and Kidney Disease: Improving Global Outcomes (KDIGO) guidelines [16]. Another glucose-lowering agent with cardio-renal protective benefits is glucagon-like peptide-1 receptor agonist (GLP-1 RA). It is however not available in primary care settings in HK and requires regular subcutaneous injections, which makes it less convenient for patients. As a result, the presence of established ASCVD and/or CKD is considered a compelling indication for utilising SGLT2 I in patients with diabetes in primary care settings in HK.

However, many factors can affect doctors’ decisions. A recent nationwide cohort study in the US found that only 7.2% pharmacologically treated patients with diabetes were prescribed SGLT2 I [17]. Additionally, there was a treatment-risk paradox that patients with MI and CKD were less likely to receive SGLT2 I [17]. One study in the UK showed that although the use of SGLT2 I was increasing, the overall utilisation of SGLT2 I remained low, and the prescription was irrespective of cardiovascular status [18]. Another study in the USA also revealed that very low utilisation of SGLT2 I among patients with diabetes and proteinuric CKD [19]. SGLT2 I was introduced in 2015 in HK. Although SGLT2 I is subsidised by the government and no extra out-of-pocket payment is needed from patients in public healthcare settings, the adoption rate of SGLT2 I by clinicians and patients were still low [20]. Therefore, there is great hesitancy in the utilisation of SGLT2 I both globally and locally.

Previous study suggested that clinical inertia, limited knowledge, insufficient treatment re-evaluation, cost and competing priorities may have contributed to the low prescription rates [21]. However, few studies have explored the influencing factors of prescription from the perspective of primary care doctors, and there was a lack of understanding of the underlying reasons for their prescription behaviour. Qualitative study helps to deeply explore peoples’ beliefs, experiences, and views [22]. In HK, greater than 65% patients with diabetes are managed in public settings [6]. Therefore, this study aimed to explore the factors affecting primary care doctors in the public healthcare setting in prescribing SGLT2 I for patients with diabetes and established ASCVD /CKD in HK using a qualitative approach. An increased understanding in primary care doctors’ decision-making process will help design effective and tailored improvement strategies to achieve better management of patients with diabetes in the public healthcare setting.

Seventeen primary care doctors with diverse characteristics were interviewed (Table 3). The demographic information for individual participants is listed in Table 4. The interviews lasted from 29 min to 48 min (mean interview length 37 ± 6 min). Data saturation was reached after 13 interviews as determined by the research team. A further four interviews were carried out to verify that no more new themes emerged. These additional interviews did not contribute to the further development of new themes. No participants requested to read their transcripts.

The inductive analysis generated four themes and nine subthemes that influenced doctors’ prescription. The four themes were: 1) Knowledge and previous practice patterns influence prescription; 2) Balancing risks and benefits; 3) Doctor’s professional responsibilities and 4) System barriers. These themes were then deductively mapped to the nine specific domains of the TDF. Theses TDF domains were: knowledge; intention; memory; beliefs about capabilities; beliefs about consequences; goals; role and identity; emotion; and environmental constraints. It was determined that TDF adequately captured all perceived determinants affecting utilisation of SGLT2 I. The themes, subthemes, related TDF domains, and the quotes to support them are displayed in Table 5. Participants were identified by unique codes.

To our knowledge, this is one of few studies to apply a qualitative method to assess factors influencing primary care doctors’ prescriptions of SGLT2 I for patients with diabetes and ASCVD/CKD. Using TDF as a theoretical framework and in-depth semi-structured interviews, we were able to obtain a comprehensive and systematic understanding of the determinants affecting doctors’ prescriptions.

Generally, most participants were familiar with the cardio-renal advantages of SGLT2 I and believed it had a good safety profile. They also recognised their role in secondary prevention. However, there were a few roadblocks to prescription. Some were unique to SGLT2 I, others were common barriers encountered in general.

The most prominent barrier found in our study was a lack of understanding that the cardio-renal benefits of SGLT2 I was independent of glycaemic control. Despite acknowledging the established value and effectiveness of SGLT2 I in cardio- renal outcomes trials, most participants still prescribed it purely as a hypoglycaemic agent, ignoring this crucial tool for cardio-renal protection. Most doctors would not consider adopting SGLT2 I if the HbA1Cc was too high or within the target. This prescribing pattern reflects a lack of knowledge about updated guidelines. A recent survey also revealed a similar knowledge gap among cardiologists [30], which showed the top barrier for prescription of SGLT2 I was a lack of knowledge. Furthermore, more than half of the interviewed cardiologists did not feel it was their responsibility to prescribe anti-diabetic medications [30]. A previous qualitative study has demonstrated similar knowledge gap of under-appreciation of the cardio-renal benefits of SGLT I by general practitioners, contributing to low prescription rates [31]. However, in contrast to their findings that there was a preference for endocrinologists to initiate therapy, most of our interviewees thought it was appropriate for primary care doctors to initiate SGLT2 I. A possible explanation for the difference could be the fact that our participants were experienced in diabetes management as the majority of patients with diabetes in NTWC were under the care of the public sector. This was also echoed by the same study that general practitioners who frequently managed diabetes were more confident to prescribe SGLT2 I for patients with diabetes [31]. There is, however, a great need to enhance our frontline primary care doctors to use SGLT2 I more actively for their cardiorenal protective effects, and not just for glycaemic control.

Another main concern was renal impairment. Despite knowing the renal protective effects of SGLT2 I, many participants were reluctant to use in patients with renal impairment. This phenomenon was also observed in a cross-sectional study in Korea which showed the utilisation of SGLT2 I was significantly higher in CKD patients with better estimated Glomerular Filtration Rate (eGFR) than those with lower eGFR [32]. A recent study in the UK showed that when SGLT2 I was initiated, over 90% of patients had an eGFR≥60 mL/minute/1.73 m², while only 1.7% had an eGFR < 60 mL/minute/1.73 m² [33]. This indicates that the prescription of SGLT2 I was heavily affected by patients’ renal impairment. There were also some conflicting recommendations regarding the use of SGLT2 I in patients with CKD which may have led to hesitations in primary care doctors prescribing SGLT2 I. The approved renal threshold for eGFR for prescription of SGLT2 I was 45 mL/minute/1.73 m² at the time (January to May 2021) of this study. However, this threshold had been reduced to 30 mL/min/1.73 m² in late 2021 in HK by the pharmaceutical company following the emergence of evidence. With the results from the DAPA-CKD trials [34] and other ongoing clinical trials involving the administration of SGLT2 I in different CKD stages, the authorised renal threshold for starting SGLT2 I might even be lower. Therefore, primary care doctors need to update regularly with the latest guideline recommendations. Future strategies should be in place to address some of the common individual (including a lack of awareness of the evidence), health system (including time constraints), and contextual barriers (including a lack of agreement with the evidence) for the implementation of clinical practice guidelines [35].

Moreover, patients’ age was recurrently cited as a critical factor affecting doctors’ decision-making. Old age has always been a key concern in pharmacotherapy due to their medical complexity, multi-morbidity, frailty, and the risk of polypharmacy [36]. For example, the use of statin has been reported to be suboptimal in elderly aged 65 to 79 years with cardiovascular disease, despite recommendations by multiple international guidelines and decades of clinical evidence [37]. For SGLT2 I, the cardio-renal benefits were consistent across all age groups including those over 65 [38, 39]. The post hoc analysis of the EMPA-REG OUTCOME study, which included 35.3% of patients between 65 and < 75 years old and 9.3% ≥75 years old, found that empagliflozin reduced the risks of CV mortality, heart failure, and renal outcomes in all age groups [38]. Similarly, post hoc analysis of the DECLARE study, in which 40% of patients were between 65 and 75 years old, and 6% > 75 years old, found that dapagliflozin is effective and safe for all ages [40]. Additional efforts are still required nonetheless to address the ongoing concerns of safety of SGLT2 I in the older age groups and give confidence to doctors prescribe SGLT2I in this population.

This study also revealed that organisational constraints, such as time constraints and lack of continuity of care, and patients’ perceptions and preferences, were barriers affecting prescription of new medications. This had been echoed in the findings from prior studies on guideline adherence [41].

Based on the findings of this study, several strategies could be implemented to improve the prescription of SGLT2 I in patients with diabetes and established ASCVD/CKD. Firstly, doctors should have additional training and education for emphasising a paradigm shift away from only glycaemic management and toward cardio-renal protection, which is crucial for implementation of the evidence-based guidelines. An open and ongoing process is needed to encourage doctors to voice their concerns. Secondly, patients should have more education on cardio-renal risk control. This could be achieved through a patient empowerment program and supported by a multidisciplinary team, which has been proved to be able to reduce the incidence of diabetic complications, hospitalisations, and mortality [42]. Finally, for policymakers, they can add the prescription of SGLT2 I as a “key performance indicator” besides HbA1c control in patients with diabetes and established ASCVD/CKD. The latest ADA guidelines have suggested using SGLT2 I in this populationindependent of HbA1c status [43]. Since the HA maintains computerised data in the Clinical Management System for all patients under its care, policymakers can readily monitor the utilisation of SGLT2 I and establish objectives to gradually increase the adoption rate of SGLT2 I in patients with established ASCVD/CKD.

Our study suggests that there may be a significant knowledge gap among primary care doctors, and providers’ prescriptions are influenced by many factors, especially consideration of patients’ age, renal impairment, and patients’ perceptions and preferences. Our findings highlight the need for further interventions in HK’s public primary health care sector to address these factors to improve patients’ cardio-renal outcomes in the future.